Implicit learning refers to the acquisition of knowledge without conscious awareness. This study investigated whether implicit learning can be observed at the neural level using EEG in participants unaware of stimulus–reward contingencies. Participants completed a Pavlovian conditioning task, associating geometrical stimuli (CS+ and CS-) with varying reward probabilities, while EEG was recorded. Bayesian classification methods identified a group of unaware participants (n = 46), whose event-related potentials (ERPs) were analysed across three distinct learning phases: cue re-activity, preparatory processing and feedback evaluation. ERP components related to cue re-activity (P300), preparatory processing (CNV, SPN), and feedback evaluation (FRN) were examined across early and late learning blocks. Results showed robust and condition-sensitive neural changes over time, particularly for CS+ stimuli in the CNV, SPN, and FRN components, suggesting that anticipatory and feedback-related processes are sensitive to implicit learning. In contrast, findings related to the P300 were less consistent and did not provide strong evidence. These findings support the notion that implicit learning can modulate neural activity even in the absence of explicit awareness, with the FRN offering the strongest neural marker. This study highlights the value of combining Bayesian inference and EEG to investigate unconscious cognitive processes.

Keywords: implicit learning, EEG, Pavlovian conditioning, P300, CNV, SPN, FRN, Bayesian classification