Emotion in Chinese Words Could Not Be Extracted in Continuous Flash Suppression

Kaiwen Cheng, Aolin Ding, Lianfang Jiang, Tian Han, Hongmei YanFrontiers in Human Neuroscience 2019 · DOI

TL;DR

This study investigated whether emotional information from written Chinese words can be processed unconsciously using a technique called Continuous Flash Suppression (CFS). The researchers found that, unlike faces or pictures, emotional valence of Chinese words did not influence how quickly they broke through suppression, suggesting that complex emotional meaning might not be accessible without conscious awareness.

What Problem Does This Paper Address?

For decades, researchers have observed an "automatic vigilance effect," where the brain seems to prioritize processing of emotionally charged stimuli, especially negative ones, even when we're not consciously aware of them. This effect is well-established for evolutionarily ancient stimuli like faces and pictures. However, whether this automatic processing extends to more recently evolved stimuli, like written words, has been a subject of considerable debate. Previous studies using different languages and methodologies have yielded conflicting results, making it unclear if emotional content in words can truly be extracted unconsciously. This paper aims to resolve some of this ambiguity by investigating this phenomenon in Simplified Chinese (SC) words, using a robust experimental design.

Methods

The core technique employed here is Breaking Continuous Flash Suppression (b-CFS). In a typical b-CFS experiment, a visual stimulus (in this case, a word) is presented to one eye while a rapidly changing, high-contrast pattern (a "Mondrian mask") is presented to the other. This creates strong interocular suppression, rendering the target stimulus invisible. The time it takes for the target stimulus to "break through" this suppression and become visible is measured. A shorter breakthrough time for emotional stimuli compared to neutral ones would suggest unconscious emotional processing.

The researchers conducted two experiments:

  • Experiment 1: Used two-character SC words (e.g., "already," "annoying").
  • Experiment 2: Used single-character SC words (e.g., "sell," "dirty").

They also manipulated word form (upright vs. inverted) to probe the role of low-level visual features versus higher-level semantic processing. A crucial control condition involved a binocular (non-CFS) session, where the word was presented to both eyes, allowing for conscious or partially conscious processing, to compare against the b-CFS results. Participants also underwent a visibility test to ensure they were genuinely unaware of the suppressed stimuli.

Datasets

The study utilized custom-created word lists for both experiments.

  • Experiment 1: 80 two-character SC words, balanced for negative (emotion-label) and neutral (functional) words. These were matched for stroke count and word frequency.
  • Experiment 2: 80 one-character SC words, also balanced for negative (emotion-laden) and neutral words, and matched for stroke count and word frequency.

These word lists were carefully selected and validated through participant questionnaires to control for emotional valence and arousal.

Key Insights and Ideas

  • The "vigilance effect" might not extend to written words: Unlike faces and pictures, the emotional content of Chinese words did not seem to be automatically prioritized for processing when they were rendered invisible.
  • Language and orthography might matter: The authors hypothesize that the simplification of Simplified Chinese (SC) compared to Traditional Chinese (TC) might reduce the orthographic-semantic mapping, making unconscious emotional extraction more difficult.
  • High-level processing requires awareness: The findings suggest that complex, high-level cognitive tasks like understanding the emotional valence of words may necessitate some level of conscious or preconscious awareness.
  • Low-level features are processed unconsciously: The consistent effect of word form (upright words breaking suppression faster than inverted ones) across both CFS and non-CFS conditions suggests that basic visual features can be processed without awareness.

Key Results

  • Experiment 1 (Two-character words): In the b-CFS condition, there was no significant difference in breakthrough times between negative and neutral words. However, upright words were detected faster than inverted words.
  • Experiment 2 (Single-character words): Similar to Experiment 1, the b-CFS condition showed no significant difference in breakthrough times for negative versus neutral words, regardless of word length or proximity to the fovea. Again, upright words were detected faster than inverted words.
  • Binocular (non-CFS) control: In the conscious/partially conscious condition, some effects of emotion and word form were observed, but these differed between the two experiments and were not the primary focus of the study's conclusions about unconscious processing.

Limitations and Future Work

  • Replication of previous findings: The study failed to replicate earlier findings by Yang and Yeh (2011) using Traditional Chinese (TC). The authors propose that differences in the Chinese writing system (SC vs. TC) might explain this discrepancy, suggesting SC's simplification may hinder unconscious emotional extraction.
  • Potential for artificial feature binding: The authors acknowledge that previous studies with smaller stimulus sets and more repetitions might have led to artificial correlations between low-level features and emotional valence, rather than true semantic processing.
  • Arousal control: While valence was controlled, the arousal level of emotional words was not strictly controlled, though the authors argue this is a common practice and unlikely to invalidate their core findings.
  • Future directions: The authors suggest further research comparing languages with different orthographic depths and employing neuroimaging techniques (like fMRI or EEG) to better understand the temporal dynamics of unconscious processing. They also propose future studies that directly cross word valence and length within a single experiment.

Summary updated

Abstract

Previous studies have demonstrated the automatic vigilance effect for faces and pictures and have attributed it to the brain's prioritized unconscious evaluation of early evolutionary stimuli that are critical to survival. Whether this effect exists for evolutionarily more recent stimuli, such as written words, has become the center of much debate. Apparently contradicting results have been reported in different languages, such as Hebrew, English, and Traditional Chinese (TC), with regard to the unconscious processing of emotional words in breaking continuous flash suppression (b-CFS). Our current study used two experiments (with two-character words or single-character words) to verify whether the emotional valence or the length of Simplified Chinese (SC) words would modulate conscious access in b-CFS. We failed to replicate the findings reported in Yang and Yeh (2011) using TC, but found that complex high-level emotional information could not be extracted from interocularly suppressed words regardless of their length. Our findings comply with the distinction between subliminal and preconscious states in Global Neuronal Workspace Theory and support the current notion that preconsciousness or partial awareness may be indispensable for high-level cognitive tasks such as reading comprehension.
BibTeX
@inproceedings{cheng2019emotion,
  author = {Kaiwen Cheng and Aolin Ding and Lianfang Jiang and Tian Han and Hongmei Yan},
  title = {{Emotion in Chinese Words Could Not Be Extracted in Continuous Flash Suppression}},
  booktitle = {Frontiers in Human Neuroscience},
  year = {2019},
  doi = {10.3389/fnhum.2019.00309},
}
Figure