Slide Slam Q4
Changes in the affective representation of words in emotional contexts
Li-Chuan Ku1, Vicky Tzuyin Lai1; 1The University of Arizona
Research debates whether people prefer positive or negative information, a.k.a. 'positivity bias' and 'negativity bias' (Kauschke et al., 2019). Studies so far primarily focused on faces and single words occurring in isolation. Here we examined words in context and asked whether the affective perception of a word could be changed by context. Behaviorally, only one study, to our knowledge, found negativity bias for words in context. Specifically, even though positive words such as "grandpa" are rated as being more positive than negative words like "burglar" based on affective norms, "The grandpa is lonely" is rated as equally negative as "The burglar is lonely" (Lüdtke & Jacobs, 2015). We used EEG to examine neural changes in the affective representation of a word before and after a certain emotionally loaded context. If negative bias holds, negative contexts should lead to more negative evaluation of the target words, regardless of the words' valence in isolation. Likewise, if positive bias holds, positive contexts should result in more positive evaluations of all target words. If neither holds, the very same word before and after the emotional contexts should show the same neural representations. Due to COVID-19 remote learning, only 21 undergraduates participated and 8 did not meet inclusion criteria. The remaining 13 (9 females, Mage = 19.5) were non-depressive (Beck Depression Index-II) with normal cognitive functioning (Mini-Mental State Examination). Stimuli included 320 three-sentence vignettes with target words (in square brackets) and contexts (in angle brackets) of different valence: 80 positive target words in positive contexts (The [pianist] had a new performance. Her skills were <remarkable>. The [pianist] practiced every day.), 80 positive target words in negative contexts (The [pianist] had a new performance. Her skills were <rusty>. The [pianist] practiced every day.), 80 negative target words in positive contexts (The [dentist] often worked with children. They found him <trustworthy>. The [dentist] cared about them.), and 80 negative target words in negative contexts (The [dentist] often worked with children. They found him <formidable>. The [dentist] cared about them.). Target words were matched for length, frequency, and concreteness, and all low-arousing words, as our prior data suggested the difficulty to change affective perception towards high-arousing words. Participants did a valence judgment task while EEG was recorded. We computed and analyzed difference waves between the first and second occurrences of the target words. We found greater P2s (180-300 ms) for target words in negative contexts than in positive contexts, regardless of the word valence. Additionally, negative words elicited larger P2s than positive words, irrespective of context valence. We suggest that these reflect automatic attention towards negative content. A second finding is that target words in negative contexts showed larger LPPs (550-900 ms) than in positive contexts, regardless of the word valence. We argue that the attention underlying P2 was sustained through this time window. Altogether, these results indicate that the affective representation of a word can be changed after negative context, not positive context, supporting negativity bias. Data collection is ongoing to verify findings.