Expectation Gates Neural Facilitation of Emotional Words in Early Visual Areas

Abstract

The current study examined whether emotional expectations gate attention to emotional words in early visual cortex. Color cues informed about word valence and onset latency. We observed a stimulus-preceding negativity prior to the onset of cued words that was larger for negative than for neutral words. This indicates that in anticipation of emotional words more attention was allocated to them than to neutral words before target onset. During stimulus presentation the steady-state visual evoked potential (SSVEP), elicited by flickering words, was attenuated for cued compared to uncued words, indicating sharpened sensory activity, i.e., expectation suppression. Most importantly, the SSVEP was more enhanced for negative than neutral words when these were cued. Uncued conditions did not differ in SSVEP amplitudes, paralleling previous studies reporting lexico-semantic but not early visual effects of emotional words. We suggest that cueing mediates re-entrant engagement of visual resources by providing an early "affective gist" of an upcoming word. Consequently, visual single-word studies may have underestimated attentional effects of emotional words and their anticipation during reading.

Keywords: SPN; SSVEP; anticipation; emotion; expectation; reading; visual attention.

FIGURE 1

Schematic structure of experimental trials. In cued trials (left), after the baseline period a fixed time period of 1667 ms with a color change of the fixation dot allowed for expectations about the emotional valence of the upcoming word. Uncued trials (right) were identical except for the cue. Baseline and word stimuli flickered at 15 Hz. The response screen indicated the trial-by-trial key assignment.

FIGURE 2

(A) ERP time course at the five central electrodes chosen to analyze the SPN amplitude, indicated on the scalp electrode map at the right side of the panel. Note that condition effects continue after the cue period throughout the stimulus period. (B) Difference maps for the SPN time window (1000 ms before word onset) reveal a larger central negativity prior to emotional words, especially negative words (upper scalp map).

FIGURE 3

(A) Time courses of SSVEP amplitude after word onset. The cue period was analyzed separately (cued trials only, not depicted). (B) Grand mean SSVEP amplitude across all conditions from –350 to 1200 ms around word onset.

FIGURE 4

Response latencies in ms for valid cues (black) and no cues (gray) for neutral, negative, and positive words.