Bilateral DLPFC in turn had a significant inhibitory influence on the rAI. In addition, dACC and posterior cingulate cortex (PCC) had significant inhibitory influence, while preSMA and temporal pole had significant excitatory influence on the rAI. These results are shown in Figure 1 and Table S2. Two-sample t tests ABT-888 revealed significant differences between patients and controls in the “causal” outflow from

the rAI to the rDLPFC. In controls, the rAI exerted a significant excitatory influence on right DLPFC (t(34) = 7.42, corrected p < 0.001), while in the patients, this influence was weak (t(37) = 2.06, uncorrected p = 0.047). In addition, there was a group difference in the effect of rAI on precuneus at an uncorrected threshold (p < 0.001, k = 30), where the controls exhibited an excitatory influence (t(34) = 3.14, uncorrected p = 0.004), while the patients exhibited an inhibitory influence (t(37) = −2.18, uncorrected p = 0.036). Patients also showed a significant reduction in the “causal” influence from bilateral visual cortex and right hippocampal formation to the insula when compared to controls. These group differences are shown in Figure 2 and Table 1. In order to investigate the effects of influences of the rDLPFC

on the rest of the brain, we performed voxelwise GCA using a 6 mm spherical region of GDC-0068 interest (ROI) placed in the rDLPFC node showing the significant group difference. The SN was the primary site of dysfunctional “causal” influence on the rDLPFC in patients. Patients had a significantly reduced excitatory Thalidomide effect from the bilateral (more ventral) insula and the dACC to the rDLPFC in addition to a significant

loss of inhibitory effect of the rDLPFC on the bilateral anterior insula and dorsal ACC (Figure 2; Table 2). The results of the one-sample t tests of GCA based on the rDLPFC seed are presented in Figure 3 and Table S3. None of the x-to-y or y-to-x path coefficients from the rAI or the DLPFC seed regions showed significant correlations with antipsychotic dose equivalents (all p > 0.2). The GCA analysis using a homologous left anterior insula seed revealed that the salience-execution loop disturbances are predominantly right lateralized in schizophrenia (further details are presented in the Supplemental Information and in Figures S4 and S5 and Tables S5 and S6). To relate the illness severity to GCA coefficients in patients, we conducted three principal component analyses to extract an illness severity factor, a factor representing the integrity of “causal” interactions within the salience-execution loop (rAI, rDLPFC, and dACC), and a factor representing visual inflow to rAI. A multiple regression analysis was then conducted as described in the Experimental Procedures section. The model had a significant fit (F[3,34] = 4.03, R2 = 0.26, p = 0.015).