Sleep-dependent motor memory formation detected by fMRI

Sleep seems to be essential for the consolidation of procedural memory, such as fast and accurate tapping of a defined sequence of finger movements. Using fMRI we measured activation while subjects were trained on a sequential finger opposition task before a night of sleep or wakefulness and while they were retested 48 hours later.

Methods: 8 volunteers performed a 5 finger-to-thumb opposition task with their non-dominant hand (left) as rapidly and accurately as possible, according to a block design with 30-s of rest (8 intervals) alternating with 30-s of task performance (7 intervals). Two different motor sequences were used to allow testing of the same subject in both conditions. Sessions took place in the evening before (learning) and after (retrieval) a 48-hour retention interval. During the first post-learning night subjects either slept regularly or stayed awake. During the second night all subjects had regular sleep. Imaging was carried out at 1.5 T. Data were analyzed using SPM2 (second-level). Motor performance was measured during the scanning in terms of rate and accuracy.

Results: Initial learning of the sequence was similar in the sleep and the wake condition. Performance rate improved from initial training to retention intervals after sleep. Initial learning was linked to a specific cerebro-cerebellar activation pattern for left non-dominant hand performance. The before/after×sleep/wake interaction indicated increased activity depending on sleep or wakefulness: the left superior parietal lobe, the right nucleus caudatus and the left anterior lobe of the cerebellum were significantly larger activated during retrieval as compared to learning when subjects slept. Conversely, the left and right middle prefrontal lobe and the left putamen activity increased when subjects stayed awake. For the reverse comparison (initial learning > retrieval testing), the interaction revealed that sleep was associated with relative decreases in the right premotor, the inferior prefrontal and the right primary motor cortex as compared with wakefulness. Our present results confirm that sleep after training improves performance of a sequential finger opposition task and that sleep-dependent motor memory formation is accompanied by a specific change in the pattern of brain activation.