External reinforcement is needed for perceptual learning of global motion.


The role of external feedback as a facilitator of perceptual learning is currently not clear. While many perceptual learning studies acknowledge that external feedback can increase the rate of learning, learning also occurs in the absence of external feedback (Vaina et al, 1998, Vis Res, 95(21) 12657-12662), and can be enhanced using fake feedback (Shibata et al, 2009, Vis Res, 49(21), 2574–2585), or by including easy exemplars (Liu et al, 2012, Proc. Natl. Acad. Sci. 49(21), 2574–2585). This study evaluated the role of external trial-by-trial feedback on learning in a global motion task. Participants trained for 10 days on a 2AFC random dot global motion task, at varying levels of coherence, requiring a leftwards/rightwards direction decision. One group received trial-by-trial feedback for correct responses (high pitched tone) or incorrect trials (low pitched tone) while the control group received no feedback as to their performance. Performance was monitored daily and a 75% correct motion coherence threshold was calculated for each participant. Performance improved over the 10 days for the feedback group, but there was no change in performance for the group that received no feedback. Our results indicate that, for a global motion task, perceptual learning only occurred when external feedback was provided, even when easy exemplars were presented.