Subliminal visual oddball stimuli evoke a P300 component

Objective: To provide evidence that a P300 component can be elicited by subliminal stimuli in an oddball paradigm. Methods: The words LEFT and RIGHT were presented in a frequent-rare ratio (80-20%), counterbalanced between subjects. Stimuli were presented at the objective detection threshold (d' = 0, via unmasked 1 ms presentations), a stringent measure for detecting any conscious perception. Results: A significantly larger amplitude component was found for rare vs. frequent stimulus presentations across electrodes Fz, Cz, and Pz using both a broad 200-900 ms window (F(1,27) = 5.75, P < 0.012, (2) = 0.18; one-tailed), and a more narrow 400-760 ms window defined using principal component analysis (F(1, 27) = 10.10, P < 0.002, (2) = 0.27; one-tailed). No significant component latency effect were found. An analysis of the conscious perception index (d') and the oddball effect (rare-frequent amplitude difference) revealed a negative relationship, further supporting the contention that conscious perception does not account for the finding, and suggesting that any conscious stimulus detection may inhibit this subliminal effect. Conclusions: Results provide evidence that an endogenous component can be elicited by undetectable subliminal stimuli in an oddball paradigm. Implications are discussed fur comparing conscious and unconscious information processing, unconscious learning, and the measurement of ERPs to subliminal stimuli. (C) 2001 Elsevier Science Ireland Ltd. All rights reserved.