REM sleep recovers disrupted perceptual learning

McDevitt, E. , Duggan, K. & Mednick, S.

Department of Psychology, University of California, Riverside

New learning is vulnerable to two sources of disruption - task-specific interference resulting from learning similar information, and ?offline? interference resulting from normal mental exertion. How do these sources interact to influence what is remembered? Perceptual learning (PL), the long-term improvement on a visual task, is vulnerable to interference and depends on rapid eye movement (REM) sleep. We varied offline interference as a between-subjects factor (active wake, quiet wake, non-REM and REM naps), and task-specific interference on a texture discrimination task as a within-subjects factor. Task-specific interference conditions included: training two conditions back-to-back with targets in the same location but with different background orientations (HIGH: A-B), or training a similar stimulus set in the contralateral location with a 7-hr inter-condition-interval (LOW: C-D). Task-specific interference disrupted learning (A<B<C), with greater retroactive than proactive interference. Surprisingly, sleep was not necessary for learning with low task-specific interference (C). NREM sleep alone was sufficient to recover moderately disrupted learning (B). However, REM sleep was critical for recovery after high task-specific interference (A). We developed a model that accounts for these findings by considering how the physiological processes underlying each offline consolidation state interact with task-specific interference to predict magnitude of learning.