Homeostatic and nonhomeostatic modulation of learning in human motor cortex

Abstract

Motor learning is important throughout life for acquisition and adjustment of motor skill. The extent of motor learning may be modulated by the history of motor cortex activity, but little is known which metaplasticity rule (homeostatic vs nonhomeostatic) governs this interaction. Here, we explored in nine healthy adults the effects of three different paired associative stimulation (PAS) protocols on subsequent learning of rapid thumb flexion movements. PAS resulted in either a long-term potentiation (LTP)-like increase in excitability of the stimulated motor cortex (PAS(LTP)), or a long-term depression (LTD)-like decrease (PAS(LTD)), or no change (control condition, PAS(CON)). Learning was indexed by the increase in peak acceleration of the trained movement. Delays of 0 and 90 min between PAS and motor practice were tested. At the 0 min delay, PAS(LTD) strongly facilitated motor learning (homeostatic interaction), and PAS(LTP) also facilitated learning, although to a lesser extent (nonhomeostatic interaction). At the 90 min delay, PAS(LTD) facilitated learning, whereas PAS(LTP) depressed learning (interactions both homeostatic). Therefore, facilitation of learning by previous brain stimulation occurs primarily and most effectively through homeostatic interactions, but at the 0 min delay, nonhomeostatic mechanisms such as LTP-induced blockade of LTD and nonsaturated LTP-induced facilitation of learning might also play a role. The present findings demonstrate that motor learning in humans can be modulated by noninvasive brain stimulation and suggest the possibility of enhancing motor relearning in defined neurological patients.

Figure 1.

Time line of experiments. Three different PAS conditions are applied in separate sessions (PAS_LTP, PAS_LTD, PAS_CON) and followed by two 15 min blocks of motor practice of fastest possible thumb flexion movements of the right hand, either immediately (experiment 1) or after a waiting period of 90 min (experiment 2). The effect of PAS on corticospinal excitability is assessed by MEP amplitudes in the resting right flexor pollicis brevis muscle at baseline (time point B0), immediately after PAS (B1) and in experiment 2 again after the waiting period (B2). At B0, stimulus intensity is adjusted to produce MEP amplitudes of on average 1 mV in peak-to-peak amplitude (MEP_1mV). Practice-dependent plasticity is assessed by peak ACC measurements of the trained thumb flexion movement, before motor practice (time points B0 and B1, and in experiment 2 in addition at B2), in between the two blocks of motor practice (P1) and for 30 min after motor practice (P2–P5).

Figure 2.

Effects of PAS on MEP amplitude. a, In experiment 1, PAS_LTP (black bars) resulted in significant MEP increase (MEPs measured at time point B1 immediately after PAS, normalized to MEP baseline at B0); PAS_LTD (white bars) resulted in an MEP decrease, and PAS_CON (gray bars) led to no change. b, In experiment 2, findings were essentially identical to experiment 1 when MEP at time point B1 was compared with baseline at B0. c, In experiment 2, when MEPs after 90 min waiting (time point B2) were compared with baseline at B0, all of the changes in b were no longer present. All data are means of nine subjects; error bars are 1 SEM. Asterisks indicate significant difference from 1 (one-sample two-tailed t tests, p < 0.05).

Figure 3.

Effects of PAS on learning during motor practice. Increase in peak ACC of the trained thumb flexion movement in experiment 1 (a) and experiment 2 (b) during two blocks of 15 min motor practice (MP) normalized to baseline peak acceleration before practice (experiment 1: ACC_MP/ACC_B1; experiment 2: ACC_MP/ACC_B2) when preceded by PAS_LTP (black circles), PAS_LTD (white circles), or PAS_CON (gray circles). ACC_MP data are binned in 1 min steps and are means of nine subjects; error bars are 1 SEM. Note that in experiment 1 (a), PAS_LTD and PAS_LTP result in enhancement of learning compared with PAS_CON, whereas in experiment 2 (b), PAS_LTD results in enhancement but PAS_LTP in a trend toward depression.

Figure 4.

Effects of PAS on learning after the end of motor practice. Change in peak ACC of the trained thumb flexion movement in experiment 1 (a) and experiment 2 (c) after the end of practice at time points P2–P5 (1, 11, 21, and 31 min after end of practice, respectively) normalized to peak acceleration immediately before practice (experiment 1, ACC_P/ACC_B1; experiment 2, ACC_P/ACC_B2) when preceded by PAS_LTP (black circles), PAS_LTD (white circles), or PAS_CON (gray circles). Same data in b and d, but peak acceleration is normalized to peak acceleration in the PAS_CON condition. All data are means of nine subjects; error bars are 1 SEM. Asterisks indicate significant difference from 1 (1-sample t tests, p < 0.05). Note that PAS_LTD led to enhancement of learning in both experiments, whereas PAS_LTP resulted in no difference with PAS_CON in experiment 1 but a depression of learning in experiment 2.

Figure 5.

Effects of PAS on learning strategy. Trial-by-trial variability of peak acceleration in experiment 1 (a) and experiment 2 (b) during two blocks of 15 min motor practice when preceded by PAS_LTP (black circles), PAS_LTD (white circles), or PAS_CON (gray circles), expressed by the coefficient of variation of peak acceleration (for details, see Materials and Methods). All data are binned in 1 min steps and are means of nine subjects; error bars are 1 SEM. Note that learning strategy was not different across the three PAS conditions.