Plasticity has been clearly observed in the sensory systems in response to both acute and chronic pain, including changes in the dorsal horn, thalamus, and the somatosensory cortex,11–14
but the idea that pain may also affect the motor system is relatively new.15
Most studies that have focused on the interactions between pain and motor function have dealt with the effects of experimental acute pain on spinal-cord reflexes (see Sandrini et al.16
and Clarke and Harris17
for reviews). During the withdrawal reflex response, nociceptive information from skin, muscles, and/or joints makes synapses with motoneurons located in various spinal-cord segments, inducing a complex flexion synergy of the stimulated limb.16,18
This flexion synergy plays a protective role against potential limb damage16
and attests that interactions between pain and motor function occur as early as in the spinal cord. Interestingly, applications of previous noxious stimuli to specific regions of the limb, as well as the presence of certain injuries, have been shown to increase the magnitude of the withdrawal reflex response (see Clarke and Harris17
for a review). These increased withdrawal responses are thought to be caused by changes occurring at the sensory level (e.g., central sensitization) that would enhance the protective function of the withdrawal reflex after tissue injury.17,19
It has been demonstrated that pain leads to a reduction of maximal voluntary contraction, a decrease in endurance during submaximal contraction, and changes in coordination during dynamic tasks (see Graven-Nielsen and Arendt-Nielsen8
and Arendt-Nielsen and Graven-Nielsen9
for reviews). Moreover, recent studies using intra-muscular electromyography (EMG) recordings have shown that pain (induced in either muscular or non-muscular tissue) results in changes in the motor-unit recruitment strategy, revealing that the effect of pain is not limited to a uniform inhibition of the motoneuron pool but, rather, includes more subtle changes in the distribution of output to the motoneuron pool.20,21
However, it is still unclear whether these alterations in motor function observed at the muscular level reflect changes at the peripheral, spinal, or cortical level.
Two different models have been proposed for the interactions between pain and movement: the vicious circle model22
and the pain-adaptation model.23
The vicious circle model suggests that musculoskeletal pain is sustained by the fact that pain-related muscle spasms lead to muscle ischemia, which in turn increases pain and contributes to its maintenance.22
However, this model has not received much support from experimental data.8,23
The pain-adaptation model, on the other hand, predicts a reduction of the agonist motoneuron output and an increase in antagonist motoneuron firing during movement in the presence of pain.23
According to this model, changes in motor output in response to pain result when interneurons receive convergent afferent information and have a reciprocal effect on agonist and antagonist muscles in the spinal cord and brainstem. Two common features of these models of interaction between pain and movement are (1) that they have arisen from clinical observations and experiments focused on localized muscle pain and (2) that they focus on changes in the spinal cord and periphery, without considering any potential role for cortical mechanisms.
More recently, several studies using transcranial magnetic stimulation (TMS) have shown that pain also influences the excitability of the primary motor cortex.6
TMS is a method of stimulating the brain non-invasively. The rapid time-varying magnetic field generated by the TMS coil penetrates the scalp and skull and induces electrical currents in the area of the brain beneath the coil that activate the axons of neurons in the cortex. Stimulation of the motor cortex evokes muscle responses, termed motor evoked potentials
(MEPs), that are measured using EMG. A variety of parameters of MEPs can be studied in order to assess changes in cortico-spinal or intra-cortical excitability. The motor threshold is generally defined as the minimal intensity of stimulation required to produce an MEP of small amplitude in 50% of trials. Therefore, a decrease in motor threshold reflects an increased excitability of the cortico-spinal tract, and vice versa. The size of the MEP (amplitude, duration, or area) also reflects the excitability of the cortico-spinal pathway, which can be affected by a number of mechanisms at both cortical and spinal levels. Paired pulse stimulation, whereby a supra-threshold test stimulus is preceded by a sub-threshold conditioning stimulus, can be used to gain insight into the contribution of local inhibitory and excitatory interneurons in order to assess changes in intra-cortical facilitation or intra-cortical inhibition mechanisms. TMS can also be used to create a cortical map of a target muscle's representation by measuring MEP amplitudes evoked by TMS applied to different positions over the motor cortex. This allows researchers to study the extent and the location (often defined by the centre of gravity of the cortical map) of the cortical representation of a given muscle target.
Although the effect of pain on the motor system can vary depending on variables such as duration of the painful stimulus (phasic vs. tonic pain), submodality (deep vs. superficial pain), and location (proximal vs. distal pain), a common finding of TMS studies is that acute experimental pain exerts an inhibitory influence on cortico-spinal excitability.24–29
This inhibitory effect of experimental pain, however, was not observed by Romaniello et al.30
Changes in responses evoked by TMS do not necessarily reflect changes at the motor-cortex level; alternatively, they could be the result of changes occurring in various neural structures between the primary motor cortex and the motoneurons in the spinal cord. However, there is evidence that the origin of these effects can be at least partially attributed to the cortex. For example, laser-evoked pain was found to attenuate motor responses to TMS but not to transcranial electrical anodal stimulation (which directly activates the pyramidal tract rather than activating cortical interneurons).24,27
Another study showed that during the initial phase of tonic pain induced by injection of hypertonic (5%) saline, there was a reduction of motor responses evoked by TMS stimulation in the absence of any effect on the H-reflex (H-reflex amplitude was decreased in a later phase, about 1 minute after the peak in pain, which suggests that the change initially occurred at the motor-cortex level).25
Pain induced by application of capsaicin on the skin was also found to reduce the amplitude of motor responses evoked by TMS without alteration of spinal excitability.26
It is noteworthy that these different experimental pain models recruit different types of nociceptive afferents. For example, injection of hypertonic saline, often used to mimic musculoskeletal pain, excites nociceptive muscle afferents (groups III and IV),31,32
while capsaicin- or laser-evoked pain selectively activates Aδ and C fibres in the superficial skin layers.33,34
Even though nociceptive inputs from muscle and skin have been shown to induce distinct changes in trigeminal motoneuronal excitability,35
changes at the motor-cortex level appear to be consistently inhibitory across the different pain models (muscle vs. cutaneous pain, phasic vs. tonic pain).24–29
In patients with motor deficits who experience acute pain, the inhibitory influence of pain on the motor cortex may hamper optimal motor-cortex activation during voluntary movement and preclude motor improvement during rehabilitation. There is striking evidence supporting this view from a recent study in healthy individuals showing that acute pain can prevent motor-cortex plasticity associated with novel motor training and impair the ability to learn a new motor task.36
In this study, healthy volunteers participated in two crossover training sessions in which they were trained in a tongue-protrusion task. Prior to each training session, a cream was applied to the tongue that contained either capsaicin (inducing moderate intra-oral tonic pain) or an inert substance (control condition). Although participants' performance in the motor task was improved following training in both painful and non-painful conditions, the improvement was significantly less when the training was performed in the presence of pain (capsaicin condition). Moreover, measurements of cortico-spinal excitability with TMS showed that the presence of pain suppressed training-induced motor-plasticity effects (e.g., increased excitability) observed in the control condition, despite a similar amount of practice. Although these results were obtained in healthy individuals, they strongly suggest that pain can interfere with the effect of motor rehabilitation, both at the cortical and behavioural levels.