AbstractThe perceptual determinants of gaze aversion by humans were examined in psychotic children with autistic symptoms, who persistently avoid eye-contact, and in normal children. Previous studies of normal adults have indicated that two concentric discoid elements, schematically resembling two facing eyes, are arousing stimuli. It was of interest to determine if similar eyelike schemata, as compared with less eyelike schemata of greater or lesser complexity, would elicit the least visual inspection as an arousal-reducing cut-off act. To clarify discrepancies in the literature as to whether normal and psychotic children differ in their sensitivity to prolonged eye-contact and concomitant readiness to gaze avert, these groups were compared for differential responsiveness to the same stimuli. Measurement of different inspection of eyelike schemata was accomplished using a special model-viewing apparatus in which two models were presented side-by-side in paired comparisons. The duration of each glance directed at a particular model was recorded on video tape and later measured with a stopwatch. Three experiments were performed. The first examined 10 normal and 10 psychotic children during presentations of 5 models comprising a blank model and models with I through 4 concentric discoid elements separated by the same interpupillary distance as are real human eyes. The second experiment examined 15 psychotic children using models presenting two concentric discoid elements in vertical, diagonal, and horizontal orientations. The third experiment examined 10 normal and 10 psychotic children using 5 models comprising two schematic facing eyes as represented by concentric discoid elements, more realistic eyes with staring and adverted irises, and closed eyes. The following results were obtained: I. A model won a particular comparison bout if it was looked at longer than its comparison model. In all experiments and in both groups of children, the model presenting two horizontally placed concentric discoid elements won the fewest number of paired comparison bouts against the field of other models. The only exception occurred for normals in the first experiment in which the blank model won fewer comparison bouts than the model presenting two concentric discoid elements. 2. For the over-all duration of model inspection in the first experiment, the psychotics looked longer at the models than did the normals. However, these groups did not differ as much for the model with two concentric discoid elements. Both groups, particularly the psychotics, looked less at the model presenting two concentric discoid elements than at models presenting other arrangements of concentric discoid elements in the first experiment. Similarly in the third experiment, both groups looked less at the model with two concentric discoid elements than at models with staring and averted irises, albeit these differences were not significant. Normals and psychotics did not differ appreciably in their over-all looking at the more realistic eyelike schemata in the third experiment. With respect to the spatial orientation of two concentric discoid elements in the second experiment, the psychotics looked the least at the model presenting two discoid elements in the horizontal plane. In light of the clinical observations that normals and psychotics differ in gaze behavior, the findings that normal adults are aroused by two concentric discoid elements, and the theoretical aspects of the cut-off hypothesis, the results suggest the following: I. Two horizontally placed concentric discoid elements, schematically resembling two facing eyes, are provocative stimuli eliciting less visual inspection in accordance with the cut-off hypothesis. 2. The lack of differential gazing by normals and psychotics during presentations of models with two schematic facing eyes seems to indicate that, in the absence of supporting features of a human face, two schematic facing eyes were equally provocative to both groups of children.