HIPPOCAMPUS 23:1015–1024 (2013)
Selective Deficit in Spatial Memory Strategies Contrast to Intact
Response Strategies in Patients With Schizophrenia Spectrum
Disorders Tested in a Virtual Navigation Task
Leanne K. Wilkins,1* Todd A. Girard,1 Kyoko Konishi,2 Matthew King,1
Katherine A. Herdman,3,4 Jelena King,3 Bruce Christensen,3 and Veronique D. Bohbot2
ABSTRACT: Spatial memory is impaired among persons with schizophrenia (SCZ). However, different strategies may be used to solve most
spatial memory and navigation tasks. This study investigated the hypothesis that participants with schizophrenia-spectrum disorders (SSD)
would demonstrate differential impairment during acquisition and
retrieval of target locations when using a hippocampal-dependent spatial strategy, but not a response strategy, which is more associated with
caudate function. Healthy control (CON) and SSD participants were
tested using the 4-on-8 virtual maze (4/8VM), a virtual navigation task
designed to differentiate between participants’ use of spatial and
response strategies. Consistent with our predictions, SSD participants
demonstrated a differential deficit such that those who navigated using
a spatial strategy made more errors and took longer to locate targets. In
contrast, SSD participants who spontaneously used a response strategy
performed as well as CON participants. The differential pattern of
spatial-memory impairment in SSD provides only indirect support for
underlying hippocampal dysfunction. These findings emphasize the
importance of considering individual strategies when investigating SSDrelated memory and navigation performance. Future cognitive intervention protocols may harness SSD participants’ intact ability to navigate
using a response strategy and/or train the deficient ability to navigate
using a spatial strategy to improve navigation and memory abilities in
C 2013 Wiley Periodicals, Inc.
participants with SSD. V
KEY WORDS:
hippocampus; caudate nucleus; navigational strategies;
virtual reality; response learning
INTRODUCTION
Hippocampi and surrounding medial-temporal structures are central
to pathophysiological theories of Schizophrenia (SCZ; Christensen and
Bilder, 2000; Grace, 2000; Tseng et al., 2009). Moreover, reduced hip-
1
Department of Psychology, Ryerson University, Toronto, Ontario, Canada; 2 Department of Neurology and Neurosurgery, McGill University,
Montreal, Quebec, Canada; 3 Department of Psychiatry and Behavioural
Neurosciences, McMaster University, Hamilton, Ontario, Canada;
4
Department of Psychology, York University, Toronto, Ontario, Canada
Grant sponsors: NARSAD (The Brain and Behaviour Research Foundation), Faculty of Arts (Ryerson University).
*Correspondence to: Leanne K. Wilkins, Department of Psychology,
Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada.
E-mail: [email protected]
Accepted for publication 6 August 2013.
DOI 10.1002/hipo.22189
Published online 12 August 2013 in Wiley Online Library
(wileyonlinelibrary.com).
C 2013 WILEY PERIODICALS, INC.
V
pocampal volume (Nelson et al., 1998; McCarley
et al., 1999; Heckers, 2001) and impaired hippocampal activation during episodic memory tasks have
been reliably demonstrated (Heckers et al., 1998).
However,
characterizing
the
specificity
of
hippocampal-dependent memory deficits is important
for advancing understanding of the brain-behavior
relations affected in SCZ and for developing cognitive
remediation strategies.
In this regard, recent studies from our laboratory
(Girard et al., 2010; Wilkins et al., 2013) and others
(Hanlon et al., 2006; Weniger and Irle, 2008; Folley
et al., 2010; Spieker et al., 2012) demonstrate robust
deficits on tasks associated with hippocampaldependent learning of allocentric spatial relations
among environmental cues in participants with
schizophrenia-spectrum disorders (SSDs). These findings are in contrast to relatively spared performance
on tasks involving learning to navigate using a bodycentered egocentric approach, which involves learning
a series of left and right turns or using a stimulusresponse approach, which involves identification of a
single stimulus (i.e., identification of a landmark),
connected with a learned chain of responses, such as
left and right body turns. For instance, researchers
reported that their SCZ sample performed normally
on a virtual water maze under a visible-platform condition, but was impaired at learning to locate a hidden platform relying on knowledge of its allocentric
relation to environmental landmarks (i.e., on walls
surrounding the maze) (Hanlon et al., 2006). Deficient allocentric learning in the virtual water maze
was further correlated with hippocampal grey matter
abnormalities in persons with SCZ (Folley et al.,
2010). Others had also reported greater deficits
among persons with SCZ when learning to navigate a
virtual park that demanded formation of an allocentric cognitive map, whereas patients’ performance was
intact on a virtual maze that relied on egocentric
response-based learning (Weniger and Irle, 2008). Of
note, these studies used different tasks to assess spatial
and response learning. Thus, we recently extended
these findings to demonstrate a within-task differential
deficit in allocentric, viewpoint-independent versus
viewpoint-dependent memory in an SSD sample
1016
WILKINS ET AL.
using both a neuropsychological task (Girard et al., 2010) and
an analogous virtual-reality task (Wilkins et al., 2013). Both
tasks tested memory for object locations in a three-dimensional
spatial array. Taken together, these findings support a differential deficit in hippocampal-dependent spatial memory in SSDs.
However, both spatial and response-based systems can support successful navigation of various environments, albeit using
different strategies (Iaria et al., 2003; Bohbot et al., 2004,
2007; Etchamendy and Bohbot, 2007). The spatial memory
system involves constructing allocentric relations between landmarks to form a cognitive map. In contrast, the response memory system supports navigation through learning stimulusresponse relations such as a series of left and right body turns
from a specific starting point.
The 4-on-8 virtual maze (4/8VM) task is a human analog of
a rodent eight-arm radial maze designed for ascertaining individual differences associated with spontaneous navigational strategies
(Iaria et al., 2003; Bohbot et al., 2007; Etchamendy and Bohbot, 2007). Briefly, task trials begin with participants navigating
to target objects located at the end of four open pathways (the
other four closed). Subsequently, participants are returned to the
maze centre with all eight pathways open and required to
remember previously visited pathways to avoid them and retrieve
targets hidden down the previously closed pathways. The maze
is surrounded by extra-maze landmarks (e.g., tree, rock, mountain, valley) allowing participants to use a spatial strategy by
forming an allocentric representation of the configuration of target pathways relative to the distal landmarks. However, participants may also apply a response-based strategy such as
remembering the within-maze sequence of target pathways relative to their starting position or a single landmark. In previous
studies, approximately half of young adults spontaneously
adopted a spatial strategy and half a response strategy on this
task (Iaria et al., 2003). Those who spontaneously used a spatial
strategy had higher hippocampal gray matter and fMRI activity
while navigating a virtual environment; whereas, spontaneous
use of a response strategy was associated with higher gray matter
and sustained fMRI activity of the caudate nucleus (Iaria et al.,
2003; Bohbot et al., 2004, 2007). Patients with medial temporal
lobe lesions who also used a spatial strategy performed the 4/
8VM with longer latencies and more errors relative to healthy
participants, whereas those who used a response strategy performed similarly to controls (Bohbot et al., 2004). In summary,
spontaneous adoption of a given strategy has been associated
with different brain systems and can be an important determinant of successful navigation and spatial memory performance
in clinical populations. Thus, this study employed the 4/8VM
to assess the contribution of spontaneous strategy on spatialmemory performance in persons with SSDs.
Since individuals with SSDs are known to have structural
and functional abnormalities in the hippocampi, we predicted
that individuals with SSDs who spontaneously use a
hippocampal-dependent spatial strategy to solve the 4/8VM
would be impaired relative to those who spontaneously use a
caudate nucleus-dependent response strategy. Given the integral
role of locating oneself in space to episodic memory, navigation
Hippocampus
and efficient day-to-day functioning, understanding the role of
spontaneous strategies used by SSD individuals is important
for advancing the development of effective neurocognitive
interventions.
MATERIALS AND METHODS
Participants
SSD participants (n 5 21) were recruited through a research
database at St. Joseph’s Healthcare, Hamilton (SJHH), as well
as through referral from outpatient clinics/programs at SJHH
and the Hamilton Program for SCZ. Healthy control (CON)
participants (n 5 24) were recruited from the Hamilton
(Ontario) community via newspaper, Craigslist, and poster
advertisements. Participants were included if they were able to
provide informed consent, were 18–60 years of age, spoke English as their primary language, and had normal or corrected-tonormal vision. SSD participants met criteria for a DSM-IV
psychotic disorder (SCZ, Schizoaffective Disorder, Schizophreniform Disorder, Delusional Disorder), as ascertained using the
Mini-International Neuropsychiatric Interview (Sheehan et al.,
1998). Diagnostic interviews were conducted by trained
research assistants or trained senior graduate students. Additionally, all SSD participants were clinically and pharmacologically stable (i.e., no recent change in medication or patient
status in the past 6 weeks). Exclusion criteria consisted of a
lifetime history of a neurological condition or lifetime or current nonpsychotic Axis I psychiatric disorder (including alcohol
or substance dependence or abuse). CON participants with a
first-degree relative with a psychotic disorder were also
excluded. Four SSD participants were excluded from analyses
due to a failure to meet experimental criterion on acquisition
of the 4/8VM (detailed below). Although they failed to meet
criterion, we assessed their spontaneous strategy selection. Of
those excluded, two SSD participants self-reported using a spatial strategy and two using a response strategy. To equate sample sizes and better match samples on demographic and
cognitive measures seven CON were also omitted from the
analyses. Data from the remaining SSD (n 5 17; 13 male)
and CON (n 5 17; 8 male) participants are reported below.
Participant characteristics are summarized in Tables (1–3).
To assess group and strategy differences across demographic,
cognitive, and clinical measures, we ran independent sample t
tests and Chi-square tests for continuous and categorical variables, respectively. Overall, the SSD group reported fewer years
of education (M 5 13.3) and had fewer participants with experience playing three-dimensional games(n 5 6) than the CON
group (M 5 15.2; n 5 13), but there were no other demographic differences between groups (CON, SSD) or strategy
types (spatial, response), ps > 0.05 (see Table 1).Also, the
groups did not differ significantly across global measures of
cognition (FSIQe, estimated Full Scale Intelligence Quotient
based on Information and Matrix Reasoning subtests from the
SPATIAL STRATEGY DEFICIT IN SCHIZOPHRENIA
1017
TABLE 1.
Demographic Characteristics of CON and SSD Groups by Strategy (Spatial, Response)
CON
Characteristicsa
Demographics
Sex (n males/females)
Age (yr)
Education
SESc,d
Video game experienced
Years of playing
Hours played/week
3D (n)
SSD
Spatial
Response
Spatial
Response
6/2
36.0 (11.5)
15.9 (2.0)
44.8 (10.7)
3/6
36.4 (15.0)
14.6 (2.2)
39.9 (13.3)
7/1
44.9 (4.8)
13.1 (1.4)
41.4 (4.2)
6/3
39.6 (9.8)
13.4 (1.7)
39.5 (9.8)
11.3 (10.4)
1.3 (1.4)
5
13.7 (9.1)
14.4 (25.3)
8
13.9 (11.8)
2.4 (2.9)
3
10.9 (10.3)
7.8 (12.9)
3
Effect(s)b
h2
G
0.23
G
0.17
a
Continuous data are presented as means (standard deviation), M (SD), and evaluated using independent-samples t tests. Sex and experience with first-person
immersive three-dimensional video-game experience (3D) are reported as frequency data (n) and evaluated with v2 tests. We evaluated Group 3 Strategy ANOVAs
for continuous variables. We evaluated Group 3 Strategy for categorical variables with the Cochran–Mantel–Haenszel test.
b
G represents significant (P < 0.05) main effects of Group; there were no main effects of Strategy or Group 3 Strategy interactions.
c
Socioeconomic status (SES) calculations were based on parental occupations (Blishen et al., 1987).
d
Appropriate data were unavailable for three SCZ participants on SES (1 spatial, 2 response) and one CON (spatial) on video game experience.
Wechsler Adult Intelligence Scale, Third Edition, Sattler and
Ryan, 1998; WRAT-reading, Reading subtest from the Wide
Range Achievement Test, Fourth Edition, (Robertson, 2006))
visual-spatial skills (3D mental rotation, Vandenberg, 1978;
RBANS-visual spatial index, The Repeatable Battery for Assessment of Neuropsychological Status, (Randolph et al., 1998)),
or global functioning (WHODAS, The World Health Organization Disability Assessment Schedule, (Janca et al., 1996)),
but persons with SSDs demonstrated poorer episodic memory,
attention, and language ability on the RBANS (see Table 2 for
descriptive statistics and definitions of abbreviations). Importantly, however, these cognitive differences between groups
(CON, SSD) did not interact with navigational strategies
(Group 3; Strategy, n.s.) (see Table 2).
The SSD group comprised mainly participants with SCZ or
Schizoaffective Disorder, along with one diagnosed with
TABLE 2.
Cognitive Characteristics of CON and SSD Groups by Strategy (Spatial, Response)
CON
Characteristics
FSIQe
WRAT-reading
3D Rotationb
WHODAS
RBANS indices
Immediate memory
Visuospatial
Language
Attention
Delayed memory
SSD
Spatial
Response
Spatial
Response
113.1 (11.2)
106.8 (10.6)
14.3 (9.8)
13.6 (19.4)
110.6 (18.7)
101.6 (13.0)
9.9 (6.4)
6.3 (14.2)
98.0 (16.1)
93.9 (9.0)
5.1 (2.0)
13.8 (11.0)
109.7 (12.4)
100.4 (12.8)
9.6 (6.5)
10.6 (4.6)
108.0 (15.4)
98.9 (24.9)
108.6 (8.5)
102.4 (11.9)
106.1 (13.4)
93.4
94.0
98.7
96.6
97.0
79.9 (15.1)
104.6 (20.6)
85.1 (12.2)
78.8 (17.4)
86.1 (15.7)
89.0 (13.8)
99.1 (16.2)
86.9 (6.7)
91.4 (19.3)
93.1 (17.0)
(23.7)
(20.2)
(10.7)
(17.2)
(14.7)
Effect(s)a
h2
G
0.20
G
G
G
0.48
0.17
0.15
Data are presented as M (SD). We evaluated Group 3 Strategy interactions for continuous variables using ANOVA and for categorical variables with the Cochran–Mantel–Haenszel test.
a
G represents significant (P < 0.05) main effects of Group; there were no main effects of Strategy or Group 3 Strategy interactions.
b
Data were missing for one SCZ participant (response) and one CON (spatial) on 3D rotation.
Abbreviations: FSIQe 5 Prorated estimate of Full-Scale Intelligence Quotient derived from the Matrix Reasoning and Information subtests of the Wechsler Adult
Intelligence Scale—Third Edition, Sattler and Ryan, 1998; RBANS 5 Repeatable Battery for the Assessment of Neuropsychological Status, Randolph et al.,
1998); 3D rotation 5 Mental rotation of three-dimensional geometric shapes, Vandenberg et al., 1978; WRAT-Reading 5 Reading subtest from the Wide Range
Achievement Test–Fourth Edition, Robertson and Wilkinson, 2006; World Health Organization Short Disability Assessment Schedule, Janca et al., 1996.
Hippocampus
1018
WILKINS ET AL.
TABLE 3.
Clinical Characteristics of the SCZ-Spectrum Sample by Strategy
(Spatial, Response)
Measurea
Spatial
Response
Diagnoses (SCZ, schizoaffective,
delusional disorder)
Medicationb
CPZe*
Atypical, typical, neither, or both
antipsychotic types*
Antidepressants
Anxiolytics
PANSS T-scores
General*
Negative
Positive
3, 5, 0
4, 4, 1
300.8 (209.7)
0, 0, 0, 8
102.6 (115.1)
4, 2, 1, 2
1
3
5
3
38.8 (4.3)
39.5 (8.5)
40.0 (13.2)
33.9 (2.0)
34.7 (5.3)
35.9 (4.7)
a
Continuous data are presented as M (SD); frequency data reflect numbers of
participants (n).
b
Appropriate medication data were unavailable for two SSD participants.
Abbreviations: CPZe 5 chlorpromazine equivalents (Virani et al., 2011;
Woods, 2003), PANSS 5 Positive and Negative Symptom Scale (Kay et al.,
1987).
*P < 0.05.
Delusional Disorder (see Table 3). There were no differences
between these subgroups listed above on the experimental, cognitive, or demographic variables (ps > 0.05). Interestingly, all of the
SSD participants who used a spatial strategy were on a combination of both atypical and typical antipsychotics, whereas those
using a response strategy were more likely to be on one type of
medication (atypical or typical; see Table 3). The participant with
delusional disorder was on an antidepressant, but not antipsychotic medication. On average, participants using a spatial strategy
were on higher dosages of antipsychotics and were more symptomatic as measured by the Positive and Negative Symptom Scale
(PANSS; Kay et al., 1987),than the SSD group reporting a
response strategy (see Table 3). However, as detailed in the
Results, accounting for dose and symptomology in the analyses
did not affect the pattern of findings. Moreover, the SSD groups
were symptomatically stable, scoring below average relative to
patient norms on the PANSS (Kay et al., 1987; see Table 3).
All participants provided written, informed consent and
were compensated with a cash honorarium of $10 dollars per
hour. This study is part of a larger battery of concurrent projects regarding memory in SCZ. However, 4/8VM testing was
counterbalanced across separate days with another objectlocation memory task. This study was approved by the
Research Ethics Boards at Ryerson University and SJHH.
The 4/8VM Task
The 4/8VM paradigm has been thoroughly described previously (Iaria et al., 2003; Bohbot et al., 2004, 2007; Etchamendy
and Bohbot, 2007). Briefly, the 4/8VM is a computerized virtual
environment (created using Unreal Tournament, Epic Games
Hippocampus
Inc., Raleigh, N.C.) comprised of a virtual eight-arm radial
maze with a central starting position, surrounded by extra-maze
landmarks (e.g., tree, rock, mountain, and valley). Participants
use a keypad with forward, left turn, and right turn buttons to
move within the environment. To obtain hidden target objects,
participants are required to navigate to the end of the maze
arms and down a staircase to a small pit. Prior to testing, participants practiced navigating with the keyboard by exploring a
similar environment to the 4/8 task, but without landmarks.
Testing involved a series of two-part trials from a constant
start position. In the first part, participants visited four open
pathways to retrieve target objects (the other four were blocked).
In the second part of each trial, participants were presented with
all eight pathways opened and had to avoid the previously visited pathways to find the target objects down the previously
closed pathways. All participants completed an initial sequence
of three trials across which the configuration of pathways containing target objects changed between the first (configuration
A) and second trial (configuration B), with the third trial having
the same configuration as the first (configuration A).Thus, we
refer to this initial key sequence of trials as the ABA trials. Participants were required to perform two type-A trials without
error before proceeding with a probe test. If participants did not
meet this criterion within the first three trials, testing continued
using the A configuration until participants met this criterion.
As noted above, four SSD participants failed to reach this
criterion within 3 h and were excluded from analyses.
The probe test was administered to further confirm navigational strategies. This trial differed at test in that walls around
the radial maze were raised to conceal the landscape and the
other landmarks were removed. Testing was discontinued after
participants entered four different pathways. The probe test
dissociates between spatial and response strategies because a
higher error rate is expected among participants using a spatial
strategy given the removal of allocentric landmarks.
At the end of the experiment, participants were interviewed
to identify their spontaneous navigational strategy. Interviews
were audio taped and two independent raters confirmed agreement on the coding of participant strategies. Spatial strategies
were operationalized by indication of remembering the positions of target pathways relative to two or more landmarks and
the absence of any reference to using a sequence of open and
closed pathways from a single position. For example, a participant would be assigned to the spatial strategy group if they
mentioned remembering the target pathways in relation to the
tree, rock or mountain. In contrast, response strategies were
defined as remembering the within-maze sequence of target
pathways relative to their starting position or to a single landmark. For example, a participant using a response strategy may
report locating the tree as a starting position for their memorized clockwise sequence of open and closed pathways.
Data Analyses
To test the hypothesized Group 3 Strategy interaction of
differential 4/8VM task impairment in the SDD-Spatial group
SPATIAL STRATEGY DEFICIT IN SCHIZOPHRENIA
on each performance measure, we conducted univariate analyses on four key dependent variables (DVs): (1) Trials to Criterion: The number of additional trials required to perform two
error-free A trials following the initial ABA sequence. (2) ABA
Latency: The summed duration (in minutes) of the initial ABA
sequence of trials. (3) ABA Errors: Errors of commission at test
across the ABA trials comprised entries into a pathway without
a target (incorrect) and repeat pathway entries within a trial.
(4) Errors on the probe test: The total number of incorrect
pathways visited during the probe test (out of a maximum of
four).
We evaluated the first three DVs using a 2 (Group: SSD
and CON) 3 2 (Strategy: Spatial and Response) factorial analyses of variance (ANOVAs). However, given the discrete and
limited scale for probe-test errors, we analyzed these errors
using nonparametric statistics. Because of small cell counts
across error rates other than perfect performance, we further
dichotomized this variable prior to the analysis (error free versus at least one error).We assessed whether Group error differences were contingent on Strategy using the Cochran–Mantel–
Haenszel test; homogeneity of the odds ratios across Strategy
were confirmed with the Breslow–Day statistic. Of note, we
corroborated the probe test results using a 2 3 2 ANOVA on
the original full-scale data, but only report the more appropriate nonparametric results below. All results were evaluated at
an alpha level of 0.05.
Given overlapping variance among the three 4/8VM acquisition variables, we also applied the multivariate analysis of variance (MANOVA) approach (Grice and Iwasaki, 2007) to
assess the extent to which a more comprehensive linear combination of the acquisition parameters account for the Group 3
Strategy interaction. Moreover, this linear composite was used
to minimize redundancy in assessing correlations between task
performance and demographic, cognitive, and clinical attributes, along with follow-up covariate analyses. That is, the twofactor between-subjects MANOVA provides a single multivariate composite that maximally accounts for the Group 3 Strategy interaction across measures, allowing for unified
correlational analyses, rather than performing multiple redundant analyses for each of the variables within each subsample.
Given the different measurement scales, the dependent measures were standardized prior to analysis. Our data met the
required assumptions of independence, univariate and multivariate normality, and homogeneity of variance-covariance matrices (Grice and Iwasaki, 2007).
RESULTS
Univariate Analyses
Results revealed that SSD participants who reported spontaneous use of a spatial strategy produced longer latencies and
more errors on the initial ABA trials, and required more additional trials to meet the acquisition criterion of two correct A
1019
FIGURE 1.
Differential deficit among SSD participants reporting spontaneous use of a Spatial versus Response strategy during
acquisition of the 4/8VM task. Data represent standardized means
(and standard errors) by Group and Strategy on the measures of
Trials to criterion, latency to find targets and test errors on the
initial three ABA trials. The z-scores are signed such that negative
scores reflect poorer performance (below the overall mean for each
measure). The simple multivariate composite reflects a linear combination of the Trials to criterion and ABA Latency measures.
Please see text for further description of these measures.
trials (see Fig. 1). Following acquisition to criterion, there were
no remaining group differences on the probe test.
More specifically, an ANOVA on the trials required to reach
criterion revealed a significant Group 3 Strategy interaction,
F(1,30) 5 5.82, P 5 0.022, h2 5 0.16, as well as main effects
of Group, F(1, 30) 5 8.86, P 5 0.006, h2 5 0.23, and Strategy,
F(1,30) 5 4.77, P 5 0.037, h2 5 0.14. Follow-up analyses
revealed the SSD-Spatial group required more additional trials (M
5 5.63, SD 5 2.77) to reach criterion than their CON-Spatial
counterparts (M 5 1.38, SD 5 1.30), t(14) 5 23.92, P 5
0.002, d 5 22.10. In contrast, the SSD and CON groups using
a Response strategy did not differ statistically, t(16) 5 20.393,
P 5 0.700, d 5 20.20 (overall M 5 1.78, SD 5 2.34).
Likewise, latencies to locate the targets across the ABA trials
revealed a significant interaction of Group 3 Strategy, F(1,30)
5 4.53, P 5 0.042, h2 5 0.13, and main effects of Group,
F(1, 30) 5 11.14, P 5 0.002, h2 5 0.27, and Strategy (Spatial > Response), F(1,30) 5 7.01, P 5 0.013, h2 5 0.19.
Again, the interaction reflected that the SSD-Spatial group
took longer to locate the targets than the CON-Spatial group,
t(14) 5 24.30, P < 0.001, d 5 22.30, whereas among
Response-strategy users the Groups did not differ, t(16) 5
20.80, P 5 0.434, d 5 20.40.
At test, SSD participants made more errors than CONs,
Group, F(1,30) 5 7.23, P 5 0.012, h2 5 0.19. Overall, participants using a Response strategy made fewer errors than
those using a Spatial strategy, F(1,30) 5 6.36, P 5 0.017,
Hippocampus
1020
WILKINS ET AL.
h2 5 0.18. However, the Group 3 Strategy interaction failed
to reach significance, F(1,30) 5 1.83, P 5 0.19, h2 5 0.06.
Nonetheless, follow-up analyses revealed that the SSD-Spatial
group was impaired relative to the CON-Spatial group, t(14)
5 22.18, P 5 0.047, d 5 21.17. The Response SSD and
CON groups did not differ significantly, t(16) 5 21.45, P 5
0.166, d 5 20.73.
Finally, the probe test did not reveal differential patterns of
errors in the Group by Strategy analysis, v2CHM(1) 5 0.625, P
5 0.429. There was no overall Group difference on this measure, v2(1) 5 1.36, P 5 0.244, u 5 0.20. Across groups,
there was a moderate, but nonsignificant, tendency in the
expected direction for more participants in the Response
groups to perform error-free on the probe test, v2(1) 5 3.03,
P 5 0.081, u 5 0.30.
Multivariate Analyses
Supporting the composite univariate analyses, the MANOVA
revealed a statistically significant Group 3 Strategy interaction,
Wilks’ k 5 0.73, F(3, 28) 5 3.41, P 5 0.031. The multivariate composite of the dependent measures (Trials to criterion,
ABA latency, ABA errors) accounted for 27% of the variance
pertaining to the Group 3 Strategy interaction, according to
Roy’s greatest characteristic root, g.c.r. (s 5 1, m 5 1=2, n 5
13) 5 0.27, P < 0.05. As such, the MANOVA approach
yielded a multivariate gain of 11% in accounting for the
Group 3 Strategy interaction beyond that detected by the univariate analyses above (Grice and Iwasaki, 2007). Consistent
with the univariate results, the standardized discriminant function coefficients identified Trials to criterion (ws5 20.86) and
ABA latency (ws 5 20.74) as main contributors to the interaction effect, with ABA errors adding substantially less to the
overall function (ws 5 0.24). Following the approach outlined
by Grice and Iwasaki (2007), we created a simplified multivariate composite 5 (21) (Trials to Criterion) 1 (21) (ABA
Latency) 1 (0) (ABA errors). Conceptually, this simplified
composite reflects an equally weighted combination of the criterion and latency measures as an index of overall performance
on the 4/8VM. Further analyses of the full and simplified
models revealed near-identical results, with the latter only
resulting in a loss of 1% explanatory power. Given the greater
stability and general applicability to future studies, we present
results from the simplified composite here.
ANOVA on the simplified composite yielded a Group 3
Strategy interaction, F(1,30) 5 10.48, P 5 0.003, h2 5 0.26,
as well as main effects of Group, F(1,30) 5 20.19, P < 0.001,
h2 5 0.40, and Strategy, F(1,30) 5 11.81, P 5 0.002, h2 5
0.28. As shown in Figure 1, the interaction reflected that the
SSD-Spatial group was impaired relative to the other three
groups, t(30) 5 6.13, P < 0.001, d 5 2.31, which did not differ from one another, F(2,23) 5 0.52, P 5 0.603, h2 5 0.04.
Correlation and Covariate Analyses
The relationships between demographic, cognitive, and clinical variables (as listed in Tables (1–3), respectively) with the
Hippocampus
simplified composite measure of 4/8VM performance were
explored via bivariate correlations within the SSD and CON
groups. Better performance among CON participants was significantly associated with sex (male > female), rpb 5 0.53, P
5 0.041, 3D videogame experience, rpb 5 0.52, P 5 0.033,
and higher visual-spatial abilities (3D mental rotation, r 5
0.65, P 5 0.007; RBANS visuospatial index, r 5 0.55, P 5
0.022). Among the SSD sample, better performance was significantly related to younger age, r 5 20.50, P 5 0.041,
WRAT-Reading scores, r 5 0.54, P 5 0.026, and 3D Mental
Rotation abilities, r 5 0.52, P 5 0.041.Consistent with the
higher symptom ratings and antipsychotic dosage among SSD
participants who used a spatial strategy compared to those
using a response strategy (see Table 3), General PANSS scores,
r 5 20.53, P 5 0.044, and CPZe, r 5 20.54, P 5 0.024,
were related to worse 4/8VM performance.
In light of these correlations and group differences, the
impact of these variables (i.e., sex, 3D gaming experience,
3D mental rotation scores, RBANS visuospatial index scores,
age, and WRAT-Reading scores) on the main Group 3 Strategy interaction was explored using covariate analysis. Importantly, the Group 3 Strategy interaction remained
significant, F(1, 21) 5 15.46, P < 0.001, h2 5 0.42, as did
the main effects of Group, F(1, 21) 5 7.16, P 5 0.014, h2
5 0.25, and Strategy, F(1, 21) 5 14.80, P < 0.001, h2 5
0.41. In addition, we ran two-way ANOVAs on matched
subsamples regarding sex and 3D gaming experience: The
Group 3 Strategy interaction remained robust among
females, F(1, 9) 5 10.67, P < 0.010, h2 5 0.54, males, F(1,
17) 5 6.34, P 5 0.022, h2 5 0.27, and among those with,
F(1, 15) 5 8.96, P 5 0.009, h2 5 0.37, and without 3Dgaming experience, F(1, 11) 5 8.76, P 5 0.013, h2 5 0.44
(see Fig. 2).
Within the SSD sample, the mean deficit among SSDSpatial versus SSD-Response participants remained significant
when covarying the variance shared with antipsychotic dosage
(CPZe), F(1, 14) 5 9.52, P 5 0.008, h2 5 0.41, and differences in symptomatology (PANSS-General), F(1, 12) 5 6.98,
P 5 0.022, h2 5 0.37. Moreover, inspection of the data indicated that among SSD participants matched on symptomatology or CPZe, as well as types of medication, those in the SSDspatial subgroup consistently performed worse than their
response subgroup counterparts. These results suggest that the
Group 3 Strategy interaction cannot be fully explained by the
SSD spatial group having a higher General PANSS and CPZe.
Figure 3 illustrates the relation between dosage (CPZe) and 4/
8VM task performance.
Likewise, follow-up assessment regarding anti-depressant and
anxiolytic use (coding subgroups as using an antidepressant,
anxiolytic, neither, or both) failed to reveal any differential
effects on performance. Within the SSD group, a Drugs 3
Strategy ANOVA maintained a robust main effect of Strategy
(Response > Spatial), F(1, 11) 5 8.10, P 5 0.016, h2 5
0.42, but neither an effect of Drugs, F(3, 11) 5 0.17, P 5
0.916, h2 5 0.04, nor Drugs 3 Strategy interaction, F(1, 11)
5 0.33, P 5 0.580, h2 5 0.03.
SPATIAL STRATEGY DEFICIT IN SCHIZOPHRENIA
1021
DISCUSSION
This study provides a unique assessment of spontaneous navigation strategies in SSD and their impact on virtual maze performance. Most notably, results revealed our hypothesized
dissociation among SSD participants between those who used a
spatial strategy versus those who used a response strategy on
the 4/8VM task. Indeed, the SSD-Spatial group took significantly longer to locate all targets and made more pathwayentry errors during learning trials, and required more trials to
reach criterion than the CON-Spatial group. In stark contrast,
the SSD-Response learners were not impaired relative the
CON-Response group. This dissociation was further underscored via a multivariate composite of the learning measures
(see Fig. 1). Moreover, despite correlations with a subset of
participant characteristics, the differential impairment of the
SDD-Spatial group remained robust to covariate analyses and
further follow-up analyses assessing the influence of demographic, cognitive, and clinical variables. Overall, these behavioral findings support hypotheses of a selective hippocampaldependent memory dysfunction in SCZ with relatively intact
functioning of response-based systems (Keri et al., 2005;
FIGURE 2.
The greater impairment among SSD participants
reporting spontaneous use of a Spatial versus Response strategy
remained robust across subsamples matched by (A) sex and (B)
those with (yes) and without (no) 3D videogame experience.
We also addressed potential differences across diagnostic subtypes. Because there was only a single patient with the diagnosis of Delusional Disorder and this patient was not taking
antipsychotics, we repeated our analyses without this participant. All results and conclusions were replicated. For example,
the ANOVA on the simplified composite measure of 4/8VM
performance yielded identical P-values and effect sizes to the
analysis with this participant included (reported above), e.g.,
Group 3 Strategy interaction omitting this patient, F(1,29) 5
10.17, P 5 0.003, h2 5 0.26. This participant scored in the
middle of the overall distribution on the simplified composite
(score 5 0.08), falling close to her SCZ (M 5 0.04) Response
group counterparts and slightly below those with Schizoaffective disorder (M 5 0.80). Although patients diagnosed with
Schizoaffective disorder performed better than those with SCZ,
this difference was not significant, F(1, 12) 5 1.32, P 5
0.413, and of small magnitude, h2 5 0.06. Moreover, this
relation was consistent across Strategies, Group 3 Strategy,
F(1, 12) 5 0.12, P 5 0.801, h2 < 0.01: The SSD-Spatial
group performed significantly and substantially worse than
their SSD-Response counterparts across diagnoses, F(1, 12) 5
28.17, P 5 0.002, h2 5 0.56.
FIGURE 3.
Scatter plot of the correlations between medication
dosage (CPZe) and 4/8VM performance. Overall, greater dosage
was significantly correlated with poorer performance in the SSD
group (r 5 20.54). However, this relation failed to account for
the behavioral difference between SSD subgroups. Regression lines
for the SSD response and spatial subgroups are similar and less
robust: SSD-Spatial, r 5 20.30; SSD-Response, r 5 20.22. Comparison between SSD-spatial and SSD-response individuals with
comparable CPZe levels indicates that those with a spatial strategy
consistently perform worse (with one exception).
Hippocampus
1022
WILKINS ET AL.
Hanlon et al., 2006; Weniger and Irle, 2008; Girard et al.,
2010; Wilkins et al., 2013). This conclusion is in line with evidence that use of a spatial strategy depends on the integrity
and recruitment of the hippocampi (O’Keefe, 1978; Smith and
Milner, 1989), whereas response strategies are associated with
the caudate nucleus in humans (Hartley et al., 2003; Iaria
et al., 2003; Bohbot et al., 2004, 2007) and striatum (which
includes the caudate nucleus) in rats (Packard et al., 1989;
McDonald and White, 1994). These data further support that
the hippocampus and caudate nucleus memory systems can
function independently of each other (McDonald and White,
1994; Bohbot et al., 2004).
Beyond corroborating previous studies reporting deficient
allocentric spatial memory in SCZ, these results are consistent
with evidence of selective behavioral deficits in hippocampaldependent spatial and episodic memory and contextual processing (Boyer et al., 2007). Indeed, hippocampal abnormalities
and deficient episodic memory are reliably reported in the
SCZ literature (Heckers, 2001; Heinrichs, 2005; Bowie and
Harvey, 2006). For example, reduced hippocampal volume is a
reliable finding (Nelson et al., 1998; McCarley et al., 1999)
and individuals living with SCZ show impaired recruitment of
this brain region during performance on episodic memory tasks
(Heckers, 2001). The overall pattern of memory impairment
in some SCZ samples appears comparable to that of individuals with medial-temporal lobe (MTL) lesions (Ornstein et al.,
2008; Bartholomeusz et al., 2011). This study builds upon this
literature and our findings of episodic and allocentric spatial
memory deficits in SCZ (Girard et al., 2010; Wilkins et al.,
2013) by demonstrating the importance of strategy use.
The observation that participants with SSD demonstrated
normal performance when utilizing a strategy dependent on
the caudate nucleus memory system has important clinical
implications. Keri et al. (2005) also found evidence of spared
stimulus-response learning in SCZ. In their study, stimulusresponse learning was assessed using the Rutgers acquired
equivalence task, which requires participants learn associations
between faces and different colored fish. Using the same task,
stimulus generalization required participants to learn to equate
the value of two stimuli when they had been associated with
the same response. SSD participants performed better on
stimulus-response learning compared to more hippocampaldependent stimulus-generalization learning. These findings are
consistent with our data, which suggest that variability in
memory performance in SCZ may be related in part, to the
variability in learning strategies available to solve a given task.
In other words, the current data may help explain individual
differences in participants with SSD, in terms of autonomy to
navigate to places without getting lost.
Interestingly, our data suggest the caudate nucleus in participants with SSD is less affected by the disease when compared to
hippocampal systems. However, the extant literature shows inconsistent findings with respect to caudate nucleus abnormalities in
SSD (Beninger et al., 2003; Purdon et al., 2003). Evidence suggests that volumes of the caudate nucleus among individuals
experiencing first episode psychosis do not differ from CON volHippocampus
umes, but that delay in treatment of antipsychotics and positive
symptomology relate to reduced volume in the caudate nucleus
(Crespo-Facorro et al., 2007). Additionally, an antipsychoticna€ıve SCZ sample had reduced caudate nucleus volumes relative
to a CON group (Banner et al., 2011), indicating that duration
of untreated illness might adversely affect the caudate nucleus.
Conversely, treatment with atypical antipsychotics was associated
with increased caudate nucleus volume (Okugawa et al., 2007).
Given the particular relation of response-based strategies to the
caudate nucleus, the 4/8VM offers a valuable tool for future
investigation regarding the functional relevance of the above volumetric findings and their associations with the course of illness
and medications in SSD and SCZ.
LIMITATIONS AND FUTURE DIRECTIONS
In this study, we sampled across the SCZ spectrum and
found a consistent pattern of differential performance within a
small heterogeneous sample of SSD participants. In the future,
a larger study with a greater representation of diagnostic subgroups will be valuable to assess whether the differential pattern
remains consistent across subgroups. However, it remains interesting that half of our SSD sample used a spatial strategy,
despite their robust impairment in doing so; whereas, the other
half of our SSD sample used a response strategy and performed
as well as the CON group. It is of potential interest that the
SSD-Spatial group was slightly more symptomatic and substantially more medicated (in terms of dosage, and taking a combination of typical and atypical antipsychotics) than the SSDResponse group. Follow-up analyses indicate that these variables failed to fully account for the difference in behavioral
impairment within the SSD-Spatial group. Nonetheless, it is an
intriguing possibility that more severe disease and/or higher
medication dosage might influence spontaneous navigation
strategy. For example, higher blockage of dopamine receptors
in the striatum due to antipsychotic medication, may lead to a
shift in spontaneous adoption from a successful response
approach to a deficient hippocampal-dependent spatial
approach. In addition, we did not have the statistical power to
assess the influence of combined usage of antipsychotics and
antidepressants/anxiolytics on behavioral performance on the 4/
8VM. It remains plausible, that use of antidepressants in combination with antipsychotics could ameliorate the SSD-spatial
deficit or prevent an antipsychotic induced shift from a
response to a spatial strategy. Understanding the mechanisms
underlying idiosyncratic spontaneous strategies in CON and
SSD populations, and their relations to symptoms and medication, may shed light on important individual difference factors
that might inform early detection, monitoring disease progression, or treatment strategies for cognitive impairment in SSD.
Based on the current findings, along with its previously demonstrated sensitivity to both hippocampal and striatal system
functioning and their inter-relations, the 4/8VM promises a
valuable tool for such further investigation
SPATIAL STRATEGY DEFICIT IN SCHIZOPHRENIA
The role of prefrontal regions in regulating and coordinating
hippocampal and caudate nucleus memory systems’ involvement in navigation and memory also deserves further attention
(Brown et al., 2012; Dahmani and Bohbot, 2012). In this
regard, participants may flexibly shift their strategies on the 4/
8VM and in real life. However, individuals with SSD who
adopt a deficient spatial strategy may persevere in doing so
rather than shift to a response strategy without external cuing
or instruction. This hypothesis is in line with models supporting SSD as a prefrontal-medial temporal lobe disconnection
syndrome (Spieker et al., 2012). Future studies will be necessary to further explore the role of shifting strategies and navigation in SSD. Cognitive and brain imaging studies will also
help to uncover the extent to which differences in behavioral
performance in the SSD-Spatial and -Response groups reflect
hippocampal, caudate nucleus, and prefrontal networks.
CONCLUSION
In summary, our results reveal a selective deficit among SSD
participants using a Spatial strategy, compared to those using a
Response strategy, despite the fact that all SSD participants
were tested on the same dual-solution task with the same
instructions. These findings corroborate previous reports of
impaired episodic memory and spatial navigation abilities in
SCZ. Moreover, the results highlight the importance of assessing strategies among individuals with SSD as important determinants of navigational and memorial performance and
possibly the related functional integrity of underlying neural
mechanisms such as the caudate-nucleus and hippocampi.
Given the integral roles of spatial memory and navigation, use
of response-based versus spatial strategies may afford more
functional ability in everyday life among those with SSD.
Thus, these findings set the stage for developing interventions
that target strengthening the hippocampal system and spatialstrategies or explore methods that might train SSD participants
to adopt the intact response system during real-world navigation. An important next step, however, will be to better understand the mechanisms underlying individual differences in
spontaneous strategies and their relations to progression of psychosis and antipsychotic medication related factors. Emerging
findings indicate that it will also be important to take into consideration genetic (Banner et al., 2011), hormonal (Bohbot
et al., 2011), and lifespan development factors (Bohbot et al.,
2012; Schwabe et al., 2012).
Acknowledgments
The authors thank Michelle Marcos, Iulia Patriciu, and Carolyn Roy for their assistance with recruitment, data collection,
and data analyses. They also thank Louisa Dahmani for her
assistance with reviewing verbal strategy reports and various
drafts of the manuscript. All study procedures were approved
by the Research Ethics Boards at Ryerson University and St.
1023
Joseph’s Healthcare, Hamilton. LW was supported by a graduate scholarship from the Ontario Mental Health Foundation
(OMHF).
REFERENCES
Banner H, Bhat V, Etchamendy N, Joober R, Bohbot VD. 2011. The
brain-derived neurotrophic factor Val66Met polymorphism is associated with reduced functional magnetic resonance imaging activity
in the hippocampus and increased use of caudate nucleusdependent strategies in a human virtual navigation task. Eur J
Neurosci 33:968–977.
Bartholomeusz CF, Proffitt TM, Savage G, Simpson L, Markulev C,
Kerr M, McConchie M, McGorry PD, Pantelis C, Burger GE,
Wood SJ. 2011. Relational memory in first episode psychosis:
Implications for progressive hippocampal dysfunction after illness
onset. Aust NZ J Psychiatry 45:206–213.
Beninger RJ, Wasserman J, Zanibbi K, Charbonneau D, Mangels J,
Beninger BV. 2003. Typical and atypical antipsychotic medications
differentially affect two nondeclarative memory tasks in schizophrenic patients: A double dissociation. Schizophr Res 61:281–
292.
Bohbot VD, Iaria G, Petrides M. 2004. Hippocampal function and
spatial memory: Evidence from functional neuroimaging in healthy
participants and performance of patients with medial temporal
lobe resections. Neuropsychology 18:418–425.
Bohbot VD, Lerch J, Thorndycraft B, Iaria G, Zijdenbos AP. 2007.
Gray matter differences correlate with spontaneous strategies in a
human virtual navigation task. J Neurosci 27:10078–10083.
Bohbot VD, Gupta M, Banner H, Dahmani L. 2011. Caudate
nucleus-dependent response strategies in a virtual navigation task
are associated with lower basal cortisol and impaired episodic
memory. Neurobiol Learn Mem 96:173–180.
Bohbot VD, McKenzie S, Konishi K, Fouquet C, Kurdi V, Schachar
R, Boivin M, Robaey P. 2012. Virtual navigation strategies from
childhood to senescence: Evidence for changes across the life span.
Front Aging Neurosci 4:1–10.
Bowie CR, Harvey PD. 2006. Cognitive deficits and functional outcome in schizophrenia. Neuropsychiatr Dis Treat 2:531–536.
Boyer P, Phillips JL, Rousseau FL, Ilivitsky S. 2007. Hippocampal
abnormalities and memory deficits: New evidence of a strong pathophysiological link in schizophrenia. Brain Res Rev 54:92–112.
Brown TI, Ross RS, Tobyne SM, Stern CE. 2012. Cooperative interactions between hippocampal and striatal systems support flexible
navigation. Neuroimage 60:1316–1330.
Christensen BK, Bilder RM. 2000. Dual cytoarchitectonic trends: An
evolutionary model of frontal lobe functioning and its application
to psychopathology. Can J Psychiatry 45:247–256.
Crespo-Facorro B, Roiz-Santianez R, Pelayo-Teran JM, GonzalezBlanch C, Perez-Iglesias R, Gutierrez A, et al. 2007. Caudate
nucleus volume and its clinical and cognitive correlations in first
episode schizophrenia. Schizophr Res 91:87–96.
Dahmani L, Bohbot VD. 2012. Prefrontal cortex contributions to spatial learning in young adults tested in a virtual navigation task. Session No. 692 (Human Spatial Representation and Navigation).
Paper presented at the Society for Neuroscience Abstract, New
Orleans, USA.
Etchamendy N, Bohbot VD. 2007. Spontaneous navigational strategies and performance in the virtual town. Hippocampus 17:595–
599.
Hippocampus
1024
WILKINS ET AL.
Folley BS, Astur R, Jagannathan K, Calhoun VD, Pearlson GD.
2010. Anomalous neural circuit function in schizophrenia during a
virtual Morris water task. Neuroimage 49:3373–3384.
Girard TA, Christensen BK, Rizvi S. 2010. Visual-spatial episodic
memory in schizophrenia: A multiple systems framework. Neuropsychology 24:368–378.
Grace AA. 2000. Gating of information flow within the limbic system
and the pathophysiology of schizophrenia. Brain Res Brain Res
Rev 31:330–341.
Grice JW, Iwasaki M. 2007. A truly multivariate approach to manova.
Appl Multivariate Res 12:199–226.
Hanlon FM, Weisend MP, Hamilton DA, Jones AP, Thoma RJ,
Huang M, Martin K, Yeo RA, Miller GA, Canive JM. 2006.
Impairment on the hippocampal-dependent virtual Morris water
task in schizophrenia. Schizophr Res 87:67–80.
Hartley T, Maguire EA, Spiers HJ, Burgess N. 2003. The well-worn
route and the path less traveled: Distinct neural bases of route following and wayfinding in humans. Neuron 37:877–888.
Heckers S. 2001. Neuroimaging studies of the hippocampus in schizophrenia. Hippocampus 11:520–528.
Heckers S, Rauch SL, Goff D, Savage CR, Schacter DL, Fischman
AJ, Alpert NM. 1998. Impaired recruitment of the hippocampus
during conscious recollection in schizophrenia. Nat Neurosci 1:
318–323.
Heinrichs RW. 2005. The primacy of cognition in schizophrenia. Am
Psychol 60:229–242.
Iaria G, Petrides M, Dagher A, Pike B, Bohbot VD. 2003. Cognitive
strategies dependent on the hippocampus and caudate nucleus in
human navigation: Variability and change with practice. J Neurosci
23:5945–5952.
Janca A, Kastrup M, Katschnig H, Lopez-Ibor JJ Jr, Mezzich JE,
Sartorius N. 1996. The World Health Organization Short Disability Assessment Schedule (WHO DAS-S): A tool for the assessment
of difficulties in selected areas of functioning of patients with mental disorders. Soc Psychiatry Psychiatr Epidemiol 31:349–354.
Kay SR, Flszbein A, Opfer LA. 1987. The Positive and Negative Syndrome Scale (PANSS) for Schizophrenia. Schizophrenia Bull 13:
261–276.
Keri S, Nagy O, Kelemen O, Myers CE, Gluck MA. 2005. Dissociation between medial temporal lobe and basal ganglia memory systems in schizophrenia. Schizophr Res 77:321–328.
McCarley RW, Wible CG, Frumin M, Hirayasu Y, Levitt JJ, Fischer
IA, Shenton ME. 1999. MRI anatomy of schizophrenia. Biol Psychiatry 45:1099–1119.
McDonald RJ, White NM. 1994. Parallel information processing in
the water maze: Evidence for independent memory systems involving dorsal striatum and hippocampus. Behav Neural Biol 61:260–
270.
Hippocampus
Nelson MD, Saykin AJ, Flashman LA, Riordan HJ. 1998. Hippocampal volume reduction in schizophrenia as assessed by magnetic resonance imaging: A meta-analytic study. Arch Gen Psychiatry 55:
433–440.
O’Keefe J, Nadel L. 1978. The Hippocampus as a Cognitive Map.
New York: Oxford University Press.
Okugawa G, Nobuhara K, Takase K, Saito Y, Yoshimura M,
Kinoshita T. 2007. Olanzapine increases grey and white matter
volumes in the caudate nucleus of patients with schizophrenia.
Neuropsychobiology 55:43–46.
Ornstein TJ, Sahakian BJ, McKenna PJ. 2008. Memory and executive
impairment in schizophrenia: Comparison with frontal and temporal brain damage. Psychol Med 38:833–842.
Packard MG, Hirsh R, White NM. 1989. Differential effects of fornix
and caudate nucleus lesions on two radial maze tasks: Evidence for
multiple memory systems. J Neurosci 9:1465–1472.
Purdon SE, Woodward N, Lindborg SR, Stip E. 2003. Procedural
learning in schizophrenia after 6 months of double-blind treatment
with olanzapine, risperidone, and haloperidol. Psychopharmacology
(Berl) 169:390–397.
Randolph C, Tierney MC, Mohr E, Chase TN. 1998. The Repeatable
Battery for the Assessment of Neuropsychological Status (RBANS):
Preliminary clinical validity. J Clin Exp Neuropsychol 20:310–319.
Robertson GJ, Wilkinson. GS. 2006. Wide Range Achievement Test 4
(WRAT4). Lutz, FL: PAR, Inc.
Schwabe L, Bohbot VD, Wolf OT. 2012. Prenatal stress changes
learning strategies in adulthood. Hippocampus 22:2136–2143.
Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller
E, et al. 1998. The Mini-International Neuropsychiatric Interview
(M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 59 (Suppl 20, 22–33);quiz 34–57.
Smith ML, Milner B. 1989. Right hippocampal impairment in the
recall of spatial location: Encoding deficit or rapid forgetting? Neuropsychologia 27:71–81.
Spieker EA, Astur RS, West JT, Griego JA, Rowland LM. 2012. Spatial memory deficits in a virtual reality eight-arm radial maze in
schizophrenia. Schizophr Res 135:84–89.
Tseng KY, Chambers RA, Lipska BK. 2009. The neonatal ventral hippocampal lesion as a heuristic neurodevelopmental model of schizophrenia. Behav Brain Res 204:295–305.
Weniger G, Irle E. 2008. Allocentric memory impaired and egocentric
memory intact as assessed by virtual reality in recent-onset schizophrenia. Schizophr Res 101:201–209.
Wilkins LK, Girard T, King J, King M, Herdmand K, Christensen B,
King J. 2013. Spatial-memory deficit in schizophrenia under
viewpoint-independent demands in the virtual courtyard task.
Unpublished Observation.