Purchase this article with an account.
Fakhri Shafai, Ipek Oruc; Congruency effects in the identification of upright versus inverted faces. Journal of Vision 2012;12(9):496. doi: https://doi.org/10.1167/12.9.496.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Face inversion effect is a well-established disruption of perception when viewing faces upside down. Nevertheless, observers can learn to identify inverted faces in a forced-choice setting following brief training. To investigate any qualitative differences between processes underlying recognition of upright and inverted faces we considered two hypotheses based on template matching. One possibility is that inverted faces are compared to inverted internal templates and classified based on the best match without necessarily recognizing facial identity. Alternatively, inverted faces may be first mentally rotated to upright, then matched to upright templates. To discriminate between these alternatives we measured contrast thresholds and reaction times in a task where observers briefly viewed a stimulus and selected it from a choice screen showing all alternatives. In four conditions we tested upright and inverted faces paired with choices displayed in a congruent or incongruent orientation. We hypothesized that upright faces would be recognized through the first route resulting in impaired performance in the incongruent compared to the congruent condition. In contrast, inverted faces may go through the second route predicting a reversed pattern: the incongruent condition would require one mental rotation (test) whereas the congruent condition would require two (test and choice). The results showed a main effect of orientation in both our measures with higher contrast thresholds (p=0.028) and longer reaction times (p=.044) with inverted faces. Contrast thresholds did not vary across congruency (p>.5) suggesting that inherent difficulty of the two tasks were equivalent. For both orientations reaction times in the incongruent condition were approximately 20ms longer than congruent, although this did not reach significance (p=.24). Upon closer examination we found considerable but meaningful individual differences in the reaction time data suggesting both recognition strategies might be available depending on other external factors such as degree of familiarity with the faces.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only