Article https://doi.org/10.1038/s41467-022-33208-w A pilot study combining noninvasive spinal neuromodulation and activity-based neurorehabilitation therapy in children with cerebral palsy Received: 3 November 2021 Susan Hastings1, Hui Zhong2, Rochel Feinstein3, Gittel Zelczer3, Christel Mitrovich 4, Parag Gad 2,5 & V. Reggie Edgerton 2,5,6,7 Accepted: 8 September 2022 1234567890():,; 1234567890():,; Cerebral Palsy (CP) is the most common pediatric motor disability with mul- Check for updates tiple symptoms and etiologies. CP is exhibited through sensorimotor delays, impaired posture resulting in limited activities and participation. Our recently concluded, single arm, unblinded, pilot study (NCT04882592) explored whe- ther an intervention combining non-invasive spinal neuromodulation during an activity-based neurorehabilitation therapy (ABNT) can improve voluntary sensory-motor function captured via the Gross Motor Function Measure (GMFM-88) scores (primary outcome). Sixteen children diagnosed with CP with Gross Motor Function Classification Scale levels I-V were recruited and received the same intervention (2x/week for 8 weeks) to correct the dys- functional connectivity between supraspinal and spinal networks using the normally developed proprioception. We demonstrate that the intervention was associated with clinically and statistically significant improvement in GMFM-88 scores in all children, thus meeting the prespecified primary end- point. However, the improvement with ABNT alone needs further exploration. No serious adverse events were observed (safety endpoint). Cerebral palsy (CP) is an umbrella diagnosis that is the most common connections among the normally developed spinal networks, thus motor disability in childhood and impacts 1.5 to 4 of every 1000 generating abnormal sensory-motor responses5. Without interven- births1–3, leading to over 10,000 new cases diagnosed each year. More tions, these functionally abnormal synaptic connections are reinforced than 500,000 children under the age of 18 currently have CP in USA, further during early childhood, further contributing to the neuro- with over 17 million cases worldwide. The more common pathology muscular disorders associated with CP, such as neuromuscular spas- that leads to CP is among supraspinal networks and could be due to ticity, balance deficits and poor coordination among motor pools. multiple etiologies. In some cases, these supraspinal pathologies also Children with CP with relatively few functional limitations have been manifest as motor neuronal dysfunctions, perhaps due to aberrant reported to be substantially weaker than typically developing peers6. supraspinal-spinal connections formed early in life4. If functionally Presently, the predominant goals of current CP treatment options are abnormal connections persist significantly beyond the early develop- to manage symptoms, relieve pain, and maximize independence to mental phase, the supraspinal dysfunction can impose dysfunctional achieve a long life, despite having irreversibly debilitating conditions. 1 Susan Hastings Pediatric Physical Therapy, San Jose, CA 95128, USA. 2Rancho Research Institute, Downey, CA 90242, USA. 3OPTimal care Therapy, Lakewood, NJ 08701, USA. 4Reneu Health, San Diego, CA 92111, USA. 5SpineX Inc., Los Angeles, CA 90064, USA. 6USC Neurorestoration Center, University of Southern California, Los Angeles, CA 90033, USA. 7Institut Guttmann. Hospital de Neurorehabilitació, Institut Universitari adscrit a la Universitat Autònoma de Barcelona, Barcelona, 08916 Badalona, Spain. e-mail: parag@spinex.co Nature Communications | (2022)13:5660 1 Article https://doi.org/10.1038/s41467-022-33208-w a modality called SCiPTM (SpineX Inc., Los Angeles, CA) (Fig. 1). Pre- viously, we have demonstrated that noninvasive spinal neuromodula- tion can lead to recovery of lower extremity13–15, upper extremity16,17, and trunk function18, as well as bladder and bowel19–23 and breathing24 function after SCI. However, the use of noninvasive spinal neuromo- dulation during ABNT in CP has not been studied. Here we have used a combination of an activity-based neurorehabilitation therapy (ABNT) modality consisting child selected, play-based activities while suc- cessfully leveraging optimal external assistance, weight shifts and alignment of the body’s center of mass to accommodate gravitational vectors as needed to sustain equilibrium. These training methods are in line with the recent guidelines and recommendations from the Cerebral Palsy Alliance (CPA) and addresses all the necessary compo- nents defined by the World Health Organization’s (WHO) International Classification of Function framework for CP wherein, treatment is focused on activity, participation, environmental factors that can help the child’s function, in addition to personal factors, such as allowing the child to participate in an activity of interest. In the current study, the children were treated with ABNT appropriate to their age level, while participating in a self-chosen activity of interest to the child, b Baseline Assessment Final Assessment allowing the child to play while interacting with family and friends25. The primary objective of our study was to determine whether the Recruitment Training SCiP + ABNT combined effect of noninvasive spinal neuromodulation15,18,26 and our ABNT paradigm is associated with increased GMFM-88 scores. 0 2 4 6 8 Weeks Results Fig. 1 | Overall Experimental Design. a Experimental setup demonstrating the All children demonstrated GMFM-88 scores consistent for their age placement of non-invasive electrodes over the cervical and thoracic regions of the and GMFCS level at the start of the study period with scores ranging spinal cord while the child is actively engaged in activity based neurorehabilitation from 68 to 85 for levels I and II and 2.9 to 29 for levels IV and V therapies, b Experimental design for each child including initial screening and (historically reported scores: level. I, 76.29 ± 17.14, level II, 59.51 ± 13.49, recruitment period, 16 training sessions (red arrows) over the 8 weeks and baseline level IV 30.21 ± 10.79, and level 8.33 ± 5.1627). Positive responses were and final assessments. observed in all children during the first ABNT session with noninvasive spinal neuromodulation including improvement in posture and sen- sorimotor capabilities consistent with previous results26. For indivi- Other methods to treat spasticity associated with CP include intra- duals with levels, I and II, improvements included the ability to take muscular Botulinum ToxinA injections, medications such as Baclofen, more independent steps with their center of mass (CoM) sufficient to surgical interventions such Selective Dorsal Rhizotomies (SDR) and sustain equilibrium, upright posture and improved balance. Children orthopedic surgeries7. While these methods may provide a short-term with level IV and V demonstrated more subtle improvements including reduction in spasticity, in the long term, they may negatively impact better voluntary head control, ability to sit for longer durations with- functional changes, especially in a growing child. out external support and lower assistance needed during stepping. Standard of Care (SoC) treatment for CP often include one or Notably, all children became more interested in the chosen activity and more of the following: 1) an activity-based neurorehabilitation therapy often responded positively via facial expressions when successfully (ABNT) play-based approach, appropriate for age, with therapist interacting with cause-and-effect toys. This was observed in almost all facilitating more normal movement patterns with the goal of the children at Levels IV and V when previously, they did not seem strengthening selected groups of muscles; 2) direct orthopedic interested in interacting with any toys. strengthening exercises, often used in combination with stretching to Over the course of 8 weeks of training every one of these children maintain range of motion, and 3) increasing activity participation by demonstrated clinically significant improvement in their GMFM-88 emphasizing on movement for sustaining metabolic and cardiovas- scores (minimal clinical important difference, MCID = 5 points28). cular fitness and maintaining the ability of the child to move and Overall, the change in GMFM-88 score (primary outcome) was also function to his/her best ability, and become more independent statistically significant (P < 0.05) (Fig. 2a). Children initially diagnosed through utilizing devices to help them sit, stand, or walk. Historically, as GMFCS levels I and II improved from 70.4 ± 4.13 to 85.28 ± 2.7 the change in Gross Motor Function Measure (GMFM-88), a gold (P < 0.05), these improvements are considered to be clinically sig- standard clinical instrument to measure voluntary sensorimotor nificant with all I and II children considered to be responders. In function scores, observed for currently used interventions ranges from addition, children initially diagnosed as GMFCS levels IV and V showed 1.3 to 13.4 points over the course of 3 weeks to 18 months of clinical improvement in their GMFM-88 scores from 11.09 ± 3.10 to treatment8,9 with potential decreases after 18 to 36 months10. However, 21.92 ± 3.85 (P < 0.05) (Fig. 2). The improvements in GMFM-88 scores most of these interventions are designed to decrease spasticity, while were accompanied by spontaneous emergence of new sensorimotor having little effect in improving GMFM-88 scores and in acquiring an skills, as well as improvements in previously learned skills (supple- expanded range of self-initiated and well-controlled movements. mentary video 1). At the start of therapy, 9 children were non- Based on our pre-clinical and clinical studies after spinal cord ambulatory (needed maximal external assistance to take steps over- injury (SCI), we have identified spinal neuromodulation strategies in ground, GMFCS level IV, n = 3 and level V, n = 6) and needed assistance conjunction with engaging the proprioception generated during ABNT while stepping whereas 7 were able to take some steps (GMFCS level I to be critical components of the overall control of posture and and II, n = 7). At the end of 8 weeks of therapy, only 4 children were locomotion11,12. Over the last few years we have developed and eval- nonambulatory (GMFCS level IV, n = 0 and level V, n = 4), 3 were cap- uated a pediatric version of the noninvasive spinal neuromodulation able to stepping with minimal assistance (GMFCS level IV, n = 2 and Nature Communications | (2022)13:5660 2 Article https://doi.org/10.1038/s41467-022-33208-w All Patients level I & II level IV & V a b 80 100 ⴕ * level I & II a * ⴕ 100 100 30 * 70 80 level IV & V Before 60 GMFM88 80 80 25 After 50 GMFM88 60 PedsQL 20 60 60 40 15 40 40 40 30 10 20 20 20 20 5 10 0 0 0 0 b 0 0 20 40 60 80 100 100 PedsQL After 80 Fig. 3 | Quality of Life Changes. a mean ± SE (n = 12, GMFCS levels I and II, n = 3, GMFM88 Before GMFCS levels IV and V, n = 9) PedsQL scores before and after therapy. All data were 60 tested for normality using the Kolmogorov-Smirnov test. Based on the result of a 40 normal distribution, paired t-tests were used to compare the group mean data before and after therapy. * Significantly different from before therapy at P < 0.05. 20 GMFCS level I & II Note the max score on the PedsQL is 88 points and a decrease in score represents GMFCS level IV&V an improvement in function. b X-Y plot for correlating GMFM-88 scores at start and 0 end of therapy with the PedsQL scores demonstrating that the increase in GMFM- 0 2 4 6 8 10 12 14 16 18 Age 88 scores (improved sensorimotor function of the child) is directly linked to the c GMFCS level I & II 35 decrease in PedsQL scores (improved quality of life of parent). Please note, data for GMFCS level IV & V 30 n = 12 out of 16 are shown since 4 PedsQL data points were not collected. Source All Patients data are provided as a Source Data file. ∆GMFM88 25 20 R² = 0.0003 15 10 studies. In addition, larger randomized clinical trials with appropriate R² = 0.0026 5 R² = 0.0855 control groups including standard of care therapies such as physical MCID therapy, pharmacological therapy to manage spasticity etc., are needed 0 0 2 4 6 8 10 12 14 16 18 to better predict the magnitude of change in each age group and each Age (Years) level of severity. The current study does not include GMFCS level III and Fig. 2 | Primary outcome changes. a mean ± SE (n = 16) GMFM-88 scores before the effectiveness of the interventions for all age groups remains and after therapy; mean ± SE (n = 7) GMFM-88 scores before (black) and after (red) uncertain. In addition, the long-term (over the course of several years) therapy for GMFCS levels I and II (blue) and; mean ± SE (n = 9) GMFM-88 scores impact of spinal neuromodulation and ABNT on activity participation, before (black) and after (red) therapy for GMFCS levels IV and V (orange). joint health, potential deformities and induced pain remains unknown. b GMFM88 scores at the start (black) and end of therapy (red) relative to age of the Based on our hypothesis and experience to date with studying the child at the start of the therapy and c ΔGMFM-88 scores relative to their age at the combined effects of noninvasive spinal neuromodulation during ABNT start of the therapy. All data were tested for normality using the Kolmogorov- in individuals living with SCI11, we hypothesized that it may be possible Smirnov test. Based on the result of a normal distribution, paired t-tests were used to sustain the level of improvement in function for up to 1 to 2 months to compare the group mean data before and after therapy. * Significantly different after the termination of the intervention, although this remains to be from before therapy at P < 0.05. † Meaningful clinically improvement difference tested in this context. It seems likely, however, that one of the factors (MCID) = 5 points. Source data are provided as a Source Data file. that may define this rate of loss of function after the intervention is stopped is dependent on the child sustaining a critical level of activity level V, n = 1) and 10 were able to step independently with no external post the intervention. While the present observations were derived assistance (GMFCS level I and II, n = 7, level IV, n = 1 and level V, n = 1) from a heterogenous population based on levels of severity as well as (primary outcome). In addition, during interviews with the study team, age of the child, the results are encouraging given some improvements all parents reported that after two to four treatment sessions, their from baseline in sensorimotor functions were observed in all children, child voluntarily began to practice the newly learned motor skills at and these changes occurred within a relatively short period of time home, even in the absence of neuromodulation, thus further enhan- compared to other interventions. It is also important to note that chil- cing neuroplasticity and increasing the child’s activity level. Parents dren as young as two years up to 16 years of age and GMFCS levels I to V and caregivers also reported that their children demonstrated responded positively. Certain new skills and motor patterns, however, improvement in many of the symptoms secondary to cerebral palsy may take longer to correct in older and more involved individuals that resulting in significant improvement in their quality of life as measured have acquired certain inefficient motor patterns over a period of years. by the PedsQoL survey (Fig. 3). Multiple attempts have been made to activate the nervous system such as using central and peripheral stimulation and combining it with Discussion robotic gait training leading to improvement in GMFM-88 scores in We demonstrated the effect of noninvasive spinal neuromodulation some of the participants29. Unlike traditional approaches (such as during ABNT over a short period of time (8 weeks) with the improve- functional electrical stimulation, FES), which has been used most often ment in GMFM-88 scores being higher in magnitude (>3x) than the to directly stimulate muscles, we have previously delivered modest minimal clinically important difference (MCID = 5 points28) and every stimulation below motor threshold19,23,26,30 to neuromodulate the child being considered as responders. physiological states of the spinal networks, and perhaps even There are several limitations in this first in human, single-arm, supraspinal networks. It appears that these neuromodulated states of unblinded pilot study. While the key objective of the current study was spinal networks resulted in improved coordinated movements during to determine the combined effect of ABNT and noninvasive spinal stepping. Thus, as hypothesized, it appears that the relatively normal neuromodulation on the primary outcome (GMFM-88), the indepen- proprioceptive input provided a critical source of movement control dent effects of ABNT and spinal neuromodulation alone and the effect of the intended task as shown in adults with SCI31 and hypothesized for of sham neuromodulation needs to be evaluated in future blinded children with CP4. Nature Communications | (2022)13:5660 3 Article https://doi.org/10.1038/s41467-022-33208-w It appears that practicing motor tasks with incorrect alignment Table 1 | Demographics of participants and posture could lead to enhancement of aberrant connections Patient ID Age Gender GMFCS Level CP Diagnosis within and among the brain and spinal neural networks32. Based on our (years) previously demonstrated acute neurophysiological results26, we 1 12-18 M I Mixed hypothesize that the physiological state of predominantly supraspinal 2 7-12 M I Spastic Biplegia aberrant connections can be suppressed with spinal neuromodulation 3 2-7 M II Spastic Biplegia to a level that allows the children to voluntarily initiate the intended movement, more normally, and thus progressively generate more 4 2-7 M II R Spastic Hemiplegia normal patterns of proprioception derived from improved signals 5 2-7 M II Post Hemispherectomy generated by the spinal networks4. We hypothesize that multisite 6 12-18 M II Spastic Quad spinal neuromodulation at the cervical and thoracic levels enables 7 2-7 M II Spastic Biplegia bidirectional communication with supraspinal centers13,33, providing 8 2-7 M IV Athetoid sufficient functional reorganization of supraspinal-spinal connections 9 2-7 M IV Spastic Quad that can facilitate sensorimotor function. In essence, spinal neuro- 10 2-7 F IV Ataxic modulation appears to facilitate neuroplasticity and leads to the for- mation of more normal bidirectional interactions of the Brain-Spinal 11 2-7 F V Infantile Cord-Muscle-Spinal Cord-Brain networks. 12 2-7 M V Spastic Quad Based on numerous studies combining electrical spinal neuro- 13 7-12 F V Dyskinetic Quad modulation with task-specific training4,14,15,17, we hypothesize that 14 2-7 M V Spastic Quad appropriately trained therapists can provide the corrective support by 15 2-7 M V Spastic Quad facilitating kinematic realignment that generates proprioceptive 16 7-12 F V Dyskinetic Quad ensembles that accommodates gravitational vectors necessary for Children’s initial demographics and GMFCS levels. GMFCS Gross Motor Functional sustaining equilibrium. As hypothesized previously4, the present data Classification Scale. are consistent with the concept that the projection of these proprio- ceptive ensembles to spinal networks enables a normalizing of motor learning and network functional neural reorganization that can lead selective dorsal root rhizotomy, 2) intramuscular Botox injections or toward more effective movements. This synergistic transformation of orthopedic surgeries in the preceding 12 months, 3) current antispastic functional connectivity between the spinal networks and the improved medications, 4) unhealed fractures or contractures that would prevent proprioception that the spinal networks receive. When the ABNT them from performing functional tasks and 5) other experimental during noninvasive spinal neuromodulation is self-selected by the therapies that were judged to be conceptually inconsistent with the child, we observed more creative exploration in performing new underlying neurophysiological hypothesis of the present experiment. motor skills (in pediatrics usually play related) rather than requiring a more standard, but less interesting activity for the child. The basic Transcutaneous spinal neuromodulation rationale of this approach is to generate a wider range of patterns of Spinal neuromodulation was delivered in the clinic using a proprietary proprioception from a greater range of movements, from which spinal SCiPTM device (SpineX, Inc). The stimulation waveform consists of two networks can experience and translate to different motor tasks. The- alternating pulses of opposite polarities separated by a 1uS delay oretically, the interaction between the spinal motor output and pro- forming a delayed biphasic waveform. The pulses consisted of a high- prioception can gradually minimize the pathological inputs from the frequency biphasic carrier pulse (10KHz) combined with a low fre- injured brain in controlling posture and locomotion. quency (30 Hz) burst pulse each with a pulse width of 1 ms. Simulta- In conclusion the present results suggest that a combination of neous spinal neuromodulation was applied using two adhesive round spinal neuromodulation and ABNT is associated with improvement electrodes (1.25” dia) located between C5-6 and T11-12 serving as the in self-initiated sensorimotor functions as recorded on the clinically cathodes, and two adhesive rectangular electrodes (3×5”) over bilateral significant improvement in GMFM-88 scores. Further, these improved iliac crests as common anodes for all children (Fig. 1). The children and functions are associated with improvement in quality of life of the parents were blinded from the exact intensity of stimulation for an parents. Indirectly, the present results demonstrate noninvasive elec- unbiased outcome. Based on our previous studies that used electro- trical spinal neuromodulation combined with an ABNT strategy may myography (EMG) as a biomarker to determine motor threshold26, facilitate improved functional, bidirectional connectivity between during training, a sub-motor threshold intensity (20% below lowest spinal and supraspinal networks in individuals with CP. Finally, the lower extremity muscle motor threshold) was used to ensure no motor present data are consistent with the concept that spinal networks that evoked responses are generated (any lower extremity muscle contrac- are normally developed in individuals with CP can be engaged by tions) and to ensure no pain or discomfort is being caused to the child. proprioception to serve as a source of control of movements In this study, the threshold was defined as the amplitude at which the child first attempted to extend the cervical or thoracic regions while in a Methods seated position. Note, since the stimulation electrodes were placed at The study was approved by an external Investigational Review Board C5-6 and T11-12, the stimulation thresholds were significantly lower (Advarra IRB). The study was listed on clinicaltrials.gov compared to our previous study with electrodes over T11-12 and L1-2. (NCT04882592). Sixteen children were recruited for this single arm, Once thresholds were determined (range: C5-6: 18-22 mA, and T11-12: unblinded, non-randomized, prospective study between May 2021 16-20 mA) the stimulation intensity was set 20% below this threshold. (first enrollment) and February 2022 (last exit) (Table 1). All study The intensities over the C5-6 spine ranged between 12 and 18 mA and participants’ parents signed the informed consent form and consented over the T11-12 ranged between 10 to 16 mA based on the activity being to their data to be used for future publications and presentations. The performed. During activities involving sitting, rolling etc. amplitudes participants did not receive compensation to be part of the study. The were further lowered by 1–2 mA, whereas, during standing and step- authors affirm that a parent of the participant provided consent for ping, the intensities were increased by 2–4 mA. During the course of a publication of the video in Supplementary video 1. The inclusion cri- given activity, the intensities would be modulated ±2 mA based on teria included 1) individuals above the age of 2 years of age and 2) observed functional performance of the child. In addition, the children diagnosed with cerebral palsy (CP). The exclusion criteria included 1) often provided feedback regarding the intensity of stimulation and if it Nature Communications | (2022)13:5660 4 Article https://doi.org/10.1038/s41467-022-33208-w was sufficient or not based on the activity being performed. These study. The primary outcomes included 1) GMFM-88 scores and 2) qua- procedures are routinely performed and consistently found to be litative assessment of stepping on the treadmill with minimal support excellent subject-specific input in our lab with patients with spinal cord and assistance, as needed) and qualitative assessment of stepping injury, stroke, multiple sclerosis and cerebral palsy. overground (with minimal support and assistance, as needed). The All children were able to communicate with the parents and GMFM-88 is an 88-item measure assessing gross motor activities across research team if they experienced any pain or discomfort either due to five dimensions: A) lying and rolling, B) sitting, C) crawling and kneeling, the spinal neuromodulation or due to the locomotor procedures. D) standing, and E) walking, running and jumping and is considered the None of the children reported any pain or discomfort during the gold standard in the U.S. for measuring gross motor function change neuromodulation and were blinded from the neuromodulation para- over time in children with CP. Children were designated as responders if meters at any given time during the testing procedures. While the they realized at least a 5-point increase in GMFM-88 score from baseline. children could feel the stimulation pulses initially, they were unable to Stepping assessments addressed three specific criteria; the ability to clearly distinguish between amplitudes or even the presence of sti- take weight-bearing steps independently, the ability to take steps with mulation after a brief period of accommodation. All children tolerated some assistance (assistance for weight bearing and/or assistance to the neuromodulation well and did not report any pain or discomfort. move limbs) and the ability to take steps with full assistance (assistance In children that are non-verbal, facial expressions were observed clo- for weight bearing and assistance to move limbs). In addition, parents/ sely during treatment sessions, and stimulation intensity was con- caregivers completed a quality-of-life satisfaction survey (PedsQL sur- sistently reduced if they became irritable or agitated. No adverse vey for CP,35) before and after therapy. The primary safety endpoint was events were reported during the course of the study. the proportion of children who experience one or more serious adverse events (SAEs). Activity-based neurorehabilitation therapy (ABNT) The children were asked to take off their shoes and braces (if any) before Statistical analysis each therapy session. Each therapy session began with 1) treadmill All data are reported as mean ± SE. All data were tested for normality training to provide maximum proprioceptive input to prime the ner- using the Kolmogorov-Smirnov test. Based on the result of a normal vous system. The treadmill speeds ranged from 0.1 m/s to 0.5 m/s based distribution, paired t-tests were used to compare the group mean data on the child’s capability. Assistance was provided as needed by thera- before and after therapy. All statistical significance is reported at pist(s) including i) moving limbs while stepping with the goal being to P < 0.05. Since this was our first in human pilot study, power calcula- accurate place the heel and toes on the treadmill belt and achieve tions were not completed prior to study initiation. The study is in consistency of gait, ii) appropriate weight-shifting at the pelvis and to compliance with ICMJE guidelines on reporting. advance their lower extremity during swing out in front of the Center of Mass and iii) maintain the head, shoulders, hips and heels in alignment. Reporting summary Next, ABNT included functional activities during play activities and Further information on research design is available in the Nature walking with body support for alignment (as needed) including, 2) Research Reporting Summary linked to this article. upright sitting, with trunk control and weight shifting during reaching, 3) transitions including sit to stand, rolling prone to a quadruped Data availability position, consecutive rolling etc., 4) dynamic standing with postural and Additional video data from training and testing sessions are available. weight shifts, 5) side-stepping with frontal plane weight shifts, 6) Data can be made available on reasonable request by qualified inves- backward stepping, 7) Climbing/creeping up an incline, 8) climbing up tigators. Source data are provided with this paper. and down steps and 9) over ground walking. Children classified as GMFCS levels I and II underwent all the components of the therapy References listed above, whereas those classified as GMFCS levels IV and V focused 1. Arneson, C. L. et al. Prevalence of cerebral palsy: Autism and more on activities 1-6. All children were trained 2 sessions a week for Developmental Disabilities Monitoring Network, three sites, United 8 weeks with each session lasting for 50-60 mins. In case a child missed a States, 2004. Disabil. Health J. 2, 45–48 (2009). session, a make-up session was performed with the next 7 days to 2. Paneth, N., Hong, T. & Korzeniewski, S. The descriptive epidemiol- ensure completion of 16 sessions within 8 weeks. During all therapeutic ogy of cerebral palsy. Clin. Perinatol. 33, 251–267 (2006). activities, it was ensured that appropriate posture, midline orientation 3. Winter, S., Autry, A., Boyle, C. & Yeargin-Allsopp, M. Trends in the and head position was maintained with maximal weight-bearing to prevalence of cerebral palsy in a population-based study. Pediatrics optimize proprioceptive information being processed by the nervous 110, 1220–1225 (2002). system. Since children with GMFCS levels I and II are functional, but with 4. Edgerton, V. R., Hastings, S. & Gad, P. N. Engaging Spinal Networks limited motor patterns available, the objective was to enable self- to Mitigate Supraspinal Dysfunction After CP. Front. Neurosci. 15, generated movement patterns with more typical posture and align- 643463 (2021). ment, whereas lesser functioning children with GMFCS levels IV and V, 5. Curtis, D. J. et al. The central role of trunk control in the gross motor the objective was to allow for new motor patterns to evolve during play function of children with cerebral palsy: a retrospective cross- activities, while retraining existing patterns by using proper posture and sectional study. Developmental Med. Child Neurol. 57, 351–357 (2015). alignment. Toys with lights and sounds were used to engage children to 6. Eken, M. M. et al. Lower limb muscle fatigue during walking in facilitate movements including reaching, grasping and visual tracking children with cerebral palsy. Developmental Med. Child Neurol. 61, etc. All children were asked to discontinue other therapies, except for 212–218 (2019). aqua and hippotherapy, prior to initiation of and during the study34. 7. Chang, C. H., Albarracin, J. P., Lipton, G. E. & Miller, F. Long-term Both aqua and hippo therapies are typical activities for children, and follow-up of surgery for equinovarus foot deformity in children with they do not place excessive motor demands on the child. cerebral palsy. J. Pediatr. Orthopedics 22, 792–799 (2002). 8. Bakaniene, I., Urbonaviciene, G., Janaviciute, K. & Prasauskiene, A. Clinical and functional assessments Effects of the Inerventions method on gross motor function in All assessments were completed within a single session before and after children with spastic cerebral palsy. Neurologia i Neurochirurgia the 8 weeks of training sessions in the absence of spinal neuromodu- Pol. 52, 581–586 (2018). lation and were done by an experienced and board-certified pediatric 9. Lundkvist Josenby, A., Jarnlo, G. B., Gummesson, C. & Nordmark, E. physical therapist. The team members remained unblinded during the Longitudinal construct validity of the GMFM-88 total score and goal Nature Communications | (2022)13:5660 5 Article https://doi.org/10.1038/s41467-022-33208-w total score and the GMFM-66 score in a 5-year follow-up study. 30. Gad, P. et al. Sub-threshold spinal cord stimulation facilitates Phys. Ther. 89, 342–350 (2009). spontaneous motor activity in spinal rats. J. Neuroeng. Rehabilita- 10. Tedroff, K., Löwing, K. & Åström, E. A prospective cohort study tion 10, 108 (2013). investigating gross motor function, pain, and health-related quality 31. Gerasimenko, Y. et al. Feed-Forwardness of Spinal Networks in of life 17 years after selective dorsal rhizotomy in cerebral palsy. Posture and Locomotion. Neuroscientist 23, 441–453 (2017). Dev. Med. Child Neurol. 57, 484–490 (2015). 32. Shuman, B. R., Goudriaan, M., Desloovere, K., Schwartz, M. H. & 11. Sayenko, V. R. E. Y. P. G. P. G. D. G. Basic concepts underlying Steele, K. M. Muscle synergies demonstrate only minimal changes activity- dependent mechanisms in the rehabilitation of sensory- after treatment in cerebral palsy. J. Neuroeng. Rehabilitation 16, motor function after spinal cord injury. Spinal Cord. Med. Chap- 46 (2019). ter 54, 897–911 (2018). 33. Gerasimenko, Y. et al. Initiation and modulation of locomotor cir- 12. Taccola, G., Sayenko, D., Gad, P., Gerasimenko, Y. & Edgerton, V. R. cuitry output with multisite transcutaneous electrical stimulation of And yet it moves: Recovery of volitional control after spinal cord the spinal cord in noninjured humans. J. Neurophysiol. 113, injury. Prog. Neurobiol. 160, 64–81 (2018). 834–842 (2015). 13. Gad, P., Gerasimenko, Y. & Reggie Edgerton, V. Tetraplegia to 34. Kelly, M. & Darrah, J. Aquatic exercise for children with cerebral Overground Stepping Using Non-Invasive Spinal Neuromodulation. palsy. Developmental Med. Child Neurol. 47, 838–842 (2005). IEEE 9th International IEEE/EMBS Conference on Neural Engineer- 35. Vitale, M. G. et al. Assessment of health status in patients with ing (NER), 89-92 (2019). cerebral palsy: what is the role of quality-of-life measures? J. 14. Gerasimenko, Y. P. et al. Noninvasive Reactivation of Motor Des- Pediatr. Orthopedics 25, 792–797 (2005). cending Control after Paralysis. J. Neurotrauma 32, 1968–1980 (2015). 15. Sayenko, D. et al. Self-assisted standing enabled by non-invasive Acknowledgements spinal stimulation after spinal cord injury. J. Neurotrauma (2018). We would like to acknowledge all the children, their parents and families 16. Gad, P. et al. Non-Invasive Activation of Cervical Spinal Networks for their participation in the study. This research was funded in part by after Severe Paralysis. J. Neurotrauma 35, 2145–2158 (2018). NIH/NIDDK R44DK129164, the Dana & Albert R. Broccoli Charitable 17. Inanici, F. et al. Transcutaneous Electrical Spinal Stimulation Pro- Foundation, Nanette and Burt Forester, including matching by PwC LLP, motes Long-Term Recovery of Upper Extremity Function in Chronic Roberta Wilson, BEL13VE in Miracles Jack Jablonski Foundation, Con- Tetraplegia. IEEE Trans. neural Syst. rehabilitation Eng.: a Publ. IEEE sortium for Technology & Innovation in Pediatrics, Brain Recovery Pro- Eng. Med. Biol. Soc. 26, 1272–1278 (2018). ject (BRP), Cerebral Palsy Alliance (Australia) and Cerebral Palsy Alliance 18. Rath, M. et al. Trunk Stability Enabled by Noninvasive Spinal Elec- Research Foundation (USA). The study sponsor had no role in study trical Stimulation after Spinal Cord Injury. J. Neurotrauma 35, design, data collection, analysis or writing. 2540–2553 (2018). 19. Gad, P. N., Kreydin, E., Zhong, H., Latack, K. & Edgerton, V. R. Non- Author contributions invasive Neuromodulation of Spinal Cord Restores Lower Urinary S.H. Designed the study, designed the ABNT protocol, oversaw the Tract Function After Paralysis. Front. Neurosci. 12, 432 (2018). study, performed all assessments, determined neuromodulation para- 20. Havton, L. A., Christe, K. L., Edgerton, V. R. & Gad, P. N. Noninvasive meters, edited and approved the manuscript. H.Z.: Delivered the ABNT spinal neuromodulation to map and augment lower urinary tract protocol and spinal neuromodulation, edited and approved the manu- function in rhesus macaques. Exp. Neurol. 322, 113033 (2019). script. R.F.: Delivered the ABNT protocol and spinal neuromodulation, 21. Kreydin, E. et al. Transcutaneous Electrical Spinal Cord Neuromo- edited and approved the manuscript G.Z.: Delivered the ABNT protocol dulator (TESCoN) Improves Symptoms of Overactive Bladder. and spinal neuromodulation, edited and approved the manuscript. CM: Front. Syst. Neurosci. 14, 1 (2020). Delivered the ABNT protocol and spinal neuromodulation, edited and 22. Gad, P., Kreydin, E., Zhong, H. & Reggie Edgerton, V. Training the approved the manuscript. P.G.: Designed the study, analyzed the data, bladder how to void: A noninvasive spinal neuromodulation case wrote, edited and approved the manuscript. V.R.E.: Designed the study, study. 10th International IEEE/EMBS Conference on Neural Engi- analyzed the data, wrote, edited and approved the manuscript. neering (NER) (2021). 23. Gad, P., Zhong, H., Edgerton, V. R. & Kreydin, E. Home-Based Competing interests SCONETM Therapy Improves Symptoms of Neurogenic Bladder. VRE has shareholder interest in Onward and SpineX. PG has shareholder Neurotrauma Rep. 2, 165–168 (2021). interest in SpineX. All other authors declare no other competing interest. 24. Gad, P. N., Kreydin, E., Zhong, H. & Edgerton, V. R. Enabling respiratory control after severe chronic tetraplegia: An exploratory Additional information case study. J. Neurophysiology (2020). Supplementary information The online version contains 25. Jackman, M. et al. Interventions to improve physical function for supplementary material available at children and young people with cerebral palsy: international clin- https://doi.org/10.1038/s41467-022-33208-w. ical practice guideline. Dev. Med. Child Neurol. (2021). 26. Gad, P. et al. Transcutaneous Spinal Neuromodulation Reorganizes Neural Correspondence and requests for materials should be addressed to Networks in Patients with Cerebral Palsy. Neurotherapeutics: the journal of Parag Gad. the American Society for Experimental NeuroTherapeutics (2021). 27. Ko, J. Functional Improvement after the Gross Motor Function Peer review information Nature Communications thanks Lynsey Duffell Measure-88 (GMFM-88) item-based training in children with cere- and the other, anonymous, reviewer(s) for their contribution to the peer bral palsy. J. Kor. Phys. Ther. 29, 115–121 (2017). review of this work. Peer reviewer reports are available. 28. Storm, F. A. et al. Minimum Clinically Important Difference of Gross Motor Function and Gait Endurance in Children with Motor Reprints and permission information is available at Impairment: A Comparison of Distribution-Based Approaches. http://www.nature.com/reprints BioMed. Res. Int. 2020, 2794036 (2020). 29. Solopova, I. A. et al. Effects of spinal cord stimulation on motor Publisher’s note Springer Nature remains neutral with regard to jur- functions in children with cerebral palsy. Neurosci. Lett. 639, isdictional claims in published maps and institutional affiliations. 192–198 (2017). Nature Communications | (2022)13:5660 6 Article https://doi.org/10.1038/s41467-022-33208-w Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2022 Nature Communications | (2022)13:5660 7