Reactivation or transformation? Motor memory consolidation associated with cerebral activation time-locked to sleep spindles

Abstract

Motor memory consolidation is thought to depend on sleep-dependent reactivation of brain areas recruited during learning. However, up to this point, there has been no direct evidence to support this assertion in humans, and the physiological processes supporting such reactivation are unknown. Here, simultaneous electroencephalographic and functional magnetic resonance imaging (EEG-fMRI) recordings were conducted during post-learning sleep to directly investigate the spindle-related reactivation of a memory trace formed during motor sequence learning (MSL), and its relationship to overnight enhancement in performance (reflecting consolidation). We show that brain regions within the striato-cerebello-cortical network recruited during training on the MSL task, and in particular the striatum, were also activated during sleep, time-locked to spindles. Interestingly, the consolidated trace in the striatum was not simply strengthened, but was transformed/reorganized from rostrodorsal (associative) to caudoventral (sensorimotor) subregions. Moreover, the degree of the reactivation was correlated with overnight improvements in performance. Altogether, the present findings demonstrate that striatal reactivation linked to sleep spindles in the post-learning night, is related to motor memory consolidation.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Fig 1. Overview of experimental design.

On Day 1, subjects (n = 13, 7 female) first underwent a screening and adaptation night (bed time: 23h00) in the mock scanner under conditions similar to those experienced in both experimental and control nights to ensure that they could attain adequate sleep required for subsequent post-training sleep sessions in the MRI scanner. Subjects returned (Day 7 and Day 14) for training on either the motor sequence learning (MSL) or motor control (CTRL) task, which were administered in a random order in two separate sessions that took place on two consecutive weeks. Practice on either of these two tasks began in the evening around 22h30, and was followed by simultaneous EEG-fMRI sleep recording (starting at 23h00) lasting up to ~2.25 hours. Subjects were then allowed to sleep for the remainder of the night in the sleep lab until 07h30. This was followed by retest sessions on the same task as the previous training session (Day 8 and Day 15) comprising both behavioral and imaging data collection in the morning beginning at 09h00.

Fig 2. Behavioral data.

A: Mean (+/- SEM) performance speed (inter-key press interval) during training and retest sessions for the explicit motor sequence learning (MSL) and control (CTRL) tasks. B: Gains in performance (reflecting consolidation) was measured using the mean (+/- SEM) inter-key-press interval in the last 4 blocks of the training vs. the first 4 blocks of retest. As expected, subjects showed significant gains in performance in the MSL task only (* indicates p < 0.05).

Fig 3. Cerebral activation during training, sleep spindles and retest.

A: MSL-specific (MSL>CTRL) brain activation during the training session. (inset): Increased activation bilaterally in the striatum from the beginning of training (first 7 blocks) as compared to the end of the training session (last 7 blocks) on the MSL task. B: Activations time-locked to sleep spindles following MSL. (inset): Conjunction of within training session-related and spindle-related activations. C: MSL-specific (MSL > CTRL) brain activation during the retest session. Results displayed at p<0.001, uncorrected.

Fig 4. Cerebral activations during practice.

Change in striatal activation from the training to the retest session in the MSL vs. CTRL condition (A), was associated with overnight gains in performance (B). Results displayed at p<0.001, uncorrected. Signal intensity expressed in arbitrary units (a.u.) obtained from raw beta weights.

Fig 5. Cerebral activations during training and retest.

Striatal activity within the training session (yellow) and the retest session (red). Following sleep, striatal activity was reorganized in different sub regions, as opposed to a strengthening of activity in the same regions. Results displayed at p<0.001, uncorrected.

Fig 6. Offline gains in performance correlated with spindle activation.

Spindle activation (MSL) in the ventral right striatum (putamen) correlated positively with overnight gains. Signal intensity expressed in arbitrary units (a.u.) obtained from extracting raw beta weights.

Fig 7. Correlation between spindle-related activation and practice session activation.

A: Positive correlation (r(11) = 0.53, p = 0.061) between spindle-related activation in the right striatum and the change in activation from the whole training session to the whole retest session. B: A similar relationship (r(11) = 0.53, p = 0.065) was observed between spindle-related activations and overnight changes in the right striatum from end of training to start of retest during practice (first 7 blocks retest—last 7 blocks training). While these correlates are significant only at trend levels, note the very clear linear relationship in the scatterplots, which are consistent with our a-priori hypotheses (and would be statistically significant if a one-tailed test were employed). Signal intensity expressed in arbitrary units (a.u.) obtained from extracting raw beta weights.

Fig 8. Summary of main findings.

Behavior (upper): Sequence learning performance improved with practice over the course of the training session. Overnight gains in behavioural performance were observed from the end of the training session to the beginning of the retest session, and performance remained asymptotic at retest. Cerebral activation (lower): Increased learning-related cerebral activity was observed in the striatum from the beginning to the end of the training session. Reactivation of the striatum was time-locked to sleep spindles, and was correlated with offline gains in performance (r = 0.57) as well as with overnight changes in striatal activity from the Training to the Retest session (r = 0.53). The overnight change in striatal activity, which involved a transformation/reorganization within the striatum, was also correlated with offline gains in performance (r = .56). Note: colored arrows indicate difference measures taken between sessions to calculate overnight gains in performance and overnight changes in cerebral activity in the striatum (Retest-Training). Dashed line indicated statistical trend at p = 0.061.