Spatial Range of Contour Integration in Schizophrenia

Abstract

Perceptual organization impairments in schizophrenia have most commonly been demonstrated in contour integration (CI) tasks. Here, we tested the hypothesis that CI impairments become more pronounced as the distance between integrated elements increases. In the first experiment, patients (n = 30) and controls (n = 20) were presented an array of Gabor elements, a subset of which could be integrated into an oblong shape. The task was to determine whether the shape pointed up, down, right or left. There were two spacing conditions, corresponding to whether the Gabor elements comprising the contours were spatially separated by 4.5 or 9 wavelengths. For each condition, we employed the method of constant stimuli and modulated delta—the ratio of background element spacing to target element spacing (where higher delta creates an easier task). Accuracy was computed for each delta level and a three-parameter Weibull function was fit to the data for each spacing condition. Both groups performed better (lower delta thresholds) when the target elements were separated only by 4.5 wavelengths. The effect of group was not significant, but there was a trend for the magnitude of patients' performance decrement relative to controls to increase in the 9 wavelength condition. In the second experiment, the shapes pointed only left or right (to reduce key-press errors). In addition, the range of delta values was restricted to the steepest part of the curve (to improve threshold estimates) and the number of trials per delta level was increased (to improve fit). With a new sample (36 patients, 25 controls), we observed main effects of target-spacing and subject group, but no significant interactions. These data suggest that the previously demonstrated CI impairment in schizophrenia is due to factors other than an impairment in integrating features at larger distances, although spatial range may affect level of CI impairment under some conditions.

NIMH.