Veronica Pisu¹ (), Omer F. Yildiran^2,3, Chloe Lam¹, Saivydas Villani¹, Pascal Mamassian³, Dominik Straub⁴, Constantin Rothkopf⁴, Guido Maiello¹; ¹University of Southampton, ²New York University, ³Ecole Normale Supérieure, ⁴Technical University of Darmstadt

‘Continuous psychophysics’ yields rapid estimates of human visual sensitivity using an easy target-tracking task (Bonnen et al., 2015). The assumption is that tracking performance—e.g., evaluated from the cross-correlation between target and tracking velocities—inversely correlates with perceptual uncertainty: it should be harder to track a target when it is not clearly visible. Estimates of motor variability can also be obtained (Straub & Rothkopf, 2022). Here, across two experiments we tested whether the paradigm can tease apart visual and motor function. Experiment 1 provides a benchmark for comparing eye and mouse tracking. Participants tracked a randomly moving target on a screen using their gaze, with and without concurrent mouse-tracking. Targets were three bivariate Gaussian luminance blobs with fixed total luminance and increasing standard deviation, to manipulate simulated visual noise. Mouse-tracking was more accurate (higher peak cross-correlation, p < .001) and had a longer lag than eye-tracking (p < .001). Critically, as visual noise increased, performance decreased similarly across tracking modalities: peak cross-correlation decreased and occurred later (peak: r = 0.39, p < .05; lag: r = .81, p < .001), suggesting that these differences stem from the motor components of the tasks. In Experiment 2, we compared tracking performance under simulated visual and motor impairment. Participants used a mouse to track a randomly moving concentric Gabor under three levels of simulated visual impairment (using Cambridge Simulation Glasses; Goodman-Deane et al., 2013), using either their dominant or non-dominant hand. Tracking performance decreased with increasing simulated visual impairment (lower peak, p < .01; longer lag, p < .001) and was consistently worse when using the non-dominant hand (lower peak, p < .001; longer lag, p < .001). Taken together, our results demonstrate that continuous psychophysics yields reliable measures of visual function, while also providing estimates of oculomotor and upper-limb motor function.

Acknowledgements: This research was funded by Fight for Sight (Grant Reference: RESSGA2302, awarded to GM).