Chanelle Demeule^1,, Anthony Proulx^1,, Isabelle Charbonneau^1,, Justin Duncan^2,, Allyson Bastien^1,, Caroline Blais^1,, Daniel Fiset^1,; ¹Université du Québec en Outaouais, ²Université d'Ottawa

Previous work suggests that early perceptual brain activity in face recognition, indexed by the N170, is jointly influenced by sensitivity to the eye region and the processing of diagnostic, task-relevant information (Schyns et al., 2007; Smith et al., 2004). However, these two factors are often confounded because the eyes typically carry diagnostic information. In earlier work, we dissociated them by equating the eye region across neutral/smiling versions of the same identity and showed that N170 amplitude tracked the eyes and the mouth. Here, we asked whether making the eyes more diagnostic would further boost this effect. We recorded EEG from 29 adults (19 women) performing an expressive/neutral (EXNEX) categorization task while viewing faces revealed through randomly located Gaussian apertures (“Bubbles”; Gosselin & Schyns, 2001). The design included two stimulus sets: (1) faces with neutral eyes identical in the smiling and neutral versions, and (2) faces with a Duchenne smile in which the eyes were markedly more expressive in the smiling than in the neutral version. Time-resolved classification images were computed at PO8 from −200 to 800 ms post-stimulus to identify pixels associated with N170 amplitude on a participant basis. In both sets, pixel tests (p ≤ .025; Stat4Ci Toolbox; Chauvin et al., 2005) revealed significant associations between N170 amplitude and both eyes and the mouth. Critically, these associations did not differ between sets, even when analyses were restricted to the 12 participants who behaviorally relied most on the eyes. This pattern suggests that N170 sensitivity to the eyes may already operate near ceiling, such that increasing their diagnostic value fails to further enhance the response, whereas the strong involvement of the mouth—highly diagnostic for the present expressivity judgment but not for tasks where it is unnecessary—illustrates how goal-driven diagnostic information can shape processing without further boosting this early eye-driven activity.