The effect of pupil size on near-threshold detection is not mediated by alpha, beta, or theta power

Larger pupils are associated with improved performance in visual detection tasks. This is often attributed to changes in level of arousal as indexed by pupil size. However, changes in pupil size also affect the amount and focus of light on the retina, potentially influencing detection independently of arousal. Such optical effects of pupil size on detection are rarely studied and poorly understood. The goal of this project was to gain a better understanding of how pupil size and arousal (as measured with EEG) are related to each other and to performance in a visual detection task. We collected EEG and pupil data while participants performed a task consisting of detecting faint luminance patches briefly flashed in peripheral vision. Stimulus parameters (contrast and spatial frequency) were adjusted with a staircase procedure to fix performance at 65% accuracy. We performed a mediation analysis within a structural equation modeling framework to investigate the direct and indirect effects present in the data. The results show that larger pupils and theta suppression are both linked to better performance. Pupil size is also positively correlated with power in the alpha and beta bands, with the latter being a stronger relationship. Importantly however, the mediation analysis showed that the relationship between pupil size and accuracy is not mediated by any of the neural measures. This suggests that pupil size affects detection performance independently of the kind of arousal that is expressed as fluctuations in power in the alpha, beta, and theta bands.