Psychophysics is concerned with how we perceive the physical stimuli impinging upon our senses. The branch of psychophysics that deals with the perception of sound is psychoacoustics. Fechner (1860/1966) is known as the father of psychophysics. MEASUREMENT METHODS: Goal- To establish relationship between the sound presented and how the subject perceives it. • Measurement of the absolute hearing threshold provides some basic information about our auditory system. The tools used to collect such information are called psychophysical methods. • Through these, the perception of a physical stimulus (sound) and our psychological response to the sound is measured. • The lowest level at which the sound is heard (the transition between audibility and inaudibility) might be considered an estimate of absolute sensitivity (sensory capability). • Alternatively, two tones might be presented, one of which is varied in frequency. The subject is asked whether the varied tone is higher (or lower) in pitch, and the smallest perceivable frequency difference—the just noticeable difference (JND)—might be considered an estimate of differential sensitivity (response proclivity). Response proclivity reflects not only the subject’s sensitivity, but also the biases and criteria that affect how he responds. We therefore try to select measurement methods and techniques that minimize the effects of response bias.

1. UNIT 1
PSYCHOACOUSTICS – REVIEW:
(Psychophysical methods – Classical and Modern
Psychophysical methods)
Submitted to Submitted by
MS. VINI ABHIJITH GUPTA HIMANI BANSAL
MVSCOSH MASLP 1st year

2. INTRODUCTION
• Psychophysics is concerned with how we perceive the physical stimuli
impinging upon our senses. The branch of psychophysics that deals with
the perception of sound is psychoacoustics.
• Fechner (1860/1966) is known as the father of psychophysics.

3. MEASUREMENT METHODS
Goal- To establish relationship between the sound presented and how the subject perceives it.
CLASSICAL METHODS
ADAPTIVE METHODS
SCALING

4. CLASSICAL
METHODS OF
MEASUREMENT
1.Method of
Limits
1.Method of
Adjustment
1.Method of
constant stimuli

5. METHOD OF LIMITS
• Descending run: sound is presented at a level expected to be well above threshold, sound is decreased until the subject
no longer perceives the sound (−)
• Ascending run: sound is first presented at a level known to be below the threshold, sound is increased in magnitude
until a positive (+) response is obtained
Since the crossover between “hearing” and “not hearing” lies somewhere between the lowest audible level and the
highest inaudible one, the “threshold” for each series may be taken as the halfway point between them
• Response Bias: ANTICIPATION & HABITUATION
• Minimizing Response Bias:
1. By using an equal number of ascending and descending test runs
2. By varying the starting levels of the runs
• Also called Method of minimal change (Silman, 1991), Method of serial exploration (cited in Katz, 2002)

7. Merits:
 Reliability of the results is more.
 It involves great precision in measurement.
 Catch trials are included.
 Estimates observer bias and persistence effects.
 It allows direct estimation of guessing behaviour.
Demerits:
 Large number of trials are needed, it is a time- consuming procedure.
 Needs more efforts to gain accuracy.
• Increases subject fatigue and difficult to maintain patient’s motivation

8. METHOD OF ADJUSTMENT
• Also called method of average error
• The stimulus is controlled by the subject
• The level of the stimulus is varied continuously
• The level is adjusted downward from above threshold until it is just inaudible, or increased from
below threshold until it is just audible
Threshold is taken as the average of the just audible and just inaudible levels
• Response bias:
1. Persistence of Stimulus (anticipation)
2. Perseveration of Response phenomenon (habituation)

9. Merits:
1. Simple to be administered.
2. The subject actively participates in the procedure.
3. It is a very fast method.
4. It is easily understandable for the test subject.
Demerits:
1. This method can produce several biases.
2. Presence of error of anticipation and habituation.

10. METHOD OF CONSTANT STIMULI
• Also known as the frequency method, described by Friedrich Hegelmaier in 1852
• A nonsequential procedure, stimuli are not presented in an ascending or descending manner
• A step size is selected, stimuli presented in random order
• Absolute sensitivity (threshold) estimation: an equal number of stimuli are presented at each level. The subject
indicates whether the stimulus presentation has been perceived during each test trial
• Differential sensitivity (DL) estimation: the subject’s task would be to say whether two items are the same or
different
• Catch trials included: intervals during which the subject is asked whether a tone was heard, when no tone was
really presented- helps to reduce response bias
• Merits: greater precision of measurement, reduced response bias
• Demerits: inefficient, because a very large number of trials are needed to obtain the data- leads to subject fatigue

11. Simpson, W. A. (1988). The method of
constant stimuli is
efficient. Perception &
psychophysics, 44(5), 433-436.
• Monte Carlo simulations show that data
collected using the method of constant
stimuli yield threshold estimates with the
same variability as and less bias than
estimates based on data collected using an
adaptive trial placement rule. In both cases,
the simulation data were analyzed using
maximum likelihood; for 100 trials or less, the
method of constant stimuli is the better trial
placement rule.
Watson, A. B., & Fitzhugh, A. (1990).
The method of constant stimuli is
inefficient. Perception &
psychophysics, 47(1), 87-91.
• Simpson (1988) has argued that the method
of constant stimuli is as efficient as adaptive
methods of threshold estimation, and has
supported this claim with simulations. We
show that Simpson’s simulations are not a
reasonable model of the experimental
process, and that more plausible simulations
confirm that adaptive methods are much
more efficient than the method of constant
stimu1i.

12. ADAPTIVE METHODS OF
MEASUREMENT
In an adaptive procedure, the level at which a particular stimulus is presented to the subject depends upon how the
subject responded to the previous stimuli (Wetherhill & Lewitt, 1965); relatively efficient methods.
1.Yes/No Procedure
1.Bekesy’s tracking method
1.Simple up- down or Staircase method
1.Parameter estimation by sequential testing (PEST)
1.Block up- down Temporal Interval Forced Choice (BUDTIF)
1.Transformed up- down procedure

13. YES/NO PROCEDURE
• The subject judges the presence or absence of the signal.
• The subject has to make a yes or no response on stimulus presentation
• Can range from 0% to 100% response
• Subject has to decide whether a signal is present or not

14. BEKESY’S TRACKING METHOD
• Controlled by the subject using motor driven
attenuator having a push button
• The pushbutton causes the motor to decrease
the sound level when it is depressed and to
increase the level when it is up
• The threshold is thus tracked by the subject
• This method has the advantage of speed and
reasonable precision

15. SIMPLE UP- DOWN OR STAIRCASE METHOD
• Increasing the stimulus when the subject did not respond to
the previous stimulus presentation & decreasing the
intensity when there was a response to the prior stimulus
• It differs from the method of limits- testing does not stop
when the responses change from “yes” to “no” or from
“no” to “yes”
The procedure is continued through at least six to eight
reversals (excluding the first one), and the threshold value is
then calculated as the average of the midpoints of the runs
• Minimizing response bias: by first estimating the threshold
with a larger step size, and then using a smaller step size
(generally half that of the previous one)

16. PARAMETER ESTIMATION BY SEQUENTIAL
TESTING (PEST)
• An adaptive procedure, which uses changes in both
the direction and step size of the stimulus
• If there is a negative response on two successive
presentations in the same direction, then the step size
is doubled for the next presentation
• The direction of the trials is changed after a response
reversal. PEST procedure also halves the step size at
this point
The level at which the next stimulus would have been
presented is taken as the threshold
• Rapid and precise threshold estimation

17. BLOCK UP- DOWN TEMPORAL INTERVAL
FORCED CHOICE (BUDTIF)
• A block of several trials per level replaces
the single trail per stimulus level
• It utilizes a two- alternative forced choice
method
• Example- for 75% point on the
psychometric function, testing is terminated
when response for 3 out of 4 trials is correct
• Stimulus intensity would be raised when the
response is less than ¾ and decreases when
all 4 are correct
The subject is presented with two stimulus intervals during each
trail, and she must indicate which of the interval contains the
stimulus, hence the name ‘Forced Choice, helps to identify
‘false alarms’

18. TRANSFORMED UP- DOWN
PROCEDURE
• Method can be for loudness balances , for
testing various aspects of speech
recognition functions
1. Type 1 - simple up- down method
2. Type 2 - one-up, two-down
3. Type 3 - one-down, two-up
4. Type 4 - one-up, three-down
5. Type 5 – used in the BUDTIF method
• To increase efficiency, start with a large
step size, and then halve it in the target
range for increased precision
Up rule : (−) or (+, −) Down rule : (+, +)

19. STUDY CITATION RESULT
Plattsmier, H. S., & McFadden, D. (1988). Temporary
threshold shift measured with two psychophysical
procedures. Audiology, 27(6), 334-343.
The postexposure estimates of hearing sensitivity obtained
with the two procedures were found to differ statistically. The
values of TTS were equivalent with the two psychophysical
procedures (method of adjustments & adaptive forced-
choice method)
Herrick, R. M. (1973). Psychophysical methodology: VI.
Random method of limits. Perception &
Psychophysics, 13(3), 548-554.
The mean threshold is the same as that obtained in the ML
or in the MCS. For a given number of judgments, the
accuracy of the mean threshold is about the same as the
accuracy obtained with the ML
Campbell, R. A. (1967). BUDTIF thresholds: Variability and
bias. The Journal of the Acoustical Society of America, 42(5),
1149-1149.
Mean thresholds were directly related to block size, as one
would expect from the binomial bias. Thresholds also tended
to become lower with increased trials per run. Efficiency was
generally inversely related to trials per run and directly
related to block size, although the trends are somewhat
mixed
Taylor, M. M., Forbes, S. M., & Creelman, C. D. (1983). PEST
reduces bias in forced choice psychophysics. The Journal of
the Acoustical Society of America, 74(5), 1367-1374.
An experiment compared psychometric functions obtained
from a set of PEST runs using different targets with those
obtained from blocks of fixed‐level trials at different levels.
PEST results were more stable across observers, performance
at all but the highest signal levels was better with PEST, and
the PEST psychometric functions had shallower slopes.

20. STUDY CITATION RESULT
Linschoten, M. R., Harvey, L. O., Eller, P. M., & Jafek, B. W.
(2001). Fast and accurate measurement of taste and smell
thresholds using a maximum-likelihood adaptive staircase
procedure. Perception & psychophysics, 63(8), 1330-1347.
On comparing staircase and Maximum- likelihood PEST
method, the results indicate that the ML-PEST method gives
reliable and precise threshold measurements. Its ability to
detect malingerers shows considerable promise. It is
recommended for use in clinical testing.
Shelton, B. R., Picardi, M. C., & Green, D. M. (1982).
Comparison of three adaptive psychophysical
procedures. The Journal of the Acoustical Society of
America, 71(6), 1527-1533.
All three procedures (BUDTIF, PEST, staircase method)
produced standard errors of measurement that were nearly
equal
Kreft, H. A., DeVries, L. A., Arenberg, J. G., & Oxenham, A. J.
(2019). Comparing rapid and traditional forward-masked
spatial tuning curves in cochlear-implant users. Trends in
Hearing, 23, 2331216519851306.
Twelve postlingually-deafened adult cochlear-implant users
participated. Spatial tuning curves using the new procedure
and using a traditional forced-choice adaptive procedure
resulted in similar estimates of parameters. The Bekesy-
tracking method was almost 3 times faster than the forced-
choice procedure, but its test–retest reliability was
significantly poorer

21. SCALING
Stevens described four scales of measurement:
• lowest order of scaling, least informative
• Eg., “gender” enables us to separate people
into two categories, “male” and “female
Nominal scale
• one class is greater or lesser than another with respect
to the parameter of interest; rank- ordered
• permits the use of the median
Ordinal scale
• possible to use most mathematical operations with
interval data, do not imply a true zero reference point
• Eg., temperature, dates of calendar
Interval scale
• true zero point exists, values can be expressed as
ratios or in decibels, gives the most information about
the data and their interrelationships
• Eg., length, time intervals, loudness (sone), etc.
Ratio scale

22. CLASSES OF SCALING PROCEDURES
Discriminability
(confusion scale)
• subject is asked
to discriminate
small differences
among the
stimuli
• Example: JND or
DL
Category or
partial scale
• subject task is to
divide a range of
stimuli into
equally spaced
categories
• Example:
intensity or
frequency range
Magnitude or
ratio scale
• subject is asked
to estimate ratio
or proportional
relationship
among the
stimuli

23. DIRECT SCALING
In this procedure the subject is asked to establish a relationship between a standard stimulus and a comparison stimulus
that is, the subject must specify a perceptual continuum.
Prothetic continua
• They are additive
• Example: loudness has the characteristics of amount
Metathetic continua
• These are substitutive and not additive
• Example: azimuth has the characteristics of location

24. Ratio
estimation
Presentation of two stimuli
differing in terms of some
parameter
Subject has to express the
subjective magnitude of one
stimulus as a ratio of the other
Ratio
production
The opposite of ratio
estimation; also called as
fractionalization
Subject adjusts the magnitude
of a variable stimulus so that it
sounds like a particular ratio
(or fraction) of the magnitude
of a standard stimulus

25. MAGNITUDE ESTIMATION:
the subject assigns to
physical intensities numbers
that corresponds to their
subjective magnitudes
MAGNITUDE PRODUCTION:
the subject is presented with
numbers and must adjust the
magnitude of stimulus to
correspond to the numbers

26. CROSS
MODALITY
MATCHING
A scaling approach related to
magnitude estimation and
production is called cross-
modality matching
Subject is asked to express the
perceived magnitude for one
sense in terms of another
sensory modality
Example, loudness (an auditory
perception) might be
expressed in terms of apparent
line length (a visual
perception).
Bias effects in magnitude estimation (ME)
and magnitude production (MP) are
minimized by geometric averaging in the
method of psychological magnitude balance
(PMB). Source: Adapted from Hellman and
Zwislocki (1968) with permission of J.
Acoust. Soc. Am.

27. CATEGORY RATING OF LOUDNESS
used in loudness
measurements,
especially hearing
aids
listener assigns
numerical
loudness ratings
to pulsed warble
tone stimuli
using a seven-
point scale

28. STUDY CITATION RESULT
Stevens, S. S., & Galanter, E. H. (1957). Ratio scales and
category scales for a dozen perceptual continua. Journal of
experimental psychology, 54(6), 377.
On prothetic continua (apparent length, duration, area, etc.)
the ratio scale of subjective magnitude approximates a
power function of the physical stimulus. The category scale is
concave downward relative to the ratio scale for
discrimination which is better at one end of the continuum
than at the other. On metathetic continua (visual position,
inclination, pitch, etc.) discrimination (in subjective units) is
constant over the range, although differential familiarity may
introduce nonuniformities
Price, D. D., McGrath, P. A., Rafii, A., & Buckingham, B. (1983).
The validation of visual analogue scales as ratio scale
measures for chronic and experimental pain. Pain, 17(1), 45-
56.
Sensory Visual Analog Scale (VAS) and affective VAS
responses to these temperatures (43, 45, 47, 48, 49 and 51°C)
yielded power functions with exponents 2.1 and 3.8,
respectively; these functions were similar for pain patients
and for volunteers. The power functions were predictive of
estimated ratios of sensation or affect produced by pairs of
standard temperatures (e.g. 47 and 49°C), thereby providing
direct evidence for ratio scaling properties of VAS.

29. References:
• Stanley A. Gelfand, HEARING an introduction to psychological and physiological acoustics. (5th ed.).
• Jack Katz, Ph.D . Handbook of Clinical Audiology (7th ed.).
Questions asked in previous years:
1. Describe a psychophysical method to study difference limen for intensity. Justify your choice of method.
(1994, 1997) (16 m)
2. Short note on the study of psychophysics. (2002, 2011) (4 m)
3. Explain classical and adaptive psychophysical methods. (2011, 2012, 2014, 2017) (16 m)
4. Short note on BUDTIF. (2009, 2017) (4 m)
5. Describe transformed up- down, PEST and maximum likelihood procedures for threshold tracking and
also explain why adaptive procedures are preferred more than classical procedures? (2013, 2015) (16 m)

30. UNIT 2
THEORY OF SIGNAL DETECTION
(Basic concepts- application of signal detection
theory/neural networks)
Submitted to Submitted by
MS. VINI ABHIJITH GUPTA HIMANI BANSAL
MVSCOSH MASLP 1st year

31. DEFINITION & PROCEDURE
SDT is used to analyze data coming from experiments where the task is to categorize ambiguous stimuli which can be
generated either by a known process (called the signal) or be obtained by chance (called the noise in SDT framework).
The theory of signal detection (Swets, 1965; Greene and Swets, 1974; Egan, 1975) provides the best approach to
separate the effects of sensitivity from those of response bias.
The subject were asked to say “yes” when he hears a tone during a test trial and “no” when a tone is not heard; catch trials
included.
Four possible outcomes:
1. HIT: when the signal is present and the subject says “yes.”
2. MISS: the signal is present but the subject says “no.”
3. CORRECT REJECTION: when the signal is absent and the
subject says “no.”
4. FALSE ALARM: the signal is absent but the subject says

32. Hypothetical results in the form of
proportions for 100 test trials actually
containing stimuli and 100 test trials actually
without stimuli (“catch trials”).
The traditional formula to correct the hit rate for chance
success is:
p(hit)corrected = p(hit) − p(false alarm) /1 − p(false alarm)
p(hit)corrected = 0.78 − 0.17 /1.0 − 0.17 = 0.61/ 0.83 = 0.735
The original 78% correct thus falls to 73.5% when we account for
the proportion of the “yes” responses due to chance.

33. SEPARATION OF SIGNAL-NOISE & NOISE ALONE
• The subject must decide whether the stimulation in
the auditory system (e.g., energy) is due to N or SN
• There must always be more energy in SN than in N
• This curve is not affected by response bias
• This separation is measured in terms of a parameter
called d prime (d’ )
Where, x = mean, Ꝺ = SD

34. RECEIVER OPERATING CHARACTERISTIC OR ROC CURVE
• Plot the proportions of hits versus false alarms for each
criterion point, as in the centre of the figure
• Such a graph is called a receiver-operating
characteristic or ROC curve.
• Allows both the effects of sensitivity and response
criterion to be illustrated at the same time
• Sensitivity is shown by the area under the ROC curve
• The response criterion is indicated by the particular
point along the ROC curve

35. PSYCHOPHYSICAL METHODS IN TSD
Yes/No
Methods
• The subject’s task is to say “yes” (“they are different”) or “no” (“they are not different”)
• a single-interval forced-choice experiment
• to choose between signal-plus-noise (one of the alternatives) and noise alone (the other
alternative).
Two- Interval
and N –
Interval Forced
– Choice
Methods
• A test trial consists of two intervals, A and B, presented one after the other
• One of the intervals (SN) contains the signal and the other one (N) does not
• Subject must indicate whether the signal was presented in interval A or in interval B; N
refers to the number of choices.
Confidence
Rating
Methods
• Strict criterion: the subject must have a great deal of confidence in his decision that the
signal is present before he is willing to say “yes.”
• Lax criterion: the subject does not require as much confidence in his “yes” decision
• This method enables the experimenter to obtain several points along the ROC curve
simultaneously

36. IMPLICATIONS OF TSD
1.Audiology 1.Psychology

37. STUDY CITATION RESULT
Haboosheh, R. (2007). Diagnostic auditory brainstem response analysis: Evaluation of
signal-to-noise ratio criteria using signal detection theory (Doctoral dissertation,
University of British Columbia).
The optimal SNR criterion is slightly lower for 500-Hz recordings than for 2000- or
4000-Hz recordings. However, when high-RN recordings were excluded, a SNR
criterion of 0.98 achieved a minimum specificity of 95% for each stimulus frequency,
with sensitivity values ranging from 64%(for 500 Hz) to 79% (for 4000 Hz)
HARRIS, S. R. (1974). HARRIS, Stephen Robert, 1940-A COMPARISON OF SINGLE-
AND MULTI-BAND ATTENTION MODELS BY USE OF SHORT DURATION NOISE
PULSES.
He studied auditory threshold variance at the frequencies 250, 500, 1 K, 2 K, 4 K and
8 K Hz using routine manual audiometric technique and 2 AFC (2 alternative forced
choice ) technique in signal detection paradigm (d’ = 1 is considered as threshold).
The threshold variance was studied over a period of 4 days in the group of 38
subjects and over a period of 3 months in a group of 12 subjects. The variability
reduced when 2 AFC was used, enhancing its reliability
Lindahl, J. T., Encina-Llamas, G., & Epp, B. (2019). Analysis of a forward masking
paradigm proposed to estimate cochlear compression using an auditory nerve
model and signal detection theory. In Proceedings of the International Symposium
on Auditory and Audiological Research (Vol. 7, pp. 445-452).
The simulation results suggest that the estimate of compression based on the
behavioural experiment cannot be derived from sensitivity at the level of the AN but
requires additional contributions, consistent with physiological studies
Moon, H., Han, S. H., & Chun, J. (2015). Applying signal detection theory to
determine the ringtone volume of a mobile phone under ambient
noise. International Journal of Industrial Ergonomics, 47, 117-123.
The results showed that the ringtone volume should increase by 10–15 dB on
average when the noise level increases from 70 dB to 80 dB. When adjusting the
volume according to the ambient noise level, the volume should be changed
differently according to the frequencies of a ringtone. The ringtone should be
composed of low-frequency sounds under loud ambient noise because the subjects
were very sensitive to the pure tone with frequency of 500 Hz

38. STUDY CITATION RESULT
Callan, D. E., Lasky, R. E., & Fowler, C. G. (1999). Neural networks applied to
retrocochlear diagnosis. Journal of Speech, Language, and Hearing Research, 42(2),
287-299.
The results also demonstrated that identification accuracy could be improved by
combining the ABR with other tests (in this case contralateral acoustic reflex at 2000
Hz, ipsilateral acoustic reflex at 2000 Hz, tone decay, and word recognition score).
Further, it was demonstrated that performance could be improved over that
obtained using dichotomous test measures (i.e., positive or negative presence of
pathology) by using raw test measures in conjunction with ABR.
Kellen, D., Klauer, K. C., & Singmann, H. (2012). On the measurement of criterion
noise in signal detection theory: the case of recognition memory. Psychological
Review, 119(3), 457.
A reanalysis of Benjamin et al.'s (2009) data sets as well as the results from a new
experimental method indicate that the different forms of criterion noise proposed in
the recognition memory literature are of very low magnitudes, and they do not
provide a significant improvement over the account already given by traditional SDT
without criterion noise
Tuzlukov, V. P. (1998). A new approach to signal detection theory. Digital Signal
Processing, 8(3), 166-184.
Theoretical and experimental investigations carried out by the author lead to the
conclusion that the suggested signal processing approach allows one to formulate a
decision-making rule based on the definition of the jointly sufficient statistics of
mean and variance of the likelihood function (or functional)
Osada H, Kaku K, Masuda K, Iitsuka Y, Seki K, Sekiya S. Quantitative and qualitative
evaluations of fetal lung with MR imaging. Radiology 2004; 231:887 –892
Used ROC analysis to assess the ability of MRI to predict fetal pulmonary hypoplasia.
They imaged 87 fetuses, measuring both lung volume and signal intensity. An ROC
curve based on lung volume showed that lung volume has some ability to
discriminate between fetuses who will have good versus those who will have poor
respiratory outcome after birth. An ROC curve based on the combined information
from lung volume and signal intensity, however, has superior accuracy

39. References:
• Stanley A. Gelfand, HEARING an introduction to psychological and physiological acoustics. (5th ed.).
• Jack Katz, Ph.D . Handbook of Clinical Audiology (7th ed.).
Questions asked in previous years:
1. Describe how signal detection paradigm can be used to differentiate cochlear and retrocochlear
disorders. What are some other applications of signal detection audiometry? (1993, 1994, 1997, 1998,
2000, 2002, 2003, 2009, 2011, 2014, 2017, 2018, 2019, 2021) (16 m)
2. Short notes on: implications of signal detection theory in Audiology. (1994) (4 m)
3. Short note on ROC. (2004, 2007, 2009) (4 m)
4. Elaborate on the theory of signal detection. What is stereophonic effect? (2004) (16 m)