SlideShare a Scribd company logo
Differential Behavioral Patterns in Sequence Performance Following
either Sequential or Random Task Practice
A thesis submitted in partial fulfillment of the requirements for the degree of Bachelor of Arts with Honors
in the Psychology concentration of Brown University
April 24th
, 2015
Kathryn Nicole Graves
Advisor: David Badre, Ph.D.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
2
Acknowledgments
I have been incredibly fortunate to have been a part of the Badre Lab at Brown University
for the past two years. My time as an undergraduate research assistant, getting to learn from and
work with this incredible team of brilliant researchers, has been a distinct honor, a welcome
challenge, and the most pivotal educational experience of my life.
First and foremost, I’d like to thank Dr. David Badre for fostering this intellectual
environment and allowing me to be a part of it. His guidance and feedback over the past two
years have allowed me invaluable growth as a student and scientist.
I would also like to thank Dr. Theresa Desrochers for being the kind of mentor I aspire to
be. From teaching me how to code, to trusting me to with her experiments and ideas, to giving
me life advice, to supporting me through my own struggles both in and out of the lab, she has
made me better in more ways than I can count.
Thank you to the entire Badre Lab for easily being the coolest group of people I’ve ever
associated with, and for being there for me in all the best ways throughout my entire experience.
I’d especially like to thank Jason Scimeca for laughing at my jokes and being a source of constant
positivity.
Thank you to Perri Katzman, a Badre Lab alumna, who has been my honors thesis
guardian angel.
Finally, thank you to my friends and family for being the kind of support system that has
made me feel unashamed to both brag and cry about my research in front of them.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
3
Abstract
Sequential task performance requires cognitive control, which is observed
experimentally by measuring reaction times during the completion of simple sequences of tasks.
Previous behavioral studies have shown an elevated reaction time at the first position task of a
given sequence, indicating the presence of cognitive control on the sequence level. The current
study investigated how an element of practice prior to sequence performance would affect this
cognitive control. Across seven experimental designs, participants were given a practice section,
in which they performed either sequential or random cued tasks, and a test section, in which they
performed unpracticed, novel sequences. Collectively, the results of these experiments suggest
that sequence performance procedure can be generalized with high efficiency from practice to
test, such that no significant difference in reaction times was demonstrated between the two
conditions. Given that participants demonstrated sequence performance behavior even when
cues were provided per trial, the results also suggest that sequence structure detection ability is
sensitive to implicit sequence structure. Overall, these results suggest that sequence learning is
sensitive to neither familiarity with specific sequences of tasks, nor sequential structure in
general.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
4
Introduction
Task and sequence performance are integral parts of daily life. From turning a doorknob
to picking up a pen, there are countless tasks that must be completed throughout the day. Task
performance takes time, and the amount of time required for completion of individual tasks, or
tasks within a sequence, can depend on a number of different variables, such as task difficulty,
cognitive resources available, task order, and available responses. The current study seeks to
illuminate the intersection of four bodies of research on four underlying components of sequence
practice and performance – switch costs, task practice, sequence costs, and motor sequence
practice.
Behavioral studies of task performance have identified a certain temporal “cost” that
exists when switching from one task to another task of equal complexity (Kramer et. al, 1999;
Allport et al., 1994; Fagot, 1994; Arbuthnott & Frank, 2000; Rogers & Monsell, 1995). This
cost, termed “switch cost,” was demonstrated to be robust across extended inter-trial intervals
(ITI’s), as indicated in a study by Allport et al. (1994), in which participants performed task
switching with intervals of either 500ms or 1000ms between each trial. The results of this study
indicated that, even with the extended ITI, participants still demonstrated an elevated reaction
time when switching between two consecutive tasks versus repeating the same task. This residual
temporal cost was thus attributed as a necessary aspect of task switching behavior.
Switch cost was initially defined as being a function of one of two separate cognitive
processes, both of which gained support from multiple cognitive theorists. The first cognitive
process, a combination of task inhibition and task set reconfiguration, involved the process of
actively preparing to respond to a successive task by inhibiting the response of a previous task
(Arbuthnott & Frank, 2000; Fagot, 1994; Rogers & Monsell, 1995). The second cognitive
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
5
process, task set inertia, involved a passive dissipation of the influence of a previous task set,
which must occur before one is able to respond to a successive trial (Allport, et al., 1994).
A pivotal study conducted by Meiran and Chorev (2000) reconciled these two
explanations by having participants perform task switching after either being given an extended
Response-Cue Interval (RCI) or an extended Cue-Target Interval (CTI). Participants judged the
spatial location of a stimulus on a 2x2 grid using two key presses, alternating between
discriminating across the vertical axis (UP-DOWN) and the horizontal axis (RIGHT-LEFT).
Across multiple iterations of the experiment, participants performed the task under varying
conditions. Some trials consisted of the interval between the last response and the proceeding
cue remaining constant while the interval between the presentation of the trial cue and the
presentation of the target varied, and vice versa. Both conditions displayed a decrease, but not
complete dissipation of reaction time, indicating that task switching consists of both task inertia
and the combination of task set inhibition and task set reconfiguration, as well as a residual
component that persisted across increased RCI and CTI (Meiran & Chorev, 2000). These
findings aligned with those of Rogers and Monsell, whose task switching study showed that the
residual switch cost persisted over an extended Inter-trial Interval (ITI) (Rogers & Monsell,
1995).
This switch cost in the context of the second component, task practice, is crucial due to
the fact that task switching in the context of daily life rarely consists of purely novel tasks – most
of the tasks individuals perform throughout the day, such as the aforementioned turning of a
doorknob and lifting of a pen, involve some degree of experience or practice. Experimentally,
exhaustive task practice has been shown to decrease overall reaction times and switch costs,
though never fully abolishing them (Rogers & Monsell, 1995; Meiran, 1996; Kramer et al., 1999;
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
6
Ruthruff et al., 2001). These results provide evidence for the assertion that the cognitive
phenomena of task set inertia, task set inhibition and task set reconfiguration are not entirely
unmalleable processes, and, to a degree, may dependent on task familiarity.
These same phenomena are also present in sequence level task performance – the third
and more direct contributing aspect of sequence practice and learning. However, there are added
costs associated with sequences, supplementary to the costs on the task level, that exist as a
function of the higher order processes needed to perform tasks in-sequence.
The significance of the distinction between task and sequence performance is emphasized
when considering the real-world context, in which isolated task performance is less prevalent
than tasks performance following some type of sequential schema – most tasks, like pouring milk
and measuring out grounds, are all part of a larger sequence of steps involved in achieving a
larger goal, like making coffee. In order to perform these tasks without external cues, one must
be able to maintain the overarching goal of having a cup of coffee in the end, while also updating
contexts in order to internally guide flexibly switching between different tasks – i.e. recognizing
that the coffee maker must be started only after pouring water in the reservoir, but before pouring
coffee into the mug, and being able to cognitively ‘reset’ in order to perform each new task.
The ability to perform these two higher-order cognitive processes of goal maintenance
and context updating is a function of cognitive control on the sequence level. Experimentally,
cognitive control can be measured as participants perform simple sequences of tasks by observing
the first position reaction time in that sequence (Schneider & Logan, 2006; Barcelo et al, 2007;
Pojac et al., 2009). In a key study performed by Schneider and Logan (2006), participants
performed repetitions of two sequential structures consisting of color and shape judgments – one
sequence followed the structure “AABB,” with one switch, while the other followed the structure
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
7
“ABBA,” with two switches. The experimental results showed that participants demonstrated a
significantly higher reaction time at the first position judgment of both sequences, regardless of
whether that first position was a switch or a repeat, than the corresponding internal sequence
judgment. This elevated reaction time, termed “restart cost”, is the temporal cost associated with
task resetting on the sequence level to allow for the initiation of the next sequence, and is
evidence of cognitive control (Schneider & Logan, 2006). Functionally, the dorsolateral
frontopolar cortex has been implicated in this higher-order cognitive process (Desrochers et al.,
2013; Badre et al., 2009).
Despite the conclusive body of evidence regarding cognitive control in the context of
general sequence performance, little research has been conducted specifically concerning the
concept of sequence learning. However, studies concerning tangential aspects of sequence
practice and learning, specifically those involving motor sequence learning, provide a context for
the current experiment. Behavioral results from previous motor sequence experiments show poor
transfer of learning from a practice to test, as manifest by a significant increase in response times
from the last run of practice to the first run of test (Willingham et al., 2000; Fezzani et al, 2000).
While these results do, in a sense, speak to learning on a sequence level, they are complicated by
the motor aspect. The paradigms used muddle the component contributions of motor and
sequence learning.
Thus, the current study seeks to selectively observe the effects of sequential structure in
practice and test by removing the motor aspect. However, the experimental structure maintains
the necessary elements that preserve previously determined behavioral trademarks of task and
sequence performance, such that observed differences from practice to test can be attributed with
confidence to learning, and not considered as possible artifacts of indistinguishable behaviors.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
8
The tasks consisted of using one of two key presses to make judgments for either color
(red or blue), shape (circle of square) and size (small or large). Judgments were based on a series
of five rules presented on a screen at the beginning of each block for the sequence practice
condition, and cues provided either during or immediately before trial presentation in the cued
condition (see Figure 1).
Figure 1 A) Example simple sequence structure: two internal switches, two internal repeats, switch at the first
position. B) Example complex sequence structure: three internal switches, one internal repeat, repeat at the first
position
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
9
Over the course of seven different experimental designs, participants first conducted a
practice section, which consisted of practicing either repetitions of two five-item sequences of
tasks – one simple and one complex – or random, cued tasks (see Figure 2). After the practice
section, participants performed a test section in which they performed either practiced, familiar
sequences and multiple novel sequences, or repetitions of two novel sequences.
Reaction times and error ratios were analyzed for differences in familiar versus novel
sequence performance, cued random task versus sequence practice performance, practice versus
test sections, and test sequence performance after having practiced either random or sequential
tasks. Across all seven experiments, both behavioral measures were analyzed for two different
effects. In order to identify any main effect related to differential practice, a paradigm element
that changed with each experiment, we performed two-sample t-tests on the measures after
collapsing them across sequence position within their respective practice groups. In search of a
position related interaction, specifically one involving changes in cognitive control at the first
position, we divided each practice group’s data into first and internal position data (positions 2-
5), and performed a repeated measures ANOVA.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
10
Figure 2 A) Abbreviated run showing general sequence structure B) Sequence rules, example trial, and sequence
position question for experiments 1 and 2, practice and test sections. Participants performed sequences at practice
and test. C) Sequence rules, example trial with cue, and sequence position question for experiment 3, practice and
test sections. Participants performed cued sequences at practice, same sequences uncued at test. D) Sequence rules,
example trial with cue, and sequence position question for experiments 4, and 5, practice and test sections.
Participants were shown a masked screen and performed trials that were in a sequence (Exp. 4) or random (Exp. 5).
E) Sequence rules, cue screen, example trial, and sequence position question for experiment 6 and 7, practice and test
sections. Participants were in one of two practice groups – random trial practice or sequence practice.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
11
Experiment 1
In Experiment 1, participants first performed a practice section, consisting of seven runs
of seven blocks of repetitions of two sequences, one simple and one complex. After practice,
participants performed seven more runs at test, which consisted of the two previously practiced,
now familiar sequences as well as multiple different novel sequences. Two hypotheses were
developed for this first experiment. The first assumed a static relationship between the first
position judgment and internal trial types within each sequence, and predicted that sequence
practice would facilitate later performance of familiar sequences overall at test. The other
hypothesis supported the alternate prediction that first position reaction time could be selectively
effected by practice due to increased efficiency with the cognitive controller during performance
of familiar sequences as compared to novel sequences at test.
Results
In order to determine whether or not sequence-specific learning was taking place at
practice and persisting at test, reaction times were measured across practice and test runs for each
trial type (first positions, switches, and repeats). Across all three trial types, the reaction times
indicated that participant performance improved with practice (main effect of run: F’s > 10.99,
p’s < 3.36 x 10-10
). This improvement in sequence performance carried over to test, with T-tests
showing no significant difference from the last practice run to the first test run (p’s> 0.11) (see
Figure 3).
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
12
Figure 3 learning curves indicated by reaction time for A) Practice and B) Familiar first position trials, C) Practice
and D) Familiar repeat trials, and E) Practice and F) Familiar switch trials. Curves suggest that participants were
learning, getting faster through practice and maintaining performance levels at test.
Post hoc T-Tests determined that reaction time across later practice runs were
significantly different from reaction times at the first run for nearly all conditions (complex
sequence repeats p = 0.07, all others p’s < 0.05).
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
13
The reaction times at test, however, indicated that there was no overall effect of sequence
practice on subsequent performance of familiar versus novel sequences (no main effect of
familiarity, p < 0.05) (see Figure 4). There was also no first position-specific benefit of sequence
practice (no familiarity x position interaction: F (1, 50) = 1, p < 0.05). However, the reaction
times did indicate that, in both conditions, participants were demonstrating trademark sequence
performance behaviors (main effect of position, F (1, 50) = 100.47, p < 0.001).
Figure 4. A-B) Reaction times and error ratios collapsed across sequence position for familiar and novel conditions
and test. Results show no main effect of familiarity. C-D) Reaction times and error ratios at first and internal
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
14
sequence positions, for familiar and novel conditions at test. Results indicate no position-specific interactions as a
function of practice.
Error ratios at test illustrated even greater chance effects, as compared to reaction times,
with no clear benefit of practice overall (no main effect of familiarity, p > 0.05), or the first
position (no familiarity x position interaction: F(1,50) = 0.28, p > 0.05).
Discussion
The results of Experiment 1 did not support either of our initial hypotheses predicting
practice effects – however, the main effect of position in reaction times validated the underlying
assumption that participant behavior would reflect sequence performance across all sequences
and both experimental sections.
Taken together, these results presented two viable possibilities for what was occurring
between the practice and test sections. First, it was possible that participant performance
improved to such a rapid plateau at practice and carried over so efficiently to test such that, when
overall performance of either the practice or later familiar sequences was averaged and compared
at test performance, the two sections displayed comparable performance efficacy. Second, the
results may have been caused by participants not getting enough time to practice the familiar
sequences, such that, throughout practice and test, those sequences still carried an element of
novelty that made them comparable to the true novel sequences.
Experiment 2
Adopting the latter theory, Experiment 2 was designed such that participants were given
over twice as much practice as those in Experiment 1. The experiment was now spread over the
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
15
course of two days, with two seven-run sections on the first day and a four-run section on the
second. The test section followed the same seven-run format as that in Experiment 1 and
succeeded the last practice section on day two.
Figure 5 learning curves indicated by reaction time for A) Practice and B) Familiar first position trials, C) Practice
and D) Familiar repeat trials, and E) Practice and F) Familiar switch trials. Curves suggest that participants were
learning, getting faster through practice and maintaining performance levels at test.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
16
Results
Similarly to Experiment 1, reaction time learning curves across all runs of practice
showed a significant decrease in magnitude (F’s > 19.5, p’s < 0.001; post hoc T-tests p’s < 0.01),
and no significant change from practice to test (p’s > 0.22) (see Figure 5).
Reaction times and error ratios at test mirrored those of Experiment 1, in that there was neither
evidence of an overall practice effect nor an interaction at the first position that suggested
selective improvement in the employment of the cognitive controller (no reaction time main
effect of familiarity, p > 0.05; no error ratio main effect of familiarity, p > 0.05; no reaction time
familiarity x position interaction: F(1,20) = 0.63, p > 0.05; no error ratio familiarity x position
interaction: F(1,20) = 3.43, p > 0.05) (See Figure 6). While there was also again no effect of
position in error ratios (no main effect of position: F (1, 20) = 0.17, p > 0.05), the reaction time
main effect presented as well (main effect of position: F(1, 20) = 53.63, p < 0.001).
Discussion
Overall, the results of Experiment 2 were remarkably consistent with those of Experiment
1, despite the addition of substantially more practice runs. There was a clear reaction time
flooring effect that was rapidly attained at practice, and the learning that was manifested in this
result clearly generalized to novel sequences at test as evidenced by the lack of significant
differences in reaction time between both practice versus test sequences and familiar versus test
sequences. Of the two predictions presented in Experiment 1, the results of Experiment 2 provide
substantial support for that which stated that participant performance improved quickly and in
such a way as to generalize learning to test, as opposed to the theory that suggested having
insufficient practice overall in the first section resulted in performance of comparably novel
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
17
sequences at test – the highly pronounced reaction time flooring effect at practice makes this
highly unlikely.
Figure 6 A-D) Reaction times and error ratios collapsed across sequence position for familiar and novel conditions
and test. Results show no main effect of familiarity. C-D) Reaction times and error ratios at first and internal
sequence positions, for familiar and novel conditions at test. Results indicate no position-specific interactions as a
function of practice.
Given the unexpected participant capacity to learn flexibly, as demonstrated by
participants in both Experiment 1 and Experiment 2, it became evident that sequence-level
learning was not sensitive to specific sequences over others. Thus, we posited that learning might
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
18
instead be sensitive to the presence of sequential structure in general, revising our operational
definition of sequential learning in the following experiments. The underlying assumption now
was not that the effects of practice would manifest themselves in a comparison of different
sequences, but instead in a comparison of sequence performance overall after one of two
conditions of practice – either sequential task practice, similar to that of the previous
experiments, or practice with an analogous task that differed only in its lack of sequential
structure. The following three studies represent the intermediate steps in our determining that
analogous task practice condition.
Experiment 3
Our first attempt to extract the sequential element from our otherwise sequential task
section involved incorporating individual trial cues, supplementary to the sequence rules that
were provided at the beginning of each block. Thus, on each trial at practice, participants were
shown a one-word cue above the image to which they were responding, which directed them
either to judge for “Color,” “Shape,” or “Size.” Then, at test, participants performed the same
sequences from practice, but without the trial cues.
Previous studies of redundant cueing indicate that the provision of the aforementioned
cues leads to dramatic decreases in reaction times during task performance, as compared to
performance of the same tasks using internally generated cues (Koch, 2003; de Jong et al., 2006;
Kleinsorge et al., 2008). From these conclusions, and also considering intuitively that cues
facilitate task performance to such an extent as to invalidate the necessity of sequence rules, we
predicted that participants would be faster overall at practice, as compared to test. We also
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
19
predicted participants would not exhibit sequence performance behavior when cues were
available to them due to the decrease in cognitive demand.
Results
The reaction time and error ratio data over all of practice compared to all of test both
indicated that task performance did not benefit from the introduction of sequence structure from
practice to test (no reaction time main effect of section, p > 0.05; no error ratio main effect of
section, p > 0.05). There was also no change in behavior at the first position from practice to test
(no reaction time section x position interaction: F(1,8) = 0.02, p > 0.05; no error ratio section x
position interaction: F(1,8) = 0.23, p > 0.05).
While there was also no positional significance in error rates between the two sections (no
main effect of position: F(1,8) = 0.14, p > 0.05), there was evidence of restart costs at both
practice and test (main effect of position: F(1,8) = 0.02, p > 0.05).
Discussion
Given our predictions about the effect of adding cues at practice, the reaction time results
of Experiment 3 were surprising. Participants were not explicitly told to conceptualize the tasks
at practice as sequences, and were given the tools so as to not rely on sequential structure – yet,
there was clear evidence that participants were still utilizing sequence performance behavior, as
indicated by the presence of restart costs at practice.
The experimental structure of the practice section was meant to be analogous to that of the
previous experimental versions, but without the sequential element. Because participants were
still clearly performing sequences, however, Experiment 3 was not appropriate as the non-
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
20
sequential condition for the later task versus sequence practice experiment. Thus, in the next
iteration of this experiment, we further stripped away the sequence element in order to prevent
unsolicited sequence performance behaviors.
Figure 7 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence test
conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and
internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main
effect of position.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
21
Experiment 4
In order to encourage participants to rely on the cues in the cued task section, Experiment
4 maintained sequential structure of tasks, but masked the sequence rules such that participants
saw only a series of X’s where, in previous experiments, they saw the words “Color”, “Shape”,
and “Size.” With this further removal of sequential structure from the practice section, our
hypotheses remained consistent with that of Experiment 3.
Results
The results of Experiment 4 were similar to those of Experiment 3 – both reaction times
and error ratios were unchanged from practice to test (no reaction time main effect of section, p >
0.05; no error ratio main effect of section, > 0.05), and there was, again, no interaction at the first
position (no reaction time section x position interaction: F(1,6) = 0.7, p > 0.05; no error ratio
section x position interaction: F(1,6) = 0.7, p > 0.05) (See Figure 8). Lastly, the main effect of
position remained robust in reaction times in Experiment 4, though not in error ratios (reaction
time main effect of position: F (1,6) = 35.8, p < 0.001; no error ratio main effect of position:
F(1,6) = 3.76, p > 0.05).
Discussion
Even with masked sequential structure and, subsequently, even less motivation for
participants to perform the practice tasks as sequences, the main effect of position in reaction
times provides clear evidence of persistent restart costs. Participants demonstrated the ability to
detect sequence structure to such a sensitive degree as to be detrimental with regards to the
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
22
overall purpose of the cued tasks. Therefore, sequential structure was entirely removed in
Experiment 5 in order to abolish these restart costs.
Figure 8 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence test
conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and
internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main
effect of position.
Experiment 5
Our next attempt at finding an appropriate practice condition represented the farthest
possible departure from sequential task performance. In Experiment 5, not only were the
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
23
sequence rules screens masked, but the order of task presentation was randomized. The
parameters of the practice section now completely disallowed the sequential performance of
tasks, and we thus maintained our hypothesis that the practice section would yield no sequence
performance behavior, as well as faster reaction times as compared to the sequence test section.
Results
Consistent with the results of Experiments 3 and 4, the absence and later presence of
sequential structure had no effect on reaction times or error rates (no reaction time main effect of
section: p > 0.05; no error ratio main effect of section: p > 0.05) (See Figure 9). There was,
again, a significant difference in reaction times between first and internal positions (main effect
of position: F(1,6) = 14.43, p < 0.01), which did not extend to error ratios (no main effect of
position, F(1,6) = 2.4, p > 0.05). However, there was also a significant section-by-position
interaction (section x position interaction: F(1,6) = 7.99, p < 0.05), which was an effect not
previously produced in the earlier versions of the task practice, sequence test experiment. This
effect was not manifest in error ratios (no section x position interaction: F(1,6) = 0.63, p > 0.05).
Discussion
The results of Experiment 5, specifically the section-by-position interaction, suggest that
the random tasks constitute an appropriately analogous condition to sequential tasks without
prompting sequential behavior. However, the remaining concern with this experimental design
was the non-significant but visually evident trend in the random task reaction times to be slower
than the sequence reaction times. This was striking, considering the ample body of evidence
supporting the opposite prediction.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
24
Figure 9 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence
test conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and
internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main
effect of position, as well as a section x position interaction.
However, as previously suggested in a reaction time study by Sohn and Anderson (2001),
this effect was likely a function of a lack of preparation time between the random trials that was
present between the sequential trials. As indicated in their study, ‘foreknowledge’ of an
upcoming task is an essential aspect of sequence performance, as sequential order allows one to
prepare for each subsequent image prior to image presentation (Sohn & Anderson, 2001).
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
25
Because the Experiment 5 paradigm did not allow for the process of preparation prior to trial
presentation, instead forcing participants to initiate preparation strategies at the same time they
were expected to respond, overall reaction times in the practice section were inflated. Therefore,
while Experiment 6 employs the general paradigm from Experiment 5, the timing of cue and
stimulus presentation was altered to allow for response preparation.
Experiment 6
In order to address the final structural issue, as indicated in Experiment 5, the cued
random practice condition now consisted of a cue being shown immediately before image
presentation within the previous trial’s inter-trial interval. This manipulation allowed for
response preparation that was not present in the previous three experiments.
Because we had finally determined the analogous task condition for our between subjects
comparison, Experiment 7 also reincorporated a sequence practice section, and participants were,
therefore, assigned to one of the two practice conditions. The sequence practice condition was a
hybrid, in structure, of the previous experiments. Like the random condition it consisted of 4
practice runs and 3 test runs, and like Experiments 1 and 2, the practice sections consisted of
repetitions of practice of the same two sequences.
Like Experiments 3, 4, and 5, the test section consisted of repetitions of performance of
two unpracticed sequences, as opposed to multiple different novel sequences. However, for the
sake of increasing statistical power, the practice and novel sequence reaction time results of
Experiments 1 and 2 have been added to the sequence condition analysis. Experiments 3, 4 and 5
have been excluded due to the aforementioned flaws in the experimental design.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
26
Our hypotheses for Experiment 6 consisted of the same behavioral predictions at test as
those predicted in Experiments 1 and 2, but based on different experimental parameters. Both
overall decrease in reaction times and the interaction at the first position were predicted to be
functions of sequence practice as opposed to cued, random task practice.
Results
The behavioral results at test, comparing sequence performance after either sequence
practice (sSEQ) or random, cued task practice (rSEQ), indicated no difference in performance as
a function of practice group (no reaction time main effect of group: F(1,64) = 0.15, p > 0.05; no
error ratio main effect of group: F(1,64) = 0.0005, p > 0.05) (See Figure 10). There was also no
evidence of a selective effect on the cognitive controller (no reaction time group x position
interaction: F (1, 64) = 0.13, p > 0.05; no error ratio group x position interaction: F (1, 64) = 1.73,
p > 0.19). The evidence of restart costs across both groups was evidenced in the significant
difference in reaction times and marginal difference in error ratios between the first position and
the internal sequence positions (reaction time main effect of position: F(1,64) = 108.9, p < 0.001;
error ratio marginal effect of position: F(1, 64) = 3.5, p = 0.067).
Discussion
As was the case with Experiments 1 and 2, the results of Experiment 6 were surprising,
given the intuitive sense of the hypotheses detailed herein. The marginally significant position
effect on error ratios followed the same trend as reaction times, suggesting overall that the first
position judgment in each sequence carried the highest cognitive demand. This is consistent with
the body of research involving sequence performance, specifically referring to the robust restart
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
27
costs at the first position, and is a good check that, despite the experimental manipulations of
sequence performance context in Experimental 6, participants are still demonstrating sequence
performance behavior.
Figure 10 A-B) Reaction times and error ratios collapsed across sequence position for rSEQ and sSEQ conditions.
Results show no main effect of type of practice. C-D) Reaction times and error ratios at first and internal sequence
positions, for rSEQ and sSEQ. Reaction time and error rate results indicate a main effect of position.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
28
Nonetheless, the otherwise null results of this version prompted one final experimental
iteration, in which we made minor changes to task administration with consideration to
laboratory best practices.
Experiment 7
Experiment 7 was a continuation of Experiment 6. However, starting with this version of
the study, we began staying in the room with participants for the duration of the task in order to
prompt improved task performance. This change in protocol had the effect of improving
participant performance extensively, both in terms of reaction times and error ratios. It was
decided that the analysis for Experiment 7 would be analyzed independently of all previous
experimental versions in order to only be representative of optimal data.
Also in accordance with seeking the best possible data-collection conditions, and based
off of the results and trends from Experiment 6, four out of the ten sequence structures were
selected as optimal and were the only structures from which sequences were drawn in
Experiment 7.
Results
In order to ensure that the learning seen in Experiments 1 and 2 was also evident under
the optimal conditions of Experiment 7, reaction times over runs at practice and test were again
analyzed for evidence of a learning curve (See Figure 11). Learning was indicated in both
practice groups (sequence practice main effect of run: F(3,27) = 5.36, p < 0.01; random practice
main effect of run: F(3, 27) = 4.21, p < 0.01). T-tests indicated that sequence learning at practice
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
29
transferred to test (no main effect of test section, p > 0.05), but no significant learning transferred
from the random, cued task practice to test (main effect of test section, p < 0.05).
Figure 11 learning curves indicated by reaction times for A) Practice sequence runs, B) sSEQ runs, C) practice
random, cued tasks, and D) rSEQ runs. Curves suggest that participants in both random and sequence groups were
learning, getting faster through practice and maintaining performance levels at test.
C.
D.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
30
At test, neither reaction times nor error ratios demonstrated any overall sensitivity to
group effects (no reaction time main effect of group, p > 0.05; no error ratio main effect of group,
p > 0.05) (See Figure 12). Position-specific interactions were also not observed (no reaction
time group x position interaction: F(1,15) = 0.46, p > 0.05; no error ratio group x position
interaction: F(1,15) = 1.26, p > 0.05). While there was also no effect of position on error ratios
(no main effect of position: F(1,15) = 0.73, p > 0.05), reaction times were sensitive to positional
differences (main effect of position: F(1,15) = 50.97, p < 0.001).
Discussion
Staying in the room with participants as they performed the task had the effect of
dramatically improving overall performance. However there was still no effect of practice group,
indicating that sequence performance is not sensitive to the type of practice that preempts it. The
presence of sequence structure at practice did not facilitate sequence performance at test over
random task practice. This conclusion extends to the lack of group-by-position interaction, as
made evident specifically by the result at the first position reaction time compared to internal
position times. It is clear that practice with utilizing the cognitive controller does not facilitate its
later use – it can be employed after any kind of previous practice and perform comparably.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
31
Figure 12 A-B) Reaction times and error ratios collapsed across sequence position for rSEQ and sSEQ conditions.
Results show no main effect of type of practice. C-D) Reaction times and error ratios at first and internal sequence
positions, for rSEQ and sSEQ. Reaction time results indicate a main effect of position.
General Discussion
Across the seven experimental designs and aims, participants proved to be highly efficient
at learning task performance and generalizing the task procedure. Sequence performance was
highly flexible and adaptable to altered task set orders. These experiments indicate that sequence
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
32
learning happens so rapidly, and that sequence learning is so well generalize at test, that reaction
times of both practice and familiar sequences are comparable to novel sequences at test.
Aside from sequence learning being highly well-generalized, Sequence-detection is
highly sensitive. Experiments 3 and 4 illustrated this phenomenon – even when given cues,
participant performance indicated sequence performance, as there was evidence of restart costs in
both the sequence test section, where they were expected, and the cued task practice section,
where they were both unexpected and unsolicited.
The results of Experiments 6 and 7 support and extend the results of Experiments 1 and 2
in their indication of the lack of sensitivity of sequence performance and practice. As the first
two experiments allowed for the assertion that sequential behaviors are not effected on the level
of practice with specific sequences, the final two experiments indicated that these behaviors are
not sensitive the sequential structure at all.
While it is possible that sequential learning in the previous experiment was, in itself, an
inconsequential aspect, it must also be considered that unaccounted for elements may have added
variability to the data. One concern involves the implications of the five-item sequence structure
on individual performance, specifically with regards to chunking.
A previous study by Schneider and Logan (2006) demonstrated evidence of chunking –
the process of dividing longer sequences of items into smaller, more manageable subsequences –
in six-item sequences such that internal reaction times were comparable to their first position
analogs (Schneider & Logan, 2006). Given that little research has been done regarding five-item
sequences, variance in current data may be attributed, in part, to unexpected chunking methods.
It is much more conceivable with five-item sequences, as opposed to six-item sequences, that
different participants could employ different chunking methods (chunks of ‘three’ and ‘three’
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
33
would be the primary prediction in sequences of six, but chunks of ‘two’ and ‘three’ or ‘three’
and ‘two’ could be equally likely in sequences of five), and averaging over these different
methods could affect the group results. Thus, future data collection must seek to minimize this
effect in order to strengthen the group effect of sequence versus random task practice.
With regards to moving forward, we would like to run more participants and raise our
statistical power. In looking at the behavioral results on performance of individual sequences
(See Appendix A), power is still very low per sequential structure. While there were no
significant sequential results at the group level, which averaged over multiple different sequential
structures and trial type frequencies, it might be that these results will reveal themselves in
performance of specific sequences. Running more participants in each of the current four optimal
structures will allow us to illuminate this lingering question.
If we discover evidence of sequence learning, next steps will hopefully involve functional
experimentation on practice effects. A previous fMRI experiment conducted by Badre et al.
(2013) discovered DLPFC activity that followed a ramping pattern during the performance of
each sequence within each block. This pattern restarted at the first position of each sequence and
was consistent across both simple and complex sequences (Badre et al., 2013). One possible
explanation for this result involves the activity being a compensatory reaction to increased
response uncertainty, which occurs as one progresses through each position in a sequence. While
certainty in judgment accuracy decreases with each successive position within a sequence, it is
restored at the first position judgment of the next sequence where, statistically, there is less room
for error. The neural activity in the DLPFC appears to mirror this effect.
Given these results and conclusions as a backdrop for the current experiment, there are
two main predictions for the results of this study in fMRI. Predicting an effect of sequence
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
34
practice, as compared to task practice, it can be inferred that this form of practice may have the
effect of increasing certainty – thus, in fMRI, the pattern of activation would likely follow that of
the aforementioned sequence experiment, but extending to the fifth position and with a decrease
in compensatory DLPFC activity overall. However, if further experimentation reveals an
interaction of practice group and position, localized at the first position, then activation in fMRI
might indicate not only an increase in overall certainty, but a supplementary increase in certainty
at the first position as a function of practice with cognitive control on the sequence level.
However, given that chunking may be occurring within these five-item sequences, it is
also possible that the fMRI results will be distorted to reflect that. If participants are segmenting
the sequences into shorter subsequences, these chunks would have their own restart costs and
would likely exhibit sequence-like qualities. Thus, in fMRI, the original ramping pattern may be
interrupted by sudden decreases in activity, which would indicate an increase in confidence at the
boundaries of the subsequences.
Nonetheless, an fMRI experiment based on the current study is a long way away, given
the results at hand. We first must illuminate whether the thus-far evasive effect of sequential
practice is possible to distill from other practice elements. In the grand scheme of sequential
learning and performance, regardless of which prediction reveals itself to be valid (if any at do so
at all), sequence practice is clearly an area that must be explored further. This study merely
scratches the surface of how the brain responds to different conditions of sequence performance,
and given the sheer volume of null results under so many various practice and test conditions, it
is clear that there is more work to be done in this area.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
35
Methods
Experiment 1
Participants
28 participants (21 females, 6 males, average age 20) participated in Experiment 1. One
participant was excluded after leaving half-way through the experiment, and one was excluded
due to a coding error in initiating the experiment, which resulted in fatally flawed data collection.
Participants were recruited from the Brown University Student body, as well as from the
Providence community, for paid participation at a rate of $10/hour.
Apparatus
The experiment was conducted on a Macintosh computer in the Badre Lab at Brown
University, running PsychToolBox in Matlab. Input was registered through standard keyboard
press responses, and output was displayed on the computer screen. Trial images were
constructed from a factorial combination of 3 dimensions: color (red or blue), shape (circle or
square), and size (small or large). The three tasks consisted of color, shape, and size judgments.
The post-test questionnaire was created using the Qualtrics Research Suite.
Procedure
Participants performed the experiment in private testing rooms after signing the proper
consent forms. Participants first performed two practice sections in which they had unlimited
response time. Participants were told to place their fingers on the “J”, “K”, “L”, “;” and “ ’ ”
keys, and told that they would be using their first two fingers to make the majority of their
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
36
responses. In the first section, participants performed three practice blocks, with one dimension
rule (“Color”, “Shape”, or “Size”) presented per block, and one judgment task to perform. The
order of rule presentation was counterbalanced across participants. Each trial screen consisted of
an image of a particular color, shape, and size, as well as key press reminders on the bottom left
and right corners of the screen that indicated the current judgment-relevant response choices and
the key to which they corresponded. A white fixation cross was shown during the inter-trial
interval for a variable amount of time (500-1000ms).
During the second section, participants performed a practice block which began with a
screen indicating the five-item, task-relevant sequence. Participants were told to pay attention to
the sequence and remember it throughout the block, as it would not be shown again until the end
of the block during the sequence position question. They were also told that the key press
reminders currently visible on the wait screen would remain onscreen and would be the same
throughout the entire experiment. The reminders now consisted of a conglomeration of the three
response choices shown individually in the first section. Lastly, participants were told that the
sequence could end on any position in the block, and that, at the end of the block, they would be
shown the sequence position question in which they would indicate which item in the sequence
they would have next performed if the block had continued, by pressing the “1(J)”, “2(K)”,
“3(L)”, 4(;)” or “5(’)” keys.
Participants then performed one abbreviated block and answered the position question,
after which the experimenter checked their performance and brought to their attention any
mistakes made. The participants then performed another practice sequence.
Following these practice sequences, participants performed the third section, which
consisted of seven runs of six blocks of repetitions of two more practice sequences – one novel
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
37
and one complex. For this and the fourth section, sequences were drawn from a pool of 50
sequences structures. Each block contained 24 to 28 trials. A wait screen was shown at the
beginning of each run instructing participants to press the spacebar to continue. Doing so
prompted the “Get Ready’ screen which instructed the participants to get ready to begin, after
which the next run would begin. This constituted a break between runs.
Prior to the start of this section, the experimenter told participants that they now had four
seconds to indicate their answer when each image were shown and five seconds to answer the
sequence-position question at the end of each block. The experimenter stressed the importance of
paying attention to the sequence rules when they were shown at the beginning of each block and
responding as soon as the participant knew the answer, as opposed to second guessing or
otherwise delaying their response. They also explained to the participants that, if at any point
they forgot where they were in the sequence, they were to pick a place in the sequence and go
from there, as opposed to making random guesses. The experimenter then left the room, telling
the participant to get their fingers ready and press the spacebar to continue as soon as the door
closed behind them.
At the end of the third section, participants left the room and got the experimenter, who
checked their performance and recorded error codes displayed at the bottom right-hand side of
the screen into the lab notebook. Participants were then instructed to perform the fourth section,
in which each block consisted of presentations of the previously practiced, now familiar
sequences as well as multiple different novel sequences. The novel sequences were
counterbalanced for sequence complexity. The fourth section was of the same duration as the
third section.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
38
The task automatically closed after the participant answered the last sequence position
question in the last block of the last run. After this, the experimenter pulled up the post-test
questionnaire (PTQ), which participants completed at their own pace. Depending on how long
participants took with the task and the availability of the testing room in use, they would
complete the PTQ either on the test computer, on a lab computer, or on the experimenter’s laptop.
After completing the PTQ, participants signed a receipt form and were compensated for
their time.
Experiment 2
Participants
11 participants (9 females, 2 males, average age 22) participated in Experiment 2.
Participants were recruited from the Brown University Student body, as well as from the
Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of Experiment 1.
Procedure
The duration of Experiment 2, specifically the third section, was the only difference
between it and Experiment 1. Experiment 2 spanned two consecutive days, and the sessions for
each participant were scheduled with no less than twenty-two and no more than twenty-six hours
apart. On the first day, participants performed the first two practice sections, as well as two,
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
39
seven-run sections of practice of two sequences with a break in-between. On the second day, they
performed four more runs of the two practice sequences, followed by the seven-run test section
and a revised PTQ. While there were supposed to be breaks in-between the runs at test,
participants later informed the experimenter that there were no wait screens and that the runs
were continuous.
Experiment 3
Participants
5 participants (3 females, 2 males, average age 21) participated in Experiment 3.
Participants were recruited from the Brown University Student body, as well as from the
Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of previous experiments.
Procedure
The first practice section was the same as that of the previous two sections. During the
first practice block in the second section, however, participants were now given cues, displayed
above the image during each trial, in addition to the five-item sequence rules that were shown at
the beginning of the block. The cues consisted the one-word judgment rule relevant to the
current trial. They answer to the sequence-position question at the end of the block was provided
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
40
in parentheses on the screen below the question. The second practice sequence block was similar
to that of previous experiments, but with an “XXXX” above each image in place of the cue.
The third section was four runs of six blocks, and consisted of two cued sequences. The
fourth run was three runs of the same sequences, minus the cues. Participants performed the
revised PTQ following the fourth section.
Experiment 4
Participants
4 participants (3 females, 1 male, average age 21) participated in Experiment 4.
Participants were recruited from the Brown University Student body, as well as from the
Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of previous experiments.
Procedure
The experimental procedure was almost identical to that of Experiment 3. The sole
difference was in the sequence rule screen shown at the beginning of each block during the first
run in the second section and in the third section. Instead of displaying a sequence of rules,
participants were now shown five “XXXX”’s masking the sequence. The experimenter also took
care not to use the word “sequence” when explaining giving instructions for the masked blocks.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
41
Experiment 5
Participants
4 participants (3 females, 1 male, average age 21) participated in Experiment 5.
Participants were recruited from the Brown University Student body, as well as from the
Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of previous experiments.
Procedure
Visually, Experiment 5 was exactly the same as Experiment 4. Structurally, however, the
trials were randomized during the first practice block of the second section and in the third
section.
Experiment 6
Participants
35 participants (25 females, 10 males, average age 21) participated in Experiment 6. 15
were placed in the cued random task practice condition, and 20 were placed in the sequence
practice condition. Participants were recruited from the Brown University Student body, as well
as from the Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of previous experiments.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
42
Procedure
Participants in the cued random task practice group performed a task similar to that of
Experiment 5. The only difference was in the presentation of the cues. To allow for preparation
costs, the cues for each successive trial were presented immediately after the participant
responded to the previous trial, as opposed to being shown simultaneously with the trial image.
The cues were shown for x milliseconds. At test, the cue screen was masked with “XXXX”.
Participants in the sequence practice group performed repetitions of two sequences at
practice, and repetitions of two difference sequences at test, with the same screens as those used
in the cued-group test section. In both sections, the sequences consisted of one simple structure
and one complex structure. Sequences were pooled from a subset of the pool of sequences used
in previous experiments – those with the structure IDs 1-4, 7-9, 12, 13, and 16 were used.
Experiment 7
Participants
15 participants (8 females, 7 males, average age 21) participated in Experiment 7. Seven
were placed in the cued random task practice condition, and 8 were placed in the sequence
practice condition. Participants were recruited from the Brown University Student body, as well
as from the Providence community, for paid participation at a rate of $10/hour.
Apparatus
The apparatus were identical to that of previous experiments.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
43
Procedure
For both groups, the structure of the experiment was similar to that of Experiment 6. In
Experiment 7, however, there were four runs in both section three and four. Based on the results
of Experiment 6, sequences with the structure ID 1 and 2, complex, and 9, and 13, simple, were
selected as optimal sequence structures and were the structures from which sequences were
drawn for Experiment 7. Instead of practicing two sequences at practice and test, participants
now performed a total of six sequences at practice and six at test, with three of each of two
sequence structures – one simple, and one complex. For example, one of the two sequence
structures a participant may have performed at test would have been “AABBC” – thus, three
sequences they could have possibly performed could be “ Color Color Shape Shape Size,”
“Shape Shape Color Color Size,” and “Size Size Color Color Shape.” In the sequence practice
group, participants had exposure to all four sequence structures at some point during the
experiment.
Starting with Experiment 7, the experimenter stayed in the room with the participant,
instead of leaving the room before the beginning of section three.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
44
References
Allport, A., Styles, E. A., & Hsieh, S. (1994). Shifting Intentional Set: Exploring the Dynamic
Control of Tasks.
Arbuthnott, K., & Frank, J. (2000). Executive control in set switching: residual switch cost and
task-set inhibition. Canadian Journal of Experimental Psychology = Revue Canadienne de
Psychologie Experimentale, 54, 33–41. doi:10.1037/h0087328
Badre, D., Hoffman, J., Cooney, J. W., & Esposito, M. D. (2010). NIH Public Access, 12(4),
515–522. doi:10.1038/nn.2277.Hierarchical
Barceló, F., Periáñez, J. A., & Nyhus, E. (2007). An information theoretical approach to task-
switching: evidence from cognitive brain potentials in humans. Frontiers in Human
Neuroscience, 1.
Desrochers, T. M., & Badre, D. (2013, January). Frontal systems supporting the hierarchical
control of task sequences. In Journal of Cognitive Neuroscience (pp. 41-41). 55 Hayward
Street, Cambridge, MA 02142 USA: MIT press.
Fagot, C. (1994). Chronometric investigations of task switching. University of California, San
Diego. Retrieved from
http://scholar.google.com/scholar?hl=ca&q=Chronometric+investigations+of+task+switchin
g&btnG=Cerca#0
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
45
Fezzani, K., Thon, B., & Lagarde, J. (2000). Can different levels of S-R practice influence
sequence learning? An investigation into the4 context of a perceptual motor sequence
learning. Perceptual and Motor Skills, 91, 463–475.
de Jong, Ritske, Thomas E. Gladwin, and Bernard M't Hart. "Movement-related EEG indices of
Kleinsorge, Thomas, and Patrick D. Gajewski. "Task switching based on externally presented
versus internally generated information." Psychological research 72.5 (2008): 501-
514.preparation in task switching and motor control." Brain research 1105.1 (2006): 73-82.
Koch, I. (2003). The role of external cues for endogenous advance reconfiguration in task
switching. Psychonomic Bulletin & Review, 10(2), 488–492. doi:10.3758/BF03196511
Kramer, a F., Hahn, S., & Gopher, D. (1999). Task coordination and aging: explorations of
executive control processes in the task switching paradigm. Acta Psychologica, 101, 339–
378. doi:10.1016/S0001-6918(99)00011-6
Meiran, N. (1996). Reconfiguration of processing mode prior to task performance. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 22(6), 1423–1442.
doi:10.1037/0278-7393.22.6.1423
Meiran, N., Chorev, Z., & Sapir, a. (2000). Component processes in task switching. Cognitive
Psychology, 41, 211–253. doi:10.1006/cogp.2000.0736
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
46
Barceló, F., Periáñez, J. A., & Nyhus, E. (2007). An information theoretical approach to task-
switching: evidence from cognitive brain potentials in humans. Frontiers in Human
Neuroscience, 1.
Rogers, R. D., & Monsell, S. (1995). Costs of a Predictable Switch Between Simple Cognitive
Tasks. Journal of Experimental Psychology: General, 124(2), 207–231.
Ruthruff, E., Johnston, J. C., & Selst, M. Van. (2001). Why Practice Reduces Dual-Task
Interference. Journal of Experimental Psychology: Human Perception and Performance,
27(1), 3–21.
Schneider, D. W., & Logan, G. D. (2006a). Hierarchical control of cognitive processes: switching
tasks in sequences. Journal of Experimental Psychology. General, 135Schneid(4), 623–640.
doi:10.1037/0096-3445.135.4.623
Schneider, D. W., & Logan, G. D. (2006b). Priming cue encoding by manipulating transition
frequency in explicitly cued task switching. Psychonomic Bulletin & Review, 13(1), 145–
151. doi:10.3758/BF03193826
Sohn, M.-H., & Anderson, J. R. (2001). Task Preparation and Task Repitition: Two-Component
Model of Task Switching. Journal of Experimental Psychology: General, 130(4), 764–778.
Stoet, G., & Snyder, L. H. (2003). Executive control and task-switching in monkeys.
Neuropsychologia, 41, 1357–1364. doi:10.1016/S0028-3932(03)00048-4
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
47
Willingham, D. B., Wells, L. A., Farrell, J. M, Stemwedel, M. E. (2000). Implicit motor
sequence learning is represented in response locations. Memory & Cognition, 28(3), 366-
375. doi: 10.3758/BF03198552.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
48
APPENDIX A – STATISTICS ON INDIVIDUAL SEQUENCE STRUCTURES
Below are the relevant statistics on reaction times and error ratios for individual
sequences in the “novel vs familiar” Experiments 1 and 2, and the optimized “rSEQ vs sSEQ”
Experiment 7. There are several sequential structures without data from one of two experimental
conditions. Those structures have been omitted from the following analyses.
Figure 13 Reaction times at test for familiar versus novel A-D) complex sequence structures 1-4, and E-F) simple
sequence structures 7-9, 12, 13, and 16.
 Structure 13 main effect of familiarity: F(1,25) = 6.02, p < 0.05
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
49
Figure 14 Error ratios at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F) simple
sequence structures 7-9, 12, 13, and 16.
 Structure 7, 16 main effect of familiarity: F’s > 4.8, p’s < 0.05
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
50
Figure 15 Reaction times at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F)
simple sequence structures 7-9, 12, 13, and 16.
 Sturcture 16 main ffect of familiarity: F(1,9) = 5.67, p < 0.05
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
51
Figure 16 Error ratios at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F) simple
sequence structures 7-9, 12, 13, and 16.
 no main effect of familiarity: F’s < 0.9, p’s > 0.34
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
52
Figure 17 Reaction times at test for rSEQ versus sSEQ A-B) complex sequence structures 1, 2, and C-D) simple
sequence structures 9 and 13.
 no main effect of practice group: F’s < 0.79, p’s > 0.399
Figure 18 Error ratios at test for rSEQ versus sSEQ A-B) complex sequence structures 1, 2, and C-D) simple
sequence structures 9 and 13.
 no main effect of group: F’s < 0.79, p’s > 0.399
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
53
APPENDIX B – EVIDENCE OF TASK SYMMETRY
Figure 19 Average reaction times for all A) repeats and B) switches and error ratios for all C) repeats and D) switches at random
practice in Experiment 6.
In order to ensure that task asymmetry was not a factor in the behavioral results of this
experiment, reaction times and error ratios for every combination of switches and repeats were
analyzed for possible significant differences in magnitude. The results for this section were taken
specifically from the random task condition of Experiment 7. Reaction times across all switches
and repeats were fairly consistent (no main effect of trial type, F’s < 2.05, p’s < 0.13). There was
also no significant difference in error ratios in the switch condition (no main effect of trial type:
F(4,120) = 0.97, p = 0.44). However, the error ratios for the “Shape-Shape” repeat trial type
were shown to be significantly higher than those of the other two trial types. Though further
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
54
analysis will be needed in order to determine what this result means for the error ratios
throughout the entirety of the experiment, the integrity of the reaction times, the primary measure
in this study, appears to have been preserved.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
55
APPENDIX C – TASK INSTRUCTIONS FOR ALL EXPERIMENTS
Experiment 1 Task Instructions
PRE-TRAIN
Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings.
-----------------
First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will
have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the
entire length of this experiment.
Task Training
Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the
majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for
[dimension]. Please always keep your fingers on the keyboard.
You will now do a series of practice trials. An image will briefly be displayed and you will respond
according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to
respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will
briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1
trials]
Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any
questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials]
Sequence Example Training
Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen
at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR.
For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about
[SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE,
SIZE, COLOR, COLOR, SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the
screen.
We will now go through a practice block that you can complete at your own pace. You will first see a sequence of
rules for 4 seconds, then a white “+” on a get ready screen for 1 second, and then the images will be displayed as
before. You will not see the sequence of rules again, so you must remember it. At the end of the block, you will see a
question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st
- COLOR,
2nd
- SHAPE, 3rd
- SHAPE, 4th
– SIZE, 5th
- COLOR) by pressing the 1st
, 2nd
, 3rd
, 4th
, or 5th
key (JKL;” – emphasize
the “). The block can end on any position in the sequence, so it is important to remember where in the sequence you
are at all times. After this question you will see a white “+” on the screen for a short amount of time.
For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as
possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of
the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to
explain their reasoning on the first few trials]
Great. Was the question about what item you would perform next at the end clear? Do you have any questions?
[Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second
practice sequence will then be presented.]
Sequence Practice
Now you will practice 2 more sequences of rules like the example you just completed using the same key
presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please
respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready
to respond, and respond as soon as you know the answer. You will not receive any additional reminders as to the
sequence of rules once the blocks begin.
The beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like
before. After each block, you will again be asked to report what element of the sequence you would next perform,
with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
56
in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block
will begin with a new 5-rule sequence displayed on screen.
If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered
and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. At the end of
each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you
have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials]
After practice
 Do you have any questions about those practice blocks? [Check performance.]
 You will next be asked to complete 7 runs that have 6 blocks each.
 After this there will be a short questionnaire.
 Any questions before you begin?
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
57
Experiment 2 Task Instructions
PRE-TRAIN – DAY 1
Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings.
-----------------
First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will
have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the
entire length of this experiment.
Task Training
Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the
majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for
[dimension]. Please always keep your fingers on the keyboard.
You will now do a series of practice trials. An image will briefly be displayed and you will respond
according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to
respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will
briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1
trials]
Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any
questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials]
Sequence Example Training
Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen
at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR.
For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about
[SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE,
SIZE, COLOR, COLOR, SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the
screen.
We will now go through a practice block that you can complete at your own pace. You will first see a sequence of
rules for 4 seconds, then a white “+” on a get ready screen for 1 second, and then the images will be displayed as
before. You will not see the sequence of rules again, so you must remember it. At the end of the block, you will see a
question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st
- COLOR,
2nd
- SHAPE, 3rd
- SHAPE, 4th
– SIZE, 5th
- COLOR) by pressing the 1st
, 2nd
, 3rd
, 4th
, or 5th
key (JKL;” – emphasize
the “). The block can end on any position in the sequence, so it is important to remember where in the sequence you
are at all times. After this question you will see a white “+” on the screen for a short amount of time.
For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as
possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of
the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to
explain their reasoning on the first few trials]
Great. Was the question about what item you would perform next at the end clear? Do you have any questions?
[Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second
practice sequence will then be presented.]
Sequence Practice
Now you will practice 2 more sequences of rules like the example you just completed using the same key
presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please
respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready
to respond, and respond as soon as you know the answer. You will not receive any additional reminders as to the
sequence of rules once the blocks begin.
The beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like
before. After each block, you will again be asked to report what element of the sequence you would next perform,
with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are
in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block
will begin with a new 5-rule sequence displayed on screen.
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
58
If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered
and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. At the end of
each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you
have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials]
After practice
 Do you have any questions about those practice blocks? [Check performance.]
 You will next be asked to complete 7 more runs of the same two sequences that have 6 blocks each.
 After this there will be a short questionnaire.
 Any questions before you begin?
PRE-TRAIN – DAY 2
Sequence Practice
First, you will be asked to practice the same 2 sequences of rules that you practiced yesterday, using the
same key presses, which will not change. You will have 4 seconds to indicate your answer. Please respond both as
accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and
respond as soon as you know the answer. You will not receive any additional reminders as to the sequence of rules
once the blocks begin.
Just a reminder, at the beginning of the block you will see the words “Get ready to begin,” then the
sequence of rules like before. After each block, you will again be asked to report what element of the sequence you
would next perform, and you will have 5 seconds to answer, so it is important that you know where you are in the
sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block will
begin with a new 5-rule sequence displayed on screen.
If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered
and don’t give up. You will be asked to complete 4 runs that are composed of 6 sequence blocks each. At the end of
each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you
have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials]
After practice
 Do you have any questions about those practice blocks? [Check performance.]
 You will next be asked to complete 7 more runs of the same two sequences that have 6 blocks each.
 After this there will be a short questionnaire.
 Any questions before you begin?
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
59
Experiment 3 Task Instructions
PRE-TRAIN
Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings.
-----------------
First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will
have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the
entire length of this experiment.
Task Training
Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the
majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for
[dimension]. Please always keep your fingers on the keyboard.
You will now do a series of practice trials. An image will briefly be displayed and you will respond
according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to
respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will
briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1
trials]
Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any
questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials]
Sequence Example Training
Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen
at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR.
For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about
[SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE,
SIZE, COLOR, COLOR; SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the
screen. For each trial, the current judgment will be displayed in the middle of the screen, above the image (provide
example).
We will now go through a practice block that you can complete at your own pace. You will first see a sequence of
rules for 4 seconds, then a white “+” on a get ready screen for a short time, and then the images will be displayed as
before with the current judgment above each image. At the end of the block, you will see a question on the screen
asking which item in the sequence as a whole you would NEXT perform (i.e. 1st
- COLOR, 2nd
- SHAPE, 3rd
-
SHAPE, 4th
– SIZE, 5th
- COLOR) by pressing the 1st
, 2nd
, 3rd
, 4th
, or 5th
key (JKL;” – emphasize the “). Below this
question is a cue, which will instruct you as to which answer to select. The block can end on any position in the
sequence. After this question you will see a white “+” on the screen for a short amount of time.
For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as
possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of
the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to
explain their reasoning on the first few trials]
Great. Was the question about what item you would perform next at the end clear? Do you have any questions?
[Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second
practice sequence will then be presented.]
Now you will be asked to perform another sequence of five rules. The difference now is that you will not receive
cues during the trials that tell you which judgment to make – you must do it from memory of the sequence rules
shown at the beginning of the block. The sequence will not be shown at any other time, so you must remember it.
You also will not be provided with the answer to the question at the end of the block, so it is important to remember
where in the sequence you are at all times. Press the space bar when you’re ready to begin.
Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the
last question. A second practice sequence will then be presented.]
Sequence Practice
Now you will practice 2 more sequences of rules like the example you just completed using the same key
presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
60
respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready
to respond, and respond as soon as you know the answer.
At the beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like
before. After each block, you will again be asked to report what element of the sequence you would next perform,
with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are
in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block
will begin with a new 5-rule sequence displayed on screen.
If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered
and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. For the first 4
runs, you will be given the cues for which judgment to make on each trial, as well as the answer to the question at the
end of each block. For the last three runs, you will not. Before the 5th
run you will see the same instruction on the
screen as in the practice that you must remember each sequence. If you have any questions at that point, please come
get me. At the end of each run, you will have the opportunity to take a break. At the end of this practice section,
please get me. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice
trials]
After practice
 Do you have any questions about those practice blocks? [Check performance.]
 You will next be asked to complete 3 more runs of 6 blocks each, this time without the cues on each trial
and without the answer at the end of the block.
 After this there will be a short questionnaire.
 Any questions before you begin?
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
61
Experiment 4/Experiment 5 Task Instructions
PRE-TRAIN
Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings.
-----------------
First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will
have a particular color, shape and size. You will be making judgments about these images based on these attributes
throughout the experiment.
Task Training
Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the
majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for
[dimension]. Please always keep your fingers on the keyboard.
You will now do a series of practice trials. An image will briefly be displayed and you will respond
according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to
respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will
briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1
trials]
Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any
questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials]
Sequence Example Training
Now you will be asked to perform judgments on a series of images based one the cues provided on each trial. The
key presses remain the same and will remain on the screen. For each trial, the current judgment will be displayed in
the middle of the screen, above the image (provide example).
We will now go through a practice block that you can complete at your own pace. You will first see a screen with
five X’s on it, sequence of rules for 4 seconds, then a white “+” on a get ready screen for a short time, and then the
images will be displayed as before with the current judgment above each image. At the end of the block, you will see
a question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st
- COLOR,
2nd
- SHAPE, 3rd
- SHAPE, 4th
– SIZE, 5th
- COLOR) by pressing the 1st
, 2nd
, 3rd
, 4th
, or 5th
key (JKL;” – emphasize
the “). Below this question is a cue, which will instruct you as to which answer to select. The block can end on any
position in the sequence. After this question you will see a white “+” on the screen for a short amount of time.
For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as
possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of
the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to
explain their reasoning on the first few trials]
Great. Was the question about what item you would perform next at the end clear? Do you have any questions?
[Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second
practice sequence will then be presented.]
Now you will be asked to perform a series of judgments based on a sequence of five rules that will be displayed on
the screen at the beginning of the block. Instead of the five X’s, you could be shown, for example, the words
COLOR, SHAPE, SHAPE, SIZE, COLOR. For the first image, you will make a decision about [COLOR], for the
second image, you will make a decision about [SHAPE] and so on. You will repeat this sequence until the block
ends, for example COLOR, SHAPE, SHAPE, SIZE, COLOR, COLOR; SHAPE, SHAPE, SIZE, COLOR. The key
presses remain the same and will remain on the screen. You will not receive cues during the trials that tell you which
judgment to make – you must do it from memory of the sequence rules shown at the beginning of the block. The
sequence will not be shown at any other time, so you must remember it. You also will not be provided with the
answer to the question at the end of the block, so it is important to remember where in the sequence you are at all
times. Press the space bar when you’re ready to begin.
Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the
last question. A second practice sequence will then be presented.]
Sequence Practice
DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER
SEQUENTIAL OR RANDOM TASK PRACTICE
62
Now you will be asked to complete 4 runs of 6 blocks of trials like the first pair of examples you just
completed, using the same key presses, which will not change. The only difference is that you will have 4 seconds to
indicate your answer. Please respond both as accurately and quickly as possible. What I am interested in is your
reaction times, so please be ready to respond, and respond as soon as you know the answer. For these runs, you will
be given the cues for which judgment to make on each trial, as well as the answer to the question at the end of each
block. After this you will complete 3 runs like the last pair of examples you completed – with a sequence of rules at
the beginning and without the cues.
At the beginning of the block you will see the words “Get ready to begin,” and then either the five X’s or
the sequence of rules like before, depending on which run you’re on. After each block, you will again be asked to
report what element of the sequence you would next perform, with the difference being that you will now have 5
seconds to answer. After you respond, a white “+” will be shown for a short time and then the next block will begin
with either five X’s or a new 5-rule sequence displayed on screen.
If you do get lost at any point during the uncued portion, just decide to start at a particular place in the sequence and
go from there. Try not to get flustered and don’t give up. At the end of each run, you will have the opportunity to
take a break, and please come get me after the fourth run. Do you have any questions now? Press the space bar when
you are ready to begin. [Sequence Practice trials]
After practice
 Do you have any questions about those practice blocks? [Check performance.]
 You will next be asked to complete 3 more runs of 6 blocks each, this time without the cues on each trial
and without the answer at the end of the block.
 After this there will be a short questionnaire.
 Any questions before you begin?
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis
BrownU Honors Thesis

More Related Content

Viewers also liked

BMRCL In-Plant Training Report
BMRCL In-Plant Training ReportBMRCL In-Plant Training Report
BMRCL In-Plant Training Report
Divyashree, EIT, CMIT
 
CV of Lee Vanessa Schumacher(1)(1) (3)
CV of Lee Vanessa Schumacher(1)(1) (3)CV of Lee Vanessa Schumacher(1)(1) (3)
CV of Lee Vanessa Schumacher(1)(1) (3)
Lee Vanessa Schumacher
 
Tampa Tribune Lobbyist Series
Tampa Tribune Lobbyist SeriesTampa Tribune Lobbyist Series
Tampa Tribune Lobbyist Series
John T. Wark
 
Art album spain
Art album spainArt album spain
Art album spain
monigom1
 
The Bottom Line on Mobile Homes!
The Bottom Line on Mobile Homes!The Bottom Line on Mobile Homes!
The Bottom Line on Mobile Homes!
Michael Brown
 
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
Law Firm "Fidel Angwarmasse & Partners"
 
Ross Virtual Teams Posting Version
Ross Virtual Teams Posting VersionRoss Virtual Teams Posting Version
Ross Virtual Teams Posting Version
Randy Ross
 
Lewis Hyde CV 2015 transporter update
Lewis Hyde CV 2015 transporter update Lewis Hyde CV 2015 transporter update
Lewis Hyde CV 2015 transporter update
Lewis Hyde
 
Jornada nacional de vacunacion
Jornada nacional de vacunacionJornada nacional de vacunacion
Jornada nacional de vacunacion
zeusg
 
Citrix Security Survey Guide
Citrix Security Survey GuideCitrix Security Survey Guide
Citrix Security Survey Guide
Citrix
 

Viewers also liked (11)

BMRCL In-Plant Training Report
BMRCL In-Plant Training ReportBMRCL In-Plant Training Report
BMRCL In-Plant Training Report
 
CV of Lee Vanessa Schumacher(1)(1) (3)
CV of Lee Vanessa Schumacher(1)(1) (3)CV of Lee Vanessa Schumacher(1)(1) (3)
CV of Lee Vanessa Schumacher(1)(1) (3)
 
Tampa Tribune Lobbyist Series
Tampa Tribune Lobbyist SeriesTampa Tribune Lobbyist Series
Tampa Tribune Lobbyist Series
 
Art album spain
Art album spainArt album spain
Art album spain
 
The Bottom Line on Mobile Homes!
The Bottom Line on Mobile Homes!The Bottom Line on Mobile Homes!
The Bottom Line on Mobile Homes!
 
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
PELAKSANAAN PELAPORAN KELAHIRAN OLEH PENDUDUK BERDASARKAN UNDANG-UNDANG NOMOR...
 
Ross Virtual Teams Posting Version
Ross Virtual Teams Posting VersionRoss Virtual Teams Posting Version
Ross Virtual Teams Posting Version
 
Lewis Hyde CV 2015 transporter update
Lewis Hyde CV 2015 transporter update Lewis Hyde CV 2015 transporter update
Lewis Hyde CV 2015 transporter update
 
Eickmoment brochure
Eickmoment brochureEickmoment brochure
Eickmoment brochure
 
Jornada nacional de vacunacion
Jornada nacional de vacunacionJornada nacional de vacunacion
Jornada nacional de vacunacion
 
Citrix Security Survey Guide
Citrix Security Survey GuideCitrix Security Survey Guide
Citrix Security Survey Guide
 

Similar to BrownU Honors Thesis

Task switchingStephen MonsellSchool of Psychology Univer.docx
Task switchingStephen MonsellSchool of Psychology Univer.docxTask switchingStephen MonsellSchool of Psychology Univer.docx
Task switchingStephen MonsellSchool of Psychology Univer.docx
josies1
 
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docx
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docxChanging Circumstances, Disrupting HabitsWendy WoodDuke .docx
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docx
cravennichole326
 
CHEM495_Thesis_Final
CHEM495_Thesis_FinalCHEM495_Thesis_Final
CHEM495_Thesis_Final
Mark Nakhla
 
Cognitive Processes in the Breakfast Task Planning and Monito.docx
Cognitive Processes in the Breakfast Task Planning and Monito.docxCognitive Processes in the Breakfast Task Planning and Monito.docx
Cognitive Processes in the Breakfast Task Planning and Monito.docx
mary772
 
Theorist
TheoristTheorist
Theorist
Nikita Woodside
 
Time-Related Academic Behavior: State or Trait?
Time-Related Academic Behavior: State or Trait?Time-Related Academic Behavior: State or Trait?
Time-Related Academic Behavior: State or Trait?
Kamden Strunk
 
Executive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docxExecutive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docx
cravennichole326
 
Executive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docxExecutive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docx
elbanglis
 
The transfer or generalizability of learning
The transfer or generalizability of learningThe transfer or generalizability of learning
The transfer or generalizability of learning
관수 박관수
 
Kinneen DBA Presentation (2)
Kinneen DBA Presentation (2)Kinneen DBA Presentation (2)
Kinneen DBA Presentation (2)
Caroline Kinneen
 
47504805 model-kajian-tindakan
47504805 model-kajian-tindakan47504805 model-kajian-tindakan
47504805 model-kajian-tindakan
Nizam Zam
 
What is research
What is researchWhat is research
What is research
FJWU
 
Design
DesignDesign
Design
kus yono
 
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docxA PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
bartholomeocoombs
 
Benjamin Crabtree Regenstreif Conference Slides
Benjamin Crabtree Regenstreif Conference SlidesBenjamin Crabtree Regenstreif Conference Slides
Benjamin Crabtree Regenstreif Conference Slides
ShawnHoke
 
Translational research
Translational researchTranslational research
Translational research
Shahin Hameed
 
LEARNING OBJECTIVES· Describe single-case experimental designs.docx
LEARNING OBJECTIVES· Describe single-case experimental designs.docxLEARNING OBJECTIVES· Describe single-case experimental designs.docx
LEARNING OBJECTIVES· Describe single-case experimental designs.docx
croysierkathey
 
Ex 2factor spq
Ex 2factor spqEx 2factor spq
Ex 2factor spq
German Fong
 
Experimental Research Design - Meaning, Characteristics and Classification
Experimental Research Design - Meaning, Characteristics and ClassificationExperimental Research Design - Meaning, Characteristics and Classification
Experimental Research Design - Meaning, Characteristics and Classification
Sundar B N
 
Single Subject Design - Research Design
Single Subject Design - Research DesignSingle Subject Design - Research Design
Single Subject Design - Research Design
Psychology Pedia
 

Similar to BrownU Honors Thesis (20)

Task switchingStephen MonsellSchool of Psychology Univer.docx
Task switchingStephen MonsellSchool of Psychology Univer.docxTask switchingStephen MonsellSchool of Psychology Univer.docx
Task switchingStephen MonsellSchool of Psychology Univer.docx
 
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docx
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docxChanging Circumstances, Disrupting HabitsWendy WoodDuke .docx
Changing Circumstances, Disrupting HabitsWendy WoodDuke .docx
 
CHEM495_Thesis_Final
CHEM495_Thesis_FinalCHEM495_Thesis_Final
CHEM495_Thesis_Final
 
Cognitive Processes in the Breakfast Task Planning and Monito.docx
Cognitive Processes in the Breakfast Task Planning and Monito.docxCognitive Processes in the Breakfast Task Planning and Monito.docx
Cognitive Processes in the Breakfast Task Planning and Monito.docx
 
Theorist
TheoristTheorist
Theorist
 
Time-Related Academic Behavior: State or Trait?
Time-Related Academic Behavior: State or Trait?Time-Related Academic Behavior: State or Trait?
Time-Related Academic Behavior: State or Trait?
 
Executive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docxExecutive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docx
 
Executive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docxExecutive FunctionThe Search for an Integrated AccountMari.docx
Executive FunctionThe Search for an Integrated AccountMari.docx
 
The transfer or generalizability of learning
The transfer or generalizability of learningThe transfer or generalizability of learning
The transfer or generalizability of learning
 
Kinneen DBA Presentation (2)
Kinneen DBA Presentation (2)Kinneen DBA Presentation (2)
Kinneen DBA Presentation (2)
 
47504805 model-kajian-tindakan
47504805 model-kajian-tindakan47504805 model-kajian-tindakan
47504805 model-kajian-tindakan
 
What is research
What is researchWhat is research
What is research
 
Design
DesignDesign
Design
 
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docxA PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
A PROCEDURE FOR IDENTIFYING PRECURSORS TOPROBLEM BEHAVIOR.docx
 
Benjamin Crabtree Regenstreif Conference Slides
Benjamin Crabtree Regenstreif Conference SlidesBenjamin Crabtree Regenstreif Conference Slides
Benjamin Crabtree Regenstreif Conference Slides
 
Translational research
Translational researchTranslational research
Translational research
 
LEARNING OBJECTIVES· Describe single-case experimental designs.docx
LEARNING OBJECTIVES· Describe single-case experimental designs.docxLEARNING OBJECTIVES· Describe single-case experimental designs.docx
LEARNING OBJECTIVES· Describe single-case experimental designs.docx
 
Ex 2factor spq
Ex 2factor spqEx 2factor spq
Ex 2factor spq
 
Experimental Research Design - Meaning, Characteristics and Classification
Experimental Research Design - Meaning, Characteristics and ClassificationExperimental Research Design - Meaning, Characteristics and Classification
Experimental Research Design - Meaning, Characteristics and Classification
 
Single Subject Design - Research Design
Single Subject Design - Research DesignSingle Subject Design - Research Design
Single Subject Design - Research Design
 

BrownU Honors Thesis

  • 1. Differential Behavioral Patterns in Sequence Performance Following either Sequential or Random Task Practice A thesis submitted in partial fulfillment of the requirements for the degree of Bachelor of Arts with Honors in the Psychology concentration of Brown University April 24th , 2015 Kathryn Nicole Graves Advisor: David Badre, Ph.D.
  • 2. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 2 Acknowledgments I have been incredibly fortunate to have been a part of the Badre Lab at Brown University for the past two years. My time as an undergraduate research assistant, getting to learn from and work with this incredible team of brilliant researchers, has been a distinct honor, a welcome challenge, and the most pivotal educational experience of my life. First and foremost, I’d like to thank Dr. David Badre for fostering this intellectual environment and allowing me to be a part of it. His guidance and feedback over the past two years have allowed me invaluable growth as a student and scientist. I would also like to thank Dr. Theresa Desrochers for being the kind of mentor I aspire to be. From teaching me how to code, to trusting me to with her experiments and ideas, to giving me life advice, to supporting me through my own struggles both in and out of the lab, she has made me better in more ways than I can count. Thank you to the entire Badre Lab for easily being the coolest group of people I’ve ever associated with, and for being there for me in all the best ways throughout my entire experience. I’d especially like to thank Jason Scimeca for laughing at my jokes and being a source of constant positivity. Thank you to Perri Katzman, a Badre Lab alumna, who has been my honors thesis guardian angel. Finally, thank you to my friends and family for being the kind of support system that has made me feel unashamed to both brag and cry about my research in front of them.
  • 3. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 3 Abstract Sequential task performance requires cognitive control, which is observed experimentally by measuring reaction times during the completion of simple sequences of tasks. Previous behavioral studies have shown an elevated reaction time at the first position task of a given sequence, indicating the presence of cognitive control on the sequence level. The current study investigated how an element of practice prior to sequence performance would affect this cognitive control. Across seven experimental designs, participants were given a practice section, in which they performed either sequential or random cued tasks, and a test section, in which they performed unpracticed, novel sequences. Collectively, the results of these experiments suggest that sequence performance procedure can be generalized with high efficiency from practice to test, such that no significant difference in reaction times was demonstrated between the two conditions. Given that participants demonstrated sequence performance behavior even when cues were provided per trial, the results also suggest that sequence structure detection ability is sensitive to implicit sequence structure. Overall, these results suggest that sequence learning is sensitive to neither familiarity with specific sequences of tasks, nor sequential structure in general.
  • 4. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 4 Introduction Task and sequence performance are integral parts of daily life. From turning a doorknob to picking up a pen, there are countless tasks that must be completed throughout the day. Task performance takes time, and the amount of time required for completion of individual tasks, or tasks within a sequence, can depend on a number of different variables, such as task difficulty, cognitive resources available, task order, and available responses. The current study seeks to illuminate the intersection of four bodies of research on four underlying components of sequence practice and performance – switch costs, task practice, sequence costs, and motor sequence practice. Behavioral studies of task performance have identified a certain temporal “cost” that exists when switching from one task to another task of equal complexity (Kramer et. al, 1999; Allport et al., 1994; Fagot, 1994; Arbuthnott & Frank, 2000; Rogers & Monsell, 1995). This cost, termed “switch cost,” was demonstrated to be robust across extended inter-trial intervals (ITI’s), as indicated in a study by Allport et al. (1994), in which participants performed task switching with intervals of either 500ms or 1000ms between each trial. The results of this study indicated that, even with the extended ITI, participants still demonstrated an elevated reaction time when switching between two consecutive tasks versus repeating the same task. This residual temporal cost was thus attributed as a necessary aspect of task switching behavior. Switch cost was initially defined as being a function of one of two separate cognitive processes, both of which gained support from multiple cognitive theorists. The first cognitive process, a combination of task inhibition and task set reconfiguration, involved the process of actively preparing to respond to a successive task by inhibiting the response of a previous task (Arbuthnott & Frank, 2000; Fagot, 1994; Rogers & Monsell, 1995). The second cognitive
  • 5. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 5 process, task set inertia, involved a passive dissipation of the influence of a previous task set, which must occur before one is able to respond to a successive trial (Allport, et al., 1994). A pivotal study conducted by Meiran and Chorev (2000) reconciled these two explanations by having participants perform task switching after either being given an extended Response-Cue Interval (RCI) or an extended Cue-Target Interval (CTI). Participants judged the spatial location of a stimulus on a 2x2 grid using two key presses, alternating between discriminating across the vertical axis (UP-DOWN) and the horizontal axis (RIGHT-LEFT). Across multiple iterations of the experiment, participants performed the task under varying conditions. Some trials consisted of the interval between the last response and the proceeding cue remaining constant while the interval between the presentation of the trial cue and the presentation of the target varied, and vice versa. Both conditions displayed a decrease, but not complete dissipation of reaction time, indicating that task switching consists of both task inertia and the combination of task set inhibition and task set reconfiguration, as well as a residual component that persisted across increased RCI and CTI (Meiran & Chorev, 2000). These findings aligned with those of Rogers and Monsell, whose task switching study showed that the residual switch cost persisted over an extended Inter-trial Interval (ITI) (Rogers & Monsell, 1995). This switch cost in the context of the second component, task practice, is crucial due to the fact that task switching in the context of daily life rarely consists of purely novel tasks – most of the tasks individuals perform throughout the day, such as the aforementioned turning of a doorknob and lifting of a pen, involve some degree of experience or practice. Experimentally, exhaustive task practice has been shown to decrease overall reaction times and switch costs, though never fully abolishing them (Rogers & Monsell, 1995; Meiran, 1996; Kramer et al., 1999;
  • 6. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 6 Ruthruff et al., 2001). These results provide evidence for the assertion that the cognitive phenomena of task set inertia, task set inhibition and task set reconfiguration are not entirely unmalleable processes, and, to a degree, may dependent on task familiarity. These same phenomena are also present in sequence level task performance – the third and more direct contributing aspect of sequence practice and learning. However, there are added costs associated with sequences, supplementary to the costs on the task level, that exist as a function of the higher order processes needed to perform tasks in-sequence. The significance of the distinction between task and sequence performance is emphasized when considering the real-world context, in which isolated task performance is less prevalent than tasks performance following some type of sequential schema – most tasks, like pouring milk and measuring out grounds, are all part of a larger sequence of steps involved in achieving a larger goal, like making coffee. In order to perform these tasks without external cues, one must be able to maintain the overarching goal of having a cup of coffee in the end, while also updating contexts in order to internally guide flexibly switching between different tasks – i.e. recognizing that the coffee maker must be started only after pouring water in the reservoir, but before pouring coffee into the mug, and being able to cognitively ‘reset’ in order to perform each new task. The ability to perform these two higher-order cognitive processes of goal maintenance and context updating is a function of cognitive control on the sequence level. Experimentally, cognitive control can be measured as participants perform simple sequences of tasks by observing the first position reaction time in that sequence (Schneider & Logan, 2006; Barcelo et al, 2007; Pojac et al., 2009). In a key study performed by Schneider and Logan (2006), participants performed repetitions of two sequential structures consisting of color and shape judgments – one sequence followed the structure “AABB,” with one switch, while the other followed the structure
  • 7. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 7 “ABBA,” with two switches. The experimental results showed that participants demonstrated a significantly higher reaction time at the first position judgment of both sequences, regardless of whether that first position was a switch or a repeat, than the corresponding internal sequence judgment. This elevated reaction time, termed “restart cost”, is the temporal cost associated with task resetting on the sequence level to allow for the initiation of the next sequence, and is evidence of cognitive control (Schneider & Logan, 2006). Functionally, the dorsolateral frontopolar cortex has been implicated in this higher-order cognitive process (Desrochers et al., 2013; Badre et al., 2009). Despite the conclusive body of evidence regarding cognitive control in the context of general sequence performance, little research has been conducted specifically concerning the concept of sequence learning. However, studies concerning tangential aspects of sequence practice and learning, specifically those involving motor sequence learning, provide a context for the current experiment. Behavioral results from previous motor sequence experiments show poor transfer of learning from a practice to test, as manifest by a significant increase in response times from the last run of practice to the first run of test (Willingham et al., 2000; Fezzani et al, 2000). While these results do, in a sense, speak to learning on a sequence level, they are complicated by the motor aspect. The paradigms used muddle the component contributions of motor and sequence learning. Thus, the current study seeks to selectively observe the effects of sequential structure in practice and test by removing the motor aspect. However, the experimental structure maintains the necessary elements that preserve previously determined behavioral trademarks of task and sequence performance, such that observed differences from practice to test can be attributed with confidence to learning, and not considered as possible artifacts of indistinguishable behaviors.
  • 8. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 8 The tasks consisted of using one of two key presses to make judgments for either color (red or blue), shape (circle of square) and size (small or large). Judgments were based on a series of five rules presented on a screen at the beginning of each block for the sequence practice condition, and cues provided either during or immediately before trial presentation in the cued condition (see Figure 1). Figure 1 A) Example simple sequence structure: two internal switches, two internal repeats, switch at the first position. B) Example complex sequence structure: three internal switches, one internal repeat, repeat at the first position
  • 9. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 9 Over the course of seven different experimental designs, participants first conducted a practice section, which consisted of practicing either repetitions of two five-item sequences of tasks – one simple and one complex – or random, cued tasks (see Figure 2). After the practice section, participants performed a test section in which they performed either practiced, familiar sequences and multiple novel sequences, or repetitions of two novel sequences. Reaction times and error ratios were analyzed for differences in familiar versus novel sequence performance, cued random task versus sequence practice performance, practice versus test sections, and test sequence performance after having practiced either random or sequential tasks. Across all seven experiments, both behavioral measures were analyzed for two different effects. In order to identify any main effect related to differential practice, a paradigm element that changed with each experiment, we performed two-sample t-tests on the measures after collapsing them across sequence position within their respective practice groups. In search of a position related interaction, specifically one involving changes in cognitive control at the first position, we divided each practice group’s data into first and internal position data (positions 2- 5), and performed a repeated measures ANOVA.
  • 10. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 10 Figure 2 A) Abbreviated run showing general sequence structure B) Sequence rules, example trial, and sequence position question for experiments 1 and 2, practice and test sections. Participants performed sequences at practice and test. C) Sequence rules, example trial with cue, and sequence position question for experiment 3, practice and test sections. Participants performed cued sequences at practice, same sequences uncued at test. D) Sequence rules, example trial with cue, and sequence position question for experiments 4, and 5, practice and test sections. Participants were shown a masked screen and performed trials that were in a sequence (Exp. 4) or random (Exp. 5). E) Sequence rules, cue screen, example trial, and sequence position question for experiment 6 and 7, practice and test sections. Participants were in one of two practice groups – random trial practice or sequence practice.
  • 11. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 11 Experiment 1 In Experiment 1, participants first performed a practice section, consisting of seven runs of seven blocks of repetitions of two sequences, one simple and one complex. After practice, participants performed seven more runs at test, which consisted of the two previously practiced, now familiar sequences as well as multiple different novel sequences. Two hypotheses were developed for this first experiment. The first assumed a static relationship between the first position judgment and internal trial types within each sequence, and predicted that sequence practice would facilitate later performance of familiar sequences overall at test. The other hypothesis supported the alternate prediction that first position reaction time could be selectively effected by practice due to increased efficiency with the cognitive controller during performance of familiar sequences as compared to novel sequences at test. Results In order to determine whether or not sequence-specific learning was taking place at practice and persisting at test, reaction times were measured across practice and test runs for each trial type (first positions, switches, and repeats). Across all three trial types, the reaction times indicated that participant performance improved with practice (main effect of run: F’s > 10.99, p’s < 3.36 x 10-10 ). This improvement in sequence performance carried over to test, with T-tests showing no significant difference from the last practice run to the first test run (p’s> 0.11) (see Figure 3).
  • 12. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 12 Figure 3 learning curves indicated by reaction time for A) Practice and B) Familiar first position trials, C) Practice and D) Familiar repeat trials, and E) Practice and F) Familiar switch trials. Curves suggest that participants were learning, getting faster through practice and maintaining performance levels at test. Post hoc T-Tests determined that reaction time across later practice runs were significantly different from reaction times at the first run for nearly all conditions (complex sequence repeats p = 0.07, all others p’s < 0.05).
  • 13. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 13 The reaction times at test, however, indicated that there was no overall effect of sequence practice on subsequent performance of familiar versus novel sequences (no main effect of familiarity, p < 0.05) (see Figure 4). There was also no first position-specific benefit of sequence practice (no familiarity x position interaction: F (1, 50) = 1, p < 0.05). However, the reaction times did indicate that, in both conditions, participants were demonstrating trademark sequence performance behaviors (main effect of position, F (1, 50) = 100.47, p < 0.001). Figure 4. A-B) Reaction times and error ratios collapsed across sequence position for familiar and novel conditions and test. Results show no main effect of familiarity. C-D) Reaction times and error ratios at first and internal
  • 14. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 14 sequence positions, for familiar and novel conditions at test. Results indicate no position-specific interactions as a function of practice. Error ratios at test illustrated even greater chance effects, as compared to reaction times, with no clear benefit of practice overall (no main effect of familiarity, p > 0.05), or the first position (no familiarity x position interaction: F(1,50) = 0.28, p > 0.05). Discussion The results of Experiment 1 did not support either of our initial hypotheses predicting practice effects – however, the main effect of position in reaction times validated the underlying assumption that participant behavior would reflect sequence performance across all sequences and both experimental sections. Taken together, these results presented two viable possibilities for what was occurring between the practice and test sections. First, it was possible that participant performance improved to such a rapid plateau at practice and carried over so efficiently to test such that, when overall performance of either the practice or later familiar sequences was averaged and compared at test performance, the two sections displayed comparable performance efficacy. Second, the results may have been caused by participants not getting enough time to practice the familiar sequences, such that, throughout practice and test, those sequences still carried an element of novelty that made them comparable to the true novel sequences. Experiment 2 Adopting the latter theory, Experiment 2 was designed such that participants were given over twice as much practice as those in Experiment 1. The experiment was now spread over the
  • 15. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 15 course of two days, with two seven-run sections on the first day and a four-run section on the second. The test section followed the same seven-run format as that in Experiment 1 and succeeded the last practice section on day two. Figure 5 learning curves indicated by reaction time for A) Practice and B) Familiar first position trials, C) Practice and D) Familiar repeat trials, and E) Practice and F) Familiar switch trials. Curves suggest that participants were learning, getting faster through practice and maintaining performance levels at test.
  • 16. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 16 Results Similarly to Experiment 1, reaction time learning curves across all runs of practice showed a significant decrease in magnitude (F’s > 19.5, p’s < 0.001; post hoc T-tests p’s < 0.01), and no significant change from practice to test (p’s > 0.22) (see Figure 5). Reaction times and error ratios at test mirrored those of Experiment 1, in that there was neither evidence of an overall practice effect nor an interaction at the first position that suggested selective improvement in the employment of the cognitive controller (no reaction time main effect of familiarity, p > 0.05; no error ratio main effect of familiarity, p > 0.05; no reaction time familiarity x position interaction: F(1,20) = 0.63, p > 0.05; no error ratio familiarity x position interaction: F(1,20) = 3.43, p > 0.05) (See Figure 6). While there was also again no effect of position in error ratios (no main effect of position: F (1, 20) = 0.17, p > 0.05), the reaction time main effect presented as well (main effect of position: F(1, 20) = 53.63, p < 0.001). Discussion Overall, the results of Experiment 2 were remarkably consistent with those of Experiment 1, despite the addition of substantially more practice runs. There was a clear reaction time flooring effect that was rapidly attained at practice, and the learning that was manifested in this result clearly generalized to novel sequences at test as evidenced by the lack of significant differences in reaction time between both practice versus test sequences and familiar versus test sequences. Of the two predictions presented in Experiment 1, the results of Experiment 2 provide substantial support for that which stated that participant performance improved quickly and in such a way as to generalize learning to test, as opposed to the theory that suggested having insufficient practice overall in the first section resulted in performance of comparably novel
  • 17. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 17 sequences at test – the highly pronounced reaction time flooring effect at practice makes this highly unlikely. Figure 6 A-D) Reaction times and error ratios collapsed across sequence position for familiar and novel conditions and test. Results show no main effect of familiarity. C-D) Reaction times and error ratios at first and internal sequence positions, for familiar and novel conditions at test. Results indicate no position-specific interactions as a function of practice. Given the unexpected participant capacity to learn flexibly, as demonstrated by participants in both Experiment 1 and Experiment 2, it became evident that sequence-level learning was not sensitive to specific sequences over others. Thus, we posited that learning might
  • 18. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 18 instead be sensitive to the presence of sequential structure in general, revising our operational definition of sequential learning in the following experiments. The underlying assumption now was not that the effects of practice would manifest themselves in a comparison of different sequences, but instead in a comparison of sequence performance overall after one of two conditions of practice – either sequential task practice, similar to that of the previous experiments, or practice with an analogous task that differed only in its lack of sequential structure. The following three studies represent the intermediate steps in our determining that analogous task practice condition. Experiment 3 Our first attempt to extract the sequential element from our otherwise sequential task section involved incorporating individual trial cues, supplementary to the sequence rules that were provided at the beginning of each block. Thus, on each trial at practice, participants were shown a one-word cue above the image to which they were responding, which directed them either to judge for “Color,” “Shape,” or “Size.” Then, at test, participants performed the same sequences from practice, but without the trial cues. Previous studies of redundant cueing indicate that the provision of the aforementioned cues leads to dramatic decreases in reaction times during task performance, as compared to performance of the same tasks using internally generated cues (Koch, 2003; de Jong et al., 2006; Kleinsorge et al., 2008). From these conclusions, and also considering intuitively that cues facilitate task performance to such an extent as to invalidate the necessity of sequence rules, we predicted that participants would be faster overall at practice, as compared to test. We also
  • 19. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 19 predicted participants would not exhibit sequence performance behavior when cues were available to them due to the decrease in cognitive demand. Results The reaction time and error ratio data over all of practice compared to all of test both indicated that task performance did not benefit from the introduction of sequence structure from practice to test (no reaction time main effect of section, p > 0.05; no error ratio main effect of section, p > 0.05). There was also no change in behavior at the first position from practice to test (no reaction time section x position interaction: F(1,8) = 0.02, p > 0.05; no error ratio section x position interaction: F(1,8) = 0.23, p > 0.05). While there was also no positional significance in error rates between the two sections (no main effect of position: F(1,8) = 0.14, p > 0.05), there was evidence of restart costs at both practice and test (main effect of position: F(1,8) = 0.02, p > 0.05). Discussion Given our predictions about the effect of adding cues at practice, the reaction time results of Experiment 3 were surprising. Participants were not explicitly told to conceptualize the tasks at practice as sequences, and were given the tools so as to not rely on sequential structure – yet, there was clear evidence that participants were still utilizing sequence performance behavior, as indicated by the presence of restart costs at practice. The experimental structure of the practice section was meant to be analogous to that of the previous experimental versions, but without the sequential element. Because participants were still clearly performing sequences, however, Experiment 3 was not appropriate as the non-
  • 20. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 20 sequential condition for the later task versus sequence practice experiment. Thus, in the next iteration of this experiment, we further stripped away the sequence element in order to prevent unsolicited sequence performance behaviors. Figure 7 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence test conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main effect of position.
  • 21. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 21 Experiment 4 In order to encourage participants to rely on the cues in the cued task section, Experiment 4 maintained sequential structure of tasks, but masked the sequence rules such that participants saw only a series of X’s where, in previous experiments, they saw the words “Color”, “Shape”, and “Size.” With this further removal of sequential structure from the practice section, our hypotheses remained consistent with that of Experiment 3. Results The results of Experiment 4 were similar to those of Experiment 3 – both reaction times and error ratios were unchanged from practice to test (no reaction time main effect of section, p > 0.05; no error ratio main effect of section, > 0.05), and there was, again, no interaction at the first position (no reaction time section x position interaction: F(1,6) = 0.7, p > 0.05; no error ratio section x position interaction: F(1,6) = 0.7, p > 0.05) (See Figure 8). Lastly, the main effect of position remained robust in reaction times in Experiment 4, though not in error ratios (reaction time main effect of position: F (1,6) = 35.8, p < 0.001; no error ratio main effect of position: F(1,6) = 3.76, p > 0.05). Discussion Even with masked sequential structure and, subsequently, even less motivation for participants to perform the practice tasks as sequences, the main effect of position in reaction times provides clear evidence of persistent restart costs. Participants demonstrated the ability to detect sequence structure to such a sensitive degree as to be detrimental with regards to the
  • 22. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 22 overall purpose of the cued tasks. Therefore, sequential structure was entirely removed in Experiment 5 in order to abolish these restart costs. Figure 8 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence test conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main effect of position. Experiment 5 Our next attempt at finding an appropriate practice condition represented the farthest possible departure from sequential task performance. In Experiment 5, not only were the
  • 23. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 23 sequence rules screens masked, but the order of task presentation was randomized. The parameters of the practice section now completely disallowed the sequential performance of tasks, and we thus maintained our hypothesis that the practice section would yield no sequence performance behavior, as well as faster reaction times as compared to the sequence test section. Results Consistent with the results of Experiments 3 and 4, the absence and later presence of sequential structure had no effect on reaction times or error rates (no reaction time main effect of section: p > 0.05; no error ratio main effect of section: p > 0.05) (See Figure 9). There was, again, a significant difference in reaction times between first and internal positions (main effect of position: F(1,6) = 14.43, p < 0.01), which did not extend to error ratios (no main effect of position, F(1,6) = 2.4, p > 0.05). However, there was also a significant section-by-position interaction (section x position interaction: F(1,6) = 7.99, p < 0.05), which was an effect not previously produced in the earlier versions of the task practice, sequence test experiment. This effect was not manifest in error ratios (no section x position interaction: F(1,6) = 0.63, p > 0.05). Discussion The results of Experiment 5, specifically the section-by-position interaction, suggest that the random tasks constitute an appropriately analogous condition to sequential tasks without prompting sequential behavior. However, the remaining concern with this experimental design was the non-significant but visually evident trend in the random task reaction times to be slower than the sequence reaction times. This was striking, considering the ample body of evidence supporting the opposite prediction.
  • 24. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 24 Figure 9 A-B) Reaction times and error ratios collapsed across sequence position for task practice and sequence test conditions. Results show no main effect of sequential structure. C-D) Reaction times and error ratios at first and internal sequence positions, for familiar and novel conditions at test. Reaction time results indicate a persistent main effect of position, as well as a section x position interaction. However, as previously suggested in a reaction time study by Sohn and Anderson (2001), this effect was likely a function of a lack of preparation time between the random trials that was present between the sequential trials. As indicated in their study, ‘foreknowledge’ of an upcoming task is an essential aspect of sequence performance, as sequential order allows one to prepare for each subsequent image prior to image presentation (Sohn & Anderson, 2001).
  • 25. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 25 Because the Experiment 5 paradigm did not allow for the process of preparation prior to trial presentation, instead forcing participants to initiate preparation strategies at the same time they were expected to respond, overall reaction times in the practice section were inflated. Therefore, while Experiment 6 employs the general paradigm from Experiment 5, the timing of cue and stimulus presentation was altered to allow for response preparation. Experiment 6 In order to address the final structural issue, as indicated in Experiment 5, the cued random practice condition now consisted of a cue being shown immediately before image presentation within the previous trial’s inter-trial interval. This manipulation allowed for response preparation that was not present in the previous three experiments. Because we had finally determined the analogous task condition for our between subjects comparison, Experiment 7 also reincorporated a sequence practice section, and participants were, therefore, assigned to one of the two practice conditions. The sequence practice condition was a hybrid, in structure, of the previous experiments. Like the random condition it consisted of 4 practice runs and 3 test runs, and like Experiments 1 and 2, the practice sections consisted of repetitions of practice of the same two sequences. Like Experiments 3, 4, and 5, the test section consisted of repetitions of performance of two unpracticed sequences, as opposed to multiple different novel sequences. However, for the sake of increasing statistical power, the practice and novel sequence reaction time results of Experiments 1 and 2 have been added to the sequence condition analysis. Experiments 3, 4 and 5 have been excluded due to the aforementioned flaws in the experimental design.
  • 26. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 26 Our hypotheses for Experiment 6 consisted of the same behavioral predictions at test as those predicted in Experiments 1 and 2, but based on different experimental parameters. Both overall decrease in reaction times and the interaction at the first position were predicted to be functions of sequence practice as opposed to cued, random task practice. Results The behavioral results at test, comparing sequence performance after either sequence practice (sSEQ) or random, cued task practice (rSEQ), indicated no difference in performance as a function of practice group (no reaction time main effect of group: F(1,64) = 0.15, p > 0.05; no error ratio main effect of group: F(1,64) = 0.0005, p > 0.05) (See Figure 10). There was also no evidence of a selective effect on the cognitive controller (no reaction time group x position interaction: F (1, 64) = 0.13, p > 0.05; no error ratio group x position interaction: F (1, 64) = 1.73, p > 0.19). The evidence of restart costs across both groups was evidenced in the significant difference in reaction times and marginal difference in error ratios between the first position and the internal sequence positions (reaction time main effect of position: F(1,64) = 108.9, p < 0.001; error ratio marginal effect of position: F(1, 64) = 3.5, p = 0.067). Discussion As was the case with Experiments 1 and 2, the results of Experiment 6 were surprising, given the intuitive sense of the hypotheses detailed herein. The marginally significant position effect on error ratios followed the same trend as reaction times, suggesting overall that the first position judgment in each sequence carried the highest cognitive demand. This is consistent with the body of research involving sequence performance, specifically referring to the robust restart
  • 27. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 27 costs at the first position, and is a good check that, despite the experimental manipulations of sequence performance context in Experimental 6, participants are still demonstrating sequence performance behavior. Figure 10 A-B) Reaction times and error ratios collapsed across sequence position for rSEQ and sSEQ conditions. Results show no main effect of type of practice. C-D) Reaction times and error ratios at first and internal sequence positions, for rSEQ and sSEQ. Reaction time and error rate results indicate a main effect of position.
  • 28. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 28 Nonetheless, the otherwise null results of this version prompted one final experimental iteration, in which we made minor changes to task administration with consideration to laboratory best practices. Experiment 7 Experiment 7 was a continuation of Experiment 6. However, starting with this version of the study, we began staying in the room with participants for the duration of the task in order to prompt improved task performance. This change in protocol had the effect of improving participant performance extensively, both in terms of reaction times and error ratios. It was decided that the analysis for Experiment 7 would be analyzed independently of all previous experimental versions in order to only be representative of optimal data. Also in accordance with seeking the best possible data-collection conditions, and based off of the results and trends from Experiment 6, four out of the ten sequence structures were selected as optimal and were the only structures from which sequences were drawn in Experiment 7. Results In order to ensure that the learning seen in Experiments 1 and 2 was also evident under the optimal conditions of Experiment 7, reaction times over runs at practice and test were again analyzed for evidence of a learning curve (See Figure 11). Learning was indicated in both practice groups (sequence practice main effect of run: F(3,27) = 5.36, p < 0.01; random practice main effect of run: F(3, 27) = 4.21, p < 0.01). T-tests indicated that sequence learning at practice
  • 29. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 29 transferred to test (no main effect of test section, p > 0.05), but no significant learning transferred from the random, cued task practice to test (main effect of test section, p < 0.05). Figure 11 learning curves indicated by reaction times for A) Practice sequence runs, B) sSEQ runs, C) practice random, cued tasks, and D) rSEQ runs. Curves suggest that participants in both random and sequence groups were learning, getting faster through practice and maintaining performance levels at test. C. D.
  • 30. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 30 At test, neither reaction times nor error ratios demonstrated any overall sensitivity to group effects (no reaction time main effect of group, p > 0.05; no error ratio main effect of group, p > 0.05) (See Figure 12). Position-specific interactions were also not observed (no reaction time group x position interaction: F(1,15) = 0.46, p > 0.05; no error ratio group x position interaction: F(1,15) = 1.26, p > 0.05). While there was also no effect of position on error ratios (no main effect of position: F(1,15) = 0.73, p > 0.05), reaction times were sensitive to positional differences (main effect of position: F(1,15) = 50.97, p < 0.001). Discussion Staying in the room with participants as they performed the task had the effect of dramatically improving overall performance. However there was still no effect of practice group, indicating that sequence performance is not sensitive to the type of practice that preempts it. The presence of sequence structure at practice did not facilitate sequence performance at test over random task practice. This conclusion extends to the lack of group-by-position interaction, as made evident specifically by the result at the first position reaction time compared to internal position times. It is clear that practice with utilizing the cognitive controller does not facilitate its later use – it can be employed after any kind of previous practice and perform comparably.
  • 31. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 31 Figure 12 A-B) Reaction times and error ratios collapsed across sequence position for rSEQ and sSEQ conditions. Results show no main effect of type of practice. C-D) Reaction times and error ratios at first and internal sequence positions, for rSEQ and sSEQ. Reaction time results indicate a main effect of position. General Discussion Across the seven experimental designs and aims, participants proved to be highly efficient at learning task performance and generalizing the task procedure. Sequence performance was highly flexible and adaptable to altered task set orders. These experiments indicate that sequence
  • 32. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 32 learning happens so rapidly, and that sequence learning is so well generalize at test, that reaction times of both practice and familiar sequences are comparable to novel sequences at test. Aside from sequence learning being highly well-generalized, Sequence-detection is highly sensitive. Experiments 3 and 4 illustrated this phenomenon – even when given cues, participant performance indicated sequence performance, as there was evidence of restart costs in both the sequence test section, where they were expected, and the cued task practice section, where they were both unexpected and unsolicited. The results of Experiments 6 and 7 support and extend the results of Experiments 1 and 2 in their indication of the lack of sensitivity of sequence performance and practice. As the first two experiments allowed for the assertion that sequential behaviors are not effected on the level of practice with specific sequences, the final two experiments indicated that these behaviors are not sensitive the sequential structure at all. While it is possible that sequential learning in the previous experiment was, in itself, an inconsequential aspect, it must also be considered that unaccounted for elements may have added variability to the data. One concern involves the implications of the five-item sequence structure on individual performance, specifically with regards to chunking. A previous study by Schneider and Logan (2006) demonstrated evidence of chunking – the process of dividing longer sequences of items into smaller, more manageable subsequences – in six-item sequences such that internal reaction times were comparable to their first position analogs (Schneider & Logan, 2006). Given that little research has been done regarding five-item sequences, variance in current data may be attributed, in part, to unexpected chunking methods. It is much more conceivable with five-item sequences, as opposed to six-item sequences, that different participants could employ different chunking methods (chunks of ‘three’ and ‘three’
  • 33. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 33 would be the primary prediction in sequences of six, but chunks of ‘two’ and ‘three’ or ‘three’ and ‘two’ could be equally likely in sequences of five), and averaging over these different methods could affect the group results. Thus, future data collection must seek to minimize this effect in order to strengthen the group effect of sequence versus random task practice. With regards to moving forward, we would like to run more participants and raise our statistical power. In looking at the behavioral results on performance of individual sequences (See Appendix A), power is still very low per sequential structure. While there were no significant sequential results at the group level, which averaged over multiple different sequential structures and trial type frequencies, it might be that these results will reveal themselves in performance of specific sequences. Running more participants in each of the current four optimal structures will allow us to illuminate this lingering question. If we discover evidence of sequence learning, next steps will hopefully involve functional experimentation on practice effects. A previous fMRI experiment conducted by Badre et al. (2013) discovered DLPFC activity that followed a ramping pattern during the performance of each sequence within each block. This pattern restarted at the first position of each sequence and was consistent across both simple and complex sequences (Badre et al., 2013). One possible explanation for this result involves the activity being a compensatory reaction to increased response uncertainty, which occurs as one progresses through each position in a sequence. While certainty in judgment accuracy decreases with each successive position within a sequence, it is restored at the first position judgment of the next sequence where, statistically, there is less room for error. The neural activity in the DLPFC appears to mirror this effect. Given these results and conclusions as a backdrop for the current experiment, there are two main predictions for the results of this study in fMRI. Predicting an effect of sequence
  • 34. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 34 practice, as compared to task practice, it can be inferred that this form of practice may have the effect of increasing certainty – thus, in fMRI, the pattern of activation would likely follow that of the aforementioned sequence experiment, but extending to the fifth position and with a decrease in compensatory DLPFC activity overall. However, if further experimentation reveals an interaction of practice group and position, localized at the first position, then activation in fMRI might indicate not only an increase in overall certainty, but a supplementary increase in certainty at the first position as a function of practice with cognitive control on the sequence level. However, given that chunking may be occurring within these five-item sequences, it is also possible that the fMRI results will be distorted to reflect that. If participants are segmenting the sequences into shorter subsequences, these chunks would have their own restart costs and would likely exhibit sequence-like qualities. Thus, in fMRI, the original ramping pattern may be interrupted by sudden decreases in activity, which would indicate an increase in confidence at the boundaries of the subsequences. Nonetheless, an fMRI experiment based on the current study is a long way away, given the results at hand. We first must illuminate whether the thus-far evasive effect of sequential practice is possible to distill from other practice elements. In the grand scheme of sequential learning and performance, regardless of which prediction reveals itself to be valid (if any at do so at all), sequence practice is clearly an area that must be explored further. This study merely scratches the surface of how the brain responds to different conditions of sequence performance, and given the sheer volume of null results under so many various practice and test conditions, it is clear that there is more work to be done in this area.
  • 35. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 35 Methods Experiment 1 Participants 28 participants (21 females, 6 males, average age 20) participated in Experiment 1. One participant was excluded after leaving half-way through the experiment, and one was excluded due to a coding error in initiating the experiment, which resulted in fatally flawed data collection. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The experiment was conducted on a Macintosh computer in the Badre Lab at Brown University, running PsychToolBox in Matlab. Input was registered through standard keyboard press responses, and output was displayed on the computer screen. Trial images were constructed from a factorial combination of 3 dimensions: color (red or blue), shape (circle or square), and size (small or large). The three tasks consisted of color, shape, and size judgments. The post-test questionnaire was created using the Qualtrics Research Suite. Procedure Participants performed the experiment in private testing rooms after signing the proper consent forms. Participants first performed two practice sections in which they had unlimited response time. Participants were told to place their fingers on the “J”, “K”, “L”, “;” and “ ’ ” keys, and told that they would be using their first two fingers to make the majority of their
  • 36. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 36 responses. In the first section, participants performed three practice blocks, with one dimension rule (“Color”, “Shape”, or “Size”) presented per block, and one judgment task to perform. The order of rule presentation was counterbalanced across participants. Each trial screen consisted of an image of a particular color, shape, and size, as well as key press reminders on the bottom left and right corners of the screen that indicated the current judgment-relevant response choices and the key to which they corresponded. A white fixation cross was shown during the inter-trial interval for a variable amount of time (500-1000ms). During the second section, participants performed a practice block which began with a screen indicating the five-item, task-relevant sequence. Participants were told to pay attention to the sequence and remember it throughout the block, as it would not be shown again until the end of the block during the sequence position question. They were also told that the key press reminders currently visible on the wait screen would remain onscreen and would be the same throughout the entire experiment. The reminders now consisted of a conglomeration of the three response choices shown individually in the first section. Lastly, participants were told that the sequence could end on any position in the block, and that, at the end of the block, they would be shown the sequence position question in which they would indicate which item in the sequence they would have next performed if the block had continued, by pressing the “1(J)”, “2(K)”, “3(L)”, 4(;)” or “5(’)” keys. Participants then performed one abbreviated block and answered the position question, after which the experimenter checked their performance and brought to their attention any mistakes made. The participants then performed another practice sequence. Following these practice sequences, participants performed the third section, which consisted of seven runs of six blocks of repetitions of two more practice sequences – one novel
  • 37. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 37 and one complex. For this and the fourth section, sequences were drawn from a pool of 50 sequences structures. Each block contained 24 to 28 trials. A wait screen was shown at the beginning of each run instructing participants to press the spacebar to continue. Doing so prompted the “Get Ready’ screen which instructed the participants to get ready to begin, after which the next run would begin. This constituted a break between runs. Prior to the start of this section, the experimenter told participants that they now had four seconds to indicate their answer when each image were shown and five seconds to answer the sequence-position question at the end of each block. The experimenter stressed the importance of paying attention to the sequence rules when they were shown at the beginning of each block and responding as soon as the participant knew the answer, as opposed to second guessing or otherwise delaying their response. They also explained to the participants that, if at any point they forgot where they were in the sequence, they were to pick a place in the sequence and go from there, as opposed to making random guesses. The experimenter then left the room, telling the participant to get their fingers ready and press the spacebar to continue as soon as the door closed behind them. At the end of the third section, participants left the room and got the experimenter, who checked their performance and recorded error codes displayed at the bottom right-hand side of the screen into the lab notebook. Participants were then instructed to perform the fourth section, in which each block consisted of presentations of the previously practiced, now familiar sequences as well as multiple different novel sequences. The novel sequences were counterbalanced for sequence complexity. The fourth section was of the same duration as the third section.
  • 38. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 38 The task automatically closed after the participant answered the last sequence position question in the last block of the last run. After this, the experimenter pulled up the post-test questionnaire (PTQ), which participants completed at their own pace. Depending on how long participants took with the task and the availability of the testing room in use, they would complete the PTQ either on the test computer, on a lab computer, or on the experimenter’s laptop. After completing the PTQ, participants signed a receipt form and were compensated for their time. Experiment 2 Participants 11 participants (9 females, 2 males, average age 22) participated in Experiment 2. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of Experiment 1. Procedure The duration of Experiment 2, specifically the third section, was the only difference between it and Experiment 1. Experiment 2 spanned two consecutive days, and the sessions for each participant were scheduled with no less than twenty-two and no more than twenty-six hours apart. On the first day, participants performed the first two practice sections, as well as two,
  • 39. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 39 seven-run sections of practice of two sequences with a break in-between. On the second day, they performed four more runs of the two practice sequences, followed by the seven-run test section and a revised PTQ. While there were supposed to be breaks in-between the runs at test, participants later informed the experimenter that there were no wait screens and that the runs were continuous. Experiment 3 Participants 5 participants (3 females, 2 males, average age 21) participated in Experiment 3. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of previous experiments. Procedure The first practice section was the same as that of the previous two sections. During the first practice block in the second section, however, participants were now given cues, displayed above the image during each trial, in addition to the five-item sequence rules that were shown at the beginning of the block. The cues consisted the one-word judgment rule relevant to the current trial. They answer to the sequence-position question at the end of the block was provided
  • 40. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 40 in parentheses on the screen below the question. The second practice sequence block was similar to that of previous experiments, but with an “XXXX” above each image in place of the cue. The third section was four runs of six blocks, and consisted of two cued sequences. The fourth run was three runs of the same sequences, minus the cues. Participants performed the revised PTQ following the fourth section. Experiment 4 Participants 4 participants (3 females, 1 male, average age 21) participated in Experiment 4. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of previous experiments. Procedure The experimental procedure was almost identical to that of Experiment 3. The sole difference was in the sequence rule screen shown at the beginning of each block during the first run in the second section and in the third section. Instead of displaying a sequence of rules, participants were now shown five “XXXX”’s masking the sequence. The experimenter also took care not to use the word “sequence” when explaining giving instructions for the masked blocks.
  • 41. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 41 Experiment 5 Participants 4 participants (3 females, 1 male, average age 21) participated in Experiment 5. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of previous experiments. Procedure Visually, Experiment 5 was exactly the same as Experiment 4. Structurally, however, the trials were randomized during the first practice block of the second section and in the third section. Experiment 6 Participants 35 participants (25 females, 10 males, average age 21) participated in Experiment 6. 15 were placed in the cued random task practice condition, and 20 were placed in the sequence practice condition. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of previous experiments.
  • 42. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 42 Procedure Participants in the cued random task practice group performed a task similar to that of Experiment 5. The only difference was in the presentation of the cues. To allow for preparation costs, the cues for each successive trial were presented immediately after the participant responded to the previous trial, as opposed to being shown simultaneously with the trial image. The cues were shown for x milliseconds. At test, the cue screen was masked with “XXXX”. Participants in the sequence practice group performed repetitions of two sequences at practice, and repetitions of two difference sequences at test, with the same screens as those used in the cued-group test section. In both sections, the sequences consisted of one simple structure and one complex structure. Sequences were pooled from a subset of the pool of sequences used in previous experiments – those with the structure IDs 1-4, 7-9, 12, 13, and 16 were used. Experiment 7 Participants 15 participants (8 females, 7 males, average age 21) participated in Experiment 7. Seven were placed in the cued random task practice condition, and 8 were placed in the sequence practice condition. Participants were recruited from the Brown University Student body, as well as from the Providence community, for paid participation at a rate of $10/hour. Apparatus The apparatus were identical to that of previous experiments.
  • 43. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 43 Procedure For both groups, the structure of the experiment was similar to that of Experiment 6. In Experiment 7, however, there were four runs in both section three and four. Based on the results of Experiment 6, sequences with the structure ID 1 and 2, complex, and 9, and 13, simple, were selected as optimal sequence structures and were the structures from which sequences were drawn for Experiment 7. Instead of practicing two sequences at practice and test, participants now performed a total of six sequences at practice and six at test, with three of each of two sequence structures – one simple, and one complex. For example, one of the two sequence structures a participant may have performed at test would have been “AABBC” – thus, three sequences they could have possibly performed could be “ Color Color Shape Shape Size,” “Shape Shape Color Color Size,” and “Size Size Color Color Shape.” In the sequence practice group, participants had exposure to all four sequence structures at some point during the experiment. Starting with Experiment 7, the experimenter stayed in the room with the participant, instead of leaving the room before the beginning of section three.
  • 44. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 44 References Allport, A., Styles, E. A., & Hsieh, S. (1994). Shifting Intentional Set: Exploring the Dynamic Control of Tasks. Arbuthnott, K., & Frank, J. (2000). Executive control in set switching: residual switch cost and task-set inhibition. Canadian Journal of Experimental Psychology = Revue Canadienne de Psychologie Experimentale, 54, 33–41. doi:10.1037/h0087328 Badre, D., Hoffman, J., Cooney, J. W., & Esposito, M. D. (2010). NIH Public Access, 12(4), 515–522. doi:10.1038/nn.2277.Hierarchical Barceló, F., Periáñez, J. A., & Nyhus, E. (2007). An information theoretical approach to task- switching: evidence from cognitive brain potentials in humans. Frontiers in Human Neuroscience, 1. Desrochers, T. M., & Badre, D. (2013, January). Frontal systems supporting the hierarchical control of task sequences. In Journal of Cognitive Neuroscience (pp. 41-41). 55 Hayward Street, Cambridge, MA 02142 USA: MIT press. Fagot, C. (1994). Chronometric investigations of task switching. University of California, San Diego. Retrieved from http://scholar.google.com/scholar?hl=ca&q=Chronometric+investigations+of+task+switchin g&btnG=Cerca#0
  • 45. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 45 Fezzani, K., Thon, B., & Lagarde, J. (2000). Can different levels of S-R practice influence sequence learning? An investigation into the4 context of a perceptual motor sequence learning. Perceptual and Motor Skills, 91, 463–475. de Jong, Ritske, Thomas E. Gladwin, and Bernard M't Hart. "Movement-related EEG indices of Kleinsorge, Thomas, and Patrick D. Gajewski. "Task switching based on externally presented versus internally generated information." Psychological research 72.5 (2008): 501- 514.preparation in task switching and motor control." Brain research 1105.1 (2006): 73-82. Koch, I. (2003). The role of external cues for endogenous advance reconfiguration in task switching. Psychonomic Bulletin & Review, 10(2), 488–492. doi:10.3758/BF03196511 Kramer, a F., Hahn, S., & Gopher, D. (1999). Task coordination and aging: explorations of executive control processes in the task switching paradigm. Acta Psychologica, 101, 339– 378. doi:10.1016/S0001-6918(99)00011-6 Meiran, N. (1996). Reconfiguration of processing mode prior to task performance. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22(6), 1423–1442. doi:10.1037/0278-7393.22.6.1423 Meiran, N., Chorev, Z., & Sapir, a. (2000). Component processes in task switching. Cognitive Psychology, 41, 211–253. doi:10.1006/cogp.2000.0736
  • 46. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 46 Barceló, F., Periáñez, J. A., & Nyhus, E. (2007). An information theoretical approach to task- switching: evidence from cognitive brain potentials in humans. Frontiers in Human Neuroscience, 1. Rogers, R. D., & Monsell, S. (1995). Costs of a Predictable Switch Between Simple Cognitive Tasks. Journal of Experimental Psychology: General, 124(2), 207–231. Ruthruff, E., Johnston, J. C., & Selst, M. Van. (2001). Why Practice Reduces Dual-Task Interference. Journal of Experimental Psychology: Human Perception and Performance, 27(1), 3–21. Schneider, D. W., & Logan, G. D. (2006a). Hierarchical control of cognitive processes: switching tasks in sequences. Journal of Experimental Psychology. General, 135Schneid(4), 623–640. doi:10.1037/0096-3445.135.4.623 Schneider, D. W., & Logan, G. D. (2006b). Priming cue encoding by manipulating transition frequency in explicitly cued task switching. Psychonomic Bulletin & Review, 13(1), 145– 151. doi:10.3758/BF03193826 Sohn, M.-H., & Anderson, J. R. (2001). Task Preparation and Task Repitition: Two-Component Model of Task Switching. Journal of Experimental Psychology: General, 130(4), 764–778. Stoet, G., & Snyder, L. H. (2003). Executive control and task-switching in monkeys. Neuropsychologia, 41, 1357–1364. doi:10.1016/S0028-3932(03)00048-4
  • 47. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 47 Willingham, D. B., Wells, L. A., Farrell, J. M, Stemwedel, M. E. (2000). Implicit motor sequence learning is represented in response locations. Memory & Cognition, 28(3), 366- 375. doi: 10.3758/BF03198552.
  • 48. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 48 APPENDIX A – STATISTICS ON INDIVIDUAL SEQUENCE STRUCTURES Below are the relevant statistics on reaction times and error ratios for individual sequences in the “novel vs familiar” Experiments 1 and 2, and the optimized “rSEQ vs sSEQ” Experiment 7. There are several sequential structures without data from one of two experimental conditions. Those structures have been omitted from the following analyses. Figure 13 Reaction times at test for familiar versus novel A-D) complex sequence structures 1-4, and E-F) simple sequence structures 7-9, 12, 13, and 16.  Structure 13 main effect of familiarity: F(1,25) = 6.02, p < 0.05
  • 49. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 49 Figure 14 Error ratios at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F) simple sequence structures 7-9, 12, 13, and 16.  Structure 7, 16 main effect of familiarity: F’s > 4.8, p’s < 0.05
  • 50. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 50 Figure 15 Reaction times at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F) simple sequence structures 7-9, 12, 13, and 16.  Sturcture 16 main ffect of familiarity: F(1,9) = 5.67, p < 0.05
  • 51. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 51 Figure 16 Error ratios at test for familiar versus novel A-D) complex sequence structures 1, 2, 3, 4, and E-F) simple sequence structures 7-9, 12, 13, and 16.  no main effect of familiarity: F’s < 0.9, p’s > 0.34
  • 52. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 52 Figure 17 Reaction times at test for rSEQ versus sSEQ A-B) complex sequence structures 1, 2, and C-D) simple sequence structures 9 and 13.  no main effect of practice group: F’s < 0.79, p’s > 0.399 Figure 18 Error ratios at test for rSEQ versus sSEQ A-B) complex sequence structures 1, 2, and C-D) simple sequence structures 9 and 13.  no main effect of group: F’s < 0.79, p’s > 0.399
  • 53. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 53 APPENDIX B – EVIDENCE OF TASK SYMMETRY Figure 19 Average reaction times for all A) repeats and B) switches and error ratios for all C) repeats and D) switches at random practice in Experiment 6. In order to ensure that task asymmetry was not a factor in the behavioral results of this experiment, reaction times and error ratios for every combination of switches and repeats were analyzed for possible significant differences in magnitude. The results for this section were taken specifically from the random task condition of Experiment 7. Reaction times across all switches and repeats were fairly consistent (no main effect of trial type, F’s < 2.05, p’s < 0.13). There was also no significant difference in error ratios in the switch condition (no main effect of trial type: F(4,120) = 0.97, p = 0.44). However, the error ratios for the “Shape-Shape” repeat trial type were shown to be significantly higher than those of the other two trial types. Though further
  • 54. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 54 analysis will be needed in order to determine what this result means for the error ratios throughout the entirety of the experiment, the integrity of the reaction times, the primary measure in this study, appears to have been preserved.
  • 55. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 55 APPENDIX C – TASK INSTRUCTIONS FOR ALL EXPERIMENTS Experiment 1 Task Instructions PRE-TRAIN Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings. ----------------- First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the entire length of this experiment. Task Training Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for [dimension]. Please always keep your fingers on the keyboard. You will now do a series of practice trials. An image will briefly be displayed and you will respond according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1 trials] Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials] Sequence Example Training Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR. For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about [SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE, SIZE, COLOR, COLOR, SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the screen. We will now go through a practice block that you can complete at your own pace. You will first see a sequence of rules for 4 seconds, then a white “+” on a get ready screen for 1 second, and then the images will be displayed as before. You will not see the sequence of rules again, so you must remember it. At the end of the block, you will see a question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st - COLOR, 2nd - SHAPE, 3rd - SHAPE, 4th – SIZE, 5th - COLOR) by pressing the 1st , 2nd , 3rd , 4th , or 5th key (JKL;” – emphasize the “). The block can end on any position in the sequence, so it is important to remember where in the sequence you are at all times. After this question you will see a white “+” on the screen for a short amount of time. For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to explain their reasoning on the first few trials] Great. Was the question about what item you would perform next at the end clear? Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Sequence Practice Now you will practice 2 more sequences of rules like the example you just completed using the same key presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and respond as soon as you know the answer. You will not receive any additional reminders as to the sequence of rules once the blocks begin. The beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like before. After each block, you will again be asked to report what element of the sequence you would next perform, with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are
  • 56. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 56 in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block will begin with a new 5-rule sequence displayed on screen. If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. At the end of each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials] After practice  Do you have any questions about those practice blocks? [Check performance.]  You will next be asked to complete 7 runs that have 6 blocks each.  After this there will be a short questionnaire.  Any questions before you begin?
  • 57. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 57 Experiment 2 Task Instructions PRE-TRAIN – DAY 1 Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings. ----------------- First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the entire length of this experiment. Task Training Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for [dimension]. Please always keep your fingers on the keyboard. You will now do a series of practice trials. An image will briefly be displayed and you will respond according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1 trials] Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials] Sequence Example Training Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR. For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about [SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE, SIZE, COLOR, COLOR, SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the screen. We will now go through a practice block that you can complete at your own pace. You will first see a sequence of rules for 4 seconds, then a white “+” on a get ready screen for 1 second, and then the images will be displayed as before. You will not see the sequence of rules again, so you must remember it. At the end of the block, you will see a question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st - COLOR, 2nd - SHAPE, 3rd - SHAPE, 4th – SIZE, 5th - COLOR) by pressing the 1st , 2nd , 3rd , 4th , or 5th key (JKL;” – emphasize the “). The block can end on any position in the sequence, so it is important to remember where in the sequence you are at all times. After this question you will see a white “+” on the screen for a short amount of time. For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to explain their reasoning on the first few trials] Great. Was the question about what item you would perform next at the end clear? Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Sequence Practice Now you will practice 2 more sequences of rules like the example you just completed using the same key presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and respond as soon as you know the answer. You will not receive any additional reminders as to the sequence of rules once the blocks begin. The beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like before. After each block, you will again be asked to report what element of the sequence you would next perform, with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block will begin with a new 5-rule sequence displayed on screen.
  • 58. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 58 If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. At the end of each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials] After practice  Do you have any questions about those practice blocks? [Check performance.]  You will next be asked to complete 7 more runs of the same two sequences that have 6 blocks each.  After this there will be a short questionnaire.  Any questions before you begin? PRE-TRAIN – DAY 2 Sequence Practice First, you will be asked to practice the same 2 sequences of rules that you practiced yesterday, using the same key presses, which will not change. You will have 4 seconds to indicate your answer. Please respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and respond as soon as you know the answer. You will not receive any additional reminders as to the sequence of rules once the blocks begin. Just a reminder, at the beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like before. After each block, you will again be asked to report what element of the sequence you would next perform, and you will have 5 seconds to answer, so it is important that you know where you are in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block will begin with a new 5-rule sequence displayed on screen. If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered and don’t give up. You will be asked to complete 4 runs that are composed of 6 sequence blocks each. At the end of each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials] After practice  Do you have any questions about those practice blocks? [Check performance.]  You will next be asked to complete 7 more runs of the same two sequences that have 6 blocks each.  After this there will be a short questionnaire.  Any questions before you begin?
  • 59. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 59 Experiment 3 Task Instructions PRE-TRAIN Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings. ----------------- First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will have a particular color, shape and size. You will be given a set of rules based on these attributes that apply for the entire length of this experiment. Task Training Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for [dimension]. Please always keep your fingers on the keyboard. You will now do a series of practice trials. An image will briefly be displayed and you will respond according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1 trials] Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials] Sequence Example Training Now you will be asked to perform judgments according to a sequence of 5 rules that will be displayed on the screen at the beginning of the block. For example, you could be shown the words COLOR, SHAPE, SHAPE, SIZE, COLOR. For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about [SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE, SIZE, COLOR, COLOR; SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the screen. For each trial, the current judgment will be displayed in the middle of the screen, above the image (provide example). We will now go through a practice block that you can complete at your own pace. You will first see a sequence of rules for 4 seconds, then a white “+” on a get ready screen for a short time, and then the images will be displayed as before with the current judgment above each image. At the end of the block, you will see a question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st - COLOR, 2nd - SHAPE, 3rd - SHAPE, 4th – SIZE, 5th - COLOR) by pressing the 1st , 2nd , 3rd , 4th , or 5th key (JKL;” – emphasize the “). Below this question is a cue, which will instruct you as to which answer to select. The block can end on any position in the sequence. After this question you will see a white “+” on the screen for a short amount of time. For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to explain their reasoning on the first few trials] Great. Was the question about what item you would perform next at the end clear? Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Now you will be asked to perform another sequence of five rules. The difference now is that you will not receive cues during the trials that tell you which judgment to make – you must do it from memory of the sequence rules shown at the beginning of the block. The sequence will not be shown at any other time, so you must remember it. You also will not be provided with the answer to the question at the end of the block, so it is important to remember where in the sequence you are at all times. Press the space bar when you’re ready to begin. Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Sequence Practice Now you will practice 2 more sequences of rules like the example you just completed using the same key presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please
  • 60. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 60 respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and respond as soon as you know the answer. At the beginning of the block you will see the words “Get ready to begin,” then the sequence of rules like before. After each block, you will again be asked to report what element of the sequence you would next perform, with the difference being that you will now have 5 seconds to answer, so it is important that you know where you are in the sequence at all times. After you respond, a white “+” will be shown for a short time and then the next block will begin with a new 5-rule sequence displayed on screen. If you do get lost, just decide to start at a particular place in the sequence and go from there. Try not to get flustered and don’t give up. You will be asked to complete 7 runs that are composed of 6 sequence blocks each. For the first 4 runs, you will be given the cues for which judgment to make on each trial, as well as the answer to the question at the end of each block. For the last three runs, you will not. Before the 5th run you will see the same instruction on the screen as in the practice that you must remember each sequence. If you have any questions at that point, please come get me. At the end of each run, you will have the opportunity to take a break. At the end of this practice section, please get me. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials] After practice  Do you have any questions about those practice blocks? [Check performance.]  You will next be asked to complete 3 more runs of 6 blocks each, this time without the cues on each trial and without the answer at the end of the block.  After this there will be a short questionnaire.  Any questions before you begin?
  • 61. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 61 Experiment 4/Experiment 5 Task Instructions PRE-TRAIN Ask participant to SHUT OFF cell phone (all the way). Buzzes are just as distracting as rings. ----------------- First, to outline the basic structure of the task, on every trial you will be presented with an image. The image will have a particular color, shape and size. You will be making judgments about these images based on these attributes throughout the experiment. Task Training Please place your fingers on the J K L semicolon keys. You will be using your first two fingers for the majority of your responses: 1 = J, 2 = K. For this rule, [TASK] you will respond [key] for [dimension] and [key] for [dimension]. Please always keep your fingers on the keyboard. You will now do a series of practice trials. An image will briefly be displayed and you will respond according to its [TASK]. Key-press reminders will remain on the screen. You will have as long as you need to respond. Please respond both as accurately and quickly as possible. After you have responded, a white “+” will briefly appear on the screen until the next image is shown. Press the space bar when you are ready to begin. [TASK 1 trials] Great. Now you will do the same thing, but for [TASK]. For this rule [go over the key presses]. Any questions? Press the space bar when you are ready to begin. [TASK 2 trials. Same thing for TASK 3 trials] Sequence Example Training Now you will be asked to perform judgments on a series of images based one the cues provided on each trial. The key presses remain the same and will remain on the screen. For each trial, the current judgment will be displayed in the middle of the screen, above the image (provide example). We will now go through a practice block that you can complete at your own pace. You will first see a screen with five X’s on it, sequence of rules for 4 seconds, then a white “+” on a get ready screen for a short time, and then the images will be displayed as before with the current judgment above each image. At the end of the block, you will see a question on the screen asking which item in the sequence as a whole you would NEXT perform (i.e. 1st - COLOR, 2nd - SHAPE, 3rd - SHAPE, 4th – SIZE, 5th - COLOR) by pressing the 1st , 2nd , 3rd , 4th , or 5th key (JKL;” – emphasize the “). Below this question is a cue, which will instruct you as to which answer to select. The block can end on any position in the sequence. After this question you will see a white “+” on the screen for a short amount of time. For all trials, you will have as long as you need to respond. Please respond both as accurately and quickly as possible. For this practice, it is ok to ask questions if you forget the rules, particularly on the question at the end of the block. Press the space bar when you are ready to begin. [TASK sequence example trials, ask participant to explain their reasoning on the first few trials] Great. Was the question about what item you would perform next at the end clear? Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Now you will be asked to perform a series of judgments based on a sequence of five rules that will be displayed on the screen at the beginning of the block. Instead of the five X’s, you could be shown, for example, the words COLOR, SHAPE, SHAPE, SIZE, COLOR. For the first image, you will make a decision about [COLOR], for the second image, you will make a decision about [SHAPE] and so on. You will repeat this sequence until the block ends, for example COLOR, SHAPE, SHAPE, SIZE, COLOR, COLOR; SHAPE, SHAPE, SIZE, COLOR. The key presses remain the same and will remain on the screen. You will not receive cues during the trials that tell you which judgment to make – you must do it from memory of the sequence rules shown at the beginning of the block. The sequence will not be shown at any other time, so you must remember it. You also will not be provided with the answer to the question at the end of the block, so it is important to remember where in the sequence you are at all times. Press the space bar when you’re ready to begin. Do you have any questions? [Check their performance and bring it to their attention if it’s not perfect or discuss the last question. A second practice sequence will then be presented.] Sequence Practice
  • 62. DIFFERENTIAL BEHAVIOR PATTERNS IN SEQUENCE PERFORMANCFOLLOWING EITHER SEQUENTIAL OR RANDOM TASK PRACTICE 62 Now you will be asked to complete 4 runs of 6 blocks of trials like the first pair of examples you just completed, using the same key presses, which will not change. The only difference is that you will have 4 seconds to indicate your answer. Please respond both as accurately and quickly as possible. What I am interested in is your reaction times, so please be ready to respond, and respond as soon as you know the answer. For these runs, you will be given the cues for which judgment to make on each trial, as well as the answer to the question at the end of each block. After this you will complete 3 runs like the last pair of examples you completed – with a sequence of rules at the beginning and without the cues. At the beginning of the block you will see the words “Get ready to begin,” and then either the five X’s or the sequence of rules like before, depending on which run you’re on. After each block, you will again be asked to report what element of the sequence you would next perform, with the difference being that you will now have 5 seconds to answer. After you respond, a white “+” will be shown for a short time and then the next block will begin with either five X’s or a new 5-rule sequence displayed on screen. If you do get lost at any point during the uncued portion, just decide to start at a particular place in the sequence and go from there. Try not to get flustered and don’t give up. At the end of each run, you will have the opportunity to take a break, and please come get me after the fourth run. Do you have any questions now? Press the space bar when you are ready to begin. [Sequence Practice trials] After practice  Do you have any questions about those practice blocks? [Check performance.]  You will next be asked to complete 3 more runs of 6 blocks each, this time without the cues on each trial and without the answer at the end of the block.  After this there will be a short questionnaire.  Any questions before you begin?