Abstract
Background. Contour integration (CI) is a visual process that combines spatially aligned edge elements into unified boundaries or shapes. We have shown that CI is compromised in chronic schizophrenia, especially for higher spatial frequency displays, but it remains unclear when this dysfunction arises in the illness. Methods. To consider the issue, we behaviorally compared chronic schizophrenia patients (SZs; N=25), healthy controls (N=24), and first episode psychosis patients (FEs; N=19) on a CI task. Groups were well-matched on a number of variables including visual acuity and gender. On each trial, subjects identified the screen quadrant thought to contain an integrated target shape, and task difficulty depended on the number of randomly oriented noise elements that were co-presented with the target. Spatial frequency was modulated by scaling the entire screen display (4 or 12 cycles/deg). To consider possible group differences in attention or motivation, certain “catch” trials were presented in which the target was shown by itself without any noise. Results. For catch trials, the FE group did not differ from the controls or schizophrenia group, and all subject groups performed above 95% accuracy. Replicating our earlier study, controls integrated contours under noisier conditions than SZs and this effect was strongest at 12 cycles/deg (ps< .002). The FE group integrated marginally better than the chronic SZ group (p=.06), and significantly worse than the control group (p=.02). The FE group differences did not depend on spatial frequency, although there was a trend for SZs to become worse relative to FEs at 12 cycles/deg (p=.07). Conclusions. Contour integration dysfunction arises by the first episode of psychosis and may worsen as schizophrenia unfolds over time, especially for high spatial frequency displays. Schizophrenia gradually alters the mechanisms underlying fine-grained perceptual organization, perhaps via changes in V1/V2 connectivity.
Meeting abstract presented at VSS 2015