Abstract
When subjects are shown visual stimuli and then have to perform a task involving these stimuli after they have disappeared from the screen, they still tend to fixate the regions where those stimuli had been located, an effect called ‘looking-at-nothing’. We conducted three experiments to examine whether this effect could act as an implicit index of face recognition in both control subjects (n = 48) and individuals with developmental prosopagnosia (n = 8). On each trial, a subject saw a 3-second video of a person’s face and then saw a choice screen of four faces for 1.2 seconds. This was followed by a response screen with empty boxes, at which point they had to respond whether the face in the video had been present in the preceding choice screen. We analyzed the fixations made while subjects were viewing the response screen (without faces present). Control subjects were more likely to fixate the empty box where the target face was presented, The frequency of this looking-at-nothing effect was greater on hit than on miss trials. Conversely, the odds of a correct response were increased if the first fixation was made within the empty target box. The temporal dynamics of the looking-at-nothing effect showed that it was present for the first fixation only and then transitioned to reduced fixations on the empty target box. Across subjects there was a positive correlation between discriminative sensitivity and the frequency of the looking-at-nothing effect. Developmental prosopagnosic subjects showed a similar effect but reduced in magnitude, with the looking-at-nothing evident on hit but not miss trials. We conclude that the looking at nothing effect can index rapid face recognition in both control and developmental prosopagnosic subjects, and that it is correlated with explicit discriminative performance.