The effect of fearful expressions on recognition memory for faces: Behavioral and electrophysiological data

Abstract

Facial expressions affect memory for face identity. We tested how fearful expressions modulate recognition memory for faces. In two studies, participants completed a continuous recognition task with fearful and neutral faces while their electroencephalogram was recorded. Each face stimulus was presented twice and participants were instructed to indicate whether it was presented for the first ('new') or second time ('old'). The false alarm rate was higher for fearful than neutral faces, which is opposite to the emotion enhancement effect on memory but in line with the liberal response bias for emotional information. There was no evidence of emotional modulation of the N400 old/new effect, which suggests that the sense of familiarity was not affected by fearful facial expressions. The LPC old/new effect, however, was modulated by facial expression, as it was absent for fearful faces because of a greater positivity in response to new fearful than new neutral faces. This LPC old/new effect finding may reflect that the emotional salience of fearful new faces is mistaken for a sense of recollection, resulting in an increased false alarm rate. In short, people seem more likely to (mistakenly) think that they have encountered a person before when the person looks scared compared to non-emotional, which has relevance for daily life and forensic situations such as police lineups.

Preschoolers’ Sensitivity to Negative and Positive Emotional Facial Expressions: An ERP Study

“…Statistical analyses were performed using R-Studio ( R Core Team, 2020 ). For all analyses, we excluded trials with reaction times (RTs) lower than 250 ms ( Johnston et al, 2011 ) and incorrect trials ( Langeslag et al, 2020 ). Neither the final number of trials for each facial expression [happy: M = 33.5, SD = 6.7, angry: M = 32.7, SD = 7.3, neutral: M = 33.3, SD = 8.1; F (2, 87) = 0.1, p = 0.91] nor for repeated ( M = 50.1, SD = 14.6) vs. novel trials ( M = 49.4, SD = 13.7) differed significantly [ t (29) = 0.2, p = 0.9].…”