Transcutaneous spinal cord stimulation neuromodulates pre- and postsynaptic inhibition in the control of spinal spasticity

Abstract

Aside from enabling voluntary control over paralyzed muscles, a key effect of spinal cord stimulation is the alleviation of spasticity. Dysfunction of spinal inhibitory circuits is considered a major cause of spasticity. These circuits are contacted by Ia muscle spindle afferents, which are also the primary targets of transcutaneous lumbar spinal cord stimulation (TSCS). We hypothesize that TSCS controls spasticity by transiently strengthening spinal inhibitory circuit function through their Ia-mediated activation. We show that 30 min of antispasticity TSCS improves activity in post- and presynaptic inhibitory circuits beyond the intervention in ten individuals with traumatic spinal cord injury to normative levels established in 20 neurologically intact individuals. These changes in circuit function correlate with improvements in muscle hypertonia, spasms, and clonus. Our study opens the black box of the carryover effects of antispasticity TSCS and underpins a causal role of deficient post- and presynaptic inhibitory circuits in spinal spasticity.

Keywords: human; low-frequency depression; neuroplasticity; non-invasive; postsynaptic inhibition; presynaptic inhibition; spasticity; spinal cord circuits; spinal cord injury; spinal cord stimulation.

Graphical abstract

Figure 1

Study protocol The protocol included the electrophysiological assessment of the maximum soleus (SOL)-H reflex (H_max) to maximum M wave (M_max) ratio (H/M), postsynaptic reciprocal Ia inhibition (RI), presynaptic D1 inhibition (D1), and heteronymous Ia facilitation (IaF) as well as low-frequency depression (LFD) of the SOL-H reflex in ten individuals with spinal cord injury (SCI). These electrophysiological assessments were supplemented by electromyography (EMG)-based assessments of spinal spasticity. All assessments were performed before (evaluation E0) and twice after (evaluations E1, E2) a 30-min session of antispasticity transcutaneous spinal cord stimulation (TSCS), applied at 50 Hz and at an intensity corresponding to 90% of the posterior root-muscle reflex threshold (PRMR thr.). Data in individuals with SCI were collected on 2 study days. Normative electrophysiological data were collected in 20 neurologically intact individuals. The research objective was to investigate whether antispasticity TSCS would transiently improve the electrophysiological measures of spinal inhibitory function in individuals with SCI compared to baseline. We investigated whether changes in these measures would correlate with changes in the EMG-based measures of spinal spasticity. Additionally, the relationship of the data derived in the SCI group to normative data was studied. Post-/presyn. inh., post- and presynaptic inhibition. See also Table S1 and Figures S1–S3.

Figure 2

Antispasticity transcutaneous spinal cord stimulation transiently improved post- and presynaptic inhibition in individuals with spinal cord injury (A) (i) Schematic drawing of the disynaptic spinal circuit underlying postsynaptic reciprocal Ia inhibition. For its assessment, the soleus (SOL)-H reflex was elicited by stimulation of the tibial nerve (tn) following a conditioning stimulus applied to the deep branch of the common peroneal nerve (dpn) at conditioning-test intervals (CTIs) of 1–5 ms. (ii) Scatterplots show individual levels of maximum postsynaptic reciprocal Ia inhibition in E0 and E1. Compared to baseline, inhibition was significantly improved in E1. (iii) Maximum baseline inhibition predicted the improvements in E1. In the inserted regression equation, y denotes the absolute changes observed in E1 vs. E0 and x is the maximum inhibition in E0. (B) (i) Spinal circuit underlying induced presynaptic D1 inhibition. For its assessment, the SOL-H reflex was conditioned by dpn stimulation at CTIs of 10–30 ms. (ii) Individual levels of maximum presynaptic D1 inhibition in E0 and E1. Statistically, presynaptic D1 inhibition did not change in E1 compared to baseline. (iii) Improvements in postsynaptic reciprocal Ia inhibition in E1 predicted improvements in presynaptic D1 inhibition in the same evaluation. In the inserted regression equation, y denotes the absolute change in presynaptic D1 inhibition observed in E1 vs. E0 and x is the respective change in postsynaptic reciprocal Ia inhibition. (C) (i) Spinal circuit underlying heteronomous Ia facilitation under ongoing presynaptic inhibition, assessed by applying a conditioning stimulation to the femoral nerve (fm) at CTIs of −9.0 to −5.6 ms. (ii) Individual levels of heteronymous Ia facilitation in E0 and E1. Compared to baseline, facilitation was significantly reduced in E1, reflecting increased background presynaptic inhibition. (iii) Relationship between TSCS-induced changes in presynaptic D1 inhibition and heteronymous Ia facilitation in E1 compared to E0. Both measures of presynaptic inhibition were concomitantly improved over baseline in eight of the participants. E0, pre-TSCS evaluation; E1, first post-TSCS evaluation; EMG, electromyographic; TSCS, transcutaneous spinal cord stimulation; ∗p < 0.05; ∗∗p < 0.001. See also Figures S4 and S6.

Figure 3

Between-groups comparisons show transient improvement in post- and presynaptic inhibition to normative levels after antispasticity transcutaneous spinal cord stimulation Scatterplots show individual levels of postsynaptic reciprocal Ia inhibition (RI), presynaptic D1 inhibition (D1), and heteronymous Ia facilitation (IaF) of the soleus-H reflex in the spinal cord injury (SCI) group in evaluations E0, E1, and E2 as well as in the neurologically intact group. Post- and presynaptic inhibition were weaker in the SCI than the neurologically intact group in E0 and E2, but not in E1. Post hoc Bonferroni-corrected pairwise comparisons showed lower levels of RI, D1, and IaF in E0 and of RI and D1 in E2 compared to normative levels. E0, pre-TSCS evaluation; E1, E2, post-TSCS evaluations; TSCS, transcutaneous spinal cord stimulation; ∗p < 0.05; ∗∗p < 0.001. See also Figure S4.

Figure 4

Antispasticity transcutaneous spinal cord stimulation did not modulate low-frequency depression (A) Top: schematic drawing of the repetitively simulated monosynaptic reflex circuit of soleus (SOL). Bottom: exemplary electromyographic recordings of SOL-H reflexes (i) from participant 8 with spinal cord injury (SCI) in evaluations E0, E1, and E2, and (ii) from a neurologically intact participant. Each line is the average of the 11^th–30^th H reflexes elicited at repetition rates as indicated. Insets are mean peak-to-peak (P2P) amplitudes per repetition rate normalized to the H reflexes at 0.1 Hz. (B) (i) Low-frequency depression curves of the H reflex in the SCI group in E0, E1, and E2 (diamonds) and the neurologically intact group (circles). Error bars indicate SE. (ii) Low-frequency depression did not differ between E0 vs. E1 and E0 vs. E2. Low-frequency depression differed significantly between the neurologically intact (int) and the SCI groups in each of the three evaluations. E0, pre-TSCS evaluation; E1, E2, post-TSCS evaluations; n.s., not significant; TSCS, transcutaneous spinal cord stimulation; ∗∗p < 0.001.

Figure 5

Transcutaneous spinal cord stimulation-induced improvements in spasticity measures correlate with improvements in post- and presynaptic inhibition (A) EMG-based measures of spinal spasticity were acquired from rectus femoris (RF), biceps femoris (BF), tibialis anterior (TA), and soleus (SOL) while (i) tonic stretch reflexes, (ii) cutaneous-input-evoked spasms, and (iii) Achilles clonus were evoked by an examiner. Data were collected twice, on 2 study days. (B) Exemplary recordings show the reduction of the different manifestations of spasticity following TSCS in E1 and E2. Arrowheads indicate onsets of manipulations by the examiner. (C) (i) Turquoise bars illustrate the observation frequency of improvements over baseline for each spasticity measure as indicated, shown separately for E1 and E2, across subjects and study days. (ii) Scatterplots show individual EMG-root-mean-square (RMS) values across muscles of the manipulated lower limb associated with the three spasticity measures as well as Achilles clonus durations. All measures were significantly reduced compared to E0 in both post-TSCS evaluations. Turquoise brackets and asterisks signify significant post hoc Bonferroni-corrected pairwise comparisons between spasticity measures in E0 and E1, black brackets and asterisks between E0 and E2. (D) Scatterplots show significant correlations in E1 between (i) a relative increase in postsynaptic reciprocal Ia inhibition and improvements in cutaneous-input-evoked spasms; a decrease in heteronymous Ia facilitation and (ii) tonic stretch reflexes as well as (iii) Achilles clonus. In the inserted regression equations, y denotes the relative change in the EMG-based measure of spasticity as indicated and x is the relative change in the respective electrophysiological measure. E0, pre-TSCS evaluation; E1, E2, post-TSCS evaluations; EMG, electromyography; TSCS, transcutaneous spinal cord stimulation; ∗p < 0.05; ∗∗p < 0.001. See also Figures S5.