Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Linguistic Label Context Effect in Facial Emotion Processing
Poster Session B, Wednesday, September 30, 4:30 - 6:30 pm, Wangari Maathai
Tsuyueh Hsu1, Atsunobu Suzuki2, Joshua Oon Soo Goh1, Chia-Lin Lee1; 1National Taiwan University, 2The University of Tokyo
Language shapes how people categorize and attend to their experiences, and, beyond domains of color and sound, has been argued to play a constitutive role in emotion. Theory of Constructed Emotion (Barrett, 2007) argues that emotion words provide emotion concepts that help interpret ambiguous bodily sensations and contextual cues by shaping how affective information is categorized and perceived. Consistent with this view, behavioral studies have found that linguistic emotional context affects offline judgments and memory for emotional facial expressions. However, such effects may reflect influences on decision processes rather than perception itself or memory retrieval after a delay. Given the limited existing evidence on how emotion labels modulate the instantaneous processing of facial expressions, the present study used Event-Related Potentials (ERPs) to examine how emotion words modulate early perceptual processing of facial expressions. Following Fugate et al. (2018), an emotion label word or an abstract neutral word was presented prior to an emotionally ambiguous face created by morphing two expressions (happiness, sadness, anger, surprise) at 50% each. A second face followed, which was either identical or slightly biased toward the cued emotion (60% vs. 40%). Participants judged whether the two faces matched. Emotionally ambiguous faces were generated by morphing prototypical expressions from the East Asian Facial Expression Database (Tu et al., 2018) using FantaMorph5. Data from 32 native speakers of Taiwan Mandarin (16 males; 20–30 years, M = 23.91) showed more “same” judgements for identical than biased second faces, regardless of whether the trial follows an emotion word or a neutral word. ERP data time-locked to word onset showed emotion effects, including enhanced P2 at anterior sites and a positive dip in the LPP time-window at posterior sites, in line with past literature observing these components for emotional content. Time-locked to the first faces, LPP was observed for faces following emotion words over faces following neutral words, suggesting that emotion words activate rapid processing of emotion that influences the perception of upcoming stimuli. For the second faces, repetition effects differed across word types. Following emotion words, greater repetition effects were observed for identical faces over anterior sites, with more positive responses for identical relative to biased second faces beginning in the N400 time-window and continuing throughout the epoch. In contrast, following neutral words, comparable repetition effects were observed for identical and biased faces, with an additional late anterior positivity (700-900ms) for biased relative to identical second faces. A possible explanation for the different repetition is that, relative to neutral words, the attention elicited by emotion words led to deeper processing of the first face, leading to brain responses discriminating identical and biased face types at an earlier stage. Taken together, the present findings suggest that emotion words act as primes for emotion-related concepts and enhance early and sustained attentional allocation in the processing of subsequent faces despite subtle changes in stimulus form. These findings support the view that language provides a conceptual anchor for emotion perception, exerting rapid influences on early perceptual processing.
Topic Areas: Multisensory or Sensorimotor Integration, Meaning: Lexical Semantics