The rapid emergence of stimulus-specific perceptual learning

Is stimulus-specificity the outcome of extended practice or can it occur with a relatively small amount of practice? Separate groups of observers practiced a 10AFC face identification task over two consecutive days, with different amounts of practice on day 1: either 40 or 5 trials per condition. On day 2, half the observers from each group transferred to novel faces. Stimuli were presented using method of constant stimuli, at one of three levels of external noise, and one of seven different contrasts (21 stimulus conditions). Learning was assessed by measuring identification accuracy across bins of trials in each session. Consistent with earlier findings, small amounts of practice on day 1 improved performance on day 2, although more learning was obtained in the 40-trials condition (Hussain et al, 2009 VisionResearch 49). Crucially, between-session learning in both conditions was stimulus-specific: performance dropped for novel stimuli regardless of whether observers received 40 or 5 trials per condition on day1. These results suggest that, at least for some tasks, a relatively small number of trials is sufficient to produce perceptual learning that is stimulus specific.