Pre-stimulus pupil-linked arousal enhances initial stimulus availability and accelerates decay in iconic memory

Pupil diameter reflects not only reflexive responses to light but also spontaneous fluctuations in arousal. Such pre-stimulus fluctuations may shape visual performance, yet it remains unclear whether they only affect perceptual sensitivity or decision processes in near-threshold tasks, or instead influence the temporal availability of sensory information even for highly visible stimuli. We tested whether pupil-linked arousal modulates iconic memory, a high-capacity but short-lived visual store bridging sensory encoding and short-term memory. Based on a model in which a brief stimulus evokes an internal response decaying over a few hundred milliseconds, we hypothesized that increased arousal enhances performance either by amplifying the initial response or by slowing its decay. Thirty-seven participants viewed brief displays of six oriented stimuli, followed by a cue after a variable stimulus-cue onset asynchrony indicating which item to report. We continuously recorded pre-stimulus pupil diameter and fitted behavioral data with an exponential decay function to estimate initial stimulus availability, decay rate, and asymptotic performance. Larger pre-stimulus pupil dilation was associated with higher initial stimulus availability but faster decay of the iconic trace, while information transferred into short-term memory and confidence ratings remained unaffected. These findings demonstrate that spontaneous fluctuations in arousal shape the temporal dynamics of sensory persistence. Elevated arousal boosts the initial strength of visual representations but shortens their duration, suggesting that arousal adaptively prioritizes rapid updating of visual input over prolonged stability. OPEN PRACTICES STATEMENT: The study was pre-registered as part of a larger project at the Open Science Framework ( https://www.osf.io/u6h8d ). The analyses of pupil data reported here were conducted exploratory and were not part of the original pre-registered plan. The data are available at the Open Science Framework ( https://www.osf.io/6rwvt ). The code is available at https://www.github.com/pauljcs/alpha-iconic .