Modulatory Effect of Peripheral Magnetic and Neuromuscular Electrical Stimulation on Cortical Excitability: A Functional Near-Infrared Spectroscopy Study

17 Background: The present study was designed to investigate the effects of 18 neuromuscular electrical stimulation (NMES) and peripheral magnetic stimulation 19 (PMS) applied to the wrist extensor muscle on the cortical activity of healthy adults by 20 using fNIRS. 21 Methods: Fifteen healthy adult subjects (7 males, mean age: 27.13 ± 4.52 years) all 22 received two different conditions of peripheral muscle stimulation in random order: (1) 23 NMES and (2) PMS. The sessions were separated by at least 48 h as a washout period. 24 During muscle stimulation, the motor evoked potential (MEP) of the left primary motor 25 cortex (M1) was measured by transcranial magnetic stimulation (TMS) and the 26 concentration of oxygenated (HbO) and deoxygenated (HbR) hemoglobin detected by 27 fNIRS were used to evaluate the excitability and the activity of the cortex. 28 Results: After the stimulation of the wrist extensor, the MEP amplitude in the left M1 29 area did not change in both conditions, and there was no difference between NMES and 30 PMS condition. NMES reduced HbO values of several channels in the Prefrontal cortex 31 (PFC), Somatosensory motor cortex (SMC) and Occipital cortex (OC), and HbR valus 32 of several channels in the PFC and SMC. During the PMS stimulation period, the HbO 33 value of all brain areas did not change significantly, while the HbR value of the SMC 34 area decreased. The HbO and HbR value of the channels in the SMC did not differ 35 between NMES and PMS. Inter-region of interest and inter-channel analysis between 36 NMES and PMS showed no difference in functional connectivity. 37 Conclusions: In the case of wrist extensor muscle stimulation, both NMES and PMS 38 can induce cortical activation. PMS targeted to increases the activity of the contralateral 39 SMC, while NMES increased contralateral SMC activity and negatively activated the 40 PFC and OC. 41


Introduction 48
Neuromuscular electrical stimulation (NMES) is a classic non-invasive peripheral 49 stimulation (NIPS) method. It is performed by applying an electric current to the muscle 50 or peripheral nerve. In general, NMES has been applied alone or in combination with 51 other rehabilitation measures for rehabilitation after stroke[1, 2], chronic obstructive 52 pulmonary disease [3], muscle weakness, and musculoskeletal diseases (low back pain, 53 hip and knee arthroplasty, anterior cruciate ligament) [4]. In essence, the mechanism of 54 NMES is that electrical current delivery to neuromuscular tissue causes the 55 depolarization of the motor axons to indirectly activate fiber contraction, When the 56 intensity of NMES exceeds the motor threshold (MT), an upward afferent signal is 57 generated, and then the muscle contraction induced by the electrical stimulation causes 58 a re-afferent. Transcranial magnetic stimulation (TMS) [5] and 59 Electroencephalography (EEG) studies [6] have found that NMES can affect the 60 excitability of the primary sensory (S1) and motor cortex (M1) when applied to the first 61 dorsal interosseous (FDI) or abductor pollicis brevis (APB) muscle. This excitatory 62 change is generally believed to reflect the restoration of brain function and 63 reorganization of brain networks [7]. Previous studies demonstrated that peripheral 64 stimulation may eventually affect cerebral functional recovery and reconfiguration of 65 brain networks [8,9], thereby improving motor performance in patients with brain 66 injury [10]. 67 Peripheral magnetic stimulation (PMS) is a new NIPS technique that applies high-68 intensity magnetic field to the periphery. The application of its magnetic coil to the 69 spinal root, nerve, or muscle belly has a similar effect to NMES [11]. Moreover, PMS 70 does not require skin contact and does not cause pain during the procedure, which 71 makes it applicable to patients with paresthesia and to perform deep stimulation. These 72 unique advantages of PMS make it an alternative to NMES. Moreover, PMS can cause 73 changes in cortical excitability by inducing proprioceptive input to the central nervous 74 system (CNS) through magnetic stimulation. It has two different mechanisms: 1) the 75 rhythmic contraction and relaxation of muscles induced by indirect stimulation lead to 76 adequate activation of mechanoreceptors (fiber groups: Ia, Ib, II), and 2) direct stimulus 77 of sensory motor fibers induce inadequate activation of sensorimotor nerve fibers [12]. 78 Considering the after-effect and no pain in clinical application, PMS is a new 79 rehabilitation technology with more potential than NMES [13,14]. 80 Functional near-infrared spectroscopy (fNIRS) is a non-invasive, real-time, and 81 continuous optical technique that is used to measure cortical activities by measuring 82

oxygenated ([HbO]) and deoxygenated ([HbR]) hemoglobin concentrations during task. 83
That is, neural activity rapidly increases local blood flow to meet transient changes in 84 local brain energy requirements [15]. As a new detection method, fNIRS has higher 85 temporal resolution and higher tolerance to motion artifacts than fMRI, but very low 86

Measurement of motor evoked potentials 154
Transcranial magnetic stimulation (TMS) was performed with an OSF-pTMS magnetic 155 stimulator (O.SELF Company, Wuhan, China) with a figure-of-eight-shaped coil, 156 which can be used in the single-pulse assessment paradigm and rTMS paradigm. To 157 assess cortical excitability, a pair of Ag/AgCl surface electrodes were placed on the 158 belly of the FDI muscle of the right hand, and the surface electromyography signals can 159 be observed on a computer screen. The coil was positioned at a 45° tangent to the skull 160 in the left M1, and the center of the coil was moved within a range of 0.5 cm each time 161 in the motor cortex until we found the optimal site that could induce the maximum MEP 162 amplitude. The resting motor threshold (RMT) and motor evoked potentials (MEPs) 163 were examined by single-pulse TMS parameters. The RMT was defined as the minimal 164 stimulation intensity that can induce at least five trials with MEP peak-peak wave 165 amplitude > 50 μV when the FDI muscles were continuously stimulated for 10 trials. 166 In both conditions, MEPs amplitude was recorded before resting-state fNIRS 167 monitoring and immediately after muscle stimulation. The MEP measured intensity 168 was the intensity with peak-peak wave value at 1 mV intensity before the intervention. 169 Ten consecutive TMS pulses were spaced by at least 5 s.  HomER 197 provides the user a wide selection of function processing tools to choose from, 198 depending on their needs. In this study, we first used the hmrIntensity2OD Utility 199 function to converts the raw optical intensity into OD optical density data. Then, the 200 hmrMotionArtifactByChannel tool was used to identify motion artifacts in the data 201 matrix. STDev-thresh was set at 10, and AMP-thresh was set at 5. Motion artifacts were 202 removed by using filtering methods based on spline interpolation. Bandpass filtering 203 was used to remove unwanted specific frequency content. According to the muscle 204 stimulation protocol frequency, we set the high pass filter at 0.01 Hz and the low pass 205 filter at 0.1 Hz. The hmrOD2Conc function was used to convert the signals into 206 oxyhemoglobin and deoxyhemoglobin via the Beer-Lambert equation, partial 207 pathlength factors for each wavelength was 6.0. Finally, the hmrBlockAvg function 208 was used to average the time series data at -5 to 40 s, the baseline of the average is set 209 to 0 by subtracting the mean of the average for -5 to 0 s. (SMC) (ch20, ch22, ch23, ch24, ch25, ch26, ch35, ch36), right SMC (ch17, ch18, ch27, 225 ch28, ch29, ch30, ch31, ch33), left occipital cortex (OC) (ch51, ch54, ch56, ch61, ch62, 226 ch63), and right OC (ch43, ch45, ch47, ch49, ch50, ch58). NirSpark's network analysis 227 maps the connections of inter-ROIs and inter-channel (similarity threshold was set as 228 0.5, 0.6, 0.7, and 0.8, respectively) during different stimulation conditions. And the 229

392
The purpose of this study was to explore the after-effects and potential mechanisms of 393 NMES and PMS on cortical activation when applied to the dominant wrist extensor 394 muscles. In the current study, cortical excitability was not changed when NMES or 395 PMS was applied to the forearm muscles to induce wrist extension, however, changes 396 in cortical activation were observed during the stimulation. NMES causes a larger area 397 of negative activation in non-stimulated brain areas, and the effect of activating the 398 corresponding cortex is weak, while PMS focuses on activating the cortex 399 corresponding to the stimulated area.

400
NMES is often applied to finger and wrist muscles to induce repetitive movements 401 to improve the efficiency of the hand in performing motor tasks by modulating the 402 cortical activity or excitability of the brain the left S1, and the results showed a decrease in oxygenated hemoglobin and an 459 inhibition of MEP amplitude in both the right M1 and S1. Our current study shows that 460 both NMES and PMS applied to the wrist extensor muscles will increase the MEP 461 amplitude of left M1. However, there was no statistical difference compared with pre-462 intervention, nor was there any significant difference in cortical excitability between 463 conditions. Corresponding to changes in cortical excitability, during repeated passive In the present study, the intensity of NMES was above the MT, and the intensity of 475 PMS was at a high frequency of 10Hz. Our study did not provide a significant 476 regulatory effect of NMES and PMS on the cortex, which could be related to a 477 combination of many factors, such as the anatomic site of the stimulus, the stimulus 478 parameters, and the timing of the test [32,39]. Referring to previous studies, we 479 hypothesized that the differences in the parameters used in the study may lead to a 480 discrepancy between the results of our study and those of previous studies [40,41]. So 481 far, there is no consensus on the best parameters for NMES and PMS application. important factor affecting its effect on cortical excitability. According to a study, 20 500 and 40 min of NMES at 30Hz intensity were strong enough to produce a "voluntary" 501 contraction of the muscles, resulting in cortical excitability facilitation [46]. A short 10-502 min NMES intervention in our experiment could temporarily alter HbO levels and 503 activate brain regions, but it had no lasting effect. More studies have shown that 2 h of 504 supra-motor threshold intensity NMES can not only increase the signal intensity of S1, 505 M1, and PMd of the brain, but also last for 60min after the stimulation is stopped [47]. 506 In previous PMS studies, different frequencies were used. Most studies agree that 507 higher high-frequency can produce stronger and lasting effects than lower high- There are still some limitations in our research. First, after the intervention, fNIRS was 527 not used to observe cerebral blood flow, and TMS was only used to assess the 528 immediate effects of the stimulus. According to previous studies, MEP amplitude, HbO 529 and HbR concentration also changed with time after stimulation. Second, magnetic 530 stimulation equipment has the function of protecting the brain and preventing the coil 531 from overheating that limits our choice of optimal parameters for peripheral stimulation. 532 Also, while the parameters of NMES and PMS need to be similar, real-time fNIRS 533 measurement is also required. The short stimulus time and insufficient intensity in our 534 study were the main reasons for the absence of observed cortical excitability in TMS 535 assessment. In our research, we have obtained some meaningful results, and we firmly 536 believe that this is important to the promotion of NMES and PMS in the field of brain 537 rehabilitation. In addition, we will perfect the experimental design to further explore 538 the effects of NMES and PMS on the cortical activity and motor function of patients 539 with brain injury. The study design was approved by Institutional Review Board of Huashan Hospital, 588 Fudan University (reference number: #2019-609). All subjects gave written informed 589 consent prior to their participation in the study. TMS-MEP and resting state fNIRS were assessed at the beginning of each condition, then one of the two muscle stimulation interventions (i.e. NMES or PMS) was applied, and fNIRS was also assessed during stimulation. After the intervention, cortical excitability was reassessed. Probes placement. The locations of fNIRS detectors and sources are indicated by the blue and red circles, respectively, and the numbers between the circles indicate the channel numbers. The distance between the luminous source and the detector is 3 cm. According to the MNI spatial coordinates, the channels in the green region are located in the PFC, the channels in the blue region are located in the SMC, and the channels in the yellow region are located in the OC. Channels 23 and 35 correspond to the left forearm motor cortexchannels18 and 30 correspond to the right forearm motor cortex. There are 64 channels in total, and only 40 channels in the color covered area are used for observation and analysis.   Averaged Hemodynamics response (0-40 s) for HbO (red) and HbR (blue) of whole channels. a during NMES condition, HbO in PFC, right SMC and OC regions were decreased; after multiple comparisons correction, there were signi cant differences in ch8 (pcorrected = 0.042), ch13 (pcorrected = 0.031), ch11 (pcorrected = 0.048) of the left PFC; ch1 (pcorrected = 0.047), and ch14 (pcorrected = 0.044) of the right PFC; ch20 (pcorrected = 0.027), and ch24 (pcorrected = 0.048) of the left SMC; ch17 (pcorrected = 0.032), ch28 (pcorrected = 0.032),ch29 (pcorrected = 0.049), and ch33 (pcorrected = 0.044) of the right SMC; ch51 (pcorrected = 0.035), ch56 (pcorrected = 0.049), ch61 (pcorrected = 0.043), ch62 (pcorrected = 0.024), and ch63 (pcorrected = 0.049) of the left OC; ch43 (pcorrected = 0.047), ch45 (pcorrected = 0.049), ch47 (pcorrected = 0.045), ch49 (pcorrected = 0.034), ch50 (pcorrected = 0.028), and ch58 (pcorrected = 0.043) of the right OC. b during PMS condition, HbO increased in the left forearm motor cortex, while decreased in right SMC and OC regions. However, there was not signi cant after Benjamini-Hochberg multiple comparisons correction (left SMC: ch35 (p = 0.014, pcorrected = 0.540); right SMC: ch29 (p = 0.038, pcorrected = 0.303), ch30 (p = 0.029, pcorrected = 0.292); left OC: ch51 (p = 0.026, pcorrected = 0.341), ch56 (p = 0.016, pcorrected = 0.314)); The HbR value of ch23 (pcorrected = 0.014), ch26 (pcorrected = 0.012), ch27 (pcorrected = 0.011), ch31 (pcorrected = 0.005), and ch35 (pcorrected = 0.005)). According to the MNI coordinates, channels without channel labels are not in our observation area and are not used for statistical analysis. P-values were adjusted for multiple comparisons using the Benjamini-Hochberg correction method. *pHbO < 0.05, #pHbR < 0.05.   Cortical activation maps. HbO activation (beta scores) maps during a NMES and b PMS tasks. The picture comes from the group GLM analysis of the fNIRS data during stimulation task using Nirspark.