How the brain signals certainty during decision-making has long been studied through reaction time — the gap between a stimulus appearing and a response beginning. A growing body of cognitive neuroscience, however, suggests this single metric may be capturing only part of the story. New findings from PNAS point to a largely overlooked dimension of behavioral timing: how long a response itself takes to execute, not merely when it starts.
The research distinguishes between two temporal components of behavioral responses — the classical reaction time (onset latency) and response duration (the time taken to complete the physical act of responding). Across a series of decision tasks, response duration emerged as a reliable, independent signal of decision confidence: when participants were more certain of their judgments, their responses were completed more quickly and with less temporal variability. Critically, this effect held even after controlling for reaction time, suggesting the two measures capture dissociable aspects of cognitive processing. The findings implicate motor execution pathways — not just pre-motor deliberation — as active participants in the confidence signaling cascade.
This result carries meaningful implications for how researchers and clinicians interpret behavioral data. Most cognitive paradigms in psychology and neuroscience record only response onset, discarding the temporal profile of the motor act itself. If response duration independently encodes metacognitive states like confidence, decades of experiments may have been systematically discarding a rich data channel. From a practical standpoint, this could eventually inform the design of brain-computer interfaces, clinical assessments of cognitive function, or real-world confidence monitoring systems that use keystroke or touchscreen dynamics rather than neural recording.
The primary limitation is that task and response-modality specificity remain unclear — it is unknown whether these patterns generalize across diverse motor actions (voice, gesture, gaze) or clinical populations. As a first clear dissociation of these two timing signals, this is an incremental but methodologically significant finding that warrants replication across broader paradigms.