Enhanced visual working memory for angry faces

Document Type



Does the presence of emotional expression influence visual working memory (WM) capacity for faces? Previous research has shown that capacity for faces was enhanced for familiar versus unfamiliar faces (Jackson & Raymond, 2004). This suggests that face capacity may be modulated by some facial attributes. It is well documented that emotional faces, especially angry and fearful faces, attract attention better than neutral faces, and that attention is necessary for working memory processes. Could visual WM for angry faces be enhanced relative to happy or neutral faces? We measured visual WM capacity for angry, happy, and neutral faces. Stimuli were 18 different male (Ekman) faces: six individuals each expressed the three emotions. On each trial, between 1 and 4 faces displaying the same emotion were presented for 2000 ms in a two-by-two matrix. When less than four faces were presented, all other grid locations were occupied by a scrambled face. A 1000 ms blank retention interval followed, succeeded by a single face probe. Participants stated whether the probe was present or absent in the previous display (identity task). Emotion was blocked and counterbalanced within subjects and a verbal suppression task was administered. Task performance was significantly enhanced for angry compared to happy and neutral faces (p < .01), and capacity estimates (k) mirrored the angry face advantage (p < .01). Happy and neutral face performance did not differ. Thus, negative emotional expression appears to enhance visual WM for faces. Control experiments revealed that this effect was not attributable to arousal.

Publication Date

June 2006

Publication Title

Journal of Vision





First Page






Repository URL