Eye Fixation-Related Potentials during Visual Search on Acquaintance and Newly-Learned Faces

Abstract

Searching familiar faces in the crowd may involve stimulus-driven attention by emotional significance, together with goal-directed attention due to task-relevant needs. The present study investigated the effect of familiarity on attentional processes by exploring eye fixation-related potentials (EFRPs) and eye gazes when humans searched for, among other distracting faces, either an acquaintance's face or a newly-learned face. Task performance and gaze behavior were indistinguishable for identifying either faces. However, from the EFRP analysis, after a P300 component for successful search of target faces, we found greater deflections of right parietal late positive potentials in response to newly-learned faces than acquaintance's faces, indicating more involvement of goal-directed attention in processing newly-learned faces. In addition, we found greater occipital negativity elicited by acquaintance's faces, reflecting emotional responses to significant stimuli. These results may suggest that finding a familiar face in the crowd would involve lower goal-directed attention and elicit more emotional responses.

Keywords: eye‐fixation related potential; eye‐tracking; face; familiarity; visual search.

Figure 1

An experimental task and an electroencephalography (EEG) montage. (a) Timeline of stimulus presentation. The face images were blurred for privacy reason. (b) The EEG montage used in this study.

Figure 2

Behavior results of the target face detection task. (a) Correct rates for all the conditions (a double star denotes $p_b$ < 0.01). (b) Error 1 and Error 2 rates of the AQ (Acquaintances faces) and NL (Newly-learned) groups (n.s. denotes not significant).

Figure 3

The distribution of fixation duration. The solid lines represent fixation durations (FD) defined as the time elapsed from the first entry to an AOI to the first exit from that AOI. The dashed lines stand for first fixation durations (FFD) defined as the duration of a first fixation point between saccades in an AOI. The onset of fixation in target trials is the first fixation on an AOI and that in non-target trials is the longest fixation on non-target faces.

Figure 4

Eye-tracking results of the target face detection task. (a) Fixation onset, (b) fixation duration, and (c) dwell time on the target faces (Target) or non-target faces with the longest fixation (Non-garget) for the AQ (Acquaintances faces) and NL (Newly-learned) groups. The stars indicate the significance from a multiple comparison test: Black stars denote $p_b$ < 0.05 and a gray one denotes $p_b$ < 0.1.

Figure 5

The grand average eye fixation-related potentials (EFRPs) of searching faces by AQ (Acquaintances faces, solid line) and NL group (Newly-learned faces, dash line). Red line represents when they looked at target faces and blue when looking at non-target faces. The bars below EFRPS indicates significant main effect of target presence (dark gray for p < 0.05 and light gray for p < 0.01) and group (orange for p < 0.01) resulting from two-way mixed repeated measures ANOVA (rmANOVA) for averaged amplitude for every 20 ms time window.

Figure 6

The scalp distributions and the mean amplitude of EFRPs: P300 and late positive potential (LPP). (a) Scalp distributions for the target and non-target trials of the average of AQ (Acquaintances faces) and NL (Newly-learned). Black dots indicate the channels where a significant main effect of target presence was found resulting from two-way mixed rmANOVA on P300. (b) Mean amplitude of the channels indicated in (a). (c) Scalp distributions for the difference between AQ and NL group of target and non-target trials. Black dots indicate the channels where a significant main effect of group resulting from two-way mixed rmANOVA on LPP. (d) Mean amplitude of the channels indicated in (c). A star in the bar plots represents $p_b$ < 0.05.

Figure 7

The amplitude at bilateral occipital area. (a) AQ (Acquaintances faces) group had significantly negative value in target trials then non-target trials at O1 during 600–800 ms. (b) at O2 during 300–800 ms. A star represents $p_b$ < 0.05.