Neurotherapeutics (2021) 18:1953–1962 https://doi.org/10.1007/s13311-021-01087-6 ORIGINAL ARTICLE Transcutaneous Spinal Neuromodulation Reorganizes Neural Networks in Patients with Cerebral Palsy Parag Gad1,2,3 · Susan Hastings4 · Hui Zhong1 · Gaurav Seth5 · Sachin Kandhari6 · V. Reggie Edgerton1,7,8,9 Accepted: 30 June 2021 / Published online: 9 July 2021 © The Author(s) 2021 Abstract Spinal neuromodulation and activity-based rehabilitation triggers neural network reorganization and enhances sensory-motor performances involving the lower limbs, the trunk, and the upper limbs. This study reports the acute effects of Transcutaneous Electrical Spinal Cord Neuromodulation (SCONE™, SpineX Inc.) on 12 individuals (ages 2 to 50) diagnosed with cerebral palsy (CP) with Gross Motor Function Classification Scale (GMFCS) levels ranging from I to V. Acute spinal neuromodu- lation improved the postural and locomotor abilities in 11 out of the 12 patients including the ability to generate bilateral weight bearing stepping in a 2-year-old (GMFCS level IV) who was unable to step. In addition, we observed independent head-control and weight bearing standing with stimulation in a 10-year-old and a 4-year old (GMFCS level V) who were unable to hold their head up or stand without support in the absence of stimulation. All patients significantly improved in coordination of flexor and extensor motor pools and inter and intralimb joint angles while stepping on a treadmill. While it is assumed that the etiologies of the disruptive functions of CP are associated with an injury to the supraspinal networks, these data are consistent with the hypothesis that spinal neuromodulation and functionally focused activity-based therapies can form a functionally improved chronic state of reorganization of the spinal-supraspinal connectivity. We further suggest that the level of reorganization of spinal-supraspinal connectivity with neuromodulation contributed to improved locomo- tion by improving the coordination patterns of flexor and extensor muscles by modulating the amplitude and firing patterns of EMG burst during stepping. Introduction Cerebral palsy (CP) is one of the most common childhood * Parag Gad motor disorders in the USA with an estimated 1.5 to 4 out paraggad@gmail.com of 1000 individuals born with CP [1–3] and over 10,000 1 Department of Neurobiology, University of California, new cases diagnosed each year. There are approximately Los Angeles, CA 90095, USA 500,000 children under the age of 18 in USA and approxi- 2 Rancho Research Institute, Downey, CA 90242, USA mately 17 million globally currently living with CP. Medical 3 costs are around 10 times higher for children with CP and SpineX Inc, Los Angeles, CA 91324, USA 26 times higher for children with CP with an intellectual 4 Susan Hastings Pediatric Physical Therapy, San Jose, disability. The total lifetime care costs currently exceed $1 CA 95125, USA million [4]. CP is generally considered to be an umbrella 5 Indian Institute of Technology, Banaras Hindu University, diagnosis that includes a wide range of symptoms with het- Uttar Pradesh, Varanasi 221005, India erogeneous etiologies of neural and cardiovascular origins 6 Institute of Brain and Spine, New Delhi 110024, India [5]. It is generally assumed that the primary pathology of 7 Department of Neurosurgery, University of California, the nervous system that leads to CP is located within and Los Angeles, CA 90095, USA among different combinations of supraspinal networks, and 8 Brain Research Institute, University of California, these pathologies can be due to multiple etiologies. In most Los Angeles, CA 90095, USA cases, however, it appears that these supraspinal patholo- 9 Institut Guttmann, Hospital de Neurorehabilitació, Institut gies also will be necessarily manifested as spinally mediated Universitari Adscrit a La Universitat Autònoma de dysfunctions, affecting multiple peripheral sensory-motor Barcelona, 08916 Badalona Barcelona, Spain 13 Vol.:(0123456789) 1954 P. Gad et al. systems including midline orientation, equilibrium, pos- present study, has been shown to have a greater effect. In the ture (including trunk and head control), locomotion, and present study, we have merged ideas and concepts derived trunk and head control [6]. Also, the visual system often is from both recent clinical observations of CP and decades impacted in children with CP, because normally, the brain of studies in SCI patients [16, 26–28] including activity- and spinal networks provide a key, coherent driving factor to based mechanisms during locomotor tasks and transcutane- accommodate a 1G environment in controlling posture. But ous spinal neuromodulation [15, 27, 29]. We hypothesized children with CP often learn faulty trunk and head alignment that a single (acute) transcutaneous spinal neuromodulation as a result of the dysfunctional input from the supraspinal session could facilitate locomotor function of individuals networks to the largely normal spinal networks that can be with CP by reorganizing spinal neural networks to func- driven by the proprioceptive system [7]. tional states which enable spinal interneurons to translate The majority of children classified as Gross Motor Func- ensembles of proprioception to temporal patterns of activity tion Classification Scale (GMFCS) Level I are expected to that generates highly coordinated movements. How electri- reach their motor potential between 7 and 9 years and there- cal neuromodulation transforms the networks to functional after remain stable until age 21 when they may experience states which enable them to generate these improved coor- functional decline due to pain, weakness, and stiffness [8, 9]. dinated patterns remains unknown, but the present observa- Presently, all available interventions are designed to manage tions provide a very important target given its significance and minimize symptoms rather than correcting the underly- to the level of plasticity of spinal and supraspinal networks ing dysfunctions [10, 11]. While some commonly preferred of individuals with CP. Thus, the objective of this study is treatments such as intramuscular injection of onabotulinum- to determine whether the acute, system-level physiological toxinA (BotoxA) reduce symptoms of spasticity, which ini- effects of transcutaneous spinal cord neuromodulation has tially may improve function, but rarely lead to significant the potential to improve locomotor function in individuals long-term functional changes [12] and may severely limit the with CP as has been shown to be effective in individuals eventual level of recovery that theoretically is possible [13]. with spinal cord injuries. We have demonstrated that noninvasive stimulation of the spinal cord can lead to recovery of lower extremity [14, 15], upper extremity [16], trunk [17], breathing and coughing Methods [18], and bladder function [19, 20] after spinal cord injury (SCI). Some features of the physiological dysfunctions The study was approved by an external Investigational seen in individuals with SCI are similar to those commonly Review Board (Integreview IRB). All study participants (or observed in individuals with CP, i.e., poor coordination of their parents in case of a minor) signed the informed con- motor pools during stepping and spasticity. Multiple neu- sent form and consented to their data to be used for future romodulation modalities have been attempted in children publications and presentations. The patient demographic with CP for recovery of sensorimotor function includ- and pathologies are summarized in Table 1. The inclusion ing combined locomotor therapy and functional electrical criteria included (1) individuals above the age of 2 years of stimulation [21, 22] and epidural [23, 24], but transcuta- age and (2) diagnosed with cerebral palsy (CP). The exclu- neous spinal stimulation [25], similar to that used in the sion criteria included (1) selective dorsal root rhizotomy, (2) Table 1  Summary of the patient demographic and severity of the pathology Pt. ID Age (years) Gender Pathology GMFCS level Parameters T11/L1 P1 2.66 M CP spastic quadriplegia IV 15 mA/20 mA P2 6.25 M CP L spastic hemiplegia II 30 mA/40 mA P3 14 M CP spastic diplegia II 40 mA/50 mA P4 50 F CP spastic diplegia I 40 mA/50 mA P5 2.25 M CP spastic quadriplegia IV 20 mA/30 mA P6 16 F CP spastic quadriplegia—post hemispherectomy I 40 mA/50 mA P7 10 F CP-dyskinetic quadriplegia V 40 mA/50 mA P8 4 M CP spastic diplegia II 25 mA/30 mA P9 3 M CP L spastic quadriplegia—post hemispherectomy I 30 mA/30 mA P10 3 M CP spastic quadriplegia V 30 mA/40 mA P11 3 M CP R spastic quadriplegia—post hemispherectomy I 30 mA/45 mA P12 4 M CP-spastic triplegia II 30 mA/40 mA 13 Transcutaneous Spinal Neuromodulation Reorganizes Neural Networks in Patients with Cerebral… 1955 intramuscular Botox injection in the preceding 12 months, Locomotor Assessment (3) current antispastic medications, (4) unhealed fractures or contractures that would prevent them from performing For patients who were unable to perform functional tasks functional tasks, and (5) tendon-lengthening surgeries. such as standing or stepping, the primary outcome was the ability to voluntarily initiate these activities without any external assistance. For patients who were capable of step- SCONE™ Therapy ping on a treadmill, the primary outcome was to assess EMG and kinematic patterns representing improved coordination Spinal stimulation was delivered over the course of a sin- and reduced co-contractions. Each patient was tested on a gle 30-min session using a proprietary device SCONE™ treadmill at the same speed and body weight support (patient (SpineX Inc, Los Angeles, CA) (Parag Gad 2019). The 1 and 5) without and with stimulation. The speed of the stimulation waveform consisted of two alternating pulses treadmill was set to ensure patients were capable of stepping of opposite polarities separated by a 1-µS delay to form a at a comfortable speed. Since patients 7 and 10 were unable delayed biphasic waveform. The pulses consisted of a high- to hold their head up or stand upright without external sup- frequency biphasic carrier pulse (10 kHz) combined with a port, they were only tested for the ability to sit upright with low-frequency (30 Hz) burst pulse each with a pulse width the head erect and when transitioning from sit to stand. of 1 ms. Stimulation was applied using an adhesive hydrogel electrode between T11–12 and L1–2 serving as the cath- Data Analysis and Statistics ode and two adhesive hydrogel electrodes over bilateral iliac crests as the anodes (Supplemental Fig. 1). Patients Joint angle excursions were calculated by identifying the were assessed both without and with neuromodulation at stance and swing phases. Integrated EMG data are reported different parameters to test the changes in stepping pat- by calculating the area under the curve of the filtered, recti- terns. However, at all times, the stimulation intensity was fied signal for each step cycle. To calculate the level of co- maintained at ~ 20–25% below the threshold that induced contraction, we quantified the joint probability distribution a motor response [15, 17, 29–31]. All patients were able (JPD) plots of all data points (resolution 0.1 ms) from two to communicate with the parents and research team if they antagonistic muscles (TA and Sol) of the filtered EMG dur- experienced any pain or discomfort either due to the spi- ing a series of 15 consecutive step cycles. These data points nal neuromodulation or due to the locomotor procedures. were divided into 4 quadrants with the axes drawn at 20% of None of the patients reported any pain or discomfort dur- the max amplitude (of the linear envelope of rectified EMG ing the stimulation and were blinded from the stimulation trace) for each muscle. The total number of data points in parameters at any given time during the testing procedures. quadrants 2 and 4 representing alternating activity between While the patients could feel the stimulation pulses initially, antagonistic muscles were summed and compared between they were unable to clearly distinguish between intensities stimulation off vs on. The coefficient of correlation was cal- or even the presence of stimulation after a brief period of culated by measuring the correlation between joint angles of accommodation. The stimulation intensities used for each each normalized step cycle with the normalized mean step patient are listed in Table 1. cycle as previously published [32]. Paired t-tests were used to compare the group mean data without and with stimula- Data Collection tion. All statistical significances are reported at P < 0.05. Surface electromyography (EMG) was recorded from nine out of the twelve participants from select lower extrem- Results ity muscles using surface EMG electrodes (LabChart and PowerLab; ADInstruments). Electronic Goniometers were The patients were first tested without stimulation for either bilaterally placed on hip and/or knee joints to record joint stepping on the treadmill at their comfortable speed, sitting angular displacements. Video recordings were completed upright, or transitioning from sitting to standing. After a in all 12 patients and video plus EMG in 5 patients. EMG brief rest, they were then tested with stimulation at T11–12 recordings were not performed in 3 patients (P5, P7, and and L1–2 for the same tasks. The intensity of the stimula- P10) due to the visible apprehension seemingly attributable tion was unknown to the patient and was maintained at a to the size of the recording electrodes and wires as well as sub-motor threshold intensity [30]. Patient P1 was unable to their low level of postural stability. Data were recorded and step on the treadmill without stimulation and demonstrated sampled at a frequency of 10 kHz and were analyzed using occasional and uncoordinated bursts in his lower extremity LabChart software. Video data were recorded on LabChart muscles. He could however step bilaterally with stimula- and synchronized with the EMG files. tion on with robust alternating EMG bursting activity in all 13 1956 P. Gad et al. his lower extremity muscles (Fig. 1, Supplementary video G). Patient 7 who was only tested for upright sitting and 1). Patients (P2-P6, P8, P9, P11, and P12) were capable of standing due to complete lack of head control was capable stepping without stimulation and demonstrated consistent of performing the functional tasks for longer periods of time bursting activity in all lower extremity muscles with signifi- with increased levels of control (10–12 s vs 45–60 s) with cant levels of co-contraction of agonist-antagonistic mus- stimulation on compared to off (Supplementary video 2). cles (Fig. 2, Supplementary Fig. 2). The overall kinetics and kinematics represented by the change in EMG activity and joint angle excursions were similar between stimulation off Discussion and on (Fig. 3A). A key difference was noted in the inter- limb and intralimb coordination and the consistency of step- The present data suggest that the supraspinal connectivity ping as reflected in the increased coefficient of correlation to the spinal neural networks in individuals with CP has with fewer variations in joint angles (hip and knee) between developed into an incomplete and/or abnormal physiologi- consecutive cycles with stimulation on vs off (P < 0.05) cal state [33, 34]. It seems rather clear that the abnormality (Fig. 3B, C). While the area under the curve for joint angles in the brain sends disruptive signals to the spinal networks (hip-knee and knee-knee) showed a trend toward a decrease that, in turn, generates poorly coordinated movements. We with stimulation on vs off; these changes were not signifi- hypothesized that this abnormal supraspinal-spinal loop cant (Fig. 3D), perhaps, because the overall angular excur- can be functionally minimized, largely via proprioception sions remained the same (Fig. 3A). However, the patients projecting normal signals to spinal networks enabling the were able to take longer steps in the presence vs absence of automaticity and feedforwardness [35] of the spinal cord stimulation (0.88 ± 0.005 s vs 0.973 ± 0.036 s, P < 0.05). In which translates the sensory input into a more normal motor addition, antagonistic muscles (TA vs soleus) showed sig- action (Fig. 4) [36]. Finally, we predict that this more normal nificantly lower levels of co-contraction (Fig. 3E). The mean action derived from proprioception in the presence of spi- percent data points representative of alternating activity nal neuromodulation can be learned via activity-dependent (Quadrant 2 + 4) was significantly higher with stimulation changes (rehabilitation). One could question, or at least on vs off representing a decreased level of co-contraction be surprised, as to how the outcomes of the interventional between antagonistic muscles, a representation of decreased strategies used for CP are similar to those observed with spasticity as well as considerably improved levels of coordi- spinal cord injury (SCI). Both dysfunctions, however, have nation of agonist-antagonistic motor pools (Fig. 3E, F, and evolved from abnormal spinal-supraspinal connections, but Fig. 1  Patient no. 1 (age 2) step- ping on a treadmill without and with spinal neuromodulation. Note that the patient is unable to step with the stimulation off whereas he begins to step voluntarily once the stimulation is turned on 13 Transcutaneous Spinal Neuromodulation Reorganizes Neural Networks in Patients with Cerebral… 1957 Fig. 2  Patient no. 2 (age 7) step- ping on a treadmill without and with spinal stimulation both can be acutely driven toward normality by proper pos- spinal neuromodulation can lead to positive chronic effects. ture, proprioception, and spinal neuromodulation [15–17, It is interesting to note that the greatest improvements in 19, 36–38]. functional performance during acute spinal neuromodulation CP patients demonstrate improper and perhaps incom- were observed in either the youngest individuals or ones that plete functional connectivity between the brain and the were most severely affected (GMFCS IV and V) whereas spinal cord (Fig. 4). Since a relatively “normal” level and the oldest reported minimal change in sensation and perfor- target-specific connectivity is needed to accommodate a 1-G mance with a single exposure to neuromodulation. environment, the dysfunctional brain-spinal cord connec- The importance of these results is that it provides a tivity results in the patient learning an abnormal gait and lead to a novel strategy with a strong biological logic that posture to accommodate gravity. Spinal neuromodulation might make it possible to realize a transformation similar changes the physiological states and the functional connec- to that which is occurring with spinal cord injury. Thus, we tions of the brain and spinal networks. The present observa- hypothesized that spinal neuromodulation in concert with tions of patients with CP over a wide range of age groups (2 proprioceptive-driven activity-dependent mechanisms of to 50 years) and severity of dysfunction (GMFCS level I to spinal networks can transform the supraspinal-spinal dys- V) are important, first, because the similarities at systems- functional connectivity of CP into highly functional connec- level mechanisms of acutely reorganizing spinal-supraspinal tions, characterized by a more normal agonist–antagonist networks generated improved coordination of motor pools coordination pattern. Unlike traditional approaches (such in stepping are similar to that observed with SCI for multi- as functional electrical stimulation, which are designed to ple motor functions. Secondly, the similarities of the acute directly induce muscle contractions), spinal neuromodulation effects of both SCI and CP provide strong evidence that sim- can be used to modulate the excitability of spinal networks ilar mechanisms could be relevant in achieving long-term below the motor threshold to a more functional physiologi- neuroplasticity with the repetitive presentation of similar cal state when engaged primarily by more normal patterns interventional parameters [36]. The acute changes in coor- of proprioception. We propose that these continuous ensem- dination of motor pools and reduced variability in stepping bles of proprioception can be translated automatically with reported here provide strong evidence that noninvasive considerable precision by the spinal networks. Based on the 13 1958 P. Gad et al. 13 Transcutaneous Spinal Neuromodulation Reorganizes Neural Networks in Patients with Cerebral… 1959 ◂Fig. 3  A Mean ± SE (n = 9 patients) angular excursion of the hip and which further restores [23] sensorimotor functions (Fig. 4). knee joints without and with stimulation. B Interlimb coordination Although improvement of multiple behaviors derived from has shown from a representative patient (P2, n = 15 step cycles) by the knee-knee joint angle plots and intralimb coordination shown by supraspinal networks following spinal injury has been clearly the hip-knee plots without and with stimulation. Note the shaded area manifested physiologically, the spinal-supraspinal networks represents the variation over the 15-step cycle. C Mean ± SE (n = 9) through which this effect is mediated have only recently coefficient of correlation of the trajectories shown in B with respect become of great importance. In essence, transcutaneous to the mean trajectory. D Mean ± SE (n = 9) area under the curve cal- culated for the plot in B. E Joint probability density (JPD) distribu- spinal neuromodulation appears to facilitate neuroplasticity tion plot of filtered rectified EMG amplitudes of the TA vs the Sol and can lead to bidirectional transformations of neural con- muscles without and with stimulation derived from an average of nections among brain-spinal cord-muscle-spinal cord-brain 15 step cycles for a representative patient (P2). F Mean ± SE (n = 9) connectomes [39]. percent data points in Quadrants (1 + 3) and (2 + 4) without vs with stimulation and G the mean percent data points for each quadrant. One of the most common spontaneous perceptions *Statistically significant at P < 0.05 expressed by patients with CP (including the ones reported here) when initially receiving epidural or transcutaneous more normal signals generated by these spinal networks. spinal neuromodulation, regardless of whether it was an With repetitive treatments, more normal signals are pro- individual with a SCI, stroke, or CP and regardless of the jected supraspinally to movement control centers in the brain, task being attempted, was that it was “easier” to perform and Fig. 4  Hypothetical schematic of the brain, spinal cord, and muscles in children with cerebral palsy without and with acute spinal neuromodulation. F flexor motor pool, E extensor motor pool. Note the symbolic reduced lesion size to reflect a lesser dysfunctional supraspinal impact during spinal neuro- modulation 13 1960 P. Gad et al. was able to perform tasks for longer periods of time with are performed repetitively, a learning phenomenon can emerge, less fatigue. Another self-reported comment has been an which facilitates regaining a range of movements requiring increase in level of awareness and “sensation of stepping,” greater skill, power, and endurance. Also, the present obser- “feeling lighter” while walking on the treadmill with stimu- vations are consistent with the neuromodulation changing the lation. Interestingly, in the first published report by Herman physiological states of spinal and supraspinal networks in a and colleagues in 2002 [40] and subsequently by Harkema way that enables the nervous system to choose from an infinite and colleagues in 2011 [28], exploring the possibility of number of potential movements. In essence, this enables the restoring motor function after a severe spinal cord injury, person with CP to engage strict guidance from proprioceptive the patient’s perception was that it was easier to walk in the input as occurs normally, to drive the spinal networks [47]. presence of epidural stimulation. The mechanisms by which We emphasize this latter point because the more effective the widely different means of neuromodulation are applied spinal neuromodulation strategy after SCI has been to use and the types of patients with different dysfunctions and the below motor threshold levels of neuromodulation, as much regaining of an array of sensation need to be further explored as possible to enable the patient to perform a wide range of more quantitatively. There has been strikingly little effort to movements, by choice, not by electrically induced movements attempt to understand the mechanisms of recovering sensa- defined by the experimenter or the technical device used [43]. tions, although there are increasingly important clues as to Supra-motor threshold levels of stimulation reduce the control its fundamental importance in recovering motor function. It by the spinal interneuronal networks, the sites at which most is also important to emphasize that while the brain-spinal of the coordination of motor pools and spinal learning nor- cord re-connectivity can be functionally (behaviorally) sub- mally occurs. The acute reduction in co-contraction between stantial, the anatomical equivalent is completely unknown. antagonistic muscles with neuromodulation was observed in Clinically, these results suggest that the supraspinal and all patients. This acute effect highlights the potential impor- spinal networks are highly responsive to activity-dependent tance of engaging activity-dependent mechanisms to minimize mechanisms that help drive more aligned posture, particularly the impact of functionally aberrant connections that may be when administered in concert with spinal neuromodulation formed in the brain and spinal cord [48, 49]. In addition, it [41] [42, 43]. We propose that acute spinal neuromodulation highlights the need for more long-term activity-based neu- in children with CP can enable a persistent drive for function- rorehabilitation therapy especially for the older individuals for ally useful reorganization of spinal and supraspinal networks whom there may be a greater persistence of aberrant connec- to emerge, similar to that which has been reported for step- tions which often lead to persistent spasticity, a manifestation ping [31], standing [15], upper limb [16, 44], trunk stability of poor coordination. [17], normalization of blood pressure function [45], breathing [18], bladder, bowel, and sexual function [19, 20, 46]. These Supplementary Information The online version contains supplemen- tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 13311-0​ 21-0​ 1087-6. observations have occurred in almost every patient who had been spinally injured for at least 1 year with complete paraly- Acknowledgements To perform the experiments and observations, sis before receiving spinal neuromodulation and some form extraordinary dedication, compliance, and accommodations of the of activity-dependent intervention. It is interesting to note the patients and parents, siblings, and therapist were necessary. We would similarity in response between CP and SCI patients considering also like to acknowledge the team at Abilities Recovery Center (ARC) for letting us use their facilities. the difference in pathologies. This is supported by the facts that the electrophysiological basis of our approach is similar, i.e., Required Author Forms Disclosure forms provided by the authors are (1) enabling the proprioceptive system to drive motor function available with the online version of this article. and (2) re-establish appropriate connections between spinal and supraspinal centers. The results observed in CP may be Funding The authors declare that this study received funding from further enhanced considering a stronger anatomical connection Walkabout Foundation, Dana & Albert R. Broccoli Charitable Foun- dation and Nanette, Burt Forester, including matching by PwC LLP, between brain and spinal cord. However, the apparent problem Roberta Wilson, The West Coast Consortium for Technology & with the greater connectivity in CP compared to SCI is the high Innovation in Pediatrics (CTIP) and the Brain Recovery Project. degree of functionally aberrant supraspinal-spinal connections, given the spastic-like behavior and the abnormal EMG patterns Declarations (Capellini et al., 2020 and Edgerton et al., 2021). While spinal neuromodulation has been explored in chil- Competing Interests V.R.E, holds shareholder interest in NeuroRecov- dren with CP, these data are unique as these is the first to dem- ery Technologies and hold certain inventorship rights on intellectual property licensed by The Regents of the University of California to onstrate the electrophysiological changes possible within one NeuroRecovery Technologies and its subsidiaries. V.R.E and PG holds session of stimulation to reorganize a nonfunctional or dys- shareholder interest in SpineX Inc. and hold certain inventorship rights functional brain-spinal network to a new functional physiologi- on intellectual property licensed by The Regents of the University of cal state. In addition, it lays the foundation for when these tasks California to SpineX Inc. 13 Transcutaneous Spinal Neuromodulation Reorganizes Neural Networks in Patients with Cerebral… 1961 Disclaimer The funders were not involved in the study design, collec- 15. Sayenko D, Rath M, Ferguson AR, Burdick J, Havton L, Edgerton tion, analysis, interpretation of data, the writing of this article or the VRPD, et al. Self-assisted standing enabled by non-invasive spinal decision to submit it for publication. stimulation after spinal cord injury. Journal of neurotrauma. 2018. 16. Gad P, Lee S, Terrafranca N, Zhong H, Turner A, Gerasimenko Y, et al. Non-Invasive Activation of Cervical Spinal Networks after Open Access This article is licensed under a Creative Commons Attri- Severe Paralysis. J Neurotrauma. 2018;35(18):2145-58. bution 4.0 International License, which permits use, sharing, adapta- 17. Rath M, Vette AH, Ramasubramaniam S, Li K, Burdick J, tion, distribution and reproduction in any medium or format, as long Edgerton VR, et al. Trunk Stability Enabled by Noninvasive as you give appropriate credit to the original author(s) and the source, Spinal Electrical Stimulation after Spinal Cord Injury. J Neu- provide a link to the Creative Commons licence, and indicate if changes rotrauma. 2018;35(21):2540-53. were made. The images or other third party material in this article are 18. Gad PN, Kreydin E, Zhong H, Edgerton VR. Enabling respiratory included in the article’s Creative Commons licence, unless indicated control after severe chronic tetraplegia: An exploratory case study. otherwise in a credit line to the material. If material is not included in Journal of neurophysiology. 2020. the article’s Creative Commons licence and your intended use is not 19. Gad PN, Kreydin E, Zhong H, Latack K, Edgerton VR. Non- permitted by statutory regulation or exceeds the permitted use, you will invasive Neuromodulation of Spinal Cord Restores Lower Urinary need to obtain permission directly from the copyright holder. To view a Tract Function After Paralysis. Front Neurosci. 2018;12:432. copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. 20. Kreydin E, Zhong H, Latack K, Ye S, Edgerton VR, Gad P. Trans- cutaneous Electrical Spinal Cord Neuromodulator (TESCoN) Improves Symptoms of Overactive Bladder. Frontiers in systems References neuroscience. 2020;14:1. 21. Armstrong EL, Boyd RN, Kentish MJ, Carty CP, Horan SA. 1. 1.Arneson CL, Durkin MS, Benedict RE, Kirby RS, Yeargin-Allsopp M, Effects of a training programme of functional electrical stimula- Van Naarden Braun K, et al. Prevalence of cerebral palsy: Autism and tion (FES) powered cycling, recreational cycling and goal-directed Developmental Disabilities Monitoring Network, three sites, United exercise training on children with cerebral palsy: a randomised States, 2004. Disability and health journal. 2009;2(1):45-8. controlled trial protocol. BMJ open. 2019;9(6):e024881. 2. Paneth N, Hong T, Korzeniewski S. The descriptive epidemiology 22. Nikityuk IE, Moshonkina TR, Shcherbakova NA, Vissarionov SV, of cerebral palsy. Clinics in perinatology. 2006;33(2):251-67. Umnov VV, Rozhdestvenskii VY, et al. [Effects of locomotor training 3. Winter S, Autry A, Boyle C, Yeargin-Allsopp M. Trends in the and functional electrical stimulation on postural function in children prevalence of cerebral palsy in a population-based study. Pediat- with severe cerebral palsy]. Fiziologiia cheloveka. 2016;42(3):37-46. rics. 2002;110(6):1220-5. 23. Hugenholtz H, Humphreys P, McIntyre WM, Spasoff RA, Steel 4. Kruse M, Michelsen SI, Flachs EM, Bronnum-Hansen H, Madsen M, K. Cervical spinal cord stimulation for spasticity in cerebral palsy. Uldall P. Lifetime costs of cerebral palsy. Developmental medicine Neurosurgery. 1988;22(4):707-14. and child neurology. 2009;51(8):622-8. 24. Dekopov AV, Shabalov VA, Tomsky AA, Hit MV, Salova EM. 5. Sanger T. Movement disorders in cerebral palsy. Journal of Pedi- Chronic Spinal Cord Stimulation in the Treatment of Cerebral atric Neurology. 2015. and Spinal Spasticity. Stereotactic and functional neurosurgery. 6. Smith AT, Gorassini MA. Hyperexcitability of brain stem pathways 2015;93(2):133-9. in cerebral palsy. Journal of neurophysiology. 2018;120(3):1428-37. 25. Solopova IA, Sukhotina IA, Zhvansky DS, Ikoeva GA, Vissarionov 7. Curtis DJ, Butler P, Saavedra S, Bencke J, Kallemose T, Sonne-Holm SV, Baindurashvili AG, et al. Effects of spinal cord stimulation S, et al. The central role of trunk control in the gross motor function on motor functions in children with cerebral palsy. Neuroscience of children with cerebral palsy: a retrospective cross-sectional study. letters. 2017;639:192-8. Developmental medicine and child neurology. 2015;57(4):351-7. 26. Edgerton VR, Tillakaratne NJ, Bigbee AJ, de Leon RD, Roy RR. 8. Findlay B, Switzer L, Narayanan U, Chen S, Fehlings D. Investigating Plasticity of the spinal neural circuitry after injury. Annual review the impact of pain, age, Gross Motor Function Classification System, of neuroscience. 2004;27:145-67. and sex on health-related quality of life in children with cerebral palsy. 27. Gad P, Roy RR, Choe J, Zhong H, Nandra MS, Tai YC, et al. Electro- Developmental medicine and child neurology. 2016;58(3):292-7. physiological mapping of rat sensorimotor lumbosacral spinal net- 9. Rosenbaum PL, Walter SD, Hanna SE, Palisano RJ, Russell DJ, Raina works after complete paralysis. Prog Brain Res. 2015;218:199-212. P, et al. Prognosis for gross motor function in cerebral palsy: creation 28. Harkema S, Gerasimenko Y, Hodes J, Burdick J, Angeli C, of motor development curves. Jama. 2002;288(11):1357-63. Chen Y, et al. Effect of epidural stimulation of the lumbosa- 10. Choi JY, Kim SK, Park ES. The Effect of Botulinum Toxin Injec- cral spinal cord on voluntary movement, standing, and assisted tions on Gross Motor Function for Lower Limb Spasticity in Chil- stepping after motor complete paraplegia: a case study. Lancet. dren with Cerebral Palsy. Toxins. 2019;11(11). 2011;377(9781):1938-47. 11. Dewar R, Love S, Johnston LM. Exercise interventions improve 29. Gad P, Gerasimenko Y, Zdunowski S, Turner A, Sayenko D, Lu postural control in children with cerebral palsy: a systematic review. DC, et al. Weight Bearing Over-ground Stepping in an Exoskel- Developmental medicine and child neurology. 2015;57(6):504-20. eton with Non-invasive Spinal Cord Neuromodulation after Motor 12. Tedroff K, Hagglund G, Miller F. Long-term effects of selective Complete Paraplegia. Front Neurosci. 2017;11:333. dorsal rhizotomy in children with cerebral palsy: a systematic review. 30. Gad P, Choe J, Shah P, Garcia-Alias G, Rath M, Gerasimenko Y, Developmental medicine and child neurology. 2020;62(5):554-62. et al. Sub-threshold spinal cord stimulation facilitates spontane- 13. Graham D, Aquilina K, Mankad K, Wimalasundera N. Selective ous motor activity in spinal rats. Journal of neuroengineering and dorsal rhizotomy: current state of practice and the role of imaging. rehabilitation. 2013;10:108. Quantitative imaging in medicine and surgery. 2018;8(2):209-18. 31. Parag Gad YG, V Reggie Edgerton. Tetraplegia to Overground 14. Gerasimenko YP, Lu DC, Modaber M, Zdunowski S, Gad P, Stepping Using Non-Invasive Spinal Neuromodulation. 2019 9th Sayenko DG, et al. Noninvasive Reactivation of Motor Descending International IEEE/EMBS Conference on Neural Engineering Control after Paralysis. J Neurotrauma. 2015;32(24):1968-80. (NER). 2019:89–92. 32. Shah PK, Garcia-Alias G, Choe J, Gad P, Gerasimenko Y, Tillakaratne N, et al. Use of quadrupedal step training to re- engage spinal interneuronal networks and improve locomotor 13 1962 P. Gad et al. function after spinal cord injury. Brain : a journal of neurology. 43. Taccola G, Sayenko D, Gad P, Gerasimenko Y, Edgerton VR. 2013;136(Pt 11):3362-77. And yet it moves: Recovery of volitional control after spinal cord 33. Shuman BR, Goudriaan M, Desloovere K, Schwartz MH, Steele injury. Prog Neurobiol. 2018;160:64-81. KM. Muscle synergies demonstrate only minimal changes after 44. Inanici F, Samejima S, Gad P, Edgerton VR, Hofstetter CP, Moritz treatment in cerebral palsy. Journal of neuroengineering and reha- CT. Transcutaneous Electrical Spinal Stimulation Promotes Long- bilitation. 2019;16(1):46. Term Recovery of Upper Extremity Function in Chronic Tetraple- 34. Cappellini G, Sylos-Labini F, MacLellan MJ, Assenza C, gia. IEEE Trans Neural Syst Rehabil Eng. 2018;26(6):1272-8. Libernini L, Morelli D, et al. Locomotor patterns during obstacle 45. Phillips AA, Squair JW, Sayenko DG, Edgerton VR, Gerasimenko avoidance in children with cerebral palsy. Journal of neurophysiol- Y, Krassioukov AV. An Autonomic Neuroprosthesis: Noninvasive ogy. 2020;124(2):574-90. Electrical Spinal Cord Stimulation Restores Autonomic Cardio- 35. Gerasimenko Y, Sayenko D, Gad P, Liu CT, Tillakaratne NJK, vascular Function in Individuals with Spinal Cord Injury. Journal Roy RR, et al. Feed-Forwardness of Spinal Networks in Posture of neurotrauma. 2018;35(3):446-51. and Locomotion. Neuroscientist. 2017;23(5):441-53. 46. Gad P, Evgeniy Kreydin, Hui Zhong, and V. Reggie Edgerton. 36. V Reggie Edgerton SH, Parag Gad. Engaging spinal networks to Training the bladder how to void: A noninvasive spinal neuro- mitigate supraspinal dysfunction after CP. Frontiers in Systems modulation case study. 10th International IEEE/EMBS Confer- Neurosciences. 2021. ence on Neural Engineering (NER). 2021. 37. Reid LB, Rose SE, Boyd RN. Rehabilitation and neuroplasticity in 47. Taccola G, Gad P, Culaclii S, Wang PM, Liu W, Edgerton children with unilateral cerebral palsy. Nature reviews Neurology. VR. Acute neuromodulation restores spinally-induced motor 2015;11(7):390-400. responses after severe spinal cord injury. Experimental neurol- 38. Morgan C, Novak I, Dale RC, Badawi N. Optimising motor learning ogy. 2020;327:113246. in infants at high risk of cerebral palsy: a pilot study. BMC pediatrics. 48. de Leon RD, Tamaki H, Hodgson JA, Roy RR, Edgerton VR. 2015;15:30. Hindlimb locomotor and postural training modulates glycinergic 39. Edgerton VR, Gad P. Is the vagus nerve our neural connectome? inhibition in the spinal cord of the adult spinal cat. Journal of Elife. 2018;7. neurophysiology. 1999;82(1):359-69. 40. Herman R, He J, D’Luzansky S, Willis W, Dilli S. Spinal cord 49. Chen B, Li Y, Yu B, Zhang Z, Brommer B, Williams PR, et al. stimulation facilitates functional walking in a chronic, incomplete Reactivation of Dormant Relay Pathways in Injured Spinal Cord spinal cord injured. Spinal cord. 2002;40(2):65-8. by KCC2 Manipulations. Cell. 2018;174(6):1599. 41. Fong AJ, Roy RR, Ichiyama RM, Lavrov I, Courtine G, Gerasimenko Y, et al. Recovery of control of posture and locomotion after a spinal Publisher’s Note Springer Nature remains neutral with regard to cord injury: solutions staring us in the face. Progress in brain research. jurisdictional claims in published maps and institutional affiliations. 2009;175:393-418. 42. Sayenko VREYPGPGDG. Basic concepts underlying activity- dependent mechanisms in the rehabilitation of sensory-motor function after spinal cord injury. Spinal Cord Medicine. 2018;Chapter 54:897 to 911. 13