Exploring the Role of Spatial Frequency Information during Neural Emotion Processing in Human Infants

Abstract

Enhanced attention to fear expressions in adults is primarily driven by information from low as opposed to high spatial frequencies contained in faces. However, little is known about the role of spatial frequency information in emotion processing during infancy. In the present study, we examined the role of low compared to high spatial frequencies in the processing of happy and fearful facial expressions by using filtered face stimuli and measuring event-related brain potentials (ERPs) in 7-month-old infants (N = 26). Our results revealed that infants' brains discriminated between emotional facial expressions containing high but not between expressions containing low spatial frequencies. Specifically, happy faces containing high spatial frequencies elicited a smaller Nc amplitude than fearful faces containing high spatial frequencies and happy and fearful faces containing low spatial frequencies. Our results demonstrate that already in infancy spatial frequency content influences the processing of facial emotions. Furthermore, we observed that fearful facial expressions elicited a comparable Nc response for high and low spatial frequencies, suggesting a robust detection of fearful faces irrespective of spatial frequency content, whereas the detection of happy facial expressions was contingent upon frequency content. In summary, these data provide new insights into the neural processing of facial emotions in early development by highlighting the differential role played by spatial frequencies in the detection of fear and happiness.

Keywords: EEG; emotion perception; face processing; infancy; spatial frequencies.

Figure 1

Examples of stimulus material. Images of happy (top row) and fearful (bottom row) faces were filtered to contain only spatial frequencies below 0.4 cycles/° (low spatial frequencies (LSF), left column) or above 0.6 cycles/° (high spatial frequencies (HSF), right column).

Figure 2

Event-related brain potential (ERP) response at frontal electrodes (F3, Fz, F4). (A) shows mean responses to images containing HSF while (B) displays responses to images containing LSF (blue/green = happy expression, red/orange = fearful expression; displayed are mean responses ± within-subject standard errors). Topographic representations show the difference in activation following happy and fearful faces between 500 ms and 600 ms, corresponding to the time-window used in the statistical analysis and marked in gray.

Figure 3

ERP response at occipital electrodes (O1, O2, P7, P8). Shows mean responses at occipital electrodes included in the analysis of the P400 (time-window marked in gray; blue/green = happy expression, red/orange = fearful expression; displayed are mean responses ± within-subject standard errors). The topographic representation show the difference in brain responses to low compared to high spatial frequencies irrespective of emotional expression between 350 ms and 600 ms, corresponding to the time window used in the statistical analysis marked in gray.