Experiments using transcranial magnetic brain stimulation in man could reveal important new mechanisms in motor control

Since its introduction 15 years ago, transcranial magnetic stimulation (TMS) has been widely used to investigate the human motor system. Our knowledge of human corticospinal function is based largely on TMS studies. Application of TMS to the motor areas of the cerebral cortex evokes short-latency responses in the EMG recorded from a large variety of human muscles. The shortest possible route for the production of these responses involves direct cortico-motoneuronal projections to the spinal motoneurones innervating the muscle. There is evidence that mechanisms operating at both cortical and spinal level can affect the size and nature of the evoked response, and a number of procedures have been found which have profound long-term effects on the responses. These include training on a skilled task, fatiguing contractions, reversible interruption of somatosensory input or prolonged periods of repetitive TMS. The duration of these effects ranges from a few minutes up to several hours or even days.

The site of action of these effects has been considered to be firmly located in the motor cortex, since controls using H-reflexes and other techniques imply that spinal motoneurone excitability is not changed by the procedure. The paper by Gandevia et al. in this issue of The Journal of Physiology changes this picture substantially. They have compared the long-term effects of a period of maximal voluntary contraction (MVC) on EMG responses evoked by TMS of the motor cortex with those on EMG responses evoked by transmastoid electrical stimulation, which has been shown to activate the corticospinal tract at a lower level, close to the junction between brainstem and spinal cord. Gandevia et al. (1999) clearly show that there is a substantial, long-lasting depression of EMG responses to the transmastoid stimulation following maximal voluntary contractions. This depression, which lasted for more than 2 min, was not present at short intervals after the MVC in TMS-evoked responses. Although transmastoid stimulation probably activates other brainstem ascending and descending pathways, Gandevia et al. (1999) have also demonstrated, using a collision method, that both TMS and transmastoid responses are mediated by the same population of corticospinal axons. Since transmastoid stimuli activate axons in the medulla, they are unlikely to have been influenced by excitability changes in the motor cortex itself. In other words, the site of depression is not in the cortex, but at a subcortical site.

Significant changes in EMG responses occur peripherally following the MVC (Taylor et al. 1999), and the contribution of such effects has to be excluded before any central mechanism for the effects described by Gandevia et al. (1999) can be identified. In this study the authors excluded these peripheral effects, first by normalising their EMG responses to the ‘M-wave’ evoked from the peripheral nerve and second by demonstrating that the depression after an MVC was not observed when the muscle was fatigued by tetanic stimulation of the motor nerve. Thus the depression can be safely ascribed to a central rather than a peripheral mechanism.

Gandevia et al. (1999) initially describe a reduction in size of transmastoid responses after very long duration maximal efforts (2 min), but later describe a very similar effect of much shorter (5 s) contractions implying that the processes underlying the change undergo saturation. Responses to TMS during the same protocol showed a very different time course of change: immediately after the MVC, responses were facilitated, followed by a depression that developed progressively to a maximum after about 2 min. This depression also seems different in that it was increased with longer-duration contractions. The facilitation of the TMS response is all the more striking since it occurs when the depression of transmastoid-evoked responses is maximal.

This study reveals an intriguing phenomenon and begs important questions. If responses to transmastoid stimulation are evoked by direct activation of corticospinal axons in the medulla, then the depression is best explained either (a) by changes in corticospinal axonal excitability after MVC; a sort of ‘history effect’ on the axon brought on by long and sustained activation during the MVC, or (b) by depression at the corticospinal terminals on motoneurones. The latter are thought not to be subject to classical presynaptic inhibition (Nielsen & Petersen, 1994; see also Rudomin & Schmidt, 1999), although effects at such long duration have not been investigated at all in non-human animal experiments (they should be!)

If transmastoid stimulation activates the corticospinal axons directly, it should produce a single action potential in each activated axon, giving rise to a synchronous descending corticospinal volley. This is very different from the effects of TMS which are known to produce a much more complex discharge pattern, with repetitive firing in corticospinal axons, corresponding to the D- and I-waves (Rothwell, 1991; Di Lazzaro et al. 1998). A single corticospinal neurone can respond to a single TMS shock with both D and I responses (Edgley et al. 1997). The compound motor-evoked potential produced by TMS reflects the summation of the excitation due to these repetitive effects. Thus, the differences in post-MVC effects may be sought in this very different pattern of corticospinal activation; possibly the repetitive bombardment of a given motoneurone by the TMS-evoked volley can overcome the depression at the cortico-motoneuronal axon and/or synapse. It is also remarkable that Gandevia et al. (1999) were able to collide the response to TMS with a single transmastoid shock.

Thus, these experiments open the door to a number of future investigations and stress that besides the complexity and plasticity of the cerebral cortex, important new mechanisms may also be operating at the spinal level of the human motor system.