Influence of Perceptual and Conceptual Information on Fear Generalization: A Behavioral and Event-Related Potential Study

Jinxia Wang, Mei E, Qi Wu, Tao Xie, Haoran Dou, Yi LeiCognitive 2021 · DOI

Abstract

Fear generalization allows organisms to apply learned threat responses to novel but similar stimuli. While perceptual similarity (e.g., color or shape) is a well-studied driver of this process, the role of conceptual information (e.g., semantic categories) remains less clear. This study employed a "de novo" fear conditioning paradigm combined with high-resolution EEG to investigate how these two streams of information interact. Results indicate that while perceptual cues facilitate rapid initial learning, conceptual categories dominate the generalization of fear. Electrophysiological data revealed that the N400 component tracks conceptual threat, while the Late Positive Potential (LPP) serves as a marker for active safety inhibition.

Introduction

Fear generalization is an evolutionary mechanism essential for survival, enabling individuals to avoid potential threats based on past experiences. However, overgeneralization—where fear is extended to safe stimuli—is a core feature of anxiety disorders and PTSD. Traditional models of fear generalization often emphasize perceptual gradients, such as the frequency of a tone or the dimensions of a geometric shape. Yet, human environments are semantically rich, requiring the brain to categorize threats conceptually (e.g., "all predators") regardless of their visual appearance. This study explores the neural and behavioral dynamics of how perceptual and conceptual information compete or cooperate during fear generalization.

Methods

The researchers recruited 23 healthy volunteers for a two-phase experimental task using a custom-curated set of 100 images (animals, furniture, and abstract shapes). In the Fear Acquisition Phase, participants learned to associate specific categories and colors with an Unconditioned Stimulus (US), which consisted of a mild electrical shock paired with a fearful image.

  • Conceptual Cues: Animals (C+, dangerous) vs. Furniture (C−, safe).
  • Perceptual Cues: Purple shapes (P+, dangerous) vs. Blue shapes (P−, safe).

In the Fear Generalization Phase, participants were presented with "crossed" stimuli—new exemplars that combined these traits in novel ways (e.g., a blue animal or purple furniture). Behavioral measures included US-expectancy ratings and reaction times. Simultaneously, Event-Related Potentials (ERPs) were recorded to track neural processing at different stages: early sensory processing (P1, EPN), semantic integration (N400), and late emotional evaluation (LPP).

Experimental paradigm showing the fear acquisition and generalization phases, including stimulus types and trial structure.

Results

The behavioral data revealed a clear distinction between learning and generalization. During acquisition, perceptual cues (color) led to faster reaction times (~1285ms) and higher fear ratings than conceptual cues (~1475ms), suggesting a rapid "low road" for sensory-driven threat detection. However, during the generalization phase, conceptual information became the primary driver of fear. Participants rated animal exemplars (C+) as significantly more dangerous than furniture (C−), even when the furniture possessed the "dangerous" purple color (P+).

Electrophysiological findings supported this conceptual dominance. The N400 component, typically associated with semantic processing, showed significantly higher amplitudes for stimuli representing conceptual danger (C+P+ and C+P−). Furthermore, the Late Positive Potential (LPP) was enhanced for conceptually safe stimuli (C−). The researchers interpret this LPP enhancement as a neural signature of active safety inhibition—the brain's effort to suppress a fear response upon recognizing a safe category.

Discussion

These findings suggest that the human brain prioritizes high-level conceptual categories over simple perceptual features when navigating complex environments. While we may react instinctively to visual cues, our final assessment of threat is deeply rooted in semantic understanding. The identification of the LPP as a marker for safety inhibition is particularly noteworthy, as it suggests that "safety" is not merely the absence of fear but an active cognitive process.

Future research should investigate these mechanisms in clinical populations. For individuals with anxiety disorders, the "safety brake" represented by the LPP may be impaired, leading to the characteristic overgeneralization of fear. Additionally, using stronger unconditioned stimuli could reveal whether perceptual cues regain dominance under conditions of extreme stress.

Conclusion

By integrating behavioral and ERP data, this study clarifies the hierarchy of information processing in fear generalization. Conceptual categories provide a robust framework for evaluating danger, overriding simpler perceptual signals to guide adaptive behavior in uncertain environments.

Summary updated

BibTeX
@inproceedings{wang2021influence,
  author = {Jinxia Wang and Mei E and Qi Wu and Tao Xie and Haoran Dou and Yi Lei},
  title = {{Influence of Perceptual and Conceptual Information on Fear Generalization: A Behavioral and Event-Related Potential Study}},
  booktitle = {Cognitive},
  year = {2021},
  doi = {10.3758/s13415-021-00912-x},
}
Figure