Sensorimotor deprivation induces interdependent changes in excitability and plasticity of the human hand motor cortex

Abstract

Prolonged limb immobilization deprives sensorimotor cortical areas of an important source of excitatory input, as well as of motor output. Previous work has described effects on motor excitability but it is unclear whether motor plasticity is also influenced. In two groups of eight healthy human subjects, the left hand was immobilized for 8 h to induce sensorimotor deprivation of the cortical representation of the abductor pollicis brevis muscle. We used transcranial magnetic stimulation protocols to evaluate motor excitability with motor-evoked potentials, input-output (IOcurve) and short-latency intracortical inhibition (SICI) recruitment curves, as well as long-term potentiation (LTP)/long-term depression (LTD)-like plasticity with paired-associative stimulation (PAS) of the median nerve and motor cortex using an interstimulus interval of 25 ms (PAS25) or 10 ms (PAS10), respectively, in two sessions at least 7 d apart (baseline and after immobilization). After immobilization, the slope of the IOcurve decreased, and SICI at lower conditioning pulse intensities was reduced. The LTP-like effects of PAS25 and the LTD-like effect of PAS10 were both significantly enhanced. The effects differed among individuals: the more IOslope decreased after immobilization, the greater the increase of PAS25 and the smaller the increase of PAS10 effects. We suggest that sensorimotor deprivation has two effects. It increases the sensitivity to remaining sensory inputs and therefore increases the effectiveness of both PAS protocols. In addition, it reduces neuronal excitability to an individually different level, as reflected in the reduced IOslope and leads to an interdependent modulation of synaptic plasticity as such as it shifts the threshold of LTP/LTD-like plasticity induction.

Keywords: human; plasticity; sensorimotor deprivation; transcranial magnetic stimulation.

Copyright © 2014 the authors 0270-6474/14/347375-08$15.00/0.

Figures

Figure 1.

Experimental protocol. The subjects participated first in the baseline session, which was followed by the interventional session within 7 d minimal.

Figure 2.

IOcurves and IOcurve slopes in the APB. The mean MEP amplitude (in mV± SE) as given on the y-axis against the stimulus intensity given on the x-axis (in percentage of SI1mV). A, B, IOcurves measured at baseline and after immobilization in the PAS25 group (A) and the PAS10 group (B). In both groups, the IOcurves are less steep after immobilization compared with baseline. C, The slopes of the IOcurves have been calculated for the approximately linear part between 90 and 130% SI1mV. The decrease of IOcurve slopes after immobilization is significant in both groups (paired t test; p < 0.001).

Figure 3.

Mean MEP (±SE) in the APB in the PAS25 and PAS10 groups. A, B, The mean MEP in millivolts measured at baseline and after immobilization; the MEPs measured before PAS are shown in gray columns and after PAS in colored columns (green for PAS25; red for PAS10). There was no difference in MEP size before PAS in the experiments performed at baseline and after immobilization, neither in the PAS25 nor in the PAS10 group. Therefore, the MEPs were normalized (MEP after PAS/ MEP before PAS) and expressed as percentages (C). After immobilization, both PAS protocols were significantly more effective: PAS25 increased whereas PAS10 decreased MEPs more than at baseline. Statistical results are given in the figure (paired t tests; *p < 0.05; **p < 0.01).

Figure 4.

Changes and correlations of IOslopes and PAS effects after immobilization in the individual subjects. A, B, Plot for each individual subject of the PAS25 (A) and PAS10 (B) groups the IOslope change (IOslope after immobilization: IOslope at baseline; percentage) on the left y-axis, as well as the change of the normalized PAS effect (PAS effect after immobilization: PAS effect at baseline; percentage) on the right y-axis, which displays the relative PAS effect change in percentage with ticks indicating steps of 10%. In both groups, the IOslope decreases after immobilization below the values measured at baseline (all black dots are <100%). However, the extent of the decrease varied in the individual subjects. The change in PAS effect is plotted as arrows going up (PAS25) or down (PAS10) to indicate the physiological direction of the PAS effects (LTP-like effect of PAS25; LTD-like effect of PAS10) which both show an increase after immobilization (note: subject 8 in the PAS10 group shows a decreased effect). In the PAS25 group, stronger decrease of the IOslope was associated with a stronger increase of the PAS25 effect after immobilization; whereas in the PAS10 group, stronger decrease of the IOslope was associated with a weaker increase of the PAS10 effect after immobilization (except subject 8 in the PAS10 group). C, The linear regression between the IOslope change (IOslope after immobilization: IOslope at baseline; percentage; x-axis) and the PAS effect change (PAS effect after immobilization: PAS effect at baseline; percentage; y-axis) for the PAS25 (green symbols) and PAS10 (red symbols) groups. In both groups, the immobilization-induced changes of IOslope and PAS effects are significantly correlated (PAS25 group: p < 0.002; PAS10 group: p < 0.013); Pearson's r is given in the figure.

Figure 5.

SICI and correlation of SICI and IOslope changes after immobilization. A, SICI obtained with a conditioning pulse intensity of 70, 80, and 90% aMT. The y-axis plots the amplitude of the conditioned MEP as the percentage of MEP evoked by the test pulse alone (±SE). With a conditioning stimulus of 70 and 80% aMT, SICI was significantly lower after immobilization than at baseline (paired t tests; *p < 0.02). B, Correlation of the changes in IOslope (x-axis) and SICI (y-axis) induced by immobilization. The SICI changes measured with 70 and 80% aMT are significantly correlated with the change of IOslope: the more the IOslope decreases after immobilization, the smaller is the reduction in SICI. Pearson's r is given for the significant correlations.