Abstract
The composite face effect has traditionally been taken as strong evidence for holistic perceptual encoding of faces. Recently, however, based on signal detection theory, some authors have suggested that the effect is driven almost entirely by decisional, rather than perceptual, factors (Richler et al., 2008). To dissociate the roles of perceptual and decisional factors in this effect, we recorded ERPs to composite faces while participants performed a go-no-go behavioral task in an oddball paradigm. Twenty participants were asked to lift their finger when the top half of a frequent face changed, but not when the bottom of the face changed. Compared to the presentation of the frequently repeated face stimulus, we found a larger ERP signal starting at the N170 face-sensitive component for the infrequent trials in which the face top changed (release from identity adaptation, Jacques et al., 2007) at right occipito-temporal electrode sites (150 ms to 170 ms). Importantly, at the same latency we also found an effect when the unattended face bottom changed (composite face illusion), indicating the perceptual nature of the effect. Participants performed half of the trials with the right or left hand, allowing to measure the lateralized readiness potential (LRP), reflecting motor preparation activation. We hypothesized that if decisional factors play a role in influencing subjects‘ responses, we would observe an LRP on the bottom-change trials, provoked by the initial reaction to the composite illusion, that would resolve when the stimulus was thoroughly processed allowing the subject to determine the no-go trial status. However, we found no indication of indecision in the response, as evidenced by the absence of any LRP deflection for the critical bottom-change trial. These observations provide direct evidence for a perceptual composite face effect measured on electrophysiological signals, well before and independently of any decisional or motor response processes.