Abstract: This study investigated cognitive neural mechanisms underlying the generalization, both perceptual and conceptual, of fear under conscious and non-conscious conditions. Sixty-seven healthy individuals participated as valid subjects, randomly assigned to either a “conscious” or “non-conscious” group and following a classical fear-conditioning procedure. The experimental design incorporated three independent variables: two within-subject factors (conceptual generalization stimulus type [C+ vs. C-] and perceptual generalization stimulus type [P+ vs. P-]) as well as one between-subjects factor (conscious vs. non-conscious condition).
The experiment unfolded over three distinct phases. In the first, participants underwent a habituation phase to ensure their familiarity with all stimuli. Next, during the acquisition phase, perceptual stimuli (navy blue and olive green colors) and conceptual stimuli (animal- and furniture-related words) served as conditioned stimuli (CS), with one category designated as CS+ (e.g., navy blue as P+ and animal words as C+) and the other as CS-. An electric shock served as the unconditioned stimulus (US). In the subsequent generalization phase, four types of generalization stimuli (GS) were used: navy blue animal words (C+P+), olive green animal words (C+P-), olive green furniture words (C-P-), and navy blue furniture words (C-P+). Critical temporal parameters differed between groups: to the conscious group, GS were presented for 30 ms, followed by a 200-ms blank screen and a 100-ms mask; to the non-conscious group, GS were presented for 30 ms, followed immediately by a 100-ms mask, then a 200-ms blank screen. Behavioral and electrophysiological data were recorded throughout.
The behavioral data indicated significantly higher US expectancy ratings for P+ than C+ during acquisition. The electrophysiological data revealed two critical patterns: 1) the non-conscious group exhibited more negative frontal N1 amplitudes in response to GS than the conscious group; and 2) in the non-conscious group, C+ elicited more negative N400 amplitudes than C-, whereas the reverse pattern emerged in the conscious group.
These results support three main conclusions. First, perceptual stimuli may more strongly encourage threat-learning to relative to conceptual stimuli. Second, non-conscious GS induce enhanced early attentional vigilance compared to conscious group, as indexed by N1 modulation. Third, distinct neural signatures appear during conceptual processing, where non-conscious conditions facilitate threat-detection through N400 modulation for threat-related concepts. Under conscious conditions, conversely, N400 modulation for safety-related concepts is inconsistent with the threat expectation generated while the individual is in a vigilant state.
Discrepancies between early attentional (N1) and later semantic (N400) processing, as evidenced in this study, suggest stage-specific neural correlates of fear generalization across states of consciousness and pointœ toward compelling possible neural mechanisms underlying anxiety disorders.

Key words: fear generalization, non-conscious, perception, concept, N1, N400