Reduced susceptibility to interference in the consolidation of motor memory before adolescence

Abstract

Are children superior to adults in consolidating procedural memory? This notion has been tied to "critical," early life periods of increased brain plasticity. Here, using a motor sequence learning task, we show, in experiment 1, that a) the rate of learning during a training session, b) the gains accrued, without additional practice, within a 24 hours post-training interval (delayed consolidation gains), and c) the long-term retention of these gains, were as effective in 9, 12 and 17-year-olds and comparable to those reported for adults. However, a follow-up experiment showed that the establishment of a memory trace for the trained sequence of movements was significantly more susceptible to interference by a subsequent motor learning experience (practicing a reversed movement sequence) in the 17-year-olds compared to the 9 and 12-year-olds. Unlike the 17-year-olds, the younger age-groups showed significant delayed gains even after interference training. Altogether, our results indicate the existence of an effective consolidation phase in motor learning both before and after adolescence, with no childhood advantage in the learning or retention of a motor skill. However, the ability to co-consolidate different, successive motor experiences, demonstrated in both the 9 and 12-year-olds, diminishes after puberty, suggesting that a more selective memory consolidation process takes over from the childhood one. Only the adult consolidation process is gated by a recency effect, and in situations of multiple, clashing, experiences occurring within a short time-interval, adults may less effectively establish in memory experiences superseded by newer ones.

Figure 1

Within-session and between-sessions gains in performance in the 3 age-groups. Inset: the two finger-to-thumb opposition movement sequences used in the study. (a) Mean number of correct sequences, and (b) mean number of errors performed in each test interval (block) during the training session and at 24, 48 hours and 6 weeks post training. Bars–standard error.

Figure 2

Age dependent effect of post-training interference. (a) Speed and (b) accuracy gains with (♦) and without (□) interference training in the three age-groups. Interference training was given at 2 hours after the termination of the initial training session. Average performance in the initial (init) and the final (end) four blocks of the initial training session, and in four consecutive blocks at 24 hours post-training (24hr post) is shown. Bars–standard error; black arrow - significant interaction; white arrows - no interaction (significant gains in both experiments). Comparison between the three experimental groups' performance at the end of training and at 24 hours post-training in the two experiments, without and with interference (repeated measures ANOVA) showed a significant main effect for time-point for both the number of sequences (speed) and the number of errors (accuracy) (speed: F _(1,110) = 152.06, P<.001 accuracy: F _(1,110) = 7.11, P<.01) and for age (speed: F _(2,110) = 60.01, P<.001 accuracy: F _(2,110) = 6.89, P<.001). The only significant interaction (age-group×assessment time×experiment) was for performance speed (F _(2,110) = 6.82, P<.05) with the 17-year-olds showing less improvement in the interference condition. There was no significant difference between the two experiments for the end time-point in the 17-year-olds (t ₍₃₇₎ = 0.81, P = .42). An analysis of variance for repeated measures, conducted for each age-group separately (with time-points as within-subject factor and age-group and experiment as between-subject factors), showed a significant difference in between-session gains accrued for the initially trained sequence (task A) between the two experiments only in the 17-year-olds (interaction of time-point×experiment, F _(1,37) = 10.62, P<.001). The 9 and 12-year-olds improved to a similar degree with and without interference (no interaction of time-point×experiment, F _(1,35) = 2.89, P = .1; F _(1,38 = 2.21, P = .15, 9 and 12-year-olds respectively).

Figure 3

Between-session (delayed) gains with (□) and without (▪) interference training in the three age-groups. The absolute gains in terms of mean number of correct sequences at 24 hours post-training compared to the end of the training session. Bars–standard error. There was a significant interaction of condition by age-group for the mean between-session gains (F _(2,110) = 6.82, P<.05). Independent-samples t-tests conducted for each age group separately showed a significant advantage of the no interference over the interference condition only in the 17-year-olds (t ₍₃₇₎ = 3.26, P = 0.02). There were no significant differences in delayed gains in the two conditions for the 9 and 12-year olds (t ₍₃₅₎ = 1.7, P = 0.1; t ₍₃₈₎ = 1.49, P = 0.002, 9 and 12-year-olds respectively).