JOURNAL OF NEUROTRAUMA 36:1451–1460 (May 1, 2019) Mary Ann Liebert, Inc. DOI: 10.1089/neu.2018.5921 Electrophysiological Guidance of Epidural Electrode Array Implantation over the Human Lumbosacral Spinal Cord to Enable Motor Function after Chronic Paralysis Jonathan S. Calvert,1 Peter J. Grahn,2 Jeffrey A. Strommen,3 Igor A. Lavrov,2 Lisa A. Beck,3 Megan L. Gill,3 Margaux B. Linde,3 Desmond A. Brown,2 Meegan G. Van Straaten,3 Daniel D. Veith,3 Cesar Lopez,3 Dimitry G. Sayenko,4,12 Yury P. Gerasimenko,4,5 V. Reggie Edgerton,4,6–10 Kristin D. Zhao,3,11 and Kendall H. Lee2,3,11 Abstract Epidural electrical stimulation (EES) of the spinal cord has been shown to restore function after spinal cord injury (SCI). Characterization of EES-evoked motor responses has provided a basic understanding of spinal sensorimotor network activity related to EES-enabled motor activity of the lower extremities. However, the use of EES-evoked motor responses to guide EES system implantation over the spinal cord and their relation to post-operative EES-enabled function in humans with chronic paralysis attributed to SCI has yet to be described. Herein, we describe the surgical and in- traoperative electrophysiological approach used, followed by initial EES-enabled results observed in 2 human subjects with motor complete paralysis who were enrolled in a clinical trial investigating the use of EES to enable motor functions after SCI. The 16-contact electrode array was initially positioned under fluoroscopic guidance. Then, EES-evoked motor responses were recorded from select leg muscles and displayed in real time to determine electrode array proximity to spinal cord regions associated with motor activity of the lower extremities. Acceptable array positioning was determined based on achievement of selective proximal or distal leg muscle activity, as well as bilateral muscle activation. Motor response latencies were not significantly different between intraoperative recordings and post-operative recordings, in- dicating that array positioning remained stable. Additionally, EES enabled intentional control of step-like activity in both subjects within the first 5 days of testing. These results suggest that the use of EES-evoked motor responses may guide intraoperative positioning of epidural electrodes to target spinal cord circuitry to enable motor functions after SCI. Keywords: electrically evoked spinal motor potentials; epidural electrical stimulation, spinal cord injury; neuromodulation; spinal cord intraoperative electrophysiology Introduction and from early-phase clinical studies have reported positive out- comes such as decreased initial injury severity by pharmacological interventions,3 repair of damaged spinal cord tissues by implanta- S evere spinal trauma may result in spinal cord injury (SCI) with permanent, devastating loss of motor, sensory, and auto- nomic functions.1,2 Scientific evidence from animal models of SCI tion of biomaterials and engineered cells at the injury site,4–6 or direct signal transmission through or around the injury using state- 1 Mayo Clinic Graduate School of Biomedical Sciences, Mayo Clinic, Rochester, Minnesota. 2 Department of Neurologic Surgery, Mayo Clinic, Rochester, Minnesota. 3 Department of Physical Medicine and Rehabilitation, Rehabilitation Medicine Research Center, Mayo Clinic, Rochester, Minnesota. 4 Department of Integrative Biology and Physiology University of California Los Angeles, Los Angeles, California. 5 Pavlov Institute of Physiology, St. Petersburg, Russia. 6 Department of Neurobiology, University of California Los Angeles, Los Angeles, California. 7 Department of Neurosurgery, University of California Los Angeles, Los Angeles, California. 8 Brain Research Institute, University of California Los Angeles, Los Angeles, California. 9 Institut Guttmann, Hospital de Neurorehabilitació, Institut Universitari adscrit a la Universitat Autònoma de Barcelona, Barcelona, Badalona, Spain. 10 Centre for Neuroscience and Regenerative Medicine, Faculty of Science, University of Technology Sydney, Ultimo, New South Wales, Australia. 11 Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, Minnesota. 12 Center for Neuroregeneration, Department of Neurosurgery, Houston Methodist Research Institute, Houston, Texas.  Jonathan S. Calvert et al., 2018; Published by Mary Ann Liebert, Inc. This Open Access article is distributed under the terms of the Creative Commons License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. 1451 1452 CALVERT ET AL. of-the-art technology to interface with other functional neural Subject descriptions pathways.7–12 Despite these promising results, translational efforts Two males diagnosed with a sensorimotor complete, American have yet to achieve significant functional gains across the popula- Spinal Injury Association Impairment Scale Grade A (AIS-A) tion of humans with SCIs.13 SCI38 were enrolled in this clinical trial (NCT02592668), and both Initial reports using epidural electrical stimulation (EES) applied provided written, informed consent to experimental procedures. to the dorsal surface of the lumbosacral spinal cord demonstrated Subject 1 was a 26-year-old male who sustained a T6 SCI during a that EES could induce tonic and rhythmic leg muscle activity both snowmobile accident 3 years before study enrollment. Subject 2 in animal models of complete SCI14–18 and in humans with chronic was a 37-year-old male who sustained a T3 SCI from a fall 6 years paraplegia attributed to severe SCI.19–21 The capacity of EES to before study enrollment. Data from subject 1 have been previously facilitate intentional movement in an individual with motor com- reported12; however, all data and analysis demonstrated in this plete SCI was first reported by Harkema and colleagues in 201122 report were not previously published. and subsequently replicated by the same investigators in 3 addi- tional subjects.11 In total, 4 males diagnosed with motor complete Motor training before to epidural electrical stimulation paraplegia at least 2 years before study enrollment were implanted system implantation with an EES system over the lumbosacral spinal cord. After several In order to confirm that rehabilitation alone would not result in months of locomotor training with EES, all 4 subjects could in- functional recovery, both subjects performed motor training for 6 tentionally control leg movements while supine with EES on. months before implantation of the EES system. A detailed de- Additionally, EES enabled subjects to stand for prolonged periods scription of pre-surgical motor training for subject 1 was previ- of time. While attempting to replicate the results reported by ously reported.12 Briefly, subjects underwent approximately 60 Harkema and colleagues,11,22 we reported that EES enabled in- motor training sessions performed over 22 weeks for approxi- tentional control of lower extremity movements and standing mately three sessions per week by a team of physical therapists within the first 2 weeks of EES in a subject diagnosed with motor and kinesiologists. Motor training sessions consisted of approxi- complete paraplegia,12 and over the course of 43 weeks of multi- mately 15 min of lower extremity stretching, 45 min of locomotor training on a treadmill, and 30 min of balance and task-specific modal rehabilitation, he achieved independent stepping in the strengthening exercises. All motor training activities were per- presence of EES.23 formed with body weight support and trainer assistance provided Detailed descriptions of the motor functions that were enabled as needed. by EES have been reported by studies of humans with chronic, motor complete,11,12,22 and incomplete24 paraplegia, as well as Surgical approach to the spinal canal motor complete tetraplegia.25 Despite small sample sizes, these positive outcomes, combined with investigations of EES to re- After general anesthesia, subjects were positioned prone us- porting improved cardiovascular,26,27 respiratory,28–31 and auto- ing a Jackson surgical table and Mayfield head clamp. General nomic32 function as well as improved body composition33 after anesthesia was maintained by intravenous delivery of propofol SCI, demonstrate that EES holds considerable potential as a ther- and fentanyl. Inhalation agents and neuromuscular blocking agents were avoided for purposes of intraoperative electro- apeutic intervention after SCI. To date, the scientific literature has physiological monitoring of spinal sensorimotor activity during described EES-enabled functions and the stimulation patterns, EES trialing. electrode configurations applied, and motor training regimens used A midline incision that spanned vertebrae T11–L2 was made enable functions.22,34–36 based on palpation of anatomical landmarks and fluoroscopy. Previous investigations using EES in humans with SCI described After initial exposure, surgical dissection of back musculature the use of intraoperative X-ray guidance to identify vertebral levels continued through the lumbodorsal fascia to the level of the T11–L1 as the target location for insertion of an electrode array.22 spinous processes in an avascular plane using electrocautery. However, X-ray imaging does not provide information regarding Once the spinous processes were exposed, vertebral level was electrode proximity to spinal cord circuitry that enables function.37 confirmed intraoperatively by inserting a radio-opaque marker in Previous reports have briefly described the use of intraoperative proximity to the spinous process followed by fluoroscopy. Next, subperiosteal dissection of the muscle from the spinous process electrophysiological monitoring to guide electrode array implan- down to the lamina and then laterally toward the facets was tation over the lumbosacral spinal cord.12,22 However, the surgical performed. Then, the T11–12 and L1–2 interspinous ligaments approach to position the epidural electrode array, intraoperative were severed and the spinous processes of T12 and L1 were electrophysiological techniques used to guide electrode position- removed with a rongeur. The facet joints of both locations were ing, and electrophysiological outcomes generated have yet to be preserved to ensure spinal stability. The laminae were then described in detail. Herein, we provide a detailed description of the thinned with a 4-mm diamond drill bit and removed by rongeurs. surgical approach with results from intraoperative electrophysi- Finally, the ligamentum flavum was dissected free from the un- ology of EES-evoked motor responses from leg muscles indicating derlying dura. appropriate positioning of the EES electrode array to enable motor functions, such as step-like lower extremity movement, after re- Initial positioning of the epidural electrical stimulation covery from EES implantation. electrode array A 16-contact epidural electrode array (Specify 5-6-5; Medtronic, Methods Fridley, MN) was passed through the T12–L1 laminectomy and directed rostrally along the dorsal midline of the dura to span the This study was performed under the approval of the Mayo Clinic T11–L1 vertebrae (Fig. 1). Insertion of the electrode to the adequate Institutional Review Board with a U.S. Food and Drug Adminis- level was visualized through the L1–2 opening. T12–L1 exposure tration Investigational Device Exemption (IDE G150167), and all provided access to the caudal portion of the array while the L1–2 experiments were performed in accord with the regulations and opening provided access to the rostral portion, allowing array vi- guidelines of these regulatory bodies. sualization and mediolateral array manipulation. To determine GUIDED PLACEMENT OF EPIDURAL ELECTRODE ARRAY 1453 FIG. 1. Surgical implantation of the EES electrode array spanning the lumbosacral spinal cord. (A) Intraoperative image of the location of the EES array at the T11–L1 vertebral levels. (B) Anterior-posterior X-ray of each subject before and after EES electrode array implantation. Subject 1 imaging was captured in a seated position. Subject 2 imaging was captured while lying supine. Inserts depict zoomed in view of the EES array. EES, epidural electrical stimulation. array integrity and compliance with surrounding tissue, the array Lead and battery placement was connected to an external pulse generator to capture impedance values at each electrode contact. Upon initial implantation, in- After confirmation of electrode position by electrophysiology, traoperative fluoroscopy was used to confirm the electrode array the electrode array was secured to the fascia and leads were tu- spanned the appropriate vertebral levels. nneled subcutaneously to the right upper abdomen where a small subcutaneous pocket was made to house the lead wires. After meticulous hemostasis and copious irrigation with a bacitracin Intraoperative electrophysiology solution, the posterior wounds were closed in anatomical layers. Subjects were then positioned on the operating table in a lateral After induction but before surgical exposure, 13-mm subdermal decubitus position. After prepping and draping the side and lateral electrodes were placed over selected muscles of the lower limbs in abdomen, the lead wires were externalized and a pocket was made a muscle to tendon recording montage. The active electrode was to insert the implantable pulse generator (RestoreUltra SureScan placed over the midpoint of the appropriate muscle with the ref- MRI Neurostimulator, Model 97712; Medtronic) in a pre- erence electrode over the distal tendon. A ground electrode was determined area that would least interfere with daily activities. The placed in the anterior thigh. neurostimulator was connected to the lead wires and inserted into Recording electrodes were placed bilaterally into the rectus fe- the subcutaneous pocket. The anterior wound was then irrigated moris (RF), vastus lateralis (VL), medial hamstring (MH), tibialis with bacitracin solution and closed in anatomic layers. anterior (TA), medial gastrocnemius (MG), and soleus (SOL). These muscles were selected to assess selectivity of EES to activate Post-operative electrophysiology proximal or distal muscle activation as well as to assess symmetry of activation between the legs.34 Once the array was inserted, minor After 3 weeks of rest, subjects returned to the laboratory to changes in position were guided based on leg muscle intramuscular record EES-evoked motor responses from leg muscles while lying electromyography (EMG) recordings of EES-evoked motor re- supine using EES parameters that were applied during in- sponses. EES-evoked motor response recordings were captured traoperative recordings (0–10 V, 0.5–1.0 Hz). during multiple EES configurations and stimulation parameters (0– 10 V, 0.5–1.0 Hz), progressing from rostral to caudal and wide field Epidural electrical stimulation–enabled motor activity to local. Additionally, EMG recordings were captured from the paraspinal muscles to record a stimulation artifact, which acted as During the first 5 days of testing with EES, both subjects at- the sweep trigger for real-time display of EES-evoked motor re- tempted to perform motor tasks in the presence of EES. Tasks sponses. included attempts to stand, control flexion/extension lower 1454 CALVERT ET AL. extremity movements while supine or side-lying, and move their Statistical analysis legs in response to an audio or visual cue. All tasks were performed with trainer assistance provided as needed. These tasks were per- EMG and goniometry data were collected at a sampling rate formed with the primary goal of identifying EES configurations and of 4 kHz (PowerLab; ADInstruments, Austin, TX) and analyzed stimulation parameters that enabled intentional control of lower using custom code written in MATLAB (Version R2015a; The extremity motor activity.12 MathWorks, Inc., Natick, MA). Notch (60 Hz) and bandpass (20– To determine whether EES could enable intentional control of 1000 Hz) filters were applied to EMG recordings to reduce envi- coordinated, robust leg muscle activity, subjects were positioned ronmental artifacts. EES-evoked motor response latencies were side lying with the top leg suspended using non-elastic nylon fabric calculated as the time from stimulus pulse to the onset of the motor support slings to allow gravity-neutral movement to identify EES response of each muscle; defined as the time at which the response parameters that enabled intentional control over robust rhythmic amplitude exceeded –3 standard deviations (SDs) from baseline muscle activity to generate step-like movements.39,40 A range of EMG values during quiescent recordings. Mean and SD values electrode configurations and voltage intensities were tested based were calculated from five consecutive stimuli. Root mean square on past literature in combination with intra- and post-operative (RMS) envelopes were calculated from full-wave rectified EMG interpretations of EES-evoked motor responses. EES voltage in- using a moving window of 200 samples and an overlap of 50 tensity was increased until volitional control over leg muscle ac- samples. Area under the curve values were calculated by taking the tivity and movement was observed.11,20,22,41 Skin-surface EMG trapezoidal integral of the calculated RMS EMG data. The Shapiro- recordings were captured over the same leg muscles as intra- and Wilk method was used to test for normal distributation of data. For post-operative EES-evoked motor response recordings. For subject normally distributed data, statistically significant differences be- 2, goniometer recordings of knee joint changes were synchronized tween motor response amplitudes were calculated using standard to EMG recordings. parametric t-tests to compare muscle activity. Data that were not FIG. 2. Intraoperative EES-evoked motor response recordings demonstrates selective activation of rostral and caudal spinal circuitry. (A) EES-evoked responses during rostral electrode array configurations demonstrate proximal muscle activation (rectus femoris) and during caudal electrode array configurations demonstrate distal muscle activation (medial gastrocnemius). Each line represents the average evoked response to stimulation over five stimulations. Gray, green, and orange lines represent stimulation at 3, 4, and 5 volts, respectively. Stimulation occurs at the 0 time point of each plot. Stimulation configuration is shown in the upper left of each figure; black = cathode, red = anode. (B) Bar plots displaying maximum evoked response in given muscles from (A) with amplitude calculated as maximum – minimum response. Blue bars indicate rostral configurations and red bars indicate caudal configurations. * = <0.05; ** = <0.01; *** = <0.001. EES, epidural electrical stimulation; NS, not significant; R, right; L, left. GUIDED PLACEMENT OF EPIDURAL ELECTRODE ARRAY 1455 normally distributed were analyzed using the Wilcoxon signed- portion of the array resulted in significantly higher amplitude MH rank test. To determine statistically significant differences in motor activity compared to all five other muscles (Fig. 3B,C). Stimulation activity across muscles, a one-way analysis of variance (ANOVA) of the middle portion of the array resulted in significantly higher was followed by a multiple comparisons test. response amplitudes in the RF and MH (Fig. 3B,C). Stimulation at the caudal portion of the array resulted in significantly higher re- sponses in the distal muscles compared to the proximal muscles, Results with significantly higher responses in the MG and SOL compared Electrode array positioning guided by intraoperative to the TA (Fig. 3B,C). epidural electrical stimulation–evoked To determine whether the array was positioned to allow sym- motor responses metrical activation of bilateral leg muscles, electrodes that were midline on the array were activated. In subject 2, initial positioning Initial positioning of the electrode array within the T11–L1 of the electrode array resulted in EES-evoked motor responses that vertebral region was performed using anatomical landmarks and were asymmetric in response amplitude. Asymmetry was charac- fluoroscopy (Figure 1). Activation of select electrodes on the array terized by increasing motor response amplitudes in right leg muscle evoked distinct motor responses. Specifically, positioning the ac- recordings as EES voltages increased. However, unremarkable tive electrodes at the rostral region of the array resulted in signifi- changes in motor response amplitudes were observed in left leg cantly higher EES-evoked motor response amplitudes in proximal muscle recordings as EES voltages were incrementally increased leg muscles compared to EES-evoked motor response amplitudes (Fig. 4A). In response to this observed asymmetry, the electrode generated when active electrodes were selected at the caudal region array was shifted to the left by 2 mm. Lateral adjustment in array of the array (Fig. 2; subject 1, p = 0.0002 and p = 0.0042 for left and positioning resulted in both the left and right leg muscles becoming right leg, respectively; subject 2, p = 0.0101 and p = 0.0110 for left active at lower EES voltages with significantly greater evoked motor and right leg, respectively). Stimulation through caudal electrode response amplitudes at multiple stimulation voltages in all left and configurations resulted in distal muscle EES-evoked motor re- right leg muscles, with the exception of the right SOL (Fig. 4B). sponse amplitudes that were significantly higher than rostral posi- tioning of active electrodes (Fig. 2; subject 1, p = 0.0055 and Characteristics of epidural electrical stimulation–evoked p = 0.0222 for left and right leg, respectively; subject 2, p = 0.0038 motor responses intraoperatively and at 3 weeks and p = 0.0017 for left and right leg, respectively). post-surgery Stimulating midline electrodes on the array at the rostral, middle, and caudal regions resulted in voltage-dependent selective activa- Three weeks after EES implantation, subjects were positioned tion of all six recorded muscles (Fig. 3A). Stimulation at the rostral supine and instructed to relax while EES-evoked motor responses FIG. 3. EES-evoked motor responses activate specific muscle circuitry intraoperatively. (A) EES-evoked responses during stimulation of the rostral, intermediate, and caudal portions of the EES array are demonstrated in six muscles (rectus femoris, vastus lateralis, medial hamstring, tibialis anterior, medial gastrocnemius, and soleus) from subject 1. Stimulation occurs at the start of each EES-evoked response trace as indicated by the gray dashed line. Dark traces are average of five individual responses that are shown in light traces. Data shown are from motor threshold responses. EES electrode array configuration is shown in the upper left; black = cathode, red = anode. (B) Bar plots displaying maximum EES-evoked responses in given muscles (RF = rectus femoris, VL = vastus lateralis, MH = medial hamstring, TA = tibialis anterior, MG = medial gastrocnemius, and SOL = soleus) from (A) with amplitude calculated as maximum – minimum EES-evoked response. Statistical significance was calculated by a one-way ANOVA followed by a multiple comparisons test and is shown above the bar plots. * = <0.05; ** = <0.01; *** = <0.001; no stars indicates not significant. ANOVA, analysis of variance; EES, epidural electrical stimulation. 1456 CALVERT ET AL. FIG. 4. Electrode location adjustment guided by intraoperative EES-evoked motor response recordings. (A) Intraoperative data from subject 2 using a caudal, symmetric (-10/+8) configuration as displayed. Bilateral electromyography (EMG) data from three bilateral distal muscles are shown (TA = tibialis anterior, MG = medial gastrocnemius, and SOL = soleus). Each line is an average of five motor- evoked potentials where stimulation occurs at the start of each trace. Data are shown while increasing the stimulation intensity incrementally from 5.5 to 6.3 V before and after shifting of the array during surgery. (B) Area under the curve of the EES-evoked responses at the four different voltages. * = <0.05; ** = <0.01; *** = <0.001; NS = not significant. Red indicates data before array shift. Blue indicates data after shift. EES, epidural electrical stimulation. were recorded using intraoperative electrode configurations. Leg rhythmic step-like activity was enabled. As shown in Figure 6, both muscle activity was recorded using EES voltages that were sub- subjects were able to make five consecutive steps while attempting threshold, threshold, and suprathreshold with respect to evoked step-like movements of the right leg. These consecutive steps were motor responses (Fig. 5A). EES-evoked motor responses differed characterized by reciprocal, EMG activity in all muscle recordings in deflection shape, peak amplitude, and number of peaks between of the right leg which was free to move in a gravity-neutral plane by intra- and post-operative recording time points. In the RF, one peak the support sling (Fig. 6B). For subject 2, right knee goniometer in motor response amplitude was observed intraoperatively, and recordings were synchronized to EMG, and demonstrated move- two distinct peaks were observed post-operatively in both subjects ment patterns that coincided with bursting EMG activity in the leg (Fig. 5B). However, latencies from the stimulus pulse to motor muscles. Supplementary Video 1 (see online supplementary ma- response onset in the RF (subject 1, mean = 11.60 – 0.38 and terial at http://www.liebertpub.com) shows the range of movement 10.85 – 1.07 ms for intra- and post-operative, respectively; subject and muscle activity observed during these attempts. These data and 2, mean = 11.10 – 0.14 and 11.15 – 0.14 ms for intra- and post- video are from the fifth and second day of EES testing for Subjects operative, respectively) were statistically insignificant between 1 and 2, respectively. Without EES, subjects were unable to in- intra- and post-surgical time points across time (subject 1, p = 0.30; tentionally generate any sustained, coordinated muscle activity in subject 2, p = 0.62; Fig. 5C). their legs. Intentional control of epidural electrical Discussion stimulation–enabled step-like activity while side-lying EES-enabled return of intentional leg movement, standing, and When positioned side-lying with one leg suspended in a gravity- independent stepping has been described in individuals with SCI neutral position by a sling system, both subjects demonstrated the who have performed motor training in the presence of ability to intentionally initiate, maintain, and terminate EES- EES.11,12,23,35,36,42,43 These scientific observations have created a enabled robust, rhythmic leg muscle activity, and step-like leg large interest within the SCI community and a sense of urgency in movements. First, motor threshold was detected as the voltage that translating EES into an effective rehabilitation tool for individuals resulted in EMG activity, which was achieved at 3.5 V for subject 1 with SCI-related deficits. As the scientific understanding of EES- and 2.5 V for subject 2. In both subjects, when voltage increased, enabled function continues to improve, along with the development GUIDED PLACEMENT OF EPIDURAL ELECTRODE ARRAY 1457 FIG. 5. Comparison of intraoperative and post-operative EES-evoked motor responses. (A) EES electrode array configuration used intraoperatively and post-operatively; black = cathode, red = anode. (B) EMG (electromyography) data are shown from the left rectus femoris of both subjects recorded intraoperatively and post-operatively using the same electrode configuration for each subject after 3 weeks of recovery from surgery. Each trace represents the average of five consecutive evoked motor responses at each EES voltage intensity. Data are shown at subthreshold levels of stimulation when no response was observed, at motor threshold where the first appearance of motor activity was observed, and at the maximum level of stimulation. These voltage values ranged from 3 to 6 V. (C) Latency of the suprathreshold evoked response in the left rectus femoris in both subjects both intraoperatively and at 3 weeks post- operatively. No significant difference was found between intraoperative and post-operative latency for either subject. EES, epidural electrical stimulation; NS, not significant. of next-generation neurotechnology to interface with the spinal terpreted from EES-evoked motor response recordings. Rostral cord, the therapeutic potential of EES continues to increase. Here, configurations preferentially activated proximal muscles, and we have detailed the surgical procedures of EES implantation for caudal configurations activated distal muscles in both subjects (see use as a foundation for future clinical investigations seeking to Fig. 2). Targeting EES to selectively activate these motor pools implant EES systems after SCI. We also provided the approach of proved useful to confirm that the electrode array spanned the in- using intraoperative EES-evoked motor responses to guide elec- tended region of the lumbosacral spinal cord in subject 1. Further, trode array positioning and compared those motor response char- when stimulation was then applied midline at rostral, intermediate, acteristics to post-surgical EES-evoked motor responses. and caudal portions of the array, specific spinal segments were Intraoperative electrophysiological monitoring is a valuable likely activated and resulted in higher amplitude EES-evoked re- technique for real-time assessment of nervous system health during sponses (see Fig. 3). Previous reports indicate that stimulation of surgery.44 Intraoperative monitoring in subjects with SCI has been these regions activates L1–S1 spinal segments.12,34 Stimulation at primarily used as a technique to identify impending or ongoing the rostral portions of the cord resulted in statistically larger re- injury and to decrease the risk of spinal cord insult during surgical sponses in the MH compared to all other muscles. Stimulation at the procedures. A systematic review has found that the procedure is intermediate portions of the array resulted in increased response sensitive and specific for detecting intraoperative injury, and a amplitude in the MH and RF muscles compared to the distal separate Monte Carlo simulation study found a 49% reduction in muscles. Finally, stimulation at the caudal portion resulted in in- risk in post-operative complications.45,46 Here, we demonstrated creased response amplitude in the distal muscles (TA, MG, and the utility of intraoperative monitoring to guide EES electrode SOL). Based on the array location, the rostral, intermediate, and positioning for post-surgical EES-enabled motor activity in 2 caudal stimulations were likely over the L1/L2, L3/L4, and L5/S1 subjects undergoing EES procedures. spinal segments, respectively. These results suggest that spatially Animal models have shown that localization of epidural elec- restricted stimulation ensures that localization of active electrodes trodes relative to the level of the dorsal roots plays a critical role in on the array can activate the most proximal or distal muscles of the determining the intensity of stimulation required to activate select leg, or excite select intermediate muscles. spinal circuitry, and this dependency can vary at a submillimeter In conjunction with the use of spatially restricted stimulation to resolution.37 Monitoring intraoperative electrophysiological re- target specific leg motor activity, the symmetry of the responses sponses in real time allowed for precise placement of the epidural was also investigated. In subject 2, the electrode array was initially electrode array. By applying stimuli at multiple different regions of found to be off-set to the right as determined by a symmetric the spinal cord using specific electrode configurations, selective electrode configuration, which generated activity in only the right activation of spinal sensorimotor networks was achieved and in- leg muscles (see Fig. 4). This asymmetry was detected and 1458 CALVERT ET AL. FIG. 6. Intentional control of rhythmic movement while side-lying. (A) EMG (electromyography) and goniometer data from both subjects are shown while each subject intentionally attempted to generate EES-enabled step-like movements of their right leg. EMG data were recorded bilaterally from rectus femoris (RF), medial hamstring (MH), tibialis anterior (TA), and medial gastrocnemius (MG). Subject 2 wore sagittal knee goniometers during testing to quantify leg motion. Stimulation configurations were chosen that allowed for optimal movement. Pulse width and frequency were held constant at 210 ls and 40 Hz, respectively. Voltage was incre- mentally increased until subjects displayed ability to intentionally control leg movements. White background indicates the leg that was supported in a gravity-neutral position in order to move freely. Gray background indicates the leg that was resting on a table and limited with respect to movement capability. (B) Muscle coordination plots from the same data as part (A). Root mean square (RMS) envelopes of the EMG data were calculated and antagonistic muscles are plotted against one another to demonstrate patterns of coordination. Note that a more normal L-shaped (reciprocal) coordination patterned was generated when the leg was suspended and freed of surface tension. EES, epidural electrical stimulation. quantified using intraoperative EES-evoked motor response re- after SCI, and they demonstrate the potential of EES to achieve cording equipment, and required a lateral shift of 2 mm by the effective recruitment patterns of muscle activity. Further, we neurosurgeon in order to increase symmetry of evoked motor re- demonstrate that within 5 days of testing, it is possible to achieve sponse amplitudes across left and right leg muscles. coordinated and robust motor output to generate leg muscle activity In addition to providing real-time monitoring of the EES-evoked that can be intentionally modulated to create step-like movements motor responses during surgery, intraoperative monitoring predicts (see Fig. 6). This demonstrates a fundamentally important result in post-operative evoked motor response characteristics and estab- the rehabilitative strategy for paralyzed individuals. Specifically, lishes the groundwork for optimizing EES stimulation parameters the unweighted condition reflects the ability of a novel supraspinal- and active electrode configurations to be used during post-surgical spinal connectome to generate an impressive degree of agonist- EES-enabled motor activities (see Fig. 5). EES-evoked motor re- antagonist coordination at such an early stage after spinal networks sponses varied in amplitude and shape, possibly attributed to a are neuromodulated. Theoretically, then, the challenge during re- range of factors including depth of anesthesia, intramuscular versus habilitation is to progressively improve the ability of the muscular skin electrode, and body positioning. However, evoked motor re- component to generate the desired torques in the presence of a sponse latencies were statistically similar across intra- and post- coordinated pattern of activation to achieve a given movement, and operative environments, indicating that intraoperative recordings that this may be achieved through task-specific training. of EES-evoked motor responses can be used to establish charac- Although the results we presented from these 2 subjects suggest teristics of sensorimotor networks activated by EES. that intra-operative EES-evoked motor responses foreshadow EES- The evoked potentials observed during surgical procedures and enabled motor activity performance post-surgery, they do not di- on the first day of stimulation after surgery provide considerable rectly demonstrate how these single evoked potentials can serve as clarity of the functional connectivity from the dorsum of the spinal a biomarker of the most effective stimulation parameters to perform cord to ventral motor pools and downstream neuromuscular cir- step-like behavior. There are multiple factors that contribute to the cuitry. These electrically evoked motor responses are a critical complexity of interactions between EES and the dynamic state of component used to assess sensorimotor spinal network activity sensorimotor spinal network excitability. This complex state of GUIDED PLACEMENT OF EPIDURAL ELECTRODE ARRAY 1459 excitability prohibits a direct correlation of EES-evoked motor 3. Bydon, M., Lin, J., Macki, M., Gokaslan, Z.L., and Bydon, A. (2014). potential characteristics to EES-enabled motor functions. For ex- The current role of steroids in acute spinal cord injury. World Neu- rosurg. 82, 848–854. ample, changing the stimulation from a single pulse to a train of 4. Chen, B.K., Madigan, N.N., Hakim, J.S., Dadsetan, M., McMahon, pulses or adjustment of stimulus frequency cannot readily define S.S., Yaszemski, M.J., and Windebank, A.J. (2018). GDNF Schwann the optimal stimulation parameters for generating a desired motor cells in hydrogel scaffolds promote regional axon regeneration, re- function. Another factor that makes the translation from evoked myelination and functional improvement after spinal cord transection potentials to motor outcomes difficult is that the responsiveness of in rats. J. Tissue Eng. Regen. Med. 12, e398–e407. 5. Anderson, M.A., Burda, J.E., Ren, Y., Ao, Y., Coppola, G., Khakh, the spinal network intraoperatively and soon after implantation is B.S., Deming, T.J., Michael, V., and Angeles, L. (2016). Astrocyte likely to change substantially as motor function improves. One of scar formation aids CNS axon regeneration. Nature 532, 195–200. the most obvious differences when stimulating single evoked re- 6. Lang, B.T., Cregg, J.M., Depaul, M.A., Tran, A.P., Xu, K., Dyck, sponses versus continuous stimulation is that the spinal networks S.M., Madalena, K.M., Brown, B.P., Weng, Y.L., Li, S., Karimi- Abdolrezaee, S., Busch, S.A., Shen, Y., and Silver, J. (2015). Mod- activated in these two conditions are dramatically different, quan- ulation of the proteoglycan receptor PTPr promotes recovery after titatively and qualitatively. For example, the behavioral responses spinal cord injury. Nature 518, 404–408. reflect the engagement of proprioception and a return of supraspinal 7. Capogrosso, M., Milekovic, T., Borton, D., Wagner, F., Moraud, connectivity. With our currently limited understanding of the E.M., Mignardot, J.B., Buse, N., Gandar, J., Barraud, Q., Xing, D., mechanisms underlying the functional connections that generate a Rey, E., Duis, S., Jianzhong, Y., Ko, W.K.D., Li, Q., Detemple, P., Denison, T., Micera, S., Bezard, E., Bloch, J., and Courtine, G. (2016). motor behavior, it cannot be expected that the response of the final A brain-spine interface alleviating gait deficits after spinal cord injury common pathways to each motor pool would be similar to a single in primates. Nature 539, 284–288. evoked response compared to the ensembles of proprioceptive and 8. Wenger, N., Moraud, E.M., Gandar, J., Musienko, P., Capogrosso, M., supraspinal inputs projecting to spinal sensorimotor networks. Baud, L., Le Goff, C.G., Barraud, Q., Pavlova, N., Dominici, N., Minev, I.R., Asboth, L., Hirsch, A., Duis, S., Kreider, J., Mortera, A., EES has shown promise to improve function previously lost Haverbeck, O., Kraus, S., Schmitz, F., DiGiovanna, J., Van Den attributed to SCI, and this study provides further evidence of the Brand, R., Bloch, J., Detemple, P., Lacour, S.P., Bézard, E., Micera, ability of EES to enable motor function in 2 subjects with motor S., and Courtine, G. (2016). Spatiotemporal neuromodulation thera- complete SCI. Here, we demonstrate that intraoperative EMG re- pies engaging muscle synergies improve motor control after spinal cordings can be used to achieve adequate array positioning by cord injury. Nat. Med. 22, 138–145. 9. Flesher, S.N., Collinger, J.L., Foldes, S.T., Weiss, J.M., Downey, J.E., spatially restricted stimulation, that similar spinal structures are Tyler-Kabara, E.C., Bensmaia, S.J., Schwartz, A.B., Boninger, M.L., stimulated intraoperatively, and post-operatively, and that stimu- and Gaunt, R.A. (2016). Intracortical microstimulation of human so- lation of these structures can enable coordinated motor function matosensory cortex. Sci. Transl. Med. 8, 1–11. within the first few days of EES. The surgical procedures and 10. Collinger, J.L., Wodlinger, B., Downey, J.E., Wang, W., Tyler- Kabara, E.C., Weber, D.J., McMorland, A.J.C., Velliste, M., Bo- electrophysiological results presented lay the groundwork for fu- ninger, M.L., and Schwartz, A.B. (2013). High-performance neuro- ture investigations using EES in humans with SCI. prosthetic control by an individual with tetraplegia. Lancet 381, 557– 564. Acknowledgments 11. Angeli, C.A., Edgerton, V.R., Gerasimenko, Y.P., and Harkema, S.J. (2014). Altering spinal cord excitability enables voluntary movements We thank the subjects for participating in this study. We also after chronic complete paralysis in humans. Brain 137, 1394–1409. thank Dr. Jingfei Chen, Dr. Beth Cloud-Biebl, Dina Drubach, Terri 12. Grahn, P.J., Lavrov, I.A., Sayenko, D.G., Van Straaten, M.G., Gill, Gardner, Deb Hare, Yulun Li, Dr. Aldo Mendez, Andrew Thor- M.L., Strommen, J.A., Calvert, J.S., Drubach, D.I., Beck, L.A., Linde, M.B., Thoreson, A.R., Lopez, C., Mendez, A.A., Gad, P.N., Ger- eson, Andy Schmeling, Tyson Scrabeck, Marc Shaft, Cindy Stop- asimenko, Y.P., Edgerton, V.R., Zhao, K.D., and Lee, K.H. (2017). pel, Bambi Wessel, and the surgical support team. Enabling task-specific volitional motor functions via spinal cord This work was supported by the Jack Jablonski Bel13ve in neuromodulation in a human with paraplegia. Mayo Clin. Proc. 92, Miracles Foundation, Craig H. Neilsen Foundation, Mayo Clinic 544–554. 13. Dietz, V., and Fouad, K. (2014). Restoration of sensorimotor functions Center for Regenerative Medicine, Mayo Clinic Graduate School after spinal cord injury. Brain 137, 654–667. of Biomedical Sciences, Mayo Clinic Transform the Practice, 14. Ichiyama, R.M., Gerasimenko, Y.P., Zhong, H., Roy, R.R., and Minnesota Government Office of Higher Education–SCI/TBI Edgerton, V.R. (2005). Hindlimb stepping movements in complete Grant, Regenerative Medicine Minnesota, The Grainger Founda- spinal rats induced by epidural spinal cord stimulation. Neurosci. Lett. tion, Mayo Clinic Rehabilitation Medicine Research Center, Dana 383, 339–344. 15. Lavrov, I., Gerasimenko, Y.P., Ichiyama, R.M., Courtine, G., Zhong, and Albert R. Broccoli Charitable Foundation, and Walkabout H., Roy, R.R., Edgerton, V.R., Gerasimenko, Y.P., Ichiyama, R.M., Foundation. Courtine, G., Zhong, H., Roy, R.R., and Reggie, V. (2006). Plasticity of spinal cord reflexes after a complete transection in adult rats : Author Disclosure Statement relationship to stepping ability. J. Neurophysiol. 96, 1699–1710. 16. Gerasimenko, Y.P., Ichiyama, R.M., Lavrov, I.A., Courtine, G., Cai, V.R.E. and Y.P.G. are shareholders in NeuroRecovery Tech- L., Zhong, H., Roy, R.R., and Edgerton, V.R. (2007). Epidural spinal nologies. V.R.E. is president and chair of the company’s board of cord stimulation plus quipazine administration enable stepping in complete spinal adult rats. J. Neurophysiol. 98, 2525–2536. directors. V.R.E. and Y.P.G. hold investorship rights on intellectual 17. Lavrov, I., Dy, C.J., Fong, A.J., Gerasimenko, Y., Courtine, G., property licensed by the regents of the University of California to Zhong, H., Roy, R.R., and Edgerton, V.R. (2008). Epidural stimula- NeuroRecovery Technologies and its subsidiaries. tion induced modulation of spinal locomotor networks in adult spinal rats. J. Neurosci. 28, 6022–6029. References 18. Ichiyama, R.M., Courtine, G., Gerasimenko, Y.P., Yang, G.J., van den Brand, R., Lavrov, I.A., Zhong, H., Roy, R.R., and Edgerton, V.R. 1. Harvey, L.A. (2016). Physiotherapy rehabilitation for people with (2008). Step training reinforces specific spinal locomotor circuitry in spinal cord injuries. J. Physiother. 62, 4–11. adult spinal rats. J. Neurosci. 28, 7370–7375. 2. Sekhon, L.H., and Fehlings, M.G. (2001). Epidemiology, demo- 19. Dimitrijevic, M.R., Gerasimenko, Y., and Pinter, M.M. (1998). Evi- graphics, and pathophysiology of acute spinal cord injury. Spine (Phila dence for a spinal central pattern generator in humans. Ann. N. Y. Pa 1976) 26, 24 Suppl., S2–S12. Acad. Sci. 860, 360–376. 1460 CALVERT ET AL. 20. Minassian, K., Jilge, B., Rattay, F., Pinter, M.M., Binder, H., Ger- 34. Sayenko, D.G., Angeli, C., Harkema, S.J., Edgerton, V.R., and Ger- stenbrand, F., and Dimitrijevic, M.R. (2004). Stepping-like move- asimenko, Y.P. (2014). Neuromodulation of evoked muscle potentials ments in humans with complete spinal cord injury induced by epidural induced by epidural spinal-cord stimulation in paralyzed individuals. stimulation of the lumbar cord: electromyographic study of compound J. Neurophysiol. 111, 1088–1099. muscle action potentials. Spinal Cord 42, 401–416. 35. Rejc, E., Angeli, C.A., Atkinson, D., and Harkema, S.J. (2017). Motor 21. Minassian, K., Persy, I., Rattay, F., Pinter, M.M., Kern, H., and Di- recovery after activity-based training with spinal cord epidural stim- mitrijevic, M.R. (2007). Human lumbar cord circuitries can be acti- ulation in a chronic motor complete paraplegic. Sci. Rep. 7, 13476. vated by extrinsic tonic input to generate locomotor-like activity. 36. Rejc, E., Angeli, C., and Harkema, S. (2015). Effects of lumbosacral Hum. Mov. Sci. 26, 275–295. spinal cord epidural stimulation for standing after chronic complete 22. Harkema, S., Gerasimenko, Y., Hodes, J., Burdick, J., Angeli, C., paralysis in humans. PLoS One 10, e0133998. Chen, Y., Ferreira, C., Willhite, A., Rejc, E., Grossman, R.G., and 37. Cuellar, C.A., Mendez, A.A., Islam, R., Calvert, J.S., Grahn, P.J., Edgerton, V.R. (2011). Effect of epidural stimulation of the lumbo- Knudsen, B., Pham, T., Lee, K.H., and Lavrov, I.A. (2017). The role sacral spinal cord on voluntary movement, standing, and assisted of functional neuroanatomy of the lumbar spinal cord in effect of stepping after motor complete paraplegia: a case study. Lancet 377, epidural stimulation. Front. Neuroanat. 11, 82. 1938–1947. 38. Kirshblum, S.C., Waring, W., Biering-Sorensen, F., Burns, S.P., Jo- 23. Gill, M.L., Grahn, P.J., Calvert, J.S., Linde, M.B., Lavrov, I.A., hansen, M., Schmidt-Read, M., Donovan, W., Graves, D.E., Jha, A., Strommen, J.A., Beck, L.A., Sayenko, D.G., Van Straaten, M.G., Jones, L., Mulcahey, M.J., and Krassioukov, A. (2011). Reference for Drubach, D.I., Veith, D.D., Thoreson, A.R., Lopez, C., Gerasimenko, the 2011 revision of the international standards for neurological Y.P., Edgerton, V.R., Lee, K.H., and Zhao, K.D. (2018). Neuromo- classification of spinal cord injury. J. Spinal Cord Med. 34, 547–554. dulation of lumbosacral spinal networks enables independent stepping 39. Gerasimenko, Y.P., Lu, D.C., Modaber, M., Zdunowski, S., Gad, P., after complete paraplegia. Nat. Med. 24, 1677–1682. Sayenko, D.G., Morikawa, E., Haakana, P., Ferguson, A.R., Roy, 24. Carhart, M.R., He, J.P., Herman, R., D’Luzansky, S., and Willis, W.T. R.R., and Edgerton, V.R. (2015). Noninvasive reactivation of motor (2004). Epidural spinal-cord stimulation facilitates recovery of func- descending control after paralysis. J. Neurotrauma 32, 1968–1980. tional walking following incomplete spinal-cord injury. Ieee Trans. 40. Gerasimenko, Y., Gorodnichev, R., Moshonkina, T., Sayenko, D., Neural Syst. Rehabil. Eng. 12, 32–42. Gad, P., and Reggie Edgerton, V. (2015). Transcutaneous electrical 25. Lu, D.C., Edgerton, V.R., Modaber, M., AuYong, N., Morikawa, E., spinal-cord stimulation in humans. Ann. Phys. Rehabil. Med. 58, 225– Zdunowski, S., Sarino, M.E., Sarrafzadeh, M., Nuwer, M.R., Roy, 231. R.R., and Gerasimenko, Y. (2016). Engaging cervical spinal cord 41. Danner, S.M., Hofstoetter, U.S., Freundl, B., Binder, H., Mayr, W., networks to reenable volitional control of hand function in tetraplegic Rattay, F., and Minassian, K. (2015). Human spinal locomotor control patients. Neurorehabil. Neural Repair 30, 951–962. is based on flexibly organized burst generators. Brain 138, 577–588. 26. Harkema, S.J., Wang, S., Angeli, C.A., Chen, Y., Boakye, M., Ugi- 42. Rejc, E., Angeli, C.A., Bryant, N., and Harkema, S.J. (2017). Effects liweneza, B., and Hirsch, G.A. (2018). Normalization of blood pres- of stand and step training with epidural stimulation on motor function sure with spinal cord epidural stimulation after severe spinal cord for standing in chronic complete paraplegics. J. Neurotrauma 34, injury. Front. Hum. Neurosci. 12, 1–11. 1787–1802. 27. West, C.R., Phillips, A.A., Squair, J.W., Williams, A.M., Walter, M., 43. Angeli, C.A., Boakye, M., Morton, R.A., Vogt, J., Benton, K., Chen, Lam, T., and Krassioukov, A.V. (2018). Association of epidural Y., Ferreira, C.K., and Harkema, S.J. (2018). Recovery of over-ground stimulation with cardiovascular function in an individual with spinal walking after chronic motor complete spinal cord injury. N. Engl. J. cord injury. JAMA Neurol. 75, 630–632. Med. 379, 1244–1250. 28. Hachmann, J.T., Grahn, P.J., Calvert, J.S., Drubach, D.I., Lee, K.H., 44. Stecker, M.M. (2012). A review of intraoperative monitoring for and Lavrov, I.A. (2017). Electrical neuromodulation of the respiratory spinal surgery. Surg. Neurol. Int. 3, Suppl. 3, S174–S187. system after spinal cord injury. Mayo Clin. Proc. 92, 1401–1414. 45. Fehlings, M.G., Brodke, D.S., Norvell, D.C., and Dettori, J.R. (2010). 29. Hachmann, J.T., Calvert, J.S., Grahn, P.J., Drubach, D.I., Lee, K.H., The evidence for intraoperative neurophysiological monitoring in and Lavrov, I.A. (2017). Review of epidural spinal cord stimulation spine surgery: does it make a difference? Spine (Phila. Pa. 1976). 35, 9 for augmenting cough after spinal cord injury. Front. Hum. Neurosci. Suppl., S37–S46. 11, 144. 46. Ney, J.P., van der Goes, D.N., and Watanabe, J.H. (2012). Cost- 30. Kowalski, K.E., Hsieh, Y.-H., Dick, T.E., and DiMarco, A.F. (2013). effectiveness of intraoperative neurophysiological monitoring for Diaphragm activation via high frequency spinal cord stimulation in a spinal surgeries: beginning steps. Clin. Neurophysiol. 123, 1705– rodent model of spinal cord injury. Exp. Neurol. 247, 689–693. 1707. 31. Kowalski, K.E., Romaniuk, J.R., Brose, S., Richmond, M.A., Ko- walski, T., and DiMarco, A.F. (2016). High frequency spinal cord stimulation—new method to restore cough. Respir. Physiol. Neuro- Address Correspondence to: biol. 232, 54–56. Kendall H. Lee, MD, PhD 32. Gad, P.N., Roy, R.R., Zhong, H., Lu, D.C., Gerasimenko, Y.P., and Department of Neurologic Surgery Edgerton, V.R. (2014). Initiation of bladder voiding with epidural Mayo Clinic stimulation in paralyzed, step trained rats. PLoS One 9, e108184. 200 First Street Southwest 33. Terson de Paleville, D.G.L., Harkema, S.J., and Angeli, C.A. (2018). Epidural stimulation with locomotor training improves body composi- Rochester, MN 55901 tion in individuals with cervical or upper thoracic motor complete spinal cord injury: a series of case studies. J. Spinal Cord Med. 14, 1–7. E-mail: lee.kendall@mayo.edu