ARTICLE IN PRESS
2 J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx
reason for this discrepancy may relate to the specific type cation-naïve patients have been reported to have extra-
of IL being studied such as priming, visuomotor, and pyramidal symptoms, suggesting intrinsic striatal
problem solving tasks. Sequence learning tasks appear dysfunction (Caligiuri and Lohr, 1997; Chatterjee et al.,
most likely to reveal deficient learning in schizophrenia 1995; Cortese, 2001; Puri et al., 1999). Further, Granholm
(Bressi et al., 1998; Green et al., 1997; Kumari et al., 2002; et al. (1993) demonstrated that the severity of tardive
Schwartz et al., 2003). Consequently, implicit sequence dyskinesia in schizophrenia was associated with both
learning may be dependent on a particular neural substrate shortened caudate T2 relaxation times and decreased IL.
that is impaired in schizophrenia. More recently, Giménez et al. (2003) found an inverse
It is also possible that the periodically observed IL correlation between caudate metabolism (NAA/Cho
deficits may relate to the effect of dopamine antagonist ratio) and IL in schizophrenia.
antipsychotic medication rather than the underlying illness. There has been one fMRI investigation to date that has
Impaired IL has been reported in healthy participants evaluated IL in participants with schizophrenia (Kumari
administered haloperidol (Kumari et al., 1997; Peretti et al., et al., 2002). In this study, using a serial reaction time task,
1997), and chlorpromazine (Danion et al., 1992), with six healthy controls demonstrated significantly greater
enhanced IL occurring with the dopamine agonist, amphet- activations in the striatum, thalamus, precuneus, and sen-
amine (Kumari et al., 1997). Investigations have demon- sorimotor regions than the six schizophrenia participants
strated largely intact IL in schizophrenia with patients who were taking typical antipsychotic medication. Impor-
taking atypical antipsychotic medication, but impaired IL tantly, while the controls demonstrated robust evidence of
with the older, typical class of antipsychotic medication IL, the patients did not whatsoever. This marked dis-
(Paquet et al., 2004; Schérer et al., 2004; Stevens et al., crepancy in task performance limits the interpretation of
2002). Moreover, IL has improved in schizophrenia neural activation differences (Rauch et al., 2001) because
patients after being switched to an atypical antipsychotic it can be argued that the two groups were not engaged in
(Purdon et al., 2000). the same cognitive processes while undergoing fMRI, i.e.
Appropriate functioning of the striatum and its connec- the controls were learning but the patients were not.
ted neural substrates are thought to be critical for adequate Given these discrepant findings on IL in schizophre-
performance in IL tasks (Curran, 1998). Individuals with nia, the current study evaluates the neural basis of an
striatal pathology, such as in Huntington's (Knopman and implicit sequence learning task in schizophrenia partici-
Nissen, 1991; Willingham and Koroshetz, 1993) and pants taking atypical antipsychotics and who demonstrat-
Parkinson's disease (Doyon et al., 1997; Pascual-Leone ed equivalent learning to controls. We chose to utilize a
et al., 1993), demonstrate impaired IL. Functional imaging sequence learning task, as this type of IL has been the
studies have reported striatal activation in participants most likely to reveal differential performance in patients
who displayed IL (Grafton et al., 1995; Peigneux et al., with schizophrenia as compared to healthy controls
2000; Martis et al., 2004; Rauch et al., 1995, 1997a,b, (Bressi et al., 1998; Green et al., 1997; Kumari et al.,
1998), and a recent paper has further consolidated this 2002; Schwartz et al., 2003). In addition to a whole-brain
relationship with the use of a control group (Reiss et al., exploratory analysis, we conducted an a priori region of
2005). Other motor-associated areas have been correlated interest analysis of the striatum, given its pivotal role in IL
with IL including primary motor cortex, premotor cortex, and postulated role in schizophrenia. We hypothesized
supplementary motor area and the cerebellum (Grafton that the schizophrenia group would demonstrate less
et al., 1992, 1995; Jueptner et al., 1997). More perceptually striatal activation than the control group, despite similar
oriented areas have also been associated with IL such as behavioral performance on the IL task.
parietal and temporal cortices, as well as the hippocampus
(Curran, 1998; Schendan et al., 2003), though all these 2. Materials and methods
areas are less consistently reported than the striatum.
Typical antipsychotics bind to striatal dopamine recep- 2.1. Participant selection
tors thereby potentially impairing IL, whereas atypical
antipsychotics have much less effect on the striatum (Blin, We examined 20 right-handed (Oldfield, 1971) volun-
1999; Chiodo and Bunney, 1983; Grace, 1992; Kapur teers, 10 with schizophrenia, and 10 age, gender, and
et al., 1995; Kapur, 1998). Yet there is reason to suspect education matched healthy controls, through outpatient
striatal abnormalities exist in schizophrenia itself. Struc- clinic contacts and poster advertisements. Task perfor-
tural imaging studies with medication-naïve patients have mance was comparable between the two groups (Table 1),
reported striatal abnormalities (Corson et al., 1999; Gur and so no participant selection actions were needed. The
et al., 1998; Lang et al., 1999, 2001). Moreover, medi- study and all procedures were conducted with the approval
ARTICLE IN PRESS
J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx 3
of the University of Manitoba and National Research an asterisk. A 25-trial practice run was followed by two
Council, Institute for Biodiagnostics, research ethics 312-trial runs (each lasting 6.25 min). Each run consisted of
boards. All participants provided signed, informed consent. alternating blocks of random (R) and IL conditions brack-
eted by a 19.2-second cross-hair fixation block. In the IL
2.2. Sample characteristics condition, the stimuli followed a 12-item sequence (posi-
tion: 1-2-1-4-2-3-4-1-3-2-4-3) that repeated six times. In
Administration of the MINI International Diagnostic the R condition, 24 trials were pseudo-randomly presented
Interview version 5.0 (Sheehan et al., 1998) for DSM-IV with no immediately repeated locations. No signs were
(American Psychiatric Association, 1994) confirmed that given to distinguish between R and IL blocks. For each
all clinical participants had a diagnosis of schizophrenia participant, a median reaction time (RT) score was calcu-
with no co-morbid conditions. Clinical participants were lated for each condition, R and IL, and an IL reaction time
assessed with the Positive and Negative Syndrome Scale advantage (RTa) score [100 ⁎ (R - IL) /R]. RT estimates
for Schizophrenia (PANSS) (Kay et al., 1987) which were based on correct trial responses.
provided a total, as well as subscale scores for positive,
negative, and general psychopathology symptoms. 2.4. Debriefing
Schizophrenia illness duration ranged from 3 to
41 years (mean and SD in Table 1). All schizophrenia Following the scan, participants were informed of and
participants were on stable doses of atypical antipsychotic asked to reproduce the repeating sequence by making 15
medication, which were converted to chlorpromazine key presses (Martis et al., 2004; Rauch et al., 1997b). The
equivalent doses (Woods, 2002) (see Table 1). Partici- longest string that matched the 12-item implicit sequence
pants with schizophrenia did not differ significantly from estimated the degree of explicit learning. Scores ≥ 7 were
healthy controls on age, gender, or education (Table 1). categorized as explicit learning (see Martis et al., 2004).
2.3. Task 2.5. MRI acquisition
We used the serial reaction time (SRT) task (Martis Images were acquired on a General Electric 1.5 Tesla
et al., 2004, Rauch et al., 1997a,b) for which MRI-com- MRI system with a clinical transmit/receive head coil.
patible eye-pieces displayed an asterisk in one of four Anatomical scans for each individual used a T1-weighted
horizontally arranged boxes for 1.0 s, followed by a 0.2 fast spin-echo sequence (TR/TE/FOV = 400/8 ms/25 cm,
second empty-boxes interval. Using separate fingers, parti- 2 mm slice thickness) for a 2 × 1 × 1 mm voxel image.
cipants pressed one of four keys in immediate response to Functional MRI data used a single-shot gradient–recalled-
Sociodemographic, clinical status, and behavioral performance of the participants
Schizophrenia (n = 10) Control (n = 10) t(18) p Grand mean
mean (SD) mean (SD) (n = 20)
Gender: female/male 1/9 1/9 – – 2/18
Age (years) 29.1 (9.4) 26.1 (7.1) 0.80 .43 27.6 (8.3)
Education (years) 12.8 (2.5) 13.4 (2.4) −0.54 .59 13.1 (2.4)
Medication dosage (chlorpromazine equivalents mg/d) 330 (180) n/a
PANSS 69.0 (8.2)a n/a
Illness duration (years) 11.5 (12) n/a
Serial reaction time task behavior performance
Reaction time (ms)b
Random 415.5 (77.0) 411.7 (53.9) −0.13 .90 413.6 (64.7)
Implicit learning 398.7 (75.6) 389.7 (55.3) −0.31 .76 394.2 (64.6)
Reaction time advantage (%)c 4.1 (2.9) 5.4 (4.7) −0.75 .46 4.8 (3.9)
Error rates 4.9 (4.1) 2.9 (1.9) −1.4 .19 3.9 (3.3)
Explicit learning scores 3.7 (0.7) 4.3 (1.3) 1.3 .22 4.0 (1.1)
n = 9; bReaction time based on the mean of median reaction times in milliseconds with the SD in brackets; c% RT advantage = 100 ⁎ (baseline RT − implicit
RT)/ baseline RT.
ARTICLE IN PRESS
4 J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx
echo echo planar imaging sequence with TR/TE/flip and group differences in IL-related BOLD responses
angle/FOV = 2400/60 ms/45°/25 cm. Approximately 80% (group by condition interactions). ANOVA cluster aver-
of the brain was imaged using 23 contiguous axial slices ages were imported into SPSS for descriptive statistic and
and 3.9 mm isotropic voxels excluding the most superior simple effect analyses. To supplement these analyses, we
and inferior brain regions. In each experimental run, tested for degree of IL-related (RTa) BOLD responses
172 volumes were acquired (344 volumes in total). within each group as well as RTa by group interactions. All
areas of activation are reported in Montreal Neurological
2.6. fMRI individual analyses Institute coordinates with positive values as right, anterior,
and superior. For each IL-related activation cluster iden-
Image analyses were performed with Analysis of tified, we tested for correlations between BOLD responses
Functional NeuroImages (AFNI) software (Cox, 1996). and PANSS scores and medication dosage.
The fMRI data were slice time-corrected. Functional vol-
umes were co-registered to the volume that minimized 3. Results
translation and rotation relative to all other volumes. The
first 3 volumes were excluded to allow for equilibrium of 3.1. Task performance
the MRI signal. One additional volume (out of 344) with
visually detectable movement (0.3 degrees of pitch ro- Schizophrenia patients and controls did not differ on
tation and 0.4 mm) was excluded in one person with mean RT or RTa (Table 1). Each group showed faster
schizophrenia. Run 2 was concatenated to Run 1. The mean RT during the IL condition compared to the random
block design time-series was convolved to account for the condition [schizophrenia, t(9) = 4.5, p b .001, and controls,
hemodynamic response function. Multiple regression an- t(9) = 3.6, p b .006]. Total error rates (commissions and
alysis included nine regressors: 2 convolved IL and R omissions) and mean explicit learning estimates did not
condition regressors, 3 head-motion covariates (i.e., roll, differ between groups (Table 1). There was no correlation
pitch, and yaw), 2 slope regressors (1 per Run), and between symptom severity, i.e. PANSS total and subscale
2 intercept (fixation cross-hair blocks) regressors (1 per scores, and task performance measures, i.e. RTa and per-
Run). The beta coefficient for each condition (IL and R) cent correct (all p valuesN .70). As well, medication dos-
was divided by the mean fixation cross-hair coefficient to age was not correlated with task performance RTa scores,
estimate percent blood oxygen level dependent (BOLD) p N .70, or percent correct, p N .35.
signal change for each condition (Cohen and DuBois,
1999). Beta values represent partial signal change when 3.2. Common IL vs. R BOLD responses
simultaneously also taking into account the baseline, linear
drift, and movement nuisance regressors. To account for Three brain regions were identified showing greater
anatomical variations, an 8 mm full-width half-maximum BOLD response to IL trials compared to R trials in both
Gaussian filter was applied. Data were transformed into groups (Table 2). Post hoc simple effect analyses
the Montreal Neurological Institute (MNI) coordinates.
2.7. Alpha control criteria Table 2
Brain regions showing greater blood oxygen level dependent (BOLD)
responses to the IL minus R condition in both groups (common
For the whole brain analyses, Monte Carlo simula- condition effect)
tions estimated that a voxel-wise probability of p b 0.02
Brain region BA x y z Vol. F Mn IL - R
and a minimum cluster volume of 640 μl would main- (μl) (1,18)
tain an overall p b 0.05. For the striatum (defined via the HC Sz
AFNI software atlas) Region of Interest (ROI) analyses, L middle 22 − 50 − 34 3 2432 16.0 0.11⁎ 0.09⁎
a more liberal criterion was used. ROI Monte Carlo temporal
simulations estimated that a voxel-wise probability of R superior 38 38 4 −14 1856 18.7 0.15⁎ 0.18⁎
p b 0.04 and a minimum cluster volume of 448 μl would temporal
L inferior 47 − 38 20 −8 1344 20.3 0.08⁎ 0.12⁎
maintain an overall p b 0.05.
2.8. fMRI between- and within-group analyses BA = Brodman Area, x, y, z = Montreal Neurological Institute
coordinates of maximally activated voxel, Vol. = volume, L = left,
R = right, Mn IL − R = mean difference of implicit learning minus
A 2 × 2 repeated measures ANOVAwas used to identify random condition % BOLD scores. ⁎ = Significantly different from
common IL-related BOLD responses (condition effect) zero, p b 05.
ARTICLE IN PRESS
J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx 5
Table 3 (Table 3b). For the schizophrenia group, these regions
Brain regions showing group differences in blood oxygen level did not differ significantly in the IL vs. R contrast.
dependent (BOLD) responses to the IL minus R contrast (group by
condition interaction effect)
3.5. Striatal region of interest group differences
Brain region BA x y z F Vol. Mn IL − R
HC Sz The left caudate showed a significant group-by-task
condition interaction in BOLD response (Table 4a and
a. Group by condition interactions
L medial 8 − 6 38 46 14.0 1792 0.11⁎ − 0.08⁎ Fig. 1a). Simple effect post hoc analyses determined that
frontal the control group exhibited significantly higher BOLD
R/L 2 − 6 17 16.3 1280 0.13⁎ − 0.13⁎ responses to the IL condition than the R condition, while
thalamus the schizophrenia group exhibited a non-significant de-
crease to the IL compared to the R condition (Table 4a).
L precentral 6 − 42 − 7 43 20.9 1088 0.06⁎ − 0.09⁎ This region partially overlaps with the thalamic/caudate
R superior 41 59 −26 12 19.7 704 0.10⁎ − 0.04 region identified in the whole brain interaction analysis.
temporal The control group showed a significantly greater BOLD
L culmen 30 − 6 −46 3 16.4 640 0.12⁎ − 0.10⁎ response to the IL vs. R contrast in a second caudate region
L posterior 30 − 14 −63 10 22.7 640 0.10⁎ − 0.10
(Table 4b and Fig. 1b). This region showed no significant
R cingulate 24 14 − 3 43 12.3 640 0.07⁎ − 0.08⁎ difference among the schizophrenia group (Table 4b). In
R precentral 4 38 −20 42 12.8 640 0.09⁎ − 0.08⁎ these two regions, the control group showed greater
absolute BOLD responses to the IL condition (p b .01) as
b. Group-specific condition effect well as greater IL vs. R relative BOLD responses (Fig. 1).
L cuneus 19 − 2 −85 35 1280 16.9 0.21⁎ 0.06
R caudate 22 −47 15 832 13.9 0.10⁎ 0.06
tail 3.6. Region of interest RTa-BOLD associations
BA = Brodman area, x, y, z = Montreal Neurological Institute
coordinates of maximally activated voxel, IL = implicit learning, R = In the voxel-by-voxel ROI analysis, among the control
random condition, Vol. = volume, L = left, R = right, Mn IL − R = mean group, degree of IL was associated with greater activation
difference of IL minus R % BOLD scores. ROI = region of interest in the left striatum (volume = 1088 μL), r = .87, p b .001,
analysis, ⁎ = significantly different from zero, p b .05. with the maximum intensity voxel located in the left
putamen (x = −18, y = 12, z = −4, t(8) = 3.1, pb .02). In the
confirmed that each group separately exhibited a voxel-by-voxel ROI analyses for the schizophrenia group
significant BOLD response in these three brain regions. only a trend level cluster was identified (overall p b .1).
For this trend level cluster, degree of IL was associated
3.3. Group differences in IL vs. R BOLD responses with greater average cluster BOLD response within the
left caudate (volume = 512 μL), r = .71, p b .03, with the
Eight brain regions were identified showing significant maximum intensity voxel at the MNI coordinates of x =
group-by-task condition BOLD responses (Table 3a). −14, y = 11, z = 9 (t(8) = 2.6, p b .05).
Simple effect post hoc analyses determined that the con-
trol group exhibited significantly higher BOLD responses Table 4
to the IL condition than the R condition in all eight regions Striatal region of interest group differences in blood oxygen level
(Table 3a). In contrast, the schizophrenia group exhibited dependent (BOLD) responses to the IL minus R condition
lower BOLD responses to the IL condition than the R BA x y z Vol. F (1,18) Mn IL − R
condition in these regions with significant decreases in six (μl)
of the eight regions (Table 3a). Note the thalamic cluster HC Sz
of differential activation included the left caudate. a. Group by condition interaction
L caudate − 10 −6 17 448 8.5 0.12⁎ − 0.06
3.4. Other control group activations in IL vs R BOLD
b. Group-specific condition effect
L caudate − 14 11 18 512 8.0 0.11⁎ 0.04
Whole brain repeated measures ANOVA identified BA = Brodman area, x, y, z = Montreal Neurological Institute
coordinates of maximally activated voxel, IL = implicit learning, R =
two additional regions that showed a significant random condition, Vol. = volume, L = left, R = right, Mn IL − R = mean
BOLD response in the control group for the IL vs. R difference of IL minus R condition % BOLD scores. ROI = region of
contrast, namely the left cuneus and the right caudate tail interest analysis, ⁎ = significantly different from zero, p b .05.
ARTICLE IN PRESS
6 J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx
Fig. 1. Group differences in blood oxygen level dependent (BOLD) responses to the implicit learning (IL) condition compared to the random (R)
condition in the striatum region-of-interest analyses. The group by condition interaction is labeled “a” and the control-specific condition effect is
labeled “b”. The 95% confidence intervals in the bar charts reveal a significant IL minus R (IL − R) effect for the control group (HC) and a non-
significant IL effect for the schizophrenia group (Sz).
3.7. Other within-group RTa-BOLD associations action cluster was found that largely overlaps with the left
ACC cluster associated with greater IL in the schizophre-
In the whole brain analysis, among the control group no nia group. Thus, the positive association between degree
clusters of activation were associated with the degree of IL. of IL and left ACC BOLD response is unique to the
In contrast, among the schizophrenia group, degree of IL schizophrenia group.
was associated with BOLD responses in two regions: the
left dorsolateral prefrontal cortex (DLPFC, volume = 3.9. Schizophrenia symptom severity and medication
1216 μL), r= .95, p b .001, and the left anterior cingulate dosage correlations to BOLD responses
cortex (ACC, volume= 832 μL), r= .84, p b .002. More
specifically, degree of IL was associated with greater Schizophrenia symptom severity estimates (PANSS
BOLD responses in the left DLPFC (BA 10) with the total and subscale scores) were not correlated with any
maximum intensity voxel at the MNI coordinates of x = of the BOLD responses in the patient group, p N .20.
−34, y= 43, z =20 (t(8)= 3.0, pb .02), and the left ACC (BA Similarly, medication dosage was not correlated to any
24,33) with the maximum intensity voxel at the MNI of the BOLD responses, p N .15.
coordinates of x =−2, y = 14, z= 23 (t(8) =3.4, p b .01).
3.8. RTa-BOLD interactions
This investigation found that schizophrenia patients
Interactions between degree of IL (RTa scores) and show reduced striatal activation relative to performance-
group were tested using both the whole brain and ROI matched healthy comparison subjects during an IL task,
analyses criteria. No significant clusters were identified. which supports our hypothesis of deficient recruitment of
However, a borderline significant (overall p b .1) inter- the striatum.
ARTICLE IN PRESS
J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx 7
The controls demonstrated a distributed neural network Our patients were taking only atypical antipsychotic
of IL-related activation, invoking frontostriatal circuits, medication, with the expectation of them demonstrating
multiple temporal lobe areas, and motor cortices, all areas intact IL. Consistent with this assumption, a recent
previously reported with this type of task (Grafton et al., neuroimaging study involving IL in schizophrenia (Paquet
1995; Peigneux et al., 2000; Rauch et al., 1995, 1998; et al., 2004) used SPECT to assess the effects of two
Schendan et al., 2003; Thomas et al., 2004). In contrast, the antipsychotic drugs on IL and striatal dopamine (D2)
schizophrenia participants, despite equivalent behavioural receptor occupancy. The patients receiving haloperidol, a
performance, recruited fewer brain regions, restricted to typical antipsychotic, demonstrated impaired IL, whereas
the middle and superior temporal, and inferior frontal those receiving olanzapine, an atypical antipsychotic, per-
cortices. Of particular note, participants with schizophre- formed equivalently to controls. Additionally, in the
nia failed to activate the striatum to the same degree as haloperidol-treated patients significant correlations were
controls, this area being considered a pivotal part of the IL found between IL deficits and striatal D2 receptor occu-
neural network (Doyon et al., 1997; Peigneux et al., 2000; pancy. Similarly, the greater striatal D2 binding of
Rauch et al., 1997a,b; Reiss et al., 2005; Willingham and risperidone relative to other atypical antipsychotics has
Koroshetz, 1993). This functional neuroimaging finding been forwarded as the explanation for impaired IL on
complements previous evidence of striatal abnormalities risperidone-treated patients in studies also involving
in schizophrenia, as summarized in the Introduction. Not- olanzapine (Purdon et al., 2003) and clozapine (Schérer
withstanding, the striatal region-of-interest RTa correlation et al., 2004). It is unknown if our participants with
analysis trend-level result in the schizophrenia group is schizophrenia had significant striatal D2 blockade or to
suggestive of some degree of striatal recruitment in task what extent this might contribute to the observed relative
performance in schizophrenia. Moreover, the schizophre- decreased striatal activation. Moreover, the present investi-
nia whole-brain RTa correlation analysis results are pos- gation cannot directly address whether IL is intact in
sibly reflective of this group having their IL performance schizophrenia per se or behaviorally normalized by the
mediated through non-striatal regions (ACC and DLPFC). usage of atypical antipsychotic medication.
Taken together, in both healthy controls and especially in In an evoked potential investigation of colour patterns,
schizophrenia subjects, IL task performance is not ex- Hsieh et al. (2004) recently reported IL deficits in
clusively dependent on striatal involvement. schizophrenia which were correlated with less gating of
The finding of the schizophrenia group invoking the the P50 component. Abnormal P50 evoked potentials
ACC and DLPFC in relation to improving task perfor- have also been reported in both Parkinson's (Teo et al.,
mance is intriguing. Both the ACC (Kumari et al., 2002; 1997) and Huntington's (Uc et al., 2003) diseases, both
Peigneux et al., 2000; Rauch et al., 1997b) and the DLPFC degenerative disorders of striatal pathology. These fin-
(Grafton et al., 1995; Schendan et al., 2003) have been dings are consistent with an overall linkage between IL,
associated with the conduct of implicit sequence tasks. striatal function, and schizophrenia as reported in this
Additionally, Snitz et al. (2005) recently demonstrated study.
hypoactivation of the DLPFC and ACC in unmedicated The findings of the current study contrast with those
first-episode patients doing a cognitive task designed to of Kumari et al. (2002), the only other fMRI study of
dissociate the two regions. After 4 weeks of treatment with IL in schizophrenia to our knowledge. However, our
atypical antipsychotics, a subset of these patients was study involved a larger sample and was comprised of
rescanned and demonstrated increased/normalized acti- patients taking only atypical antipsychotics, whereas in
vation of the ACC, but not of the DLPFC. However, the earlier study all patients were on typical anti-
Honey et al. (1999) reported increased ACC/supplemen- psychotics. Conceivably, the Kumari et al. (2002)
tary motor area and DLPFC activation in a working findings may reflect an effect of typical antipsychotics
memory task when patients were switched from typical (as opposed to being an effect of illness) in that the
antipsychotic medications to risperidone. Moreover, it has schizophrenia participants performing a similar implicit
been shown that both clozapine (Lahti et al., 2004) and sequence-learning paradigm demonstrated minimal
olanzapine (Lahti et al., 2005) enhance ACC regional neural activation. However, an equally plausible ex-
cerebral blood flow when given to medication-free pa- planation would be that their schizophrenia group
tients. Hence, our schizophrenia group demonstrated showed no behavioural evidence of IL. The authors
normal behavioural performance which was conceivably acknowledged these points as well making the
partially mediated through the ACC and DLPFC, and in recommendation that further fMRI studies be done in
turn this effect may have been enabled by atypical anti- participants receiving atypical antipsychotic drugs,
psychotic medication. such as the present one.
ARTICLE IN PRESS
8 J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx
The present study may shed light on the divergent In conclusion, we found a relative lack of striatal
literature on IL in schizophrenia. Most, but not all studies activation in schizophrenia patients on atypical antipsy-
report intact IL in schizophrenia, with sequence learning chotic medication during an implicit sequence-learning
studies being the type most likely to yield differential task, despite similar behavioral performance to healthy
behavioural performance. The present investigation pro- controls. This result is consistent with convergent evi-
duced results that are suggestive of a neural inefficiency dence of striatal dysfunction in schizophrenia.
in schizophrenia during implicit sequence learning.
Despite equivalent behavioural performance, the schizo- Acknowledgments
phrenia group did not invoke an as distributed network as
the controls, and in particular failed to equally recruit the Grant support was provided by the Health Sciences
striatum. This proposed neural inefficiency might ac- Centre and the St. Boniface Hospital Research Founda-
count for findings of intact IL in those studies where task tions. The authors thank Brenda Stoesz, Stephen Hassard,
demands are relatively modest (as perhaps in the present Michael McIntyre, Kristina Malisza, Patrick Stroman,
study), but result in impaired IL when the sequencing Wolfgang Richter, Corinne LeBlanc, and Deanna Braun
task demands exceed the capability of the deficient for their contributions.
Another aspect of the discrepancies regarding IL References
findings in schizophrenia is the heterogeneity of the
schizophrenia groups between studies and of the illness American Psychiatric Association, 1994. Diagnostic and Statistical
Manual of Mental Disorders, Fourth ed. American Psychiatric
itself. It has been adduced that schizophrenia is a hetero- Association, Washington, DC.
geneous illness (Clark et al., 1995; Tsuang et al., 1990), Bazin, N., Perruchet, P., 1996. Implicit and explicit associative
such that only some patients may have IL impairment as memory in patients with schizophrenia. Schizophr. Res. 22 (3),
an endophenotype, which could account for the majority 241–248.
Bazin, N., Perruchet, P., Féline, A., 1996. Mood congruence effect in
of studies reporting normal IL.
explicit and implicit memory tasks: a comparison between
Previous studies have reported associations between depressed patients, schizophrenic patients and controls. Eur.
clinical symptoms, particularly negative symptoms and IL Psychiatry 11 (8), 390–395.
(Schroder et al., 1996; Schwartz et al., 2003). We found no Blin, O., 1999. A comparative review of new antipsychotics. Can.
correlations between symptoms and behavioural mea- J. Psychiatry 44 (3), 235–244.
sures of IL, nor with the identified neural activations. This Bressi, S., Miele, L., Bressi, C., Vita, A., Astori, S., Gimosti, E., Linciano,
A.D., Sessini, M., Sferlazza, A., Zirulia, G., 1998. Sequence skill
may be due to a lack of statistical power, a potential learning and semantic priming in schizophrenia. Evidence for a
limitation of our study. However, this would have been differential impairment of the implicit memory system. New Trends
more of a concern if there had been a failure to differen- in Experimental and Clinical Psychiatry, vol. 14 (4), pp. 179–188.
tiate the schizophrenia group from the healthy controls Caligiuri, M.P., Lohr, J.B., 1997. Instrumental motor predictors of
(Type 2 error) in neural responses. Other limitations of neuroleptic-induced parkinsonism in newly medicated schizophre-
nia patients. J. Neuropsychiatry Clin. Neurosci. 9 (4), 562–567.
this investigation are the predominance of males, and the Chatterjee, A., Chakos, M., Koreen, A., Geisler, S., Sheitman, B.,
lack of a medication-naïve or -free schizophrenia group. Woerner, M., Kane, J.M., Alvir, J., Lieberman, J.A., 1995.
Future fMRI studies should address the abovemen- Prevalence and clinical correlates of extrapyramidal signs and
tioned limitations. Ideally, an adequately powered 4-arm spontaneous dyskinesia in never-medicated schizophrenic patients.
Am. J. Psychiatry 152 (12), 1724–1729.
design with a medication-free patient group, two with
Chiodo, L.A., Bunney, B.S., 1983. Typical and atypical neuroleptics:
patients taking antipsychotic medication (typical and differential effect of chronic administration on the activity of A9 and
atypical medication groups), and a healthy control A10 midbrain dopamine neurons. J. Neurosci. 3 (8), 1607–1619.
group, would address the neural mechanisms of IL in Clare, L., McKenna, P.J., Mortimer, A.M., Baddeley, A.D., 1993.
schizophrenia while accounting for (helpful or harmful) Memory in schizophrenia: what is impaired and what is preserved?
medication effects. Additionally, in order to more fully Neuropsychologia 31 (11), 1225–1241.
Clark, L.A., Watson, D., Reynold, S., 1995. Diagnosis and
explore the neural correlates of IL in schizophrenia, classification of psychopathology: challenges to the current system
imaging studies with a broader variety of IL tasks, or and future directions. Annu. Rev. Psychol. 46, 121–153.
one of incremental difficulty could be utilized. If it is Cohen, M.S., DuBois, R.M., 1999. Stability, repeatability, and the
probable that most patients with schizophrenia are ca- expression of signal magnitude in functional magnetic resonance
pable of demonstrating some degree of IL, then by imaging. J. Magn. Reson. Imaging 10, 33–40.
Corson, P.W., Nopoulos, P., Andreasen, N.C., Heckel, D., Arndt, S.,
parametrically adjusting the task a more comprehensive 1999. Caudate size in first-episode neuroleptic-naive schizophrenic
understanding of the neural processing of IL in schizo- patients measured using an artificial neural network. Biol.
phrenia could be achieved. Psychiatry 46 (5), 712–720.
ARTICLE IN PRESS
J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx 9
Cortese, L., 2001. Movement disorders and clinical correlates in drug- Gur, R.E., Maany, V., Mozley, P.D., Swanson, C., Bilker, W., Gur, R.C.,
naive first episode patients with psychosis. Mov. Disord. 16 (suppl 1), 1998. Subcortical MRI volumes in neuroleptic-naive and treated
S51 (abstract). patients with schizophrenia. Am. J. Psychiatry 155 (12),
Cox, R.W., 1996. AFNI: software for analysis and visualization of 1711–1717.
functional magnetic resonance neuroimages. Comput. Biomed. Heinrichs, R.W., Zakzanis, K.K., 1998. Neurocognitive deficit in
Res. 29 (3), 162–173. schizophrenia: a quantitative review of the evidence. Neuropsy-
Curran, T., 1998. Implicit sequence learning from a cognitive chology 12 (3), 426–445.
neuroscience perspective. What, how, and where? In: Stadler, M.A., Honey, G.D., Bullmore, E.T., Soni, W., Varatheesan, M., Williams, S.C.R.,
Frensch, P.A. (Eds.), Handbook of Implicit Learning. Sage, Newbury Sharma, T., 1999. Differences in frontal cortical activation by a work-
Park, CA, pp. 365–400. ing memory task after substitutional of risperidone for typical anti-
Danion, J.M., Peretti, S., Grangé, D., Bilik, M., Imbs, J.L., Singer, L., psychotic drugs in patients with schizophrenia. Proc. Natl. Acad. Sci.
1992. Effects of chlorpromazine and lorazepam on explicit 96 (23), 13432–13437.
memory, repetition priming and cognitive skill learning in healthy Hsieh, M.H., Liu, K., Liu, S.K., Chiu, M.J., Hwu, H.G., Chen, A.C.,
volunteers. Psychopharmacology (Berl.) 108 (3), 345–351. 2004. Memory impairment and auditory evoked potential gating
Danion, J.M., Meulemans, T., Kauffmann-Muller, F., Vermaat, H., deficit in schizophrenia. Psychiatry Res.: Neuroimaging 130 (2),
2001. Intact implicit learning in schizophrenia. Am. J. Psychiatry 161–169.
158 (6), 944–948. Jueptner, M., Frith, C.D., Brooks, D.J., Frackowiak, R.S., Passingham,
Dominey, P.F., Georgieff, N., 1997. Schizophrenics learn surface but R.E., 1997. Anatomy of motor learning. II. Subcortical structures
not abstract structure in a serial reaction time task. NeuroReport and learning by trial and error. J. Neurophysiol. 77 (3), 1325–1337.
8 (13), 2877–2882. Kapur, S., 1998. A new framework for investigating antipsychotic
Doyon, J., Gaudreau, D., Laforce Jr., R., Castonguay, M., Bedard, P.J., action in humans: lessons from PET imaging. Mol. Psychiatry
Bedard, F., Bouchard, J.P., 1997. Role of the striatum, cerebellum, 3 (2), 135–140.
and frontal lobes in the learning of a visuomotor sequence. Brain Kapur, S., Remington, G., Zipursky, R.B., Wilson, A.A., Houle, S.,
Cogn. 34 (2), 218–245. 1995. The D2 dopamine receptor occupancy of risperidone and its
Fioravanti, M., Carlone, O., Vitale, B., Cinti, M.E., Clare, L., 2005. A relationship to extrapyramidal symptoms: a PET study. Life Sci. 57
meta-analysis of cognitive deficits in adults with a diagnosis of (10), 103–107.
schizophrenia. Neuropsychol. Rev. 15 (2), 73–95. Kay, S.R., Fiszbein, A., Olper, L.A., 1987. The positive and negative
Giménez, M., Junqué, C., Pérez, M., Vendrell, P., Baeza, I., Salamero, M., syndrome scale. Schizophr. Bull. 13 (2), 261–276.
Mercader, J.M., Bernardo, M., 2003. Basal ganglia N-acetylaspartate Kazes, M., Berthet, L., Danion, J.M., Amado, I., Willard, D., Robert, P.,
correlates with the performance in the procedural task ‘Tower of Poirier, M.F., 1999. Impairment of consciously controlled use of
Hanoi’ of neuroleptic-naive schizophrenic patients. Neurosci. Lett. memory in schizophrenia. Neuropsychology 13 (1), 54–61.
347 (2), 97–100. Keri, S., Kelemen, O., Szekeres, G., Bagoczky, N., Erdelyi, R., Antal, A.,
Goldberg, T.E., Saint-Cyr, J.A., Weinberger, D.R., 1990. Assessment Benedek, G., Janka, Z., 2000. Schizophrenics know more than they
of procedural learning and problem solving in schizophrenic can tell: probabilistic classification learning in schizophrenia.
patients by Tower of Hanoi type tasks. J. Neuropsychiatry Clin. Psychol. Med. 30 (1), 149–155.
Neurosci. 2 (2), 165–173. Kern, R.S., Green, M.F., Wallace, C.J., 1997. Declarative and
Gopal, K., 1997. Procedural and working memory in schizophrenia, procedural learning in schizophrenia: a test of the integrity of
Dissertation Abstracts International: Section B: The Sciences and divergent memory systems. Cogn. Neuropsychiatry 2 (1), 39–50.
Engineering 58 (1-B) 0435. Knopman, D., Nissen, M.J., 1991. Procedural learning is impaired in
Grace, A.A., 1992. The depolarization block hypothesis of neuroleptic Huntington's disease: evidence from the serial reaction time task.
action: implications for the etiology and treatment of schizophre- Neuropsychologia 29 (3), 245–254.
nia. J. Neural Transm. 36, 91–131. Kumari, V., Corr, P.J., Mulligan, O.F., Cotter, P.A., Checkley, S.A.,
Grafton, S.T., Mazziotta, J.C., Presty, S., Friston, K.J., Frackowiak, R.S., Gray, J.A., 1997. Effects of acute d-amphetamine and haloperidol
Phelps, M.E., 1992. Functional anatomy of human procedural on procedural learning in man. Psychopharmacology (Berl.) 129
learning determined with regional cerebral blood flow and PET. (3), 271–276.
J. Neurosci. 12 (7), 2542–2548. Kumari, V., Gray, J.A., Honey, G.D., Soni, W., Bullmore, E.T.,
Grafton, S.T., Hazeltine, E., Ivry, R., 1995. Functional mapping of Williams, S.C., Ng, V.W., Vythelingum, G.N., Simmons, A.,
sequence learning in normal humans. J. Cogn. Neurosci. 7, 497–510. Suckling, J., Corr, P.J., Sharma, T., 2002. Procedural learning in
Granholm, E., Bartzokis, G., Asarnow, R.F., Marder, S.R., 1993. schizophrenia: a functional magnetic resonance imaging investi-
Preliminary associations between motor procedural learning, basal gation. Schizophr. Res. 57 (1), 97–107.
ganglia T2 relaxation times, and tardive dyskinesia in schizophre- Lahti, A.C., Holcomb, H.H., Weiler, M.A., Medoff, D.R., Frey, K.N.,
nia. Psychiatry Res. 50 (1), 33–44. Hardin, M., Tamminga, C.A., 2004. Clozapine but not Haloperidol
Gras-Vincendon, A., Danion, J.M., Grange, D., Bilik, M., Willard- re-establishes normal task-activated rCBF patterns in schizophre-
Schroeder, D., Sichel, J.P., Singer, L., 1994. Explicit memory, nia within the anterior cingulate cortex. Neuropsychopharmacol-
repetition, priming and cognitive skill learning in schizophrenia. ogy 29, 171–178.
Schizophr. Res. 13 (2), 117–126. Lahti, A.C., Weiler, M.A., Medoff, D.R., Tamminga, C.A., Holcomb,
Green, M.F., Kern, R.S., Williams, O., McGurk, S., Kee, K., 1997. H.H., 2005. Functional effects of single dose first- and second-
Procedural learning in schizophrenia: evidence from serial reaction generation antipsychotic administration in subjects with schizo-
time. Cogn. Neuropsychiatry 2 (2), 123–134. phrenia. Psychiatry Res.: Neuroimaging 139, 19–30.
Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive Lang, D.J., Kopala, L.C., Smith, G.N., Vandorpe, R.A., Good, K.P.,
deficits and functional outcome in schizophrenia: are we Rui, Q., Butler, M.M., McSweeney, M.H., Goghari, V.M., Honer,
measuring the “right stuff”? Schizophr. Bull. 26 (1), 119–136. W.G., 1999. MRI study of basal ganglia volumes in drug-naive
ARTICLE IN PRESS
10 J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx
first-episode patients with schizophrenia. Schizophr. Res. 36, 202 Rauch, S.L., Whalen, P.J., Curran, T., McInerney, S., Heckers, S.,
(abstract). Savage, C.R., 1998. Thalamic deactivation during early implicit
Lang, D.J., Kopala, L.C., Vandorpe, R.A., Smith, G.N., Rui, Q., sequence learning: a functional MRI study. NeuroReport 9 (5),
Goghari, V.M., Hones, W.G., 2001. Basal ganglia volumes and 865–870.
treatment-response in never-medicated, first episode schizophrenia Rauch, S.L., Whalen, P.J., Curran, T., Shin, L.M., Coffey, B.J., Savage,
patients. Schizophr. Res. 49, 159 (abstract). C.R., McInerney, S.C., Baer, L., Jenike, M.A., 2001. Probing
Lussier, I., Stip, E., 2001. Memory and attention deficits in drug-naive striato–thalamic function in obsessive–compulsive disorder and
patients with schizophrenia. Schizophr. Res. 48 (1), 45–55. tourette syndrome using neuroimaging methods. Adv. Neurol. 85,
Martis, B., Wright, C.I., McMullin, K.G., Shin, L.M., Rauch, S.L., 207–224.
2004. Functional magnetic resonance imaging evidence for a lack Reiss, J.P., Campbell, D.W., Leslie, W.D., Paulus, M.P., Stroman, P.W.,
of striatal dysfunction during implicit sequence learning in Polimeni, J.O., Malcolmson, K.A., Sareen, J., 2005. The role of the
individuals with animal phobia. Am. J. Psychiatry 161 (1), 67–71. striatum in implicit learning: a functional magnetic resonance
Ninomiya, H., Sato, E., Onitsuka, T., Chen, C., 1998. Adaptation of imaging study. NeuroReport 16 (12), 1291–1295.
visually guided behavior during reversed vision in schizophrenia: a Saykin, A.J., Shtasel, D.L., Gur, R.E., Kester, D.B., Mozley, L.H.,
preliminary study. Psychiatry Res. 78 (1–2), 51–58. Stafiniak, P., Gur, R.C., 1994. Neuropsychological deficits in
Oldfield, R.C., 1971. The assessment and analysis of handedness: the neuroleptic-naive patients with first-episode schizophrenia. Arch.
Edinburgh Inventory. Neuropsychologia 9 (1), 97–113. Gen. Psychiatry 51 (2), 124–131.
Paquet, F., Soucy, J.P., Stip, E., Lévesque, M., Elie, A., Bédard, M.A., Schendan, H.E., Searl, M.M., Melrose, R.J., Stern, C.E., 2003. An
2004. Comparison between olanzapine and haloperidol on fMRI study of the role of the medial temporal lobe in implicit and
procedural learning and the relationship with striatal D2 receptor explicit sequence learning. Neuron 37 (6), 1013–1025.
occupancy in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 16 Schérer, H., Bedard, M.A., Stip, E., Paquet, F., Richer, F., Bériault, M.,
(1), 47–56. Rodriguez, J.P., Motard, J., 2004. Procedural learning in
Pascual-Leone, A., Grafman, J., Clark, K., Stewart, M., Massaquoi, S., schizophrenia can reflect the pharmacologic properties of the
Lou, J.S., Hallatt, M., 1993. Procedural learning in Parkinson's antipsychotic treatments. Cogn. Behav. Neurol. 17 (1), 32–40.
disease and cerebellar degeneration. Ann. Neurol. 34 (4), 594–602. Schmand, B., Kop, W.J., Kuipers, T., Bosveld, J., 1992a. Implicit
Peigneux, P., Maquet, P., Meulemans, T., Destrebecqz, A., Laureys, S., learning in psychotic patients. Schizophr. Res. 7 (1), 55–64.
Degueldre, C., Delfiore, G., Aerts, J., Luxen, A., Franck, G., Van Schmand, B., Brand, N., Kuipers, T., 1992b. Procedural learning of
der Linden, M., Cleeremans, A., 2000. Striatum forever, despite cognitive and motor skills in psychotic patients. Schizophr. Res.
sequence learning variability: a random effect analysis of PET 8 (2), 157–170.
data. Hum. Brain Mapp. 10 (4), 179–194. Schroder, J., Tittel, A., Stockert, A., Karr, M., 1996. Memory deficits in
Peretti, C.S., Danion, J.M., Kauffman-Muller, F., Grange, D., Patat, A., subsyndromes of chronic schizophrenia. Schizophr. Res. 21 (1), 19–26.
Rosenzweig, P., 1997. Effects of haloperidol and amisulpride on Schwartz, B.L., Howard, D.V., Howard Jr., J.H., Hovaguimian, A.,
motor and cognitive skill learning in healthy volunteers. Psy- Deutsch, S.I., 2003. Implicit learning of visuospatial sequences in
chopharmacology (Berl) 131 (4), 329–338. schizophrenia. Neuropsychology 17 (3), 517–533.
Purdon, S.E., Mintz, A.R., Labelle, A., Waldie, B.D., 2000. Improvement Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J.,
of procedural learning in schizophrenia after six months of treatment Weiller, E., Hergueta, T., Baker, R., Dunbar, G.C., 1998. The Mini-
with clozapine. Schizophr. Res. 41 (1), 184–185. International Neuropsychiatric Interview (M.I.N.I.): the develop-
Purdon, S.E., Woodward, N., Lindborg, S.R., Stip, E., 2003. ment and validation of a structured diagnostic psychiatric interview
Procedural learning in schizophrenia after 6 months of double- for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (Suppl 20), 22–33.
blind treatment with olanzapine, risperidone, and haloperidol. Snitz, B.E., MacDonald III, A., Cohen, J.D., Cho, R.Y., Becker, T.,
Psychopharmacology (Berl.) 169 (3–4), 390–397. Carter, C.S., 2005. Lateral and medial hypofrontality in first-
Puri, B.K., Barnes, T.R., Chapman, M.J., Hutton, S.B., Joyce, E.M., episode schizophrenia: functional activity in a medication-naive
1999. Spontaneous dyskinesia in first episode schizophrenia. state and effects of short-term atypical antipsychotic treatment.
J. Neurol. Neurosurg. Psychiatry 66 (1), 76–80. Am. J. Psychiatry 162 (12), 2322–2329.
Randolph, C., Gold, J.M., Carpenter, C.J., Goldberg, T.E., Weinberger, Sponheim, S.R., Steele, V.R., McGuire, K.A., 2004. Verbal memory
D.R., 1993. Implicit memory in patients with schizophrenia and processes in schizophrenia patients and biological relatives of
normal controls: effects of task demands on susceptibility to schizophrenia patients: intact implicit memory, impaired explicit
priming. J. Clin. Exp. Neuropsychol. 15 (6), 853–866. recollection. Schizophr. Res. 71 (2–3), 339–348.
Rauch, S.L., Savage, C.R., Brown, H.D., Curran, T., Alpert, N.M., Stevens, A., Schwarz, J., Schwarz, B., Ruf, I., Kolter, T., Czekalla, J.,
Kendrick, A., Fischman, A.J., Kosslyn, S.M., 1995. A PET 2002. Implicit and explicit learning in schizophrenics treated with
investigation of implicit and explicit sequence learning. Hum. olanzapine and with classic neuroleptics. Psychopharmacology
Brain Mapp. 3, 271–286. (Berl.) 160 (3), 299–306.
Rauch, S.L., Savage, C.R., Alpert, N.M., Dougherty, D., Kendrick, A., Teo, C., Rasco, L., al-Mefty, K., Skinner, R.D., Boop, F.A., Garcia-Rill,
Curran, T., Brown, H.D., Manzo, P., Fischman, A.J., Jenike, M.A., E., 1997. Decreased habituation of midlatency auditory evoked
1997a. Probing striatal function in obsessive–compulsive disorder: responses in Parkinson's disease. Mov. Disord. 12 (5), 655–664.
a PET study of implicit sequence learning. J. Neuropsychiatry Clin. Thomas, K.M., Hunt, R.H., Vizueta, N., Sommer, T., Durston, S.,
Neurosci. 9 (4), 568–573. Yang, Y., Worden, M.S., 2004. Evidence of developmental
Rauch, S.L., Whalen, P.J., Savage, C.R., Curran, T., Kendrick, A., differences in implicit sequence learning: an fMRI study of
Brown, H.D., Bush, G., Breiter, H.C., Rosen, B.R., 1997b. Striatal children and adults. J. Cogn. Neurosci. 16 (8), 1339–1351.
recruitment during an implicit sequence learning task as measured Tsuang, M.T., Lyons, M.J., Faraone, S.V., 1990. Heterogeneity of
by functional magnetic resonance imaging. Hum. Brain Mapp. schizophrenia. Conceptual models and analytic strategies. Br. J.
5 (2), 124–132. Psychiatry 156, 17–26.
ARTICLE IN PRESS
J.P. Reiss et al. / Schizophrenia Research xx (2006) xxx–xxx 11
Uc, E.Y., Skinner, R.D., Rodnitzky, R.L., Garcia-Rill, E., 2003. The Willingham, D.B., Koroshetz, W.J., 1993. Evidence for dissociable motor
midlatency auditory evoked potential P50 is abnormal in Hun- skills in Huntington's disease patients. Psychobiology 21, 173–182.
tington's disease. J. Neurol. Sci. 212, 1–5. Woods, S.W., 2002. Chlorpromazine equivalent doses for the newer
Wexler, B.E., Hawkins, K.A., Rounsaville, B., Anderson, M., Sernyak, atypical antipsychotics. J. Clin. Psychiatry 64, 663–667.
M.J., Green, M.F., 1997. Normal neurocognitive performance after Woonings, F.M.J., Appelo, M.T., Kluiter, H., Slooff, C.J., van den
extended practice in patients with schizophrenia. Schizophr. Res. 26 Bosch, R.J., 2002. Learning (potential) and social functioning in
(2–3), 173–180. schizophrenia. Schizophr. Res. 59 (2–3), 287–296.