Modulation of motor learning by a paired associative stimulation protocol inducing LTD-like effects

Abstract

BackgroundPaired associative stimulation (PAS) induces long-term potentiation (LTP)-like effects when interstimulus intervals (ISIs) between electrical peripheral nerve stimulation and transcranial magnetic stimulation (TMS) to M1 are approximately 21–25 ms (PASLTP). It was previously reported that two forms of motor learning (i.e., mode-free and model-based learning) can be differentially modulated by PASLTP depending on the different synaptic inputs to corticospinal neurons (CSNs), which relate to posterior-to-anterior (PA) or anterior-to-posterior (AP) currents induced by TMS (PA or AP inputs, respectively). However, the effects of long-term depression (LTD)-inducing PAS with an ISI of approximately 10 ms (PASLTD) on motor learning and its dependency on current direction have not yet been tested. ObjectiveTo investigate whether, and how, PASLTD affects distinct types of motor learning. MethodsEighteen healthy volunteers participated. We adopted the standard PAS using suprathreshold TMS with the target muscle relaxed, as well as subthreshold PAS during voluntary contraction, which was suggested to selectively recruit PA or AP inputs depending on the orientation of the TMS coil. We examined the effects of suprathreshold and subthreshold PASLTD on the performance of model-free and model-based learning, as well as the corticospinal excitability, indexed as the amplitudes of motor evoked potentials (MEPs). ResultsPASLTD inhibited model-free learning and MEPs only when subthreshold AP currents were applied. The PASLTD protocols tested here showed no effects on model-based learning. ConclusionsPASLTD affected model-free learning, presumably by modulating CSN excitability changes, rather than PA inputs, which are thought to be related to model-free learning.