Antipsychotic Effects on Brain Function in Schizophrenia

Patients with schizophrenia have significant problems in thinking, referred to as cognitive dysfunction, which make it difficult for them to function well, and have no effective treatment at this time. A number of existing medications currently used in other medical or psychiatric conditions have been considered, however, none of them has yet shown unequivocal signs of effectiveness for these problems in schizophrenia. One important reason for this, that has not been properly tested, is how antipsychotic treatment may affect the impact of these newer medications for cognition. In the present study, the investigators will evaluate how two different antipsychotic medications may have a different impact on the brain processes that give rise to important cognitive functions that are impaired in schizophrenia. They will conduct functional MRI in schizophrenia patients while they perform a cognitive task, both before and after 8 weeks of treatment with either aripiprazole or risperidone. Then the investigators will evaluate the effects of a new medication called modafinil, and how it may improve brain and cognitive function, depending on which antipsychotic medication the patients are taking. The results of this study may inform researchers, doctors and patients about how to understand the effects of these drugs on the brain and cognition, and which treatments might be best to improve cognition in schizophrenia, in order to improve these patients' ability to function in their lives.

40 right-handed adults with schizophrenia and no significant medical or neurological illness (18-30 years old) will 1) undergo fMRI during performance of a cognitive control task, then 2) randomize to an 8-week, double-blind antipsychotic treatment trial, followed by 3) repeat fMRI in a single-dose, counterbalanced study of Modafinil (one 200 milligram oral add-on dose) vs. Placebo (Placebo). Neuroimaging measures will be the primary endpoint, cognition secondary, and symptoms an exploratory endpoint.

Diagnoses will be made by SCID-I and DSM-IV-TR criteria, by a PhD or MD clinician (with demonstrated reliability), followed by consensus. A research pharmacist will randomly assign treatment (without stratification), and package ARI and RIS in identical-appearing capsules. A blinded clinician will measure symptoms within 3 days before antipsychotic initiation and weekly thereafter; adjust doses based on weekly evaluations of symptoms and side effects, and evaluate treatment adherence and substance use. The diagnostician, pharmacist and treating psychiatrist are otherwise uninvolved in the study. Low-dose benztropine (≤2 mg daily), lorazepam (≤1 mg daily) and antidepressant treatment will be permitted, but not adrenergic agents or anticonvulsants. Initiation, stepped, and maximal antipsychotic doses will respectively be: ARI (7.5, 7.5 and 30 mg, daily); RIS (2, 2 and 8 mg daily). If patients do not tolerate the treatment (at the minimal dose), or exhibit treatment-refractory symptoms (at the highest tolerated dose), they will discontinue the study, the blind broken, and transition into naturalistic treatment in the clinic. Receptor antagonist loads (daily dose X Haldol equivalent) will be computed for patients at unblinding, using indices for α2 and α1 receptors, D2 and muscarinic receptors. In the last half of the 8th week of antipsychotic treatment, subjects will undergo fMRI on two days, separated by one day for modafinil washout. They will receive modafinil (200 mg po) on one day and Placebo on the other (double-blind, counterbalanced), each in mid-morning, and scan 3-4 hours later, during the average peak plasma levels of modafinil 30. All subjects will provide informed consent for all procedures, approved by the UCSF IRB. This study will meet the CONSORT standard for clinical trial design, conduct and reporting.

functional MRI. 3 Tesla Siemens Timm Trio with 8 channel coil will be used for event-related fMRI, with single-shot, T2*-weighted sequence, (TR 2000 ms, TE 30 ms, flip angle 90°, FOV 220 x 220 mm, 36 contiguous axial oblique slices, 3.4 mm isotropic voxels). A structural MRI (MP-RAGE) will be acquired for normalization of EPI images. Pre-processing will include spatial realignment, slice timing correction, spatial normalization to T1 template, 8mm smoothing kernel. The GLM will be used with HRF convolved with a series of delta functions, and regressors for Task events, for each combination of Modafinil and Placebo and either Cue or Target. Errors/no-response trials are modeled separately. The functional connectivity analysis will use an LC-seeded beta series method; each event modeled uniquely and the time series of betas (segregated by task/treatment condition) correlated with the seed at every voxel. The LC will be localized in EPI images with a mask in MNI space, derived from voxels meeting these criteria: A) located in rostrodorsal pons, adjacent to 4th ventricle/sylvian aqueduct, restricted to an anatomically-defined dorsal pontine mask (e.g., not extending rostrally into the mesencephalon or superiorly into the 4th ventricle; and B) showing significant association with in-scanner pupil diameter, using this as a parametric regressor in GLM of BOLD time series, in an independent, healthy sample without treatment.

Specific Aim 1. To test whether aripiprazole preserves LC, PFC and cognitive control function to a greater degree than risperidone in schizophrenia outpatients. Hypothesis 1. ARI group will exhibit greater task-related PFC activity and LC-PFC connectivity than RIS group, after 8 weeks of antipsychotic treatment: (post-antipsychotic Placebo) vs. (pre-antipsychotic Baseline).

Specific Aim 2A. To test whether aripiprazole is superior to risperidone in permitting modafinil enhancement of LC and PFC activity and cognition. Hypothesis 2A. The ARI group, vs. RIS group, will exhibit 2A) greater effects of Modafinil (vs. Placebo) on task-related LC activity and greater cognitive control task improvement. Specific Aim 2B. To test the role of α2 autoreceptor antagonism by antipsychotic medications as a source of blunted LC response to Modafinil. Hypothesis 2B. Across groups (ARI+RIS), α2 autoreceptor antagonist load correlates with impaired Modafinil effects (vs. Placebo) on task-related LC activity and cognitive control performance. These aims are related but not interdependent: antipsychotic effects on LC/PFC/cognitive function, vs. antipsychotic effects in moderating modafinil actions on these same measures, represent distinct research questions.

Hypothesis 1 will be tested with the contrast indicated as per the Aim 1 Prediction in the table above. This is analogous to a directional test of the Group-by-Time-by-Task interaction. Hypothesis 2A will be tested as per the Aim 2 Prediction in the table. We predict that ARI treatment supports greater task-related Modafinil effects on LC activity, vs. RIS treatment. Hypothesis 2B will be tested across all patients by correlating α2 load (by antipsychotics) with mean task-related LC betas from the Treatment X Task Condition contrast. Significance for all analyses is set at p<.05, FDR-corrected, or using the LC as SVC, as appropriate. Accuracy and RT cost will be tested by ANOVA.