Different Top-Down Goals Modulate Early Stage of Attention Bias: Irrelevant Task Suppresses the N170 of Automatic Attention Allocation to Threat Faces

Haoran Dou, Limei Liang, Jie Ma, Jiachen Lu, Wenhai Zhang, Yang LiResearch 2020 · DOI

Abstract

Recent research suggests that stimulus-driven attention bias for threat can be modulated by top-down goals. However, the early neural mechanisms underlying this modulation remain unclear. This study investigated how different top-down goals affect the early stages of neural processing using electroencephalographic (EEG) data from 28 healthy volunteers. Participants performed a modified spatial cueing task with four levels of inconsistency based on cue validity and task relevance. Results indicated that when a task made emotional information irrelevant (e.g., gender identification), the typical early neural response to fearful faces (the N170 component) was suppressed. These findings suggest that top-down processing can inhibit the early, seemingly automatic allocation of attention to threatening stimuli.

Introduction

Humans have evolved a specialized "attention bias" toward threat, allowing us to detect potential dangers, such as a fearful face, more rapidly than neutral stimuli. A long-standing debate in neuroscience is whether this process is purely automatic (bottom-up) or if it can be controlled by our current conscious goals (top-down). While it is known that goals can influence behavioral outcomes, it remains uncertain if they can modulate the very first 200 milliseconds of brain activity. This study specifically examines the N170 (a face-specific neural marker) and the Vertex Positive Potential (VPP) to determine if top-down goals can filter out threat signals at the earliest stages of perception.

Methods

The researchers recruited 28 healthy college students and recorded EEG data during a modified spatial cueing paradigm. In each trial, a green frame served as a "cue," followed by a "target" consisting of four faces (one fearful or neutral face framed in green).

Figure 1: The modified spatial cueing paradigm used to manipulate task relevance and cue validity.

The study manipulated two primary variables to create four "levels of inconsistency":

  1. Cue Validity: Whether the target appeared in the same location as the cue (Valid) or a different location (Invalid).
  2. Task Relevance: Whether participants identified the emotion of the face (making fear relevant) or the gender of the face (making fear irrelevant).

These variables combined to form a spectrum of inconsistency ranging from "Weak" (Valid cue + Emotion task) to "Very Strong" (Invalid cue + Gender task). Stimuli were selected from the Chinese Facial Affective Picture System (CFAPS), controlled for valence and arousal.

Results

Behavioral Performance

Reaction times (RT) revealed that participants were significantly slower in the irrelevant (gender) task compared to the relevant (emotion) task. In the relevant task, participants found fearful faces faster than neutral ones in valid trials (703ms vs. 750ms). However, in the irrelevant task, this behavioral advantage for fearful faces disappeared, with no significant difference in RTs between fearful and neutral faces.

Electrophysiological Findings

The ERP results provided evidence for neural suppression:

  • N170 Component: In the relevant task (weak and medium inconsistency), fearful faces evoked significantly larger N170 amplitudes than neutral faces. However, in the irrelevant task (strong and very strong inconsistency), this "threat boost" was entirely suppressed.
  • VPP Component: Similar to the N170, the VPP was larger for fearful faces, but only in the "weak" inconsistency condition (Valid cue + Relevant task).

Discussion and Conclusion

The findings demonstrate that the N170—an early neural marker of face processing—is not a hard-wired, unstoppable reflex to threat. Instead, the brain's early attention allocation is highly sensitive to top-down goals. When a task renders emotional information irrelevant, the brain effectively filters out the threat signal within the first 200ms.

While this study provides high temporal resolution regarding when these processes occur, future research using Magnetoencephalography (MEG) could help pinpoint where this suppression happens, potentially involving structures like the amygdala or the Fusiform Face Area (FFA). Ultimately, these results suggest that our conscious goals play a critical role in modulating how we perceive and react to environmental threats.

Summary updated

Abstract

Abstract Recent research has evidenced that stimulus-driven attention bias for threat can be modulated by top-down goals. However, it remains unclear how different top-down goals affect the early stage of attention processing and its early neural mechanism. We collected electroencephalographic data from 28 healthy volunteers during four inconsistent levels in a modified spatial cueing task according to cue validity and task relevance. Our data revealed that reaction time(RT) to the target in the irrelevant task were much slower than that in the relevant task. In the irrelevant task, we did not find the difference between the RTs to the fearful and neutral face. In the relevant task, we found RTs of fearful face were faster than that of neutral face in valid cue condition(weak inconsistent level), whereas the RTs of fearful face were slower than that of neutral face in invalid cue condition(medium inconsistent level). The ERPs results showed that in relevant task(weak, medium inconsistent levels), fearful face in cue position of the target evoked larger N170 amplitudes than neutral face did, whereas this effect was suppressed in irrelevant task(strong, very strong inconsistent levels). Besides, fearful face in cue position of the target also evoked larger vertex positive potential (VPP) amplitudes than the neutral face did in valid cue condition( weak inconsistent level). These results suggest that irrelevant task may inhibit the early attention allocation to the fearful face. Furthermore, the top-down processing modulates the early attention bias for threatening faces.
BibTeX
@inproceedings{dou2020different,
  author = {Haoran Dou and Limei Liang and Jie Ma and Jiachen Lu and Wenhai Zhang and Yang Li},
  title = {{Different Top-Down Goals Modulate Early Stage of Attention Bias: Irrelevant Task Suppresses the N170 of Automatic Attention Allocation to Threat Faces}},
  booktitle = {Research},
  year = {2020},
  doi = {10.21203/rs.3.rs-121215/v1},
}
Figure