Abstract
We focused on the early stage of face processing, particularly on the attention mechanisms. We investigated the effect of limited eye movements and attention on the identification of faces. Subjects were asked to memorize each of 30 faces under different viewing conditions: free viewing, limited peripheral viewing, and limited central viewing of a face. They were required to fixate alphanumeric characters successively displayed on various parts of a face image. On detecting a change of the numeric to alphabetic character, subjects had to immediately press the key. In order to manipulate the extent of attention, the color and /or kind of characters was changed. After this task, subjects were tested on their recognition of the faces shown in the first session of the experiment. Reaction times and error rates were measured. We also recorded eye movements by the method of corneal reflection to monitor saccades made by the subject. The results of the recognition task show that hit rates dropped significantly more for the condition of peripheral viewing of a face than for the free viewing and central viewing conditions, and that reaction times were longer. Furthermore, the extent of attention had effects on the performances under each viewing condition. We conclude that narrowing the spatial extent of attention could lead to impairment of configural encoding.