A nap consolidates generalized perceptual learning.

Previous research has demonstrated that a night's sleep can consolidate rote and generalized perceptual learning. Over a waking retention period following training, performance gains from learning significantly decline, but sleep can restore performance to levels found immediately after learning. Furthermore, when sleep precedes a waking retention period following training, performance is protected against loss. Other research demonstrating that rote learning can be consolidated by a night's sleep has shown that a relatively brief nap can consolidate rote learning. This suggests that short periods of sleep can produce consolidation, indicating that consolidation may not require successive sleep cycles over an entire night to emerge. However, previous research has demonstrated that there can be differences in sleep-dependent consolidation for rote and generalized learning. In this study, we investigated whether an opportunity for a 90-min midday nap was sufficient to consolidate generalized perceptual learning of synthetic speech. We recruited 75 participants from the University of Chicago community (mean age of 20.83) who completed a pretest, training, and posttest in the morning on perception of synthetic speech. Training and testing in this manner are known to result in substantial generalized learning of synthetic speech. Participants then returned in the afternoon and were either given an opportunity for a 90-min nap or remained awake for 90-min. Participants were then given another posttest later that evening, never hearing the same words twice during the experiment. Results demonstrated that participants who did not nap showed significant loss of learning at the evening posttest. In contrast, individuals who napped retained what they learned, and did not show loss of learning at the evening posttest. These results are consistent with the view that an opportunity for a 90-min midday nap can consolidate generalized learning, as only individuals with consolidated learning should be able to retain what they learned despite an intervening waking retention period. This is the first demonstration that generalized skill learning is subject to sleep-dependent consolidation in short durations of sleep and does not require a full night of sleep. This work has implications for understanding the basic neural mechanisms that operate to stabilize short-term learning experiences.