Most ERP studies of face perception have focused on the N170, an ERP component that is systematically larger for faces compared to objects in the time window 140–200 ms, however, some studies have reported face-specific ERP differences as early as 80–120 ms after stimulus onset. Moreover, the N170 itself remains poorly defined, both in terms of stimulus-specificity and cortical origin. We further examine these issues in a new study, where 16 observers discriminated briefly flashed (80 ms) upright and inverted faces, houses and textures (240 trials per condition), while high-density EEG (256 electrodes) recordings were collected. All stimuli had identical amplitude spectra. To examine the stimulus-differences in ERP across time, a spatiotemporal analysis was performed. Within the time window of the classically defined N170, the N170 was larger to faces than to houses, which was in turn larger compared to textures. Surprisingly, the topographic maps during this time window show a different pattern. Although the topography for faces differed significantly from that of textures (suggesting the involvement of different cortical sources), it did not differ significantly from houses, a result at odds with recent reports of face specific N170 topographies (Itier & Taylor, 2004; Rousselet et al. 2004). Further, there was no significant difference between topographic maps obtained with upright and inverted stimuli. A significant effect of inversion on face ERP topography does emerge during the earlier time window of 80–150ms, providing evidence for face-specific processing prior to the N170. Finally, preliminary results from source analyses on individual data using a linear distributed model suggests that from 80 to 280 ms, ERPs to faces and houses are generated by a network of cortical areas including ventral and lateral occipital-temporal areas as well as parietal areas.
This work was supported by NSERC Discovery Grants 42133 and 105494 and the Canada Research Chair program.