Testing for translation invariance reveals two stages of facial adaptation

Abstract

Facial adaptation - induced by prolonged exposure to an individual face - can bias the perceived identity of a subsequently presented face. The goal of the present study is to test how presenting the adapter and test stimuli in different hemifields will affect the magnitude of the perceptual facial aftereffect as well as its ERP correlates (i.e. the increase in latency and decrease in amplitude of the N170 component).

Subjects performed a gender discrimination task for peripherically (6 deg) presented facial morphs of upright or upside down presented female and male faces.

Each trial consisted of a 5 sec adaptation period followed by a test face. During adaptation two stimuli were displayed on the two sides of the fixation: within a block they were either both Fourier phase-randomised images (control condition) or one was a Fourier image and the other was a prototypical female face. After 200 msec blank a test face image (chosen from female - male morphed facial image series) was displayed for 200 ms randomly on either side of the fixation. ERP was recorded from 23 channels.Throughout the experiments fixation was controlled by an infrared eye tracking system.

The psychophysical results showed strong adaptation effect both when the adapter and test images appeared on the same side of the fixation (SAME) as well as when they were presented in different hemifields (DIFF), compared to the control condition. However, the magnitude of adaptation was approximately twice as large in the SAME condition than in the DIFF condition. The adaptation effects on the N170 ERP component followed a similar pattern to that found in the psychophysical data. Interestingly, the behavioural and electrophysiological results were essentially the same for upright and upside down presented faces.

Our results provide evidence that facial adaptation consist of two components - one is translation invariant and the other is not - that might take place at different stages of face processing.

Footnotes

 This work was supported by grants from the Hungarian National Research and Development Program: 5/0079/2 to G.K. and 2/035 to Z.V. and the VW Foundation to AA.