Abstract
Practice in perceptual tasks over hundreds or thousands of trials often leads to long-lasting improvements in performance that generalize only partially to new stimuli. However, the time courses of the general and stimulus-specific aspects of learning are still debated. Some researchers argue that general aspects of the task are learned first in an initial rapid phase of learning and that stimulus-specificity emerges more slowly. In contrast, Hussain et al. (Front Psychol. 2012; 3:226) reported that 105 trials in a 10-AFC face identification task on Day 1 was sufficient to produce stimulus-specific learning in a test phase on Day 2, suggesting that stimulus-specific improvements can emerge rapidly. The current experiments extend these findings by examining 1) whether similar, rapid stimulus-specific learning occurs in a 10-AFC texture identification task; 2) if this rapid stimulus-specificity is long-lasting by increasing the interval between Days 1 and 2 from 24 hours to 1 week; and 3) the effects of reducing practice on Day 1 from 840 to just 21 trials. On Day 1, subjects performed a 10-AFC identification task with band-pass random textures embedded in three levels of external noise. The textures were presented at 7 contrasts that spanned the threshold range; hence the signal-to-noise ratio varied significantly across trials. On Day 2, subjects performed the task with the same or a novel set of textures. The dependent variable was response accuracy, and stimulus-specificity was measured by comparing performance with the same and novel textures on Day 2. We found stimulus-specific learning in subjects who received 840, 105, and 63 trials of practice, but not in subjects who received 21 trials of practice. Our results are consistent with the idea that stimulus-specific learning can emerge rapidly during practice and that this rapid learning is long lasting.
Meeting abstract presented at VSS 2013