The dynamics of memory consolidation in visuomotor adaptation: memory stabilization prevents decay

GONZALO LERNER; PEDRO CAFFARO; AGUSTIN SOLANO; FLORENCIA JACOBACCI; VALERIA DELLA MAGGIORE

Congreso; Congreso de la Society for the Neural Control of Movement (NCM); 2021

Memories are considered to be consolidated when they are no longer susceptible to interference by new information1,2. When learning is followed by a period of quiescence, memory traces stabilize, leading to long-term memory. Conversely, if learning is followed by exposure to conflicting information the process of consolidation is disrupted, leading to interference3,4. Recently, we showed that anterograde interference impairs the ability to adapt to a visuomotor rotation within a 6 hs time window5. Here, we used the same experimental protocol to examine the time course of motor memory consolidation and whether it relates to the rate of forgetting. To this aim, we first exposed four groups of subjects to learn two opposing 30-degree visuomotor rotations that were separated by either 5 min, 1 hs, 6 hs or 24 hs, while a control group was only exposed to the second rotation. All subjects returned 24 hs later to assess long-term memory retention. We found that even though all groups reached the same level of asymptotic performance during training (Kruskal-Wallis, X2(89) = 7.22, p = 0.13), memory retention differed across groups (X2(87) = 32.87, p < 0.001, Kruskal-Wallis). Retention was severely impaired in the 5 min and 1 h group (both different from control with p < 0.001, Dunn?s test with Bonferroni correction), but recovered to control levels in the 6 hs and 24 hs groups (not different from control with p > 0.88). This result was strengthened by a set of control experiments that ruled out potential confounds related to differences in the amount of overlearning6 and forgetting across groups. Next, we asked whether the stabilization of the memory trace observed around 6 hs post training, would impact on the rate of decay. If so, memories encoded during training in the same experimental paradigm should decay during the first 6 hs post training but not thereafter. To test this hypothesis, we exposed 7 groups of subjects to a visuomotor rotation of the same magnitude and quantified memory retention at different time intervals from 1 min to 24 hs. As predicted, we found that memory decayed (One Way ANOVA, F (147,6) = 26.3, p<0.001) monotonically but reached an asymptote at 5.5 hs post training (5.5 hs group did not differ from the 9 hs nor the 24 hs group; One way ANOVA, followed by Tukey?s test p>0.99). Altogether, our findings provide compelling behavioral evidence suggesting that visuomotor adaptation memories consolidate within a ~6 hs time window.