The Potential of Corticospinal-Motoneuronal Plasticity for Recovery after Spinal Cord Injury

Abstract

Purpose of review:

This review focuses on a relatively new neuromodulation method where transcranial magnetic stimulation over the primary motor cortex is paired with transcutaneous electrical stimulation over a peripheral nerve to induce plasticity at corticospinal-motoneuronal synapses.

Recent findings:

Recovery of sensorimotor function after spinal cord injury largely depends on transmission in the corticospinal pathway. Significantly damaged corticospinal axons fail to regenerate and participate in functional recovery. Transmission in residual corticospinal axons can be assessed using non-invasive transcranial magnetic stimulation which combined with transcutaneous electrical stimulation can be used to improve voluntary motor output, as was recently demonstrated in clinical studies in humans with chronic incomplete spinal cord injury. These two stimuli are applied at precise inter-stimulus intervals to reinforce corticospinal synaptic transmission using principles of spike-timing dependent plasticity.

Summary:

We discuss the neural mechanisms and application of this neuromodulation technique and its potential therapeutic effect on recovery of function in humans with chronic spinal cord injury.

Introduction

Spinal cord injury (SCI) results in numerous neurological deficits with devastating consequences for the quality of life. Endogenous regeneration of significantly damaged nervous tissue is limited and fails to contribute to functional recovery. To date, there is no medical treatment to improve recovery of motor function after SCI. The corticospinal tract is a major descending pathway that contributes to the control of voluntary movement [1]. The majority of spinal cord lesions in humans are contusions which typically leave a superficial rim of spared white matter that includes corticospinal axons [2, 3]. A number of individuals with the diagnosis of a clinically complete SCI present evidence of descending connectivity from the brain and show less atrophy of their spinal cords at segments above the level of injury [4]. Therefore, interventions that successfully engage the residual corticospinal tract and strengthen the connections between corticospinal drive and spinal motoneurons are crucial to increase corticospinal transmission to facilitate functional recovery.

Multiple non-invasive strategies have been used to target spinal plasticity in humans with SCI including paired corticospinal-motoneuronal stimulation (PCMS) [5–8], transcutaneous electrical stimulation [9–11], operant conditioning [12], and acute intermittent hypoxia [13]. While most of these interventions use relatively general approaches to change the state of the nervous system, operant conditioning and PCMS target specific physiological spinal connections to induce plasticity. This review will focus on PCMS. Over the last 10 years, a growing number of studies have used PCMS to assess neural plasticity and to explore its efficacy as a neuromodulatory intervention to modulate spinal plasticity in humans with and without SCI, but its effects have not been systemically discussed to date. This review first introduces the neural mechanisms and methods of PCMS as well as earlier electrophysiological studies using PCMS to modulate spinal plasticity in humans. Then, we discuss its efficacy in inducing plasticity of residual corticospinal projections to promote functional recovery in humans with incomplete SCI and its potential as a clinical therapeutic intervention.

Neural Mechanisms of PCMS

Neurostimulation can elicit changes in neuronal circuits by the release of specific signaling molecules to accommodate immediate requirements resulting in changes in synaptic plasticity. Synaptic plasticity can be defined as the ability to make experience-dependent long-lasting changes in the strength of neuronal connections enabling learning and memory [14]. Importantly, neurostimulation can be used for induction of synaptic plasticity. One of the well-known rules in synaptic plasticity is based on Hebb’s postulation [15] that repeated activation of a presynaptic cell immediately before spikes in a postsynaptic cell induces synaptic strengthening, which is referred to as timing-dependent long-term potentiation (LTP). Repeated activation in the reversed spike order, i.e., presynaptic cell immediately after a postsynaptic cell, leads to timing-dependent long-term depression (LTD). Together, these modifications in synaptic strength are known as spike timing-dependent plasticity (STDP) [16]. Evidence from in vitro [17] and human [18, 19] studies support that STDP is under neuromodulatory control involving N-methyl-D-aspartate (NMDA)-type glutamate receptors (NMDAR). Timed arrival of the antidromic postsynaptic activation removes the Mg block which enables Ca²⁺ to diffuse through the NMDAR [20]. The NMDAR thusly senses the postsynaptic depolarization and the presynaptic release of glutamate. The resulting change in Ca²⁺ influx leads to LTP or LTD induction dependent on the intracellular calcium concentration. At the molecular level, LTP and LTD rely upon similar mechanisms. Ca²⁺/calmodulin-dependent protein kinase II (CaMKII), which is a signaling enzyme that controls synaptic plasticity, is regulated by calmodulin. Post-synaptically, Ca²⁺ binds to CaMKII and calmodulin resulting in intracellular processes that lead to increased or decreased expression of AMPAR on the post-synaptic membrane [21]. Because LTP results in a persistent enhancement of synaptic transmission and LTD in a reduction in the efficacy of synaptic transmission, several hypotheses have been proposed to explain how calcium inflow triggers LTP and LTD in different situations. One of them is the amplitude hypothesis, which postulates that calcium amplitude determines the direction of plasticity, with a moderate calcium amplitude needed for LTD and a higher calcium amplitude needed for LTP [22]. The duration hypothesis postulates that the time course of calcium elevation helps to determine the direction of plasticity where a slower calcium signal results in LTD and higher and faster calcium transient results in LTP [23]. The location hypothesis proposes that the specific site of calcium entry determines the direction of plasticity by regulating the calcium binding proteins that are close to the calcium influx [24].

In humans, the first STDP-like plasticity protocol described as paired associative stimulation (PAS) targeted the primary motor cortex [19]. Here, during PAS, transcranial magnetic stimulation (TMS) was used to elicit descending volleys from the primary motor cortex and this was paired with transcutaneous peripheral nerve stimulation to activate Ia afferents to time the converging of inputs at the cortical level. PAS resulted in either facilitation [19] or suppression [25] of the size of motor evoked potentials (MEPs) depending on the specific timing of stimuli. Similarly, other studies used motor point stimulation paired with TMS to deliver PAS and reported plastic changes in motor cortical excitability [26, 27]. Notably, the ability to induce PAS aftereffects seems to have a gradient with more prominent effects observed in flexors compared with extensor arm muscles and more effective on hand muscles compared with forearm muscles [26].

More recently, PAS protocols have been applied to target the corticospinal-motoneuronal connections to induce plasticity at the spinal cord level and these studies usually refer to the PAS protocol as PMCS [5, 28–31]. During PCMS, volleys elicited by TMS over the primary motor cortex are precisely timed to reach the corticospinal-motoneuronal synapses in the spinal cord with antidromic volleys elicited in peripheral motor axons by electrical stimulation to either increase or decrease the efficacy of corticospinal-motoneuronal synapses. STDP-like changes in the spinal cord by PCMS were first demonstrated in control, non-injured, subjects by Taylor and Martin [31] who showed that when repeated corticospinal activity triggered by TMS over a region of the primary motor cortex controlling the biceps brachii was timed to arrive a few milliseconds before antidromic activation of the motoneurons caused by supramaximal peripheral nerve electrical stimulation, the size of cervicomedullary MEPs (CMEPs) increased. Whereas, when postsynaptic depolarization preceded the presynaptic activation, the size of CMEPs decreased. The precise timing of the two stimuli based on measurements of central and peripheral motor conduction time is the key to potentiate or depress corticospinal transmission in PCMS. Conduction times can be estimated in each individual from the latency of a MEP, the latency of a response elicited by cervical root stimulation, and the latency of the M-wave and F-wave. Taylor and Martin [32] also demonstrated in control subjects that the facilitation effect of PCMS is more reliable and longer lasting with more repetition of stimulation.

Bunday & Perez applied PCMS for the first time in humans with incomplete cervical SCI and demonstrated that STDP-like effects on residual corticospinal-motoneuronal synapses can be achieved similarly as in control subjects [5]. It was shown that arrival of presynaptic volleys prior to motoneuronal discharge enhances corticospinal transmission which was accompanied by improved hand function as measured by performance time on the 9-Hole Peg Test. The reverse order of volley arrival did not affect or decreased physiological outcomes without changes in voluntary motor output. Importantly, the changes in corticospinal transmission positively correlated with enhancements in voluntary motor output in both injured and control subjects, suggesting an association between motor output and strength of the induced plasticity. The effect of PCMS protocols on physiological and behavioral outcomes occurred after 100 pairs of stimuli at 0.1 Hz and lasted for up to 80 min. In a subsequent study, PCMS was used to target spinal synapses of lower-limb motoneurons in humans with chronic incomplete SCI [8]. It was found that the size of MEPs in the tibialis anterior muscle increased when TMS-induced presynaptic volleys elicited by stimulation of the leg representation of the primary motor cortex arrived before antidromic volleys elicited by electrical stimulation of the common peroneal nerve at corticomotoneuronal synapses (PCMS+). In contrast, when antidromic volleys arrived at the spinal cord before the presynaptic volleys, the size of MEPs elicited by TMS and electrical stimulation was suppressed (PCMS−). More specifically, when the common peroneal nerve was stimulated 20 ms before a TMS pulse, resulting in peripheral antidromic volleys arriving at the spinal cord 15 ms before TMS-induced presynaptic volleys, some control subjects showed suppression of tibialis anterior MEP size (responders) while others showed no MEP suppression (non-responders). These findings were consistent with previous results showing suppression of tibialis anterior MEPs following electrical stimulation of homonymous and heteronymous nerves of the knee and/or ankle and the skin (i.e. cutaneous branches of common peroneal nerve) applied 20–40 ms before a TMS pulse applied over leg motor cortex [33–36]. We further noted that a conditioning pulse to the common peroneal nerve in non-responders did cause a reduced tibialis anterior MEP when delivered with longer intervals of 30 and 40 ms prior to the TMS pulse. Similarly, we found a PCMS− protocol delivered with these longer intervals between TMS pulse and peripheral stimulation also resulted in MEP suppression in non-responders. These findings suggest that the effectiveness of PCMS− might be linked to contributions from homonymous vs. heteronymous connections to tibialis anterior motorneurons [37]. As in previous studies using PCMS in upper-limb muscles, physiological changes present after PCMS+ were accompanied by increases in voluntary motor output in tibialis anterior muscle. The effect of PCMS protocols on physiological and behavioral outcomes occurred after 200 pairs of stimuli at 0.1 Hz and lasted for up to 30–60 min. This is in contrast to the upper-limb PCMS protocols where 100 pairs prove sufficient in inducing changes in hand muscle MEPs. This might reflect some of the differences raised by engaging upper and lower limb corticospinal projections which have different locations in the brain and thresholds for activation.

Potentiating the effects of PCMS

Efforts has been made to potentiate the aftereffects of PCMS in humans with SCI. We examined the effect of PCMS on corticospinal excitability when pairs of stimuli were delivered during short-lasting low-intensity isometric voluntary contractions [6]. It was hypothesized that because spinal lesions are associated with reduced corticospinal inputs to motoneurons, increasing the number and size of descending volleys by voluntary contraction could potentially boost the spinal plasticity after SCI. This was supported by results showing that MEPs elicited by TMS and electrical stimulation at cervicomedullary junction increased to a larger extent when PCMS was applied during voluntary activity compared with rest in SCI participants. In addition, SCI participants who did not respond to PCMS at rest responded to voluntary activity and those participants who responded to both protocols showed larger increments in corticospinal transmission when PCMS was applied during voluntary activity. This is likely because PCMS predominantly affects low-threshold motoneurons [38] and the threshold of motoneurons decreased with the small levels of voluntary contractions used in this study. In contrast, the size of MEPs elicited by TMS and electrical stimulation increased to a similar extent when PCMS was applied at rest and during voluntary activity in control participants. A possibility is that this plasticity was saturated or reached ceiling effects in control participants under this protocol. This is consistent with new evidence showing that the excitability of motoneurons increases to a larger extent in control compared with SCI participants during small levels of tonic voluntary activity [39].

We then tested the clinical potential of PCMS when combined with exercise training for repeated sessions [7]. Most rehabilitation strategies in individuals with SCI rely on the use of exercise [40]. Exercise training aims to drive neural networks in an activity-dependent manner to elicit plasticity and facilitate functionally relevant muscle activity. Studies in animals [41] and humans [42] have demonstrated that physiological and functional effects of exercise can be augmented using neural stimulation, which is thought to increase the likelihood of activating spared neural pathways. Indeed, a recent study showed that even thousands of repetitions of voluntary contractions of very weak muscles for several weeks had no effect or a very small effect on voluntary strength in individuals with SCI [43]. Jo and Perez studied PCMS and exercise by enrolling individuals with chronic incomplete SCI who then underwent 10 sessions of PCMS combined with exercise training, PCMS without exercise, or sham-PCMS combined with exercise. The results showed that corticospinal drive and maximal voluntary contraction (MVC) in targeted muscles increased after PCMS with or without exercise but not after sham-PCMS with exercise, suggesting that exercise alone was not enough to elicit physiological changes. Although previous studies showed that a single session of PCMS facilitates voluntary output after SCI [5, 8], it did not increase MVC [44] and showed only a small increase (~7%) in EMG during maximal thumb adduction without changes in force. However, the results of this trial suggest that PCMS effects are potentiated during repeated sessions. Another noteworthy result was that both physiological and behavioral effects of 10 sessions were preserved at least up to 6-months in the group receiving PCMS with exercise but not sham-PCMS (Fig. 1). Thus, the effect of PCMS can be long-lasting with repeated sessions combined with exercise versus after a single session where effects are relatively transient. Therefore, targeting spinal synapses using PCMS could represent an effective strategy to facilitate exercise-mediated recovery in humans with different levels of SCI.

Figure 1. Long lasting effects of PMCS combined with exercise on Maximal voluntary contractions and functional outcomes.

(A) Raw MEP traces from four representative participants from biceps brachii and abductor pollicis brevis muscles before and after 10 sessions of PCMS and sham-PCMS combined with exercise. Graphs show subgroup data for PCMS with exercise (green bars; n=5) and sham-PCMS with exercise (orange bars; n=5) groups with 6-month follow-up assessments. The abscissa of graph shows the time of assessments (pre-assessment=PRE, post-assessment=POST, 6-month follow-up assessment=6M) and the ordinate shows the amplitude of MEP as % of MEP at pre-assessment. In this subgroup of participants that received PCMS+exercise, MEPs increased after 10 sessions and remained increased for 6 months compared with baseline but not in sham-PCMS+exercise. (B) Rectified electromyographic traces during MVCs from four representative participants from biceps brachii and first dorsal interosseous muscles before and after 10 sessions of PCMS and sham-PCMS combined with exercise. Graphs show subgroup data for PCMS with exercise (green bars; n=5) and sham-PCMS with exercise (orange bars; n=5) groups with 6-month follow-up assessments. The abscissa of graph shows the time of assessments (pre-assessment=PRE, post-assessment=POST, 6-month follow-up assessment=6M) and the ordinate shows the size of MVC as % of MVC at pre-assessment. In this subgroup of participants that received PCMS+exercise, MVCs increased after 10 sessions and remained increased for 6 months compared with baseline but not in sham-PCMS+exercise. (C) Functional measurements involved subcomponents of the Graded and Redefined Assessment of Strength, Sensibility and Prehension (GRASSP) and the 10-meter walk test. Graphs show subgroup data for PCMS with exercise (green bars) and sham-PCMS with exercise (orange bars) groups with 6-month follow-up assessments. The abscissa shows the time of assessments (pre-assessment=PRE, post-assessment=POST, 6-month follow-up assessment=6M) and the ordinate shows the time to perform tasks as % of time at pre-assessment. Functional outcomes increased after 10 sessions of PCMS+exercise and remained increased for 6 months compared with baseline while the increase present after 10 sessions of sham-PCMS+exercise did not persist 6 months later. Error bars, SDs, *p<0.05. Modified from Jo and Perez, 2020.

Cortical excitability is further augmented when PCMS is administered in combination with acute intermittent hypoxia (AIH) in both control subjects [45] and humans with SCI [46]. AIH is another non-invasive neuromodulatory technique used to facilitate the recovery of motor function in humans with SCI [13]. AIH-induced phrenic motor facilitation requires NMDA receptor activation to maintain the plasticity [47, 48] similar to STDP-like plasticity elicited by PCMS, which also depends on NMDA receptor activation [18]. Evidence for the combined use of AIH and PCMS is presently limited and further studies are needed to better understand the precise mechanistic underpinnings of the interaction between these two modalities.

Future directions

One of the most promising features of PCMS is its cumulative effects of several sessions. The next logical step would be to apply PCMS for longer time-periods and examine if the effects reach a plateau, which would help determine its optimal dose for clinical application. Another potential strategy to boost its effect is to target multiple spinal levels simultaneously with PCMS. Although its application has been restricted to one specific muscle at a time in previous studies, PCMS could target multiple muscles if each antidromic volley from different peripheral nerves is precisely timed with descending TMS volleys. The possibility that the effect of PCMS could be augmented when it is applied at multiple sites simultaneously is a clinically important question because most muscles below the level of injury after SCI are affected and functional improvements (e.g. gait) are most likely to be accompanied by improved coordinated action of multiple muscle groups.

Several practical aspects need to be considered for future translation of PCMS protocols to a clinical environment. For example, it will be important to assess the effectiveness of PCMS using lower intensities to make it more tolerable for all subjects because current PCMS protocols in humans with SCI use 100% of maximal stimulator output (MSO) of TMS. In control subjects, submaximal stimulation intensities (~70% of MSO) are successful in inducing both physiological and behavioral plasticity, suggesting that lower intensities might have beneficial effects in people with SCI. In addition, the effect of frequency needs to be explored. Current studies use PCMS at 0.1 Hz, but if the same effect could be acquired when the same number of pulses is applied at higher frequencies the duration of the protocol could be reduced making it more practical for use in clinical settings. Because the effectiveness of PCMS depends on the precise timing between TMS and peripheral nerve electrical stimulation in individual subjects to induce STDP, training and careful methodological considerations can represent the next step for bringing this strategy to the clinic. One alternative way to implement this protocol might be that TMS is replaced by natural activation. It would be also beneficial for developing more effective PCMS protocols for application in the clinic, if this technique could be reliably and efficiently applied in an animal model of SCI. Such an approach would greatly facilitate investigations of the role of stimulation parameters, such as frequency and duration, on the effect of PCMS on functional outcome. Moreover, animal models would also be beneficial to investigate pharmacological approaches, such as NMDA receptor agonists, for their effect on PCMS-based strengthening of corticospinal-motoneuronal synapses and functional improvements.

Overall, approaches to apply PCMS targeting spinal plasticity represent a new promising tool to promote functional recovery after SCI but further investigations into underlying mechanisms, optimal dose, size and duration of effects along with applicability out of laboratory settings are yet to be adequately addressed.