The task was presented using E-prime software (http://www.pstnet.com/eprime.cfm). Presentation of stimuli was pseudorandom, and the first 2 stimuli for each participant were target (AX) trials. Correct responses reflect an index-finger button press to an X probe following an A cue (AX trial). All other cues and probes should be correctly identified as nontargets and be given a middle-finger button press. Target (AX) sequence trials are frequent (70.0% of trials) and set up a prepotent tendency to make a target response when the probe letter X occurs. Consequently, trials in which the probe X is preceded by a non-A cue (eg, BX trials) are the most difficult (12.5% of trials). AY and BY trials offer additional control conditions and represent 10.0% and 7.5% of the trials, respectively. The task consisted of 4 runs of 40 trials for a total of 160 trials and a total time of 37 minutes 20 seconds.
Row 1, Patients with first-episode schizophrenia (schizophrenia) compared with control participants (control). Row 2, Patients who received atypical antipsychotics (medicated) compared with control group. Row 3, Patients who did not receive antipsychotics (unmedicated) compared with control group. Row 4, Medicated compared with unmedicated patient groups. The significance scale reflects the transformation −log10(P value).
Graphs reflect β values (AX-CPT cue B − cue A contrast) from a priori left and right dorsolateral prefrontal cortex (DLPFC) regions of interest. Error bars reflect SE. Medicated indicates individuals with first-episode schizophrenia who had been treated for a number of weeks with atypical antipsychotics; unmedicated indicates individuals with first-episode schizophrenia who were not treated with antipsychotics.
eAppendix 1. Methods
eAppendix 2. Results
eFigure 1. Regions of Uncorrected Cortical Thinning Between Groups
eFigure 2. Correlations Between d'-Context and DLPFC Activity in Patients With Schizophrenia
eTable. Regions of Significant fMRI Activity (Cluster Corrected)
Lesh TA, Tanase C, Geib BR, Niendam TA, Yoon JH, Minzenberg MJ, Ragland JD, Solomon M, Carter CS. A Multimodal Analysis of Antipsychotic Effects on Brain Structure and Function in First-Episode Schizophrenia. JAMA Psychiatry. 2015;72(3):226-234. doi:10.1001/jamapsychiatry.2014.2178
Copyright 2015 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Recent data suggest that treatment with antipsychotics is associated with reductions in cortical gray matter in patients with schizophrenia. These findings have led to concerns about the effect of antipsychotic treatment on brain structure and function; however, no studies to date have measured cortical function directly in individuals with schizophrenia and shown antipsychotic-related reductions of gray matter.
To examine the effects of antipsychotics on brain structure and function in patients with first-episode schizophrenia, using cortical thickness measurements and administration of the AX version of the Continuous Performance Task (AX-CPT) during event-related functional magnetic resonance imaging.
Design, Setting, and Participants
This case-control cross-sectional study was conducted at the Imaging Research Center of the University of California, Davis, from November 2004 through July 2012. Participants were recruited on admission into the Early Diagnosis and Preventive Treatment Clinic, an outpatient clinic specializing in first-episode psychosis. Patients with first-episode schizophrenia who received atypical antipsychotics (medicated patient group) (n = 23) and those who received no antipsychotics (unmedicated patient group) (n = 22) and healthy control participants (n = 37) underwent functional magnetic resonance imaging using a 1.5-T scanner.
Main Outcomes and Measures
Behavioral performance was measured by trial accuracy, reaction time, and d′-context score. Voxelwise statistical parametric maps tested differences in functional activity during the AX-CPT, and vertexwise maps of cortical thickness tested differences in cortical thickness across the whole brain.
Significant cortical thinning was identified in the medicated patient group relative to the control group in prefrontal (mean reduction [MR], 0.27 mm; P < .001), temporal (MR, 0.34 mm; P = .02), parietal (MR, 0.21 mm; P = .001), and occipital (MR, 0.24 mm; P = .001) cortices. The unmedicated patient group showed no significant cortical thickness differences from the control group after clusterwise correction. The medicated patient group showed thinner cortex compared with the unmedicated patient group in the dorsolateral prefrontal cortex (DLPFC) (MR, 0.26 mm; P = .001) and temporal cortex (MR, 0.33 mm; P = .047). During the AX-CPT, both patient groups showed reduced DLPFC activity compared with the control group (P = .02 compared with the medicated group and P < .001 compared with the unmedicated group). However, the medicated patient group demonstrated higher DLPFC activation (P = .02) and better behavioral performance (P = .02) than the unmedicated patient group.
Conclusions and Relevance
These findings highlight the complex relationship between antipsychotic treatment and the structural, functional, and behavioral deficits repeatedly identified in schizophrenia. Although short-term treatment with antipsychotics was associated with prefrontal cortical thinning, treatment was also associated with better cognitive control and increased prefrontal functional activity. This study adds important context to the growing literature on the effects of antipsychotics on the brain and suggests caution in interpreting neuroanatomical changes as being related to a potentially adverse effect on brain function.
In the 40 years since the classic study by Johnstone and colleagues,1 schizophrenia has been understood as a disorder characterized by alterations in brain structure, with the most robust findings including ventricular enlargement and reductions in overall gray matter volume.2 Studies of cortical thickness and surface area have also revealed consistent patterns of frontotemporal thinning in patients with chronic3- 5 and first-episode6- 10 schizophrenia. In recent years, accumulating evidence suggests that antipsychotics may contribute to observed structural brain changes, including the cortical and subcortical structures.11,12 Some studies suggest that typical antipsychotics may be associated with more dramatic structural changes (ie, larger reductions in gray matter volume, increased caudate and putamen volume) than atypical agents,5,13 whereas others fail to find a structural relationship with specific medications.3,14 In addition, studies of brain volume, cortical thickness, and surface area in antipsychotic-naive8,15 and prodromal16 individuals reveal that structural abnormalities are present before the initiation of antipsychotic therapy. These findings suggest that antipsychotics may contribute to findings of gray matter reduction in schizophrenia, but measurable changes are also caused by the illness itself.
Although antipsychotic therapy has the clear benefits of reducing psychotic symptoms and relapse in schizophrenia,17 recent findings of an association between gray matter loss and antipsychotic treatment in schizophrenia has raised concerns about the potential adverse effects of antipsychotics that may need to be taken into consideration when prescribing these agents.18 Recent studies have identified relationships among brain structure abnormalities, cognition, and brain function early in the course of schizophrenia,19 including the association of white matter volume loss and cognitive impairment.20 Therefore, examining the influence of antipsychotics on measures of brain structure and function (ie, task-related brain activity) is of critical importance to understand adequately how these factors contribute to the illness. However, to our knowledge, no studies to date have examined conjointly whole-brain cortical thickness and functional engagement of the prefrontal cortex (PFC) in groups of patients with first-episode schizophrenia who have or have not received antipsychotics. Our study seeks to address this clinically significant gap in knowledge.
More specifically, this investigation examines cortical structure and function within 1 year of psychosis onset in individuals who had not received antipsychotics (unmedicated patient group) or had been treated for a number of weeks with atypical antipsychotics (medicated patient group). Given numerous studies identifying structural alterations in the frontal regions, we hypothesized that antipsychotic treatment would be associated with a thinner cortex in the PFC. The AX version of the Continuous Performance Task (AX-CPT)21 was used to evaluate cognitive control performance and PFC recruitment because it is a robust and reliable measure of a known executive deficit in schizophrenia.22 Some studies have found marginal improvement in neuropsychological measures in patients taking atypical antipsychotics,23 and we anticipated that the medicated patient group would show AX-CPT performance intermediate to those of the healthy control group and the unmedicated patient group. In addition, animal studies showed increased prefrontal dopamine release as a consequence of antipsychotic administration24,25 and evidence of medication-related increases in the blood oxygen level–dependent (BOLD) response in the PFC after starting antipsychotic treatment or after switching to an atypical antipsychotic26- 28; we therefore anticipated that patients receiving medication would show BOLD activity intermediate to that of the control group and the unmedicated patient group. In short, we anticipated that atypical antipsychotic treatment may be associated with a net positive response in terms of higher behavioral performance and brain activity even in the context of cortical thinning.
The study was approved by the institutional review board of the University of California, Davis. Participants provided written informed consent after receiving a complete description of the study and were compensated.
Forty-five patients with first-episode schizophrenia (36 with schizophrenia, 5 with schizoaffective disorder, and 4 with schizophreniform disorder) were recruited from the Early Diagnosis and Preventive Treatment Clinic at the University of California, Davis, along with 37 healthy control participants aged 15 to 26 years. Table 1 outlines the participants’ demographic and clinical status at the time of testing. Participants with schizophrenia were outpatients within 1 year of the onset of psychotic symptoms. Twenty-three patients were currently treated with atypical antipsychotics (medicated patient group), and the remaining 22 patients were not receiving antipsychotics (unmedicated patient group). Of the 22 patients in the unmedicated patient group, 17 were antipsychotic naive and the remaining 5 had discontinued medication more than 1 month before the study. Detailed inclusion and exclusion criteria are described in eAppendix 1 in the Supplement.
The AX-CPT has been described in great detail previously,21 and the task parameters specific to this implementation are shown in Figure 1. The task was presented using E-prime software (http://www.pstnet.com/eprime.cfm). Briefly, the participants are presented with a series of cues and probes and are instructed to make a target response (pressing a button with the index finger) to the probe letter X only if it was preceded by the cue letter A. All cues and nontarget probes require nontarget responses (pressing a button with the middle finger). Target sequence trials are frequent and set up a prepotent tendency to make a target response when the probe letter X occurs. As a result, nontarget sequence trials in which any non-A cue (collectively called Bcues) is presented and followed by a probe letter X require the most cognitive control.
A specific measure of cognitive control performance, d′-context,21 was computed from AX hits and BX false alarms and analyzed using a 1-way analysis of variance (ANOVA). Analyses of AX-CPT accuracy and reaction time are reported in eAppendix 2 in the Supplement. Measures passing significance in the overall ANOVA underwent pairwise t tests (2-tailed) between groups, with the exception of comparisons between patient groups, which included duration of illness as a covariate in the analysis of covariance. We used the Holm-Bonferroni method29 to correct for multiple comparisons on post hoc pairwise tests. Group comparisons on measures that violated sphericity assumptions were adjusted using the Greenhouse-Geisser correction.30
Functional magnetic resonance imaging data were obtained using a 1.5-T scanner (Signa; GE Healthcare). For the AX-CPT, T2-weighted echoplanar imaging sessions used the following settings: repetition time, 2000 milliseconds; echo time, 40 milliseconds; flip angle, 90°; and field of view, 22 cm. Functional images consisted of 24 contiguous and interleaved 4.0-mm axial sections with a 3.4-mm2 in-plane resolution. Preprocessing steps are outlined in eAppendix 1 in the Supplement. Functional imaging analysis was performed in statistical parametric mapping software (SPM8; http://www.fil.ion.ucl.ac.uk/spm/software/spm8/) using the general linear model. All trial types were modeled, and only correct responses were included in the reported contrasts. Regressors included all cues, probes, and error trials. Translational and rotational movement data were included as covariates of noninterest. Group-level random-effects comparisons were performed between groups for the AX-CPT contrast subtracting cue A from cue B (cue B − cue A contrast) to measure activation under conditions of high vs low cognitive control. Contrasts were thresholded at the voxel level (P < .01), and clusters were considered significant if they survived cluster-level familywise error correction (P < .05).
In addition to whole-brain analyses, a priori hypotheses regarding the dorsolateral PFC (DLPFC) prompted the interrogation of left and right DLPFC regions of interest. Details concerning the selection of regions of interest and analysis are included in eAppendix 1 in the Supplement.
Spoiled gradient recalled images were collected in the same session using the following parameters: repetition time, 9 milliseconds; echo time, 2 milliseconds; flip angle, 15°; field of view, 22 cm; one hundred twenty-four 1.5-mm axial sections; and 0.86-mm2 in-plane resolution. Images were processed with the FreeSurfer software package (version 4.3; http://surfer.nmr.mgh.harvard.edu)31,32 using the processing stream described in eAppendix 1 in the Supplement.
Measurements of cortical thickness were obtained for each vertex and mapped on a common spherical coordinate system. Maps were smoothed with a 10-mm gaussian kernel, and right and left hemispheres were tested separately. To correct for multiple comparisons, a cluster analysis was conducted using a Monte Carlo simulation with 10 000 iterations. The vertexwide threshold was set at P < .01 for simulation and clustering, and clusters were considered significant if they survived a clusterwise probability of P < .05. Duration of illness was included as a covariate of noninterest in the patient subgroup comparison.
Participant demographic and clinical information is presented in Table 1. The groups did not differ significantly by age, sex, handedness, or parental educational level. We found a trend for groups to differ on participant educational level (F2,79 = 2.82; P = .07). Post hoc independent-sample t tests revealed that the controls completed significantly more years of education compared with participants in the unmedicated patient group (t57 = 2.44; P < .05); the difference between the control group and the medicated patient group had a trend for more years of education (t58 = 1.76; P = .08). The medicated and unmedicated patient groups did not differ on years of education (t43 = 0.65; P = .65). Significant group differences emerged for estimated IQ (F2,77 = 7.02; P < .01). The control group showed a significantly higher estimated IQ compared with the medicated (t58 = 3.48; P < .01) and unmedicated (t57 = 2.97; P < .01) patient groups. The IQ did not differ between the medicated and unmedicated patient groups (t43 = 0.27; P = .79). The medicated and unmedicated patient groups also did not differ on duration of illness (t43 = 1.75; P = .09) or on scores from the modified Global Assessment of Functioning (t43 = 0.75; P = .46),33 Brief Psychiatric Rating Scale (t43 = 0.35; P = .73),34 Schedule for the Assessment of Negative Symptoms (t43 = 0.09; P = .93),35 or Schedule for the Assessment of Positive Symptoms (t43 = 1.62; P = .11).36
Between-group cortical thickness comparisons are presented in Table 2 and rendered in Figure 2 (cluster-corrected) and eFigure 1 in the Supplement (uncorrected). Comparison of the control group with both schizophrenia patient groups revealed significant cortical thinning in the schizophrenia groups in the left supramarginal gyrus, right rostral middle frontal gyrus, right superior parietal cortex, right middle temporal gyrus, and right lateral occipital cortex. Compared with the control group, the medicated patient group showed cortical thinning in the bilateral rostral middle frontal gyrus (0.22- and 0.27-mm mean reduction in right and left hemisphere, respectively), left orbitofrontal cortex (0.22-mm mean reduction), left pars opercularis (0.24-mm mean reduction), left fusiform gyrus (0.22-mm mean reduction), left supramarginal gyrus (0.23-mm mean reduction), left precuneus (0.22-mm mean reduction), right superior frontal gyrus (0.25-mm mean reduction), right lateral occipital cortex (0.24-mm mean reduction), right superior parietal cortex (0.21-mm mean reduction), and right superior temporal sulcus (0.34-mm mean reduction). In contrast, the unmedicated patient group showed no significant differences in cortical thickness compared with controls. Finally, comparing the medicated with the unmedicated patient groups revealed thinner cortex in the medicated patient group in 2 regions of the left rostral middle frontal gyrus (0.26-mm mean reduction), in the left middle temporal gyrus (0.33-mm mean reduction), and in the right pars opercularis (0.16-mm mean reduction).
Table 1 provides behavioral data; eAppendix 2 in the Supplement provides an analysis of all AX-CPT conditions. However, the primary analysis focused on d′-context. One-way ANOVA of d′-context scores revealed significant group differences (F2,79 = 13.61; P < .001). Pairwise tests revealed lower d′-context scores in the unmedicated patient group compared with the control group (t57 = 5.68; P < .001) and medicated patient group (F1,42 = 5.53; P = .02) and lower d′-context scores in the medicated patient group compared with the control group (t33 = 2.17; P = .04), all of which remained significant after correcting for multiple comparisons.
Within-group results with significant clusters are described in detail in the eTable in the Supplement. Whole-brain comparison (Figure 3) of the controls with all patients with schizophrenia revealed higher activity in the controls in the right DLPFC and bilateral inferior parietal cortex under conditions requiring high cognitive control. The medicated patient group showed no significant differences compared with the control group, although the controls showed higher frontal and parietal activity at lower thresholds. In contrast, comparison of the unmedicated patient group with the control group revealed robust differences, with controls showing significantly higher activity in the bilateral DLPFC and inferior parietal cortex. Finally, a comparison of the medicated and unmedicated patient groups revealed significantly higher activity in the bilateral DLPFC of patients in the medicated group.
The ANOVA of the DLPFC regions of interest (Figure 3) revealed a significant between-group difference in cue B − cue A activity on the right (F2,78 = 13.62; P = .001) and left (F2,78 = 6.22; P = .003) DLPFC. Post hoc t tests revealed significantly higher activity in the control group compared with the medicated (t58 = 2.38; P = .02) and unmedicated (t56 = 5.50; P < .001) patient groups in the right DLPFC. Within the left DLPFC, the control group showed higher activity compared with the unmedicated patient group (t56 = 3.72; P = .001) but not the medicated patient group (t58 = 1.21; P = .23). The medicated patient group showed higher activity compared with the unmedicated patient group in the right (F1,41 = 6.34; P = .02) and left (F1,41 = 4.89; P = .03) DLPFC. All significant findings remained so after correction for multiple comparisons.
Three additional analyses were performed and are presented in the Supplement. First, follow-up cortical thickness and functional magnetic resonance imaging analyses performed only on the subgroup of 17 antipsychotic-naive patients revealed results similar to those of the whole sample (eAppendix 2 in the Supplement). Second, we explored relationships between functional and behavioral data and identified a significant positive relationship between d′-context and left DLPFC BOLD activity when looking at the patient group as a whole (eFigure 2 in the Supplement). Finally, we included DLPFC β values in a vertexwise analysis of cortical thickness and identified no significant relationships between the functional and structural variables.
As expected, the medicated patient group showed significant thinning in the PFC and middle temporal regions when compared with the unmedicated patient group and in the prefrontal, middle temporal, parietal, and occipital regions when compared with the control group. Despite these effects on cortical thickness, we saw no evidence of deleterious effects of atypical antipsychotic treatment on cognition and brain activity in this group of patients with first-episode schizophrenia. To the contrary, examination of behavioral performance and BOLD activity during the AX-CPT revealed better performance and increased DLPFC activity in the medicated compared with the unmedicated patient groups. In addition, both patient groups showed significantly reduced DLPFC activity and performance decrements compared with the control group. We also identified a significant positive relationship between behavioral performance and left DLPFC BOLD activity in the patient groups. This combination of structural, functional, and behavioral findings adds to an already substantial literature identifying DLPFC impairment in schizophrenia37,38 and contributes novel findings related to the effects of medication on the brain and behavior in this illness.
The cortical thickness results we described are consistent with those of other studies of patients who received antipsychotics and have shown cortical thinning in samples with first-episode schizophrenia.6,10 Although Narr and colleagues9 implemented different study methods, their findings of thinning are similar to ours and complement region of interest–based studies of anterior cingulate cortex thinning.7 Our findings are also generally consistent with those of studies of cortical volume in which antipsychotic treatment was associated with decreased gray matter volume in the frontal cortex and overall reduced gray matter volume.12,13,18 However, some studies have identified preserved or even increased cortical volume or thickness in patients treated with atypical antipsychotics.13,39,40
Long-term and even relatively brief exposure to antipsychotics has been linked to widespread loss of gray matter volume in macaque monkeys41 and rodents,42 with the most robust findings in the frontal and parietal cortices. Konopaske and colleagues43 demonstrated that these volume losses in nonhuman primates can be explained by fewer glial cells and higher densities of neurons.44 However, the mechanism by which antipsychotics may produce these effects remains unclear and will likely be the focus of significant future work.
Our finding of increased prefrontal activity in the present study’s medicated patient group is consistent with that of previous work by Jones and colleagues,28 who identified increased PFC activity in a small group of medicated patients compared with a drug-naive group during a verbal fluency task. A similar finding was also identified using a prospective design in which untreated patients who underwent subsequent testing after 12 weeks of quetiapine fumarate treatment showed increased PFC activity during a working memory task.26 However, both of these studies used relatively small samples, and other work using positron emission tomography45 provides conflicting evidence of the effect of antipsychotics on prefrontal cerebral blood flow.
Although the mechanism for antipsychotic treatment effects on brain structure and function is unclear, neuroinflammatory models provide a potential link. A growing body of evidence implicates neuroinflammation in the pathophysiological features of schizophrenia,46 including elevations in proinflammatory cytokine levels47 and microglia activation48 and increased extracellular volume in white and gray matter.49 However, antipsychotic treatment has been associated with an anti-inflammatory effect,50 which could promote decreases in extracellular volume and activated glia and improve neuronal function and consequently cognition. Thus, the interaction of antipsychotic treatment and neuroinflammatory processes at the first episode reflects one potential mechanism to address our findings. Another potential mechanism to explain improved cognitive performance and BOLD activity in the medicated patient group would be the effect of second-generation antipsychotics on the dopaminergic and serotonergic systems. Several studies24,25 have found increased prefrontal dopamine release as a consequence of serotonin2A and D2 receptor blockade. Given that schizophrenia has been associated with decreased dopaminergic activity within the mesocortical system,51 atypical antipsychotics could improve dopaminergic tone in the prefrontal cortex, with beneficial effects on cognition and potentially BOLD activity.26,52 The absence of a significant relationship between BOLD activity and cortical thickness suggests that thinning per se is an unlikely explanation for increases in BOLD activity in patients who use antipsychotics. The mechanism by which antipsychotics are associated with thinning and functional/behavioral performance improvements may therefore be different. Nonetheless, higher performance and greater BOLD activity in the medicated sample of patients highlight the potentially positive effects of antipsychotics in what traditionally might have been interpreted as a detrimental effect if cortical thickness was examined in isolation.
Limitations include the naturalistic between-subject design of the study. The spoiled gradient recalled images collected in a 1.5-T scanner represent lower resolution than can be obtained on 3-T scanners. Consequently, the reduced precision of these data may underestimate the differences between groups. Although this study demonstrates no detrimental effect of medication on performance of the AX-CPT, the possibility exists that medication has a different effect on other behavioral tasks. However, the importance of cognitive control for goal-directed behavior would lead to the prediction that other cognitive tasks with superordinate goals (ie, memory and language tasks) would show comparable medication effects. In addition, a prospective, within-subject, counterbalanced drug/placebo design would be preferable to the naturalistic design of the present study, but such a study would not be feasible or ethically justified. Important pathophysiological differences may exist between patients who went untreated before entering the study and those who received atypical antipsychotic treatment that could account for improved cognition and prefrontal recruitment in the medicated patient group. We believe that such differences are highly unlikely for a number of reasons. First, all patients were quite early in the course of their illness (mean duration of 6 months) and were well matched on variables such as age, sex, IQ, socioeconomic status, and clinical symptoms. Second, the mean duration of antipsychotic treatment was brief (99 days) and the mean dose was low (190-mg chlorpromazine equivalent). Finally, the only clinical difference between the medicated and unmedicated patient groups was a nonsignificant difference in positive symptoms, which would be consistent with the known clinical effects of these agents.
The present study provides important new data to inform the debate regarding the functional significance of the effects of antipsychotics on cortical gray matter. In addition to replicating previous reports of reduced cortical thickness in a cohort of patients with briefly medicated first-episode schizophrenia, we find no evidence of a deleterious effect of this treatment on higher cognition or on underlying neurophysiological features. Short-term treatment with atypical antipsychotics was associated with better cognition and functional brain activity in these individuals. Additional data regarding the longer-term effects of antipsychotics on brain structure and function are needed to inform this debate, with further data from animal models that might provide additional insights into the neurobiological features of the antipsychotic effects on brain structure and function.
Submitted for Publication: January 21, 2014; final revision received August 19, 2014; accepted August 25, 2014.
Corresponding Author: Cameron S. Carter, MD, Imaging Research Center, University of California, Davis, 4701 X St, Ste E, Sacramento, CA 95817 (email@example.com).
Published Online: January 14, 2015. doi:10.1001/jamapsychiatry.2014.2178.
Author Contributions: Dr Lesh had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Lesh, Carter.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lesh, Carter.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Lesh, Minzenberg, Ragland, Carter.
Obtained funding: Carter.
Administrative, technical, or material support: Lesh, Tanase, Geib, Niendam, Yoon, Solomon.
Study supervision: Lesh, Carter.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by grant 5R01MH059883 from the National Institutes of Health (Dr Carter).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentations: Preliminary data from this paper were presented at the 51st Annual Meeting of the American College of Neuropsychopharmacology; December 4, 2012; Hollywood, Florida and at the 14th International Congress on Schizophrenia Research; April 24, 2013; Orlando, Florida.
Additional Contributions: We thank the study participants and their families. Third- and fourth-year undergraduate interns Ashley Falzone, Samuel Crowley, Markie Benavidez, Alex Mawla, Shivani Desai, Weizhen Xie, and Agnieszka Lewandowski assisted with data collection and processing. Each undergraduate intern contributed to the project for a minimum of 1 year and a maximum of 3 years. They received no financial compensation.