Cameron Phan¹ (), Erik Van der Burg², David Alais¹; ¹University of Sydney, ²University of Amsterdam

We investigated visual working memory (WM) during locomotion, comparing capacity during two phases of stepping: stance phase vs swing phase. In a virtual reality setting, participants walked along a straight level path while viewing a circular array of 20 discs in a head-mounted display. Discs changed polarity every 150 ms (some changing from black to white, others white to black) and after a random duration, some of the discs temporarily became green (1, 3, 5, 7, or 9 discs). This cued the locations to be remembered (the encoding event) and at the trial’s end one of the 20 disc locations was probed and participants judged whether this was one of the memorized locations or not. The task was completed whilst the participants walked at a comfortable natural pace or walked slowly. Importantly, the encoding event was timed to occur either in the swing or stance phases of the participants’ step cycle and the test probe was timed to occur at the same phase two steps later. Visual WM capacity was greater for encoding in the swing phase of the step cycle than during the stance phase. This WM capacity difference did not depend on walking speed. Response accuracy was greater for swing phase encoding, but individuals were more likely to respond during the stance phase. These findings expand on the impact of active movement on visual abilities and add to recent work by our lab showing that perceptual performance and reaction times modulate over time at a rate locked to the step cycle