A female advantage in basic face recognition is absent in schizophrenia

Abstract

Healthy females outperform males on face recognition tasks. Relative to healthy individuals, schizophrenia patients are impaired at face perception. Yet, it is unclear whether the female advantage found in healthy controls is preserved in females with schizophrenia. In the present study, we compared male and female patients and healthy controls on two basic face perception tasks - detection and identity discrimination. In the detection task, subjects located an upright or inverted line-drawn face (or tree) embedded within a larger line-drawing. In the identity discrimination task, subjects determined which of two side-by-side face images matched an earlier presented face image. Healthy females were significantly more accurate than healthy males on face detection, but not on identity discrimination. However, female patients were not more accurate than male patients on either task. On both upright face detection and face identity discrimination, healthy controls significantly outperformed patients. Patients' performance on face detection was closely associated with tree detection and IQ scores, as well as level of psychosis. This pattern of results suggests that a female advantage in basic face perception is no longer available in schizophrenia, and that this absence may be related to a generalized deficit factor which acts to level performance across sexes, and putative changes in sex-related neurobiological differences associated with schizophrenia.

Figure 1

Illustration of the upright face detection task.

Figure 2

Illustration of the face identity discrimination task.

Figure 3

A. Mean accuracy of male and female patients and controls on the three conditions of the detection task. The x-axis specifies the task condition. The y-axis represents mean accuracy, measured as percent of correct trials, of each group/sex. Error bars are representative of ±1 standard error. B. Mean thresholds of male and female patients and controls on face identity discrimination. The x-axis specifies the task condition; the y-axis is the minimum difference in facial similarity required for participants to achieve 80% accuracy, otherwise known as the ‘perceptual threshold.’ The lower a threshold is, the better the performance of the individual subject. Error bars are representative of ±1 standard error.

Figure 4

A. Effect size of the sex difference in patients and controls on the three conditions of the detection task. The x-axis specifies the task condition; the y-axis is effect size of the sex difference, defined as delta averaged across presentation times (13, 26, 52, 104 msec). B. Effect size of the sex difference in patients and controls on the identity discrimination task. The y-axis represents effect size of the sex difference, defined as delta.