Event-related potentials to task-irrelevant changes in facial expressions

Abstract

Background: Numerous previous experiments have used oddball paradigm to study change detection. This paradigm is applied here to study change detection of facial expressions in a context which demands abstraction of the emotional expression-related facial features among other changing facial features.

Methods: Event-related potentials (ERPs) were recorded in adult humans engaged in a demanding auditory task. In an oddball paradigm, repeated pictures of faces with a neutral expression ('standard', p = .9) were rarely replaced by pictures with a fearful ('fearful deviant', p = .05) or happy ('happy deviant', p = .05) expression. Importantly, facial identities changed from picture to picture. Thus, change detection required abstraction of facial expression from changes in several low-level visual features.

Results: ERPs to both types of deviants differed from those to standards. At occipital electrode sites, ERPs to deviants were more negative than ERPs to standards at 150-180 ms and 280-320 ms post-stimulus. A positive shift to deviants at fronto-central electrode sites in the analysis window of 130-170 ms post-stimulus was also found. Waveform analysis computed as point-wise comparisons between the amplitudes elicited by standards and deviants revealed that the occipital negativity emerged earlier to happy deviants than to fearful deviants (after 140 ms versus 160 ms post-stimulus, respectively). In turn, the anterior positivity was earlier to fearful deviants than to happy deviants (110 ms versus 120 ms post-stimulus, respectively).

Conclusion: ERP amplitude differences between emotional and neutral expressions indicated pre-attentive change detection of facial expressions among neutral faces. The posterior negative difference at 150-180 ms latency resembled visual mismatch negativity (vMMN) - an index of pre-attentive change detection previously studied only to changes in low-level features in vision. The positive anterior difference in ERPs at 130-170 ms post-stimulus probably indexed pre-attentive attention orienting towards emotionally significant changes. The results show that the human brain can abstract emotion related features of faces while engaged to a demanding task in another sensory modality.

Figure 1

Illustration of the stimulus paradigm applied.

Figure 2

Grand averaged ERPs to the fearful deviants (blue lines) and the neutral standards (red lines) immediately preceding them. The difference ERPs (deviant minus standard) are drawn with the black lines. The time windows for extracting the mean values for the repeated measures MANOVA are marked with rectangles. The x-axis shows stimulus onset.

Figure 3

Grand averaged ERPs to the happy deviants (blue lines) and the neutral standards (red lines) immediately preceding them. The difference ERPs (deviant minus standard) are drawn with the black lines. The time windows for extracting the mean values for the repeated measures MANOVA are marked with rectangles. The x-axis shows stimulus onset.