Spontaneous eye-movements in neutral and emotional gaze-cuing: An eye-tracking investigation

Abstract

Our attention is spontaneously oriented in the direction where others are looking. This attention shift manifests as faster responses to peripheral targets when they are gazed at by a central face instead of gazed away from, and this effect is even more pronounced when the face expresses an emotion. This so called gaze-cuing effect, and its enhancement by emotion, is thought to reflect covert attention orienting. However, eye movements are typically not monitored in gaze-cuing paradigms, yet free viewing and saccadic reaction time research suggests individuals commonly and quickly look at gazed-at locations. Furthermore, in dynamic gaze-cuing studies, emotional faces differ from neutral faces in their affective content but also in their apparent facial motion, both of which could affect participants' eye-movements. We investigated the contribution of overt orienting to the gaze-cuing effect by monitoring eye-movements during emotional and neutral gaze-cuing trials. We found that eye-movements were infrequent, and when they occurred, they were directed toward the target, not toward the gazed-at location. Removing trials with eye-movements did not affect gaze-cuing much, confirming it reflects a covert attention process. However, participants were more likely to move their eyes during neutral trials, which lacked perceived face movement, than during emotion trials or neutral movement trials. Including these eye-movement contaminated trials in our analysis resulted in an impaired ability to detect the gaze-cuing variations with emotion. In contrast, removing trials with eye-movements, or including a neutral movement control such as a neutral tongue protrusion, revealed more subtle emotional modulation of gaze-cuing.

Keywords: Psychology.

Fig. 1

Sample trial. The neutral direct, neutral averted and neutral tongue/classic neutral/emotion frames were included in the face sequence analyses. The target frame was included in the target frame analyses (human image obtained from NimStim Face Stimulus Set, used with permissionhttps://www.macbrain.org/resources.htm).

Fig. 2

Comparison of mean RTs computed with trials in which eye-movements were included (left panels) and removed (right panels). a) Congruent and incongruent reaction times (RT) for each emotion condition across the four stimulus-onset asynchronies (SOA). b) Congruent and incongruent RTs for each emotion (averaged across SOA). c) Gaze-cuing effect (RT_incongruent – RT_congruent) for each emotion condition and SOA. d) Gaze-cuing effect for each emotion (averaged across SOA).

Fig. 3

a) Proportion of trials with one or more saccades made during the face sequence for each emotion condition (averaged across stimulus-onset asynchrony – SOA). b) Proportion of trials with one or more eye-movements for each SOA (averaged across emotion).

Fig. 4

Total number of leftward (left panels) and rightward (right panels) saccades (averaged across the group) made during the face sequence, and displayed for a) each stimulus-onset asynchrony (SOA; averaged across the emotions), b) each emotion condition (averaged across all SOAs), and c) each emotion condition and face gaze direction. Note the overall larger number of leftward saccades for classic neutral compared to fearful and happy expressions and the same pattern for rightward saccades but only when the face looked to the right.

Fig. 5

Proportion of trials with one or more saccades made during the target presentation, displayed a) for each SOA (averaged across emotional expression) and b) for each emotion and SOA condition.

Fig. 6

Total number of leftward (left panels) and rightward (right panels) saccades (averaged across the group) made during the target presentation, displayed a) for each emotion condition and each face gaze direction, b) for each emotion condition (averaged across SOAs), c) for each SOA (averaged across emotion).