Purchase this article with an account.
Alyson Saville, Carol Huynh, Benjamin Balas; Face animacy does not impact the N170 inversion effect. Journal of Vision 2014;14(10):127. doi: 10.1167/14.10.127.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Face orientation is known to affect visual processing such that inversion typically impairs recognition ability (Yin, 1969). While the electrophysiological component known as the N170 is generally greater for faces versus non-face stimuli (Bentin, Allison, Puce, Perez, & McCarthy, 1996), inverted faces typically exhibit higher amplitudes compared to upright faces (Anaki, Zion-Golumbic, & Bentin, 2007). Presently, we investigated the generality of this neural inversion effect by comparing the inversion effect for real and artificial faces. Artificial faces are a theoretically important example of an "other" face category and recent results indicate that face animacy impacts early visual ERPs (Balas & Koldewyn, 2013). Participants (N=15) viewed 8 real and 8 artificial faces that were presented upright and inverted in a pseudo-random order for a total of 160 experimental trials. Event-related potentials (ERPs) were recorded with a 64-channel net while participants completed an oddball detection task. We analyzed the mean amplitude and the latency of the P100 and the N170 components using a 2 x 2 x 2 repeated-measures ANOVA with animacy (real or doll), orientation (upright or inverted), and hemisphere (right or left) as within-subject factors. This revealed a main effect of orientation on the mean amplitude of the P100 (p = .004), and the N170, (p <.001), with inverted faces eliciting larger amplitudes. In addition, we observed a main effect of orientation on the latency of the P100 (p = .009). Our findings indicate that inverted faces, whether real or artificial, bilaterally enhance the P100 and N170 and delay the P100. Face animacy thus does not appear to impact the inversion effect on early visual ERPs, suggesting that real and artificial faces are potentially coded for by either a shared neural population, or at least by populations with similar tuning characteristics.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only