Consolidated learning can be susceptible to gradually-developing interference in prolonged motor learning

Abstract

When multiple items are learned in sequential order, learning for one item tends to be disrupted by subsequently learned items. Such retrograde interference has been studied with paradigms conducted over a relatively short term. Resistance to interference is generally believed to be a measure of learning or consolidation. Here, we used a finger-tapping motor sequence paradigm to examine interference in prolonged motor learning. Three groups of nine subjects participated in training sessions for 16 days, and practiced three different sequences in different orders and combinations. We found that a well-trained motor sequence was subject to a gradual interference when the subsequent learning was paired in a particular order. The results suggest that a well-learned motor memory is still susceptible to interference, and that resistance to interference in one condition does not necessarily imply full, permanent consolidation.

Keywords: consolidation; finger tapping; interference; motor learning; retrograde interference.

Figure 1

Schematic representation of the experimental procedure.

Figure 2

Performance during a finger-tapping motor sequence task for three training sequences (three groups). The mean number of correctly typed responses as a function of the training day. In the first part of the training session (Days 1–8), a single sequence was trained; two sequences (sequences B and A) were practiced in all groups in the second training period (Days 9–16). (A) Group ABA; (B) Group CBA; (C) Group BBA.

Figure 3

Relative performance improvement compared to the initial day performance by group.

Figure 4

Averaged number of correctly tapped sequences (±s.e.m.) plotted for each block and for each session. For each group, the mean numbers of correctly tapped sequences are plotted for each learned sequence.

Figure 5

Within-session learning as indicated by an increase in the number of correctly tapped sequences within a session (±s.e.m.). Asterisks indicate sessions with significant deviation from zero (p < 0.05, Bonferroni correction).

Figure 6

(A) Correlations between the initial performance on Day 1 and the extent of interference. (B) Correlations between the performance improvement during the first part of the training, and the extent of interference.