Performance-Metacognition Dissociation During Fear Memory Recall: Role of Task Sequence and Prefrontal TMS

Metacognitive processing—the ability to monitor and evaluate one's decisions—has been extensively studied in perceptual tasks but never examined in fear conditioning. We investigated whether metacognitive sensitivity can be quantified in emotional learning and how testing order and neuromodulation of the anterior prefrontal cortex (aPFC) influence metacognitive access to the memory of a shock-associated cue. Eighty-one participants (51 females and 30 males) underwent fear conditioning and received either active or sham TMS over the aPFC. They were tested with implicit (skin conductance) and explicit (two-interval forced-choice with confidence ratings) tasks in counterbalanced order. Behavioral performance (d′) and metacognitive sensitivity (meta-d′) were quantified using signal detection theory. While d′ remained stable across conditions, meta-d′ and confidence were significantly higher when the explicit task preceded the implicit task. TMS did not affect metacognitive sensitivity. To our knowledge, these results provide the first evidence that metacognitive frameworks can be applied to a fear-conditioning paradigm, revealing that confidence judgments are dynamically modulated by task sequence and remain intact even when bodily arousal is reduced. We propose that metacognitive sensitivity may decline faster than first-order performance, pointing to a functional dissociation between decision accuracy and metacognition in emotionally salient contexts. Significance Statement This study provides the first evidence that metacognitive monitoring can be quantified within a fear-conditioning paradigm, extending metacognitive research beyond perceptual decision-making into emotional learning. We show that confidence in recognizing a shock-associated cue can be dissociated from first-order performance: task sequence critically modulates metacognitive sensitivity even when first-order performance remains stable. Notably, confidence judgments persist despite reduced autonomic arousal following prefrontal stimulation, suggesting that metacognitive access does not require intact bodily signals. These findings indicate that fear conditioning offers a viable paradigm for investigating metacognition and suggest a partial dissociation between first-order performance and self-monitoring in emotional memory. This work opens new avenues for understanding confidence disruptions in anxiety disorders in which threat detection and metacognitive evaluation are impaired.