Cerebral palsy (CP) is the most common childhood motor disorder affecting 2–4 children in every 1,000 births (1–3). The affected children present with a wide range of functional disorders including inability to move voluntarily, maintain balance and posture, spasticity and abnormal sensation during early development that often worsen with age (4, 5). The primary standard of care (SoC) is physical therapy (PT) (6), potentially with subsequent medication and/or surgery to manage pain and reduce spasticity (7). For children with significant spasticity, SoC often includes selective dorsal root rhizotomy (SDR) (8) and intramuscular injections of OnabotulinumtoxinA (9). While these treatments reduce spasticity, they are invasive, may diminish muscle function, and have minimal effect on voluntary sensorimotor function. For instance, 3 months of standard PT resulted in 5.7 points increase in GMFM88 (10); and intramuscular OnabotulinumtoxinA injections result in 1.7–2.2 points increase in GMFM88 after 1–2 months (7). However, CP children that underwent SDR surgery showed a 6.5 points increase in GMFM66 at 4 months (8). More importantly, the GMFM66 decreased by 20 points 17 years post-surgery (11).

Over the last decade, we and others have extensively shown the therapeutic promise of noninvasive spinal cord neuromodulation in spinal cord injury (12–18). We have previously demonstrated the acute (19) and chronic effects (20) of spinal cord neuromodulation on improvements in sensorimotor function in children with CP. However, the effect of activity-based neurorehabilitation therapy (ABNT) alone compared to spinal neuromodulation with ABNT remains unknown. We hypothesized that children with CP who undergo SCiP™ therapy with ABNT will show greater levels of sensorimotor function improvement as assessed by GMFM88 score, compared to children with CP who undergo inactive sham neuromodulation with ABNT. To test this hypothesis, we performed a single blinded, sham-controlled, one-sided crossover study to investigate the impact of noninvasive spinal neuromodulation with ABNT to improve sensorimotor function in children with CP.

Six participants diagnosed with CP (GMFCS level I (n = 1), level II (n = 1), level III (n = 1) and level V (n = 3), aged 20 months–8 years) were enrolled in the study (Table 1). The participants demographics and baseline characteristics are described in Table 1. Participants were randomly assigned to either treatment or sham group (n = 3 each). Sham group received 8 weeks of ABNT with sham therapy (2 mA for 1 min followed by 0 mA for 60 min) whereas the treatment group received 8 weeks of ABNT with therapeutic SCiP™ therapy delivered using our proprietary SCiP™ device (SpineX Inc., Los Angeles, CA) (20). The spinal neuromodulation consists of delayed biphasic waveform formed with a carrier pulse (10 KHz) with a 1 µs delay between the two phases (positive and negative). The delayed biphasic carrier (10 KHz) was combined with a low frequency (30 Hz) burst with a pulse width of 1 ms. Neuromodulation was applied using two adhesive electrodes placed between C5–6 and T11–12 vertebral levels serving as the cathodes (1.25″ in diameter), and two adhesive electrodes over bilateral iliac crests as anodes (3 × 5″). A visible motor contraction of any muscle or any involuntary movement induced by the stimulation, identified by the therapist was used to determine thresholds for the two sites (C5–6: 18–22 mA, and T11–12: 16–20 mA). The neuromodulation intensity was initially set at 20% below the threshold for each site. The intensities over the C5–6 spine ranged between 12 and 18 mA and over the T11–12 ranged between 10 and 16 mA depending on the activity being performed by the participant. During activities involving sitting, rolling, etc., the therapist lowered the amplitudes by 2–4 mA prior to initiation of the activity. Whereas, during standing and stepping, the therapist increased the intensities by 1–2 mA prior to initiation of the activity. During the course of a given activity, the intensities would be modulated ±2 mA based on observed functional performance of the child.

Details of ABNT sessions are provided in Table 1. The treatment was administered by a trained pediatric physical therapist. The participants (and parents) were blinded to the randomization group. At the end of 8 weeks, the sham group crossed into the therapeutic group and received 8 weeks of SCiP therapy with ABNT. The treatment group continued SCiP™ therapy with ABNT for another 8 weeks (i.e., total 16 weeks; Figure 1A). Voluntary sensorimotor function was measured as the primary outcome using Gross Motor Function Measure 88 (GMFM88) and muscle tone (spasticity) was measured using the Modified Ashworth Scale as the secondary outcome (22), at baseline, 8 weeks, and 16 weeks (10). Ten weeks after the last therapeutic session, three participants were reassessed for the primary outcome. Primary end point assessment was based on improvement in GMFM88 scores at 8 weeks compared to baseline.

Eight weeks of SCiP™ therapy resulted in an increase (mean ± SD) in GMFM88 scores by 7.6 ± 2.08 points (minimal clinically important difference; MCID = 5 points) (23), compared to a 2.2 ± 1.38 points increase in the sham group (Figures 1B,C). Further, when the participants from sham arm crossed over and received 8 weeks of therapeutic SCiP™, their GMFM88 increased by 7.13 ± 0.6, equivalent to the treatment group. The participants originally randomized to treatment group continued SCiP™ therapy for another 8 weeks and achieved the ΔGMFM88 score of 9.4 ± 1.5 at week 16 compared to baseline. Interestingly, three participants (1 from treatment group and 2 from sham group) that were reassessed at 26 weeks (i.e., 10 weeks after last SCiP™ therapy session and no further intervention) showed a ΔGMFM88 score of 10.8 ± 6.3 compared to baseline, suggesting a sustained effect of SCiP™ therapy with ABNT. All participants receiving SCiP™ therapy qualified as responders at the primary efficacy endpoint (i.e., ΔGMFM88 > 5 points at 8 weeks), and showed an accelerated functional improvement, compared to the predicted GMFM88 model curve matched for age and GMFCS level (Figures 1D,E) (21). Qualitative observations by the physical therapist and parents suggested meaningful functional improvements in response to SCiP™ therapy, during and post treatment. Table 1 describes the qualitative results for each participant, along with notable feedback from parents. Eight weeks of SCiP™ therapy with ABNT reduced spasticity compared to the sham therapy with ABNT group (ΔMAS −0.06 ± 0.1 SCiP™ vs. +0.02 ± 0.09 sham). Continuation of SCiP™ therapy with ABNT for additional 8 weeks further reduced spasticity score (ΔMAS −0.1 ± 0.1). None of the participants demonstrated an increase in spasticity in response to SCiP™ (Figures 1F,G). No adverse events were reported during the course of SCiP™ therapy with ABNT.

To our knowledge SCiP™ therapy with ABNT is the first intervention to show a significant clinical improvement in sensorimotor function in children with CP within a short period of 8 weeks and be able to sustain the improvement for an extended period of time (10 weeks). Our preliminary findings demonstrate greater improvement in sensorimotor function relative to the available standard of care treatment options, reduced spasticity and increased participation in activities of daily living with SCiP™ therapy with ABNT. Although the exact mechanistic understanding of the proposed combination therapy of SCiP™ with ABNT is incomplete, insights can be gained from studies with spinal cord injury and other forms of paralysis. We hypothesize that spinal neuromodulation (SCiP™) transforms the targeted spinal-supraspinal neural networks into an activated state of plasticity, which are made functionally more competent using activity dependent guidance (ABNT), obtained from proprioception (24). The two key findings of this study are (a) the recovery in voluntary motor function even in the absence of active spinal neuromodulation, and (b) the persistence of improved function during the follow up period. While the present study did not directly test the evidence for putative neural plasticity, it has been previously documented in studies investigating neuromodulation-mediated recovery in the spinal cord injury population (25–28). However, since CP and spinal cord injury have distinct pathophysiologies, the mechanism of action responsible for neuromodulation driven changes in sensorimotor function remains unknown. Despite the lack of mechanistic evidence, our initial findings suggest that noninvasive neuromodulation (i.e., SCiP™ therapy) can be a viable option to improve sensorimotor function in CP, and warrants a comprehensive investigation using a randomized control trial with a larger sample size.