Abstract
Schizophrenia is a chronic psychiatric condition that is generally characterized by hallucinations, delusions, and trouble with speech and thinking. In addition, this condition is characterized by alterations in visual perception. As such, a number of groups have examined how perceptual training may be beneficial in helping those with schizophrenia. Here we present training data from a visual perceptual learning study examining contrast detection and contour integration; targeting low-level and mid-level vision training, respectively. Twenty-four schizophrenia patients (SP) and fifteen neurotypical (NT) individuals were trained on contrast and/or contour detection tasks for up to forty 20-minute sessions. The contrast detection task was comprised of two conditions: a Dynamic condition in which targets appeared at low contrast at different spatial frequencies (SF) and dynamically increased in number during a particular trial, and a Static condition in which 6-8 target gabors appeared at fixed contrast and SF at random locations on the screen. Similarly, for the contour detection task, participants detected contour shapes, which were manipulated by varying the Orientation Jitter (OJ) or Inducer Number (IN) of Gabors that made up the shape. Performance improved in both contrast and contour tasks across the course of training, with the NT group outperforming the SP group. We present details of how parameters of the contrast sensitivity function changed across sessions and between groups, and on performance on the different training modules for contrast and contours in each group. Overall, these results provide insight into perceptual learning in people with schizophrenia and in neurotypical individuals, as well as data-driven approaches for developing robust adaptive techniques based on different distractors and contour shapes used during training.