Interference in perceptual learning

Abstract

Perceptual learning is known to be highly specific to stimulus conditions that are very similar to those under which the subject has been trained. Does this high specificity mean that learning under one set of conditions is entirely independent of learning under another? In the present study, we found evidence that perceptual learning in one visual task can interfere with learning in another similar task, at least when the training of one task occurs immediately after the training of another. We compared the learning rates of two groups of subjects trained in a three-dot vernier acuity task for five days. While the control group was only trained on a three-dot vernier acuity task in which the offset was always to the same side, the experimental group carried out two consecutive training sessions each day: session A in which the offset was to one side, and session B in which the offset was to the opposite side. For large offsets (easier trials), both groups improved at a similar rate over the five-day period. For the smaller offsets (difficult trials), however, the experimental group showed a decrease in performance and the control group showed improvement. This discrepancy in performance suggests that learning one perceptual task can interfere with the learning of another very similar task. To test if this interference effect can be ameliorated by a consolidation period, a second study was conducted in which the A and B sessions were separated in time by more than 24 hours. In this experiment, no or markedly less interference was found between the tasks. These results suggest that learning under one set of conditions is susceptible to interference from another until it is consolidated. While interference is a well-known phenomenon in explicit memory and motor learning research, the present study is the first to show interference in perceptual learning.

Footnotes

 This work was supported by the Clara Mayo Fellowship.