Purchase this article with an account.
Isabelle Charbonneau, Stéphanie Cormier, Joël Guérette, Marie-Pier Plouffe-Demers, Caroline Blais, Daniel Fiset; Spatial frequencies for accurate categorization and discrimination of facial expressions. Journal of Vision 2018;18(10):601. doi: 10.1167/18.10.601.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Many studies have examined the role of spatial frequencies (SFs) in facial expression perception. However, most of these studies used arbitrary cut-off to isolate the impact of low and high SFs (De Cesarei & Codispoti, 2012) thus removing possible contribution of mid-SFs. This present study aims to reveal the diagnostic SFs for each basic emotion as well as neutral using SFs Bubbles (Willenbockel et al., 2010). Forty participants were tested (20 in a categorization task, 20 in a discrimination task; 4200 trials per participant). In the categorization task, subjects were asked to identify the perceived emotion among all the alternatives. In the discrimination task, subjects were asked, in a block-design setting (block order was counterbalanced across participants), to discriminate between a target emotion (e.g fear) and all other emotions. Mean accuracy was maintained halfway between chance (i.e. 12.5% and 50% correct for each task, respectively) and perfect accuracy. In both tasks, accuracy for happiness and surprise is associated with low-SFs (peaking at around 5 cycles per face (cpf); Zcrit=3.45, p< 0.05 for all analysis) whereas accuracy for sadness and neutrality is associated with mid-SFs (peaking between 11.5 and 15 cpf for both tasks). Interestingly, the facial expressions of fear and anger reveal significantly different patterns of use across task. Whereas their correct categorization is correlated with the presence of mid-to-high SFs (peaking at 14 and 20 cpf for angry and fear, respectively) their accurate discrimination is correlated with the utilization of lower SFs (peaking at 4 and 3.7 cpf). These results suggest that the visual system is able to use low-SF information to detect and discriminate social threatening cues. However, higher-SFs are probably necessary in a multiple-choices categorization task to allow fine-grained discrimination.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only