Intensity-dependent corticospinal facilitation by repetitive peripheral magnetic stimulation: Evidence for a major contribution of group I afferents

1

Repetitive peripheral magnetic stimulation (PMS) is increasingly used in 2 neurorehabilitation, yet the optimal stimulation intensity for inducing corticospinal facilitation 3 and the underlying afferent mechanisms remain unclear. We investigated the 4 intensity-dependent effects of repetitive PMS on corticospinal excitability and single motor 5 unit responses, and tested group I afferent contribution. 6

Healthy participants received repetitive PMS (25 Hz; 2-s ON/2-s OFF) over the 7 extensor carpi radialis (ECR) in a crossover design at 0.9× motor threshold (MT), 1.2×MT, and 8 high intensity sufficient to induce maximal wrist dorsiflexion (mean 1.8×MT). Motor-evoked 9 potentials (MEPs) elicited by transcranial magnetic stimulation were recorded from the ECR 10 and flexor carpi radialis (FCR) before and during the intervention (total 15 min). The lasting 11 effects were assessed after 9 min of high-intensity PMS for 50 min. To examine group I 12 afferent contribution, the same high-intensity protocol was applied during upper-arm ischemia 13 after reducing the ECR H-reflex to <10% of baseline. Sensory–motor input characteristics 14 across stimulation intensities were compared using post-stimulus time histograms of ECR 15 single motor unit firings during weak voluntary contraction. 16

High-intensity PMS significantly increased ECR MEPs after 9 min of intervention, 17 whereas 1.2×MT of PMS required 15 min to induce a marked effect. PMS at 0.9×MT did not 18 induce significant MEP changes. Across all intensities, the FCR MEPs remained unaltered. 19 ECR MEPs remained markedly elevated for up to 30 min after 9 min of high-intensity PMS. In 20 contrast, PMS delivered during ischemia produced no MEP enhancement. The motor unit 21 analysis revealed that suprathreshold PMS elicited an early peak in firing 22 probability—consistent with monosynaptic Ia excitation—whose amplitude increased with 23 stimulation intensity, whereas PMS at 0.9×MT produced no discernible peak. 24

Repetitive PMS above MT facilitates corticospinal excitability in an 25 intensity-dependent manner. Facilitation was abolished during ischemia. Together with the 26 presence of a short-latency peak in motor unit firing via a monosynaptic pathway, this finding 27 supports a major contribution of large-diameter muscle afferents, with a substantial Ia 28 component, to PMS-induced corticospinal facilitation. 29 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 2 30

Group Ia afferents, Humans, Neuromodulation, Rehabilitation, Sensory inputs, 31 Transcranial magnetic stimulation 32 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 3

33 The corticospinal tract is the primary pathway that mediates voluntary motor control in humans 34 (1, 2). Transcranial magnetic stimulation (TMS)–elicited motor-evoked potentials (MEPs) 35 provide a noninvasive measure of corticospinal excitability (CSE) (3). CSE changes have been 36 closely linked to motor impairment and recovery after central nervous system (CNS) lesions 37 (4), and to motor performance improvement (5, 6). These associations underscore CSE’s 38 central role in motor function and its modulation as a potential therapeutic target. 39 Repetitive peripheral magnetic stimulation (PMS) has emerged as a potential adjuvant 40 therapy in physical rehabilitation and is increasingly used to promote motor recovery after 41 CNS lesions (7-11). Accumulating evidence suggests that repetitive PMS facilitates motor 42 recovery by enhancing cortical excitability (7, 12, 13). However, despite these promising 43 findings, the optimal stimulation intensity for inducing neuroplastic changes remains poorly 44 understood. Previous studies have applied various PMS intensities, including levels inducing 45 palpable muscle contractions or small joint movements (7, 9, 10, 12-16) and levels producing 46 maximal joint movements (17, 18). Others have prescribed PMS as a fixed percentage of the 47 maximal stimulator output without calibrating intensity to the target muscle response (19-22). 48 Overall, despite evidence supporting contraction-inducing PMS for enhanced motor recovery after 49 CNS lesions, systematic characterization of its physiological, stimulation intensity–dependent 50 effects on CSE remains unestablished. 51 PMS physiological effects are thought to result primarily from the activation of 52 low-threshold afferents, inducing Ia afferents from muscle spindles, Ib afferents from Golgi 53 tendon organs, and cutaneous afferents (23-25). Studies using peripheral electrical stimulation 54 show that suprathreshold, contraction-inducing stimulation increases CSE, whereas 55 subthreshold stimulation can reduce it, likely through preferential activation of inhibitory 56 afferent pathways (26-28). Despite expectations of similar mechanisms with PMS, afferent and 57 efferent fiber recruitment patterns differ from those observed with peripheral electrical 58 stimulation. In general, electrical stimulation with increasing strength elicits action potentials 59 in the largest fibers first, as they have the lowest electrical resistance, followed by 60 progressively smaller fibers (23). In contrast, peripheral nerve magnetic stimulation requires a 61 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 4 higher excitation threshold for sensory fibers than for α -motor fibers, and H-reflexes are not 62 elicited at intensities below the motor threshold (MT) (29). Moreover, PMS generates eddy 63 currents that directly stimulate deeper tissues without passing through the skin, suggesting 64 reduced recruitment of cutaneous afferents (8). Collectively, these properties imply that 65 subthreshold PMS may fail to activate low-threshold muscle afferents and may therefore be 66 insufficient to enhance CSE. Clarifying subthreshold PMS effects will provide insight into the 67 stimulation intensity–dependent mechanisms of PMS-induced neuroplasticity and help 68 optimize stimulation parameters for diverse clinical applications. In addition, PMS-induced 69 CSE modulation lacks direct afferent pathway characterization. In particular, the extent of 70 group I afferent—especially muscle spindle group Ia afferent—contributions to repetitive 71 PMS–induced CSE facilitation remains unclear. 72 Therefore, this study investigated the effects of different repetitive PMS intensities 73 applied to the wrist extensor muscles on CSE. Group I afferent contributions were assessed 74 using ischemic nerve block and motor unit (MU) firing–based afferent input estimates. We 75 hypothesized that 1) suprathreshold PMS would trigger intensity-dependent CSE enhancement 76 and within a shorter intervention period than subthreshold PMS, and 2) these effects would be 77 mediated primarily by large-diameter group I afferents, with findings consistent with a major 78 contribution from Ia afferents. Four experiments were conducted: Experiment 1 tested 79 intensity-dependent time courses of MEPs; Experiment 2 examined persistence after a short 80 high-intensity protocol; Experiment 3 probed the necessity of large-diameter afferent input 81 using ischemia after marked H-reflex suppression; and Experiment 4 quantified afferent-driven 82 motoneuron facilitation using single MU recordings. 83 84

85 Participants 86 The study included 34 healthy volunteers, whose participation was distributed across four 87 experiments (20 females, aged 18–34 years, 23 ± 4 years, mean ± standard deviation [SD]). 88 The number of participants for Experiments 1, 2, 3, and 4 was 19 (10 females, 24.0 ± 4.2 89 years), 17 (9 females, 22.9 ± 4.7 years), 5 (all males, 27.4 ± 4.2 years), and 8 (3 females, 25.6 90 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 5 ± 4.1 years), respectively. Notably, six individuals participated in multiple experiments. 91 Specifically, three individuals completed all four experiments. One individual completed 92 Experiments 1, 2, and 4, while two individuals completed Experiments 2, 3, and 4. None of the 93 participants had a history of neurological disease or received any medication affecting the CNS. 94 All participants except two were considered right-handed based on Chapman’s dominant hand 95 test (30). The experimental procedures were approved by the Ethics Committee of Y amagata 96 Prefectural University of Health Science (approval number: 2308-15 ) and followed the 97 Declaration of Helsinki. Before participating in this study, all participants signed written 98 informed consent forms for the experimental procedures. 99 100 Experimental setup 101 Electromyographic recordings. Surface electromyographic (EMG) signals were recorded with 102 paired 1.0-cm diameter Ag/AgCl disk electrodes. The electrodes (1.5 cm interelectrode 103 distance) were secured to the skin overlying the right extensor carpi radialis (ECR) and flexor 104 carpi radialis (FCR) muscles. The EMG signals were amplified, bandpass filtered (15–1,000 105 Hz), and sampled at 2,000 Hz for offline analysis (Micro 1401 with Signal software, 106 Cambridge Electronic Design, Cambridge, UK). 107 MU discharges were recorded with a pair of needle electrodes (Seirin acupuncture needle, 108 0.16/i2 mm diameter, Seirin Kasei, Shizuoka, Japan) inserted into the ECR muscle belly. EMG 109 signals were amplified and processed with a low-cut filter (50/i2 Hz). Audiovisual EMG 110 potential feedback was provided to help participants maintain stable single MU activity. Single 111 MU discrimination was carefully defined using the upper and lower amplitude discrimination 112 thresholds of the recorded potential. The EMG potential shape was displayed on an 113 oscilloscope (TDS210, Tektronix, Tokyo, Japan) throughout the experiment. The MU 114 discharges were converted into standard pulses and used for offline analysis (see Experiment 115 4). 116 117 Peripheral magnetic stimulation. PMS was applied over the right ECR muscle to stimulate the 118 radial nerve innervating the ECR, as described previously (13). The forearm was fixed in a 119 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 6 pronation position with the fingers free. A biphasic pulse of PMS was delivered using a 120 figure-eight coil (Cool-B65; outer diameter 75 mm) connected to a MagPro R30 (MagVenture 121 A/S, Denmark). The coil was placed on the skin overlaying the right ECR muscle and 122 positioned perpendicular to the forearm. The stimulus intensity was expressed in multiples of 123 the MT of the direct motor response (M-wave). MT was defined as the minimal stimulus 124 intensity required to induce an M-wave ≥ 50 μ V (peak-to-peak amplitude) by a single-pulse 125 stimulus in at least three of five consecutive trials. The ECR muscle contraction was confirmed 126 by palpation, with the stimulus intensity well above the MT. 127 128 Transcranial magnetic stimulation. TMS was applied over the left primary motor cortex using 129 a figure-eight coil (loop diameter 70 mm) connected to Magstim 200 (Magstim Company, 130 Whitland, Dyfed, United Kingdom). We determined the optimal positioning to elicit MEPs in 131 the ECR muscle at rest (hot spot) by moving the coil with the handle pointing backward and 132 45° away from the midline. The hot spot was defined as the region where the largest MEP in 133 the ECR muscle could be evoked with minimal stimulus intensity (Lotze et al., 2003). Resting 134 MT was defined as the minimal stimulus intensity required to induce MEPs of ≥ 50 μ V 135 (peak-to-peak amplitude) in at least three of five consecutive trials in the relaxed muscle (5). 136 The TMS intensity was set at 120% of the resting MT to measure MEPs as a CSE indicator. A 137 total of 11 MEPs was recorded in the resting condition. Each peak-to-peak amplitude was 138 measured and averaged, and the mean value among the participants was calculated for further 139 analysis. 140 141 Electrical peripheral nerve stimulation. Rectangular electrical pulses (1.0 ms) were 142 percutaneously delivered to the radial nerve trunk using bipolar surface electrodes (1.0 cm 143 diameter, 1.5 cm interelectrode distance) positioned along the nerve trajectory at the arm’s 144 lateral intermuscular septum. The stimulus electrode was connected to a constant-current 145 stimulator (DS8R, Digitimer, Welwyn Garden City, UK). The maximum direct motor response 146 (M-max) was measured by supramaximal electrical stimulation (at an intensity of 120% to 147 induce M-max). The inter-stimulus interval was 4–5 s. 148 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 7 149 Experimental procedures 150 Participants were comfortably seated, with the examined right arm supported on an armrest; 151 the shoulder was slightly flexed (≈ 10°), followed by elbow flexion (≈ 90°), and forearm 152 pronation. 153 154 Experiment 1: Intensity-dependent effects of repetitive PMS on MEPs, M-waves 155 Repetitive PMS was delivered in a 2 s ON and 2 s OFF cycle, according to a previous report 156 (13). The stimulation frequency was set at 25 Hz for the following reasons: 1) the stimulation 157 frequency of repetitive PMS was determined by afferent input from muscle spindles (Ia 158 afferents) of the wrist extensor muscles during voluntary movement (31); 2) our previous study 159 demonstrated that 25 Hz and 50 Hz repetitive PMS produced comparable increases in MEPs 160 (13); and 3) in our preliminary experiment, application of 50 Hz repetitive PMS at higher 161 intensities caused stimulation-coil heating, which prevented some participants from completing 162 a 15-min intervention. 163 The effects of repetitive PMS intensity were assessed using conditions applied in random 164 order across three sessions on separate days. A crossover design was employed, and repetitive 165 PMS was delivered at three stimulus intensities: 0.9×MT (below the MT), 1.2×MT [(same as 166 Nito et al. (2021) (13)], and a high intensity can induce maximal dorsiflexion movement of the 167 wrist (high-intensity). The stimulus intensity required to induce maximal dorsiflexion ranged 168 from 1.4 to 2.4×MT (mean ± SD: 1.8 ± 0.3×MT). The MT for the 0.9×MT, 1.2×MT, and 169 high-intensity were 18.8 ± 2.9, 19.4 ± 2.1, and 20.1 ± 3.2, respectively (F2,36 = 2.513, p = 0.10, 170 /g2015 /g3043 /g2870 = 0.123). A minimum one-day interval separated sessions to minimize carry-over effects. 171 The intervention comprised five sessions, and a single session of the repetitive PMS was 172 delivered for 2 s ON and 2 s OFF cycle for 3 min. MEPs were measured 3 min before 173 intervention (baseline), just before repetitive PMS (T0), and after one session (T3), two 174 sessions (T6), three sessions (T9), four sessions (T12), and five sessions of repetitive PMS 175 (T15). The averaged MEP amplitude at each time point was normalized to the baseline and 176 expressed as a percentage of the baseline value. M-max was also measured at T0 and T15. 177 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 8 178 Experiment 2: Persistence of facilitation after short, high-intensity repetitive PMS 179 To investigate the lasting effect of repetitive PMS, the stimulation was administered in three 180 consecutive sessions, totaling 9 min, using the same stimulation cycle and frequency as in 181 Experiment 1. The stimulus intensity was set to induce maximal dorsiflexion of the wrist 182 (high-intensity). MEPs were measured 5 min before repetitive PMS (baseline), just before the 183 repetitive PMS (Pre), and every 10 min for 50 min after the repetitive PMS (P0, P10, P20, P30, 184 P40, and P50). The averaged MEP amplitude at each time point was normalized to the baseline 185 and expressed as a percentage of the baseline value. 186 187 Experiment 3: Ischemia to probe the contribution of large-diameter afferents 188 To investigate the contribution of Ia afferent input to MEP enhancement, we applied repetitive 189 PMS during ischemia and examined its effects on MEP amplitude. A blood pressure cuff 190 placed around the participant’s upper arm was inflated to 220 mmHg pressure (32). After the 191 inflation onset, the H-reflexes were elicited every 4 s by electrical stimulation to the radial 192 nerve trunk, and the stimulus intensity was set to elicit the maximal peak-to-peak amplitude of 193 the H-reflex. The H-reflex, which was absent at rest, was elicited during weak isometric wrist 194 dorsiflexion. After the peak-to-peak amplitude of the H-reflex decreased to less than 10% of its 195 preischemic size (6.4% ± 2.6% M-max; 12–20 min from the onset of ischemia), repetitive 196 PMS was performed during ischemia in three consecutive sessions, totaling 9 min. The 197 stimulus intensity was set to induce maximal dorsiflexion of the wrist (high-intensity). MEPs 198 were measured 5 min before repetitive PMS (baseline), just before the repetitive PMS (Pre), 199 and every 10 min for 50 min after the repetitive PMS (P0, P10, P20, P30, P40, and P50). The 200 MEP measurement at P0 was performed after confirming that M-max remained unchanged 201 compared to its size before ischemia (before ischemia, 5.9 ± 2.6 mV; after ischemia, 6.2 ± 2.0 202 mV; p = 0.79, d = 0.126). 203 204 Experiment 4: Estimation of afferent-driven motoneuron facilitation using MU recordings 205 To compare sensory input magnitude across PMS intensities, conditioning stimulation effects 206 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 9 on a single MU were investigated (33). This approach is based on the premise that group Ia 207 afferents have a monosynaptic connection to almost all homonymous motoneurons (34), 208 enabling us to estimate the level of Ia afferent input by analyzing MU firings. Poststimulus 209 time histograms (PSTHs) (bin width 0.2 ms or 0.5 ms) of the discharge of a voluntarily 210 activated MU (approximately 100-ms firing interval) were constructed for the period ranging 211 from 15 to 50/i2 ms after conditioning stimulation. ECR MU discharges were recorded during 212 isometric wrist extension at <5% maximal voluntary contraction. MU discharges were detected 213 by the computer at intervals of approximately every 0.7 s, triggering the stimulator. The 214 conditioning stimulation was triggered with a delay of about 70/i2 ms after voluntary MU 215 activation. The delay was set at a time that easily affected the next MU firing. In other words, 216 the delay was set so that the afferent volleys by the conditioning stimulation would arrive at 217 the motoneuron around the latest period of hyperpolarization due to the previous MU firing 218 and just before voluntarily driven discharge of the motoneuron. Additionally, a 219 firing-probability histogram was constructed under a no-stimulation control condition. The 220 control and stimulated situations were alternated randomly (the same number of triggers) 221 within a sequence. Each sequence generally comprised 200–800 stimulated and control 222 situations. Conditioning stimulation effects were obtained by subtracting control-condition 223 trigger counts from poststimulation values in each bin. Histogram subtraction was used to 224 generate a cumulative sum curve confirming facilitation (35). 225 Conditioning PMS was applied to the radial nerve innervating the ECR. PSTHs were 226 constructed using stimulation intensities of 0.9, 1.2, 1.5, and 1.8× MT to investigate PMS 227 intensity effects on the firing probability of ECR MUs. For each single MU, PSTHs were 228 constructed for each stimulation intensity, with the order of intensities randomized across 229 measurements. 230 231 Statistical analysis 232 The Shapiro–Wilk test was used to confirm that the normalized MEPs and M-max followed a 233 normal distribution. In Experiment 1, a two-way repeated-measures mixed model analysis of 234 variance (ANOV A) was used with INTENSITY (0.9×MT, 1.2×MT, and high-intensity) and 235 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 10 TIME (T0, T3, T6, T9, T12, T15) as factors to compare normalized MEPs. When a significant 236 interaction or main effects were determined, a two-way repeated-measures mixed model 237 ANOV A was similarly used with INTENSITY (0.9×MT, 1.2×MT, and high-intensity) and 238 TIME (T0 and T15) as factors to compare M-max. 239 Experiment 2 assessed lasting repetitive PMS effects on MEPs using a repeated-measures 240 mixed model ANOVA with TIME as the main factor (Pre, P0, P10, P20, P30, P40, and P50). 241 Owing to the small sample size (n = 5), Experiment 3 reports individual data without 242 statistical analysis. 243 In Experiment 4, effect latency and duration were defined as consecutive bins showing 244 increased firing probability and were determined from fluctuations in cumulative sum curves 245 by visual inspection from at least two investigators. Differences between poststimulation and 246 control firing probabilities across consecutive-bin time windows were assessed using χ 2-test. A 247 repeated-measures mixed model ANOVA with main factor INTENSITY (0.9×MT, 1.2×MT, 248 1.5×MT, and high-intensity) was used to compare peak amplitudes derived from cumulative 249 difference histograms. An unpaired t-test was performed to compare the peak duration induced 250 by PMS and electrical stimulation. 251 Statistical significance was set at p < 0.05 for all comparisons. When significant main 252 effects or interactions were found, post-hoc comparisons were performed using a paired t-test 253 with Bonferroni adjustments. Effect sizes were reported as partial η ² (/g2015 /g3043 /g2870) for ANOVA and 254 Cohen’s d for the t-test. All statistical analyses, except the χ 2-test, were performed using SPSS 255 30 (IBM, Armonk, NY , USA). The χ 2-test was conducted using a PSTH analysis program 256 (MTS0014, Gigatex, Japan). Group data are presented as mean ± SD in the text. 257 258 259

260 Experiment 1: Intensity-dependent effects of repetitive PMS on MEPs, M-waves 261 Figure 1 shows the time courses of the normalized MEPs. Notably, ECR MEP amplitudes were 262 differentially modulated by stimulation intensity (Figure 1A). Two-way repeated-measures 263 mixed model ANOVA revealed a significant main effect of INTENSITY (F2,36 = 11.524, p < 264 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 11 0.001, /g2015 /g3043 /g2870 = 0.192) and TIME (F5,90 = 13.312, p < 0.001, /g2015 /g3043 /g2870 = 0.470), and their interaction 265 (F10,180 = 2.688, p = 0.004, /g2015 /g3043 /g2870 = 0.167) on the MEP amplitudes of the ECR. Time-factor 266 comparisons revealed no changes following repetitive PMS at 0.9×MT, even after a total 267 15-min intervention (F5,90 = 0.885, p = 0.494, /g2015 /g3043 /g2870 = 0.047). In contrast, repetitive PMS at 268 1.2×MT increased MEPs at T15 compared to T0 (p < 0.001, d = 1.047), whereas repetitive 269 PMS at high-intensity induced an earlier facilitation, with MEP amplitudes increasing at T9 (p 270 = 0.002, d = 0.997), T12 (p < 0.001, d = 1.585), and T15 compared to T0 (p < 0.001, d = 271 1.240). Comparison of the MEPs of ECR among different conditions showed a significant 272 difference between 0.9×MT and high intensity repetitive PMS at T12 (p = 0.013, d = 0.859) 273 and T15 (p = 0.002, d = 0.800), but no significant difference was observed in other 274 combinations (p > 0.05). 275 FCR MEP amplitudes remained unaltered under any of these conditions (Figure 1B). A 276 two-way repeated-measures mixed model ANOV A revealed no significant main effects of 277 INTENSITY (F2,36 = 0.594, p = 0.104, /g2015 /g3043 /g2870 = 0.033) or TIME (F 5,90 = 0.820, p = 0.536, /g2015 /g3043 /g2870 = 278 0.062), and their interaction (F10,180 = 0.968, p = 0.473, /g2015 /g3043 /g2870 = 0.051) on the FCR MEP 279 amplitudes. 280 The M-max amplitudes remained unchanged under any of these conditions. The mean 281 values of M-max at T0 and T15 were 7.9 ± 3.2 mV and 8.2 ± 3.2 mV for 0.9×MT repetitive 282 PMS, 7.5 ± 2.6 mV and 7.5 ± 2.8 mV for 1.2×MT repetitive PMS, 7.3 ± 2.4 mV and 7.7 ± 2.7 283 mV for high-intensity repetitive PMS, respectively. A two-way repeated-measures mixed 284 model ANOVA revealed no significant main effects of INTENSITY (F2,36 = 1.242, p = 0.301, 285 /g2015 /g3043 /g2870 = 0.065) or TIME (F1,18 = 2.065, p = 0.168, /g2015 /g3043 /g2870 = 0.103), and their interaction (F 2,36 = 0.432, 286 p = 0.652, /g2015 /g3043 /g2870 = 0.023). 287 288 Experiment 2: Persistence of facilitation after short, high-intensity repetitive PMS 289 Figure 2 shows the changes in MEPs of ECR before and after 9 min of repetitive PMS. 290 Repetitive PMS markedly increased MEPs in stimulated ECR, whereas no significant changes 291 were observed in FCR. A repeated-measures mixed model ANOV A for ECR MEPs revealed a 292 significant main effect of Time (F6,96 = 5.153, p < 0.001, /g2015 /g3043 /g2870 = 0.244). Post-hoc tests revealed 293 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 12 that ECR MEPs were significantly increased at P0 (p < 0.001, d = 1.174), P10 (p 0.05). In contrast, a 296 repeated-measures mixed model ANOVA for FCR MEPs revealed no significant main effect of 297 Time (F6,96 = 0.985, p = 0.440, /g2015 /g3043 /g2870 = 0.059). 298 299 Experiment 3: Ischemia to probe the contribution of large-diameter afferents 300 Figure 3 illustrates the individual changes in ECR MEPs for five participants who participated 301 in Experiments 2 and 3. As shown in Figure 3A, which indicates the individual data underlying 302 Figure 2, a robust increase in MEPs was confirmed. Conversely, repetitive PMS applied after 303 ischemic blockade of Ia afferent input produced no observable MEP changes (Figure 3B), 304 suggesting a contribution of Ia afferent input to repetitive PMS–induced MEP augmentation. 305 306 Experiment 4: Estimation of afferent-driven motoneuron facilitation using MU recordings 307 Figure 4 shows a representative facilitation induced by PMS with various intensities. A 308 significant peak (representing an increase in the firing probability) of PSTH was observed 309 22.5–24.0 ms after PMS using 1.8×MT (p/i2 </i2 0.001). In this MU, a peak was observed at the 310 same latency across all stimulus intensities from 1.2×MT to 1.8×MT, whereas the peak size, as 311 indicated by the cumulative sum curve, decreased upon PMS intensity reduction. The PMS at 312 0.9×MT did not affect the MU firing probability (p > 0.05). A total of 27 MUs from eight 313 participants were studied. Because stable MU firing could not be maintained over prolonged 314 periods, the number of analyzed MUs varied slightly across stimulation intensities. A 315 significant peak was observed in all 25 MUs when PMS was applied at 1.8×MT (100%), in 17 316 of 24 MUs at 1.5×MT (71%), and in 14 of 25 MUs at 1.2×MT (56%). In contrast, no 317 significant peak was detected in any of the 25 MUs using PMS at 0.9×MT (0%) (Figure 4B). 318 Furthermore, the peak amplitudes obtained from the cumulative sums of the difference 319 histograms increased progressively with stimulation intensity (Figure 4C). A 320 repeated-measures mixed model ANOVA revealed a significant main effect of INTENSITY 321 (F3,66.065=23.936, p<0.001, /g2015 /g3043 /g2870=0.486). These findings suggest that higher PMS intensities are 322 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 13 associated with greater afferent input, whereas PMS delivered below the MT does not elicit 323 afferent input. 324 Figure 5 shows an example of the effects of PMS and electrical stimulation of the radial 325 nerve trunk on single MU firings. The latency of the peak was 22.0 ms for PMS and 20.4 ms 326 for electrical stimulation, and the duration was 2.4 ms and 2.4 ms, respectively. A total of 15 327 MUs from 8 participants were studied. The difference latency of the peaks was 21.4 ± 2.0 ms 328 for PMS and 20.4 ± 2.0 ms for electrical stimulation, respectively, indicating that the electrical 329 stimulation–induced peak latency was 1.0 ± 0.7 ms shorter than that of PMS. Considering the 330 distance between the two stimulation sites (8.2 ± 3.4 cm), the central latency of the 331 PMS-induced peak can be considered equivalent to that of monosynaptic Ia facilitation 332 induced by electrical stimulation (36). The peak duration was 2.4 ± 1.0 ms for PMS and 2.2 ± 333 0.8 ms for electrical stimulation, respectively, indicating no significant difference between the 334 two conditions (p = 0.402 using unpaired t-test, d = 0.157). These findings suggest that PMS 335 activates Ia afferents and generates excitatory postsynaptic potentials (EPSPs), with the initial 336 component mediated by a monosynaptic pathway. 337 338 339

340 The present study aimed to clarify how varying intensities of repetitive PMS applied to wrist 341 extensor muscles influence CSE. Our results yielded three main findings. First, repetitive PMS 342 facilitated CSE in an intensity-dependent manner, with high-intensity stimulation inducing 343 earlier facilitation than 1.2×MT, while 0.9×MT produced no effect. Second, the facilitation was 344 muscle-specific and not attributable to peripheral changes, as FCR MEPs and ECR M-max 345 remained unchanged. Third, the abolition of facilitation during ischemia together with a 346 short-latency peak in MU firing via a monosynaptic path supports a major contribution of 347 large-diameter muscle afferents, consistent with a substantial Ia component, to PMS-induced 348 corticospinal facilitation. 349 350 Group I Afferents Responsible for CSE Enhancement 351 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 14 Group I afferents from wrist extensor muscles are most likely the afferent fibers responsible for 352 CSE enhancement. In the present study, we demonstrated that a suprathreshold PMS increased 353 MEPs. Specifically, high-intensity PMS-induced increased MEPs, achieving this effect only 354 after a short duration of intervention. Furthermore, MU firing analysis showed similar PSTH 355 peak latencies across suprathreshold PMS intensities, whereas peak amplitudes increased with 356 stimulation intensity. Additionally, the PSTH duration was comparable to that induced by 357 electrical stimulation to the radial nerve trunk, which is known to elicit a monosynaptic 358 component mediated by Ia afferents (37, 38), regardless of the stimulation intensity of PMS. 359 The peak amplitude and duration of the PSTH are considered to reflect the size and rise time of 360 the composite EPSPs, respectively (39, 40). Higher PMS intensities may have recruited a 361 greater number of Ia afferents, thereby eliciting larger EPSPs. Consequently, stronger 362 intensities of PMS may have induced increased CSE within a shorter intervention period 363 compared to weaker intensities. Ischemic nerve block is known to preferentially attenuate 364 large-diameter afferents (32). Accordingly, the absence of MEP facilitation during ischemia 365 supports a major contribution of large-diameter muscle afferents (i.e., group I pathways) to 366 PMS-induced corticospinal facilitation. Although ischemia does not isolate Ia afferents 367 exclusively, the monosynaptic timing and narrow duration of MU facilitation are compatible 368 with a substantial Ia component. Because repetitive PMS was delivered after the H-reflex was 369 reduced to <10% of baseline, these findings collectively suggest that attenuation of 370 large-diameter afferent input abolished the facilitatory effect on CSE. 371 PMS at subthreshold intensities did not alter CSE or MU firings in the present study. This 372 interpretation is supported by a previous report by Panizza et al. (1992) (29), which 373 demonstrated two key findings regarding PMS: First, sensory fibers exhibit a higher threshold 374 than motor fibers. Second, the H-reflex—elicited by monosynaptic excitation from the 375 homonymous group Ia afferents—is not induced at intensities lower than the MT. Additionally, 376 previous studies on electrical stimulation have shown contrasting effects: suprathreshold 377 electrical stimulation of the median nerve trunk at the wrist increases MEPs, whereas 378 subthreshold stimulation decreases MEPs (26-28). Furthermore, cutaneous nerve stimulation 379 of the index finger originating from the median nerve also decreases MEPs (26). Comparable 380 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 15 conduction velocities between the fastest cutaneous and group Ia afferents (41) suggest similar 381 recruitment thresholds. Therefore, the inhibitory effect of subthreshold stimulation on CSE 382 may result from the preferential activation of cutaneous afferents rather than of group Ia 383 afferents. In contrast, PMS may be more effective in increasing CSE by preferentially 384 activating Ia afferents and providing stronger ascending proprioceptive input. Although PMS 385 was applied transcutaneously and may activate cu taneous afferents beneath the stimulation site, 386 magnetically induced eddy currents directly stimulate deeper tissues and recruit superficial 387 cutaneous afferents less effectively than electrical stimulation (8). Thus, cutaneous afferent 388 contributions at the stimulation site are likely smaller than those typically recruited by 389 peripheral electrical stimulation, although this study did not directly quantify such observation. 390 Therefore, these findings suggest that subthreshold PMS is insufficient to excite the sensory 391 afferents necessary to modulate CSE, thereby failing to induce either enhancement or 392 suppression of MEPs. 393 394 Changes in CSE Induced by Repetitive PMS 395 Previous studies (7, 12, 13) have revealed that repetitive PMS at suprathreshold intensity 396 increases MEPs. Moreover, reductions in short-latency intracortical inhibition and increases in 397 intracortical facilitation have been reported following repetitive PMS (12, 13). In contrast, 398 MEPs evoked by transmastoid electrical stimulation were not altered by repetitive PMS 399 (Omiya et al., under review), which is thought to reflect changes in the efficiency of synaptic 400 transmission between the cortex and spinal motor neurons or the excitability of spinal motor 401 neurons themselves (42, 43). Furthermore, the excitability of the spinal reflex loop in the 402 muscle receiving PMS appears to remain unchanged, as demonstrated using the H-reflex (13). 403 In addition, M-max remained unchanged following repetitive PMS, indicating preserved 404 neuromuscular transmission efficacy. Overall, these findings suggest that augmented MEPs are 405 predominantly attributable to changes at supraspinal levels (e.g., cortical and/or corticospinal 406 synaptic efficacy), although the present design does not fully exclude contributions from spinal 407 mechanisms. 408 Prior work demonstrated sustained MEP enhancement for up to 60 min after 15 min of 409 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 16 repetitive PMS at 1.2×MT (13). In contrast, the present study demonstrated that higher 410 intensities of PMS-induced greater increases in MEPs following only 9 min of the intervention, 411 but the effect persisted for only 30 min. These findings suggest a trade-off between the 412 intensity and duration of PMS-induced effects: higher intensities can induce greater increases 413 in CSE within a shorter intervention time, but the effects are less sustained compared with 414 those induced by lower intensities. Therefore, high-intensity PMS, capable of inducing rapid 415 increases in CSE, may provide practical advantages in clinical settings with time constraints. A 416

of the present study is that the duration of CSE changes following high-intensity 417 PMS applied over longer intervention periods (e.g., 15 min) was not investigated. It is 418 conceivable that extending the duration of high-intensity stimulation may prolong CSE 419 enhancement. Future studies are warranted to determine whether longer high-intensity PMS 420 protocols can combine the advantages of rapid induction and sustained effects. 421 422 Methodological Considerations 423 Despite these strengths, several methodological considerations warrant acknowledgment. First, 424 the ischemic nerve block procedure in Experiment 3 was performed in a small sample (n = 5). 425 This small sample size limits the generalizability of our findings that large-diameter afferent 426 input—with a substantial Ia component—mainly contributes to PMS-induced facilitation. 427 Second, despite marked ischemia-induced H-reflex suppression indicating reduced group Ia 428 afferent input, potential contributions from other large-diameter afferents, including group Ib 429 afferents, remain possible. These afferents may also have been partially affected, potentially 430 contributing to the observed increase in MEPs (23-25). Third, repetitive PMS–induced coil 431 heating constitutes a methodological limitation. At frequencies higher than 25 Hz or at higher 432 stimulation intensities, continuous intervention may become challenging due to the rapid rise 433 in coil temperature. Therefore, the parameters used in this study were selected to balance the 434 induction of corticospinal facilitation with the physical capacities of the stimulator. 435 436 Clinical Implications 437 Numerous studies have reported that repetitive PMS contributes to improved motor 438 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 17 dysfunction following CNS lesions (7, 9-11, 16, 18-20, 22), suggesting promising therapeutic 439 potential. Repetitive PMS–induced increases in CSE likely contribute, at least in part, to these 440 functional improvements (7). 441 Beyond these potential benefits for functional recovery, PMS offers practical advantages 442 for clinical application: PMS can directly stimulate deep tissues without penetrating the skin to 443 induce eddy currents (8), and is therefore considered less likely to cause discomfort during 444 stimulation. Indeed, none of the participants reported any discomfort during PMS, even at an 445 intensity sufficient to elicit maximal wrist dorsiflexion, and no participant found it difficult to 446 continue the experiment. 447 While the present findings provide evidence for the underlying mechanism of PMS, its 448 clinical applicability requires further investigation. The present study examined only the 449 immediate effects of repetitive PMS on wrist extensors in healthy individuals; whether similar 450 facilitatory effects occur in individuals with CNS lesion–related motor dysfunction or are 451 sustained across repeated, longer-term interventions remains to be determined. Furthermore, 452 although we demonstrated that repetitive PMS increases CSE, the functional implications of 453 this enhancement remain to be fully elucidated. Future studies incorporating behavioral or 454 functional outcome measures should clarify whether PMS-induced increases in excitability 455 translate into meaningful improvements in motor performance or clinical function. 456 457

458 Repetitive PMS at intensities above the MT enhances CSE, likely mediated predominantly by 459 large-diameter muscle afferents (group I afferents), consistent with a substantial Ia component. 460 Short-duration interventions at higher stimulation intensities produce greater CSE 461 enhancement than lower-intensity stimulation. In addition, stronger stimulation intensities are 462 associated with greater afferent input. These findings suggest that group Ia afferents from 463 muscle spindles markedly contribute to CSE modulation. These results provide insight into the 464 mechanisms underlying PMS-induced neuroplasticity and may help optimize stimulation 465 parameters for clinical applications in neurorehabilitation. 466 467 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 18 LIST OF ABBREVIATIONS 468 CSE: corticospinal excitability 469 CNS: central nervous system 470 ECR: extensor carpi radialis 471 EMG: electromyography 472 EPSP: excitatory postsynaptic potential 473 FCR: flexor carpi radialis 474 MEP: motor-evoked potential 475 M-max: maximum direct motor response 476 MT: motor threshold 477 MU: motor unit 478 PMS: peripheral magnetic stimulation 479 PSTH: poststimulus time histogram 480 TMS: transcranial magnetic stimulation 481 482 Data availability 483 The datasets supporting the conclusions of this article are available from the corresponding 484 author on reasonable request. 485 486 Acknowledgments 487 The authors thank all the participants who engaged in this study. 488 489 Grants 490 M.N. was supported by JSPS KAKENHI (Grant Number: 22K17628) and by a grant from the 491 Watanabe Foundation. 492 493 Disclosures 494 The authors declare that they have no competing interests. 495 496 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 19 Author contributions 497 M.N., N.M. and T.Y . conceived and designed research, K.Y ., M.N., D.M., A.O., and K.S. 498 performed experiments and analyzed data, K.Y ., M.N., H.F., and T.Y . interpreted results of 499 experiments, M.N. prepared figures, K.Y ., M.N. and T.Y . drafted manuscript, K.Y ., M.N., D.M., 500 A.O., K.S., T.K., D.K., N.M., H.F., and T.Y . approved the final version of manuscript. 501 502 503

504 1. Porter R, and Lemon R. Corticospinal Function and V oluntary Movement. Oxford 505 University Press, 1995. 506 2. Lemon RN. Descending pathways in motor control. Annu Rev Neurosci 31: 195–218, 507 2008. 508 3. Rossini PM, Burke D, Chen R, Cohen LG, Daskalakis Z, Di Iorio R, Di Lazzaro V , 509 Ferreri F, Fitzgerald PB, George MS, Hallett M, Lefaucheur JP, Langguth B, 510 Matsumoto H, Miniussi C, Nitsche MA, Pascual-Leone A, Paulus W, Rossi S, 511 Rothwell JC, Siebner HR, Ugawa Y, Walsh V , and Ziemann U. Non-invasive electrical 512 and magnetic stimulation of the brain, spinal cord, roots and peripheral nerves: Basic 513 principles and procedures for routine clinical and research application. An updated report 514 from an I.F.C.N. Committee. Clin Neurophysiol 126: 1071–1107, 2015. 515 4. Nudo RJ, Wise BM, SiFuentes F, and Milliken GW. Neural substrates for the effects of 516 rehabilitative training on motor recovery after ischemic infarct. Science 272: 1791–1794, 517 1996. 518 5. Lotze M, Braun C, Birbaumer N, Anders S, and Cohen LG. Motor learning elicited by 519 voluntary drive. Brain 126: 866–872, 2003. 520 6. Perez MA, Lungholt BK, Nyborg K, and Nielsen JB. Motor skill training induces 521 changes in the excitability of the leg cortical area in healthy humans. Exp Brain Res 159: 522 197–205, 2004. 523 7. Flamand VH, and Schneider C. Noninvasive and painless magnetic stimulation of 524 nerves improved brain motor function and mobility in a cerebral palsy case. Arch Phys 525 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 20 Med Rehabil 95: 1984–1990, 2014. 526 8. Polson MJ, Barker AT, and Freeston IL. Stimulation of nerve trunks with time-varying 527 magnetic fields. Med Biol Eng Comput 20: 243–244, 1982. 528 9. Struppler A, Havel P, and Müller-Barna P. Facilitation of skilled finger movements by 529 repetitive peripheral magnetic stimulation (RPMS) - a new approach in central paresis. 530 NeuroRehabilitation 18: 69–82, 2003. 531 10. Flamand VH, Beaulieu LD, Nadeau L, and Schneider C. Peripheral magnetic 532 stimulation to decrease spasticity in cerebral palsy. Pediatr Neurol 47: 345–348, 2012. 533 11. Beaulieu LD, and Schneider C. Effects of repetitive peripheral magnetic stimulation on 534 normal or impaired motor control. A review. Neurophysiol Clin 43: 251–260, 2013. 535 12. Gallasch E, Christova M, Kunz A, Rafolt D, and Golaszewski S. Modulation of 536 sensorimotor cortex by repetitive peripheral magnetic stimulation. Front Hum Neurosci 9: 537 407, 2015. 538 13. Nito M, Katagiri N, Yoshida K, Koseki T, Kudo D, Nanba S, Tanabe S, and 539 Y amaguchi T. Repetitive Peripheral Magnetic Stimulation of Wrist Extensors Enhances 540 Cortical Excitability and Motor Performance in Healthy Individuals. Front Neurosci 15: 541 632716, 2021. 542 14. El Nahas N, Kenawy FF, Abd Eldayem EH, Roushdy TM, Helmy SM, Akl AZ, 543 Ashour AA, Emara TH, Moawad MM, Amin RM, and Elbokl AM. Peripheral 544 magnetic theta burst stimulation to muscles can effectively reduce spasticity: a 545 randomized controlled trial. J Neuroeng Rehabil 19: 5, 2022. 546 15. Jiang YF, Zhang D, Zhang J, Hai H, Zhao YY, and Ma YW. A Randomized Controlled 547 Trial of Repetitive Peripheral Magnetic Stimulation applied in Early Subacute Stroke: 548 Effects on Severe Upper-limb Impairment. Clin Rehabil 36: 693–702, 2022. 549 16. Krewer C, Hartl S, Muller F, and Koenig E. Effects of repetitive peripheral magnetic 550 stimulation on upper-limb spasticity and impairment in patients with spastic hemiparesis: 551 a randomized, double-blind, sham-controlled study. Arch Phys Med Rehabil 95: 552 1039–1047, 2014. 553 17. Zhang W, Y amaguchi T, and Fujiwara T. Effects of Different Intensities of Repetitive 554 Peripheral Magnetic Stimulation on Spinal Reciprocal Inhibition in Healthy Persons. 555 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 21 Juntendo Iji Zasshi 70: 283–288, 2024. 556 18. Beaulieu LD, Masse-Alarie H, Camire-Bernier S, Ribot-Ciscar E, and Schneider C. 557 After-effects of peripheral neurostimulation on brain plasticity and ankle function in 558 chronic stroke: The role of afferents recruited. Neurophysiol Clin 47: 275–291, 2017. 559 19. Beaulieu LD, Masse-Alarie H, Brouwer B, and Schneider C. Noninvasive 560 neurostimulation in chronic stroke: a double-blind randomized sham-controlled testing of 561 clinical and corticomotor effects. Top Stroke Rehabil 22: 8–17, 2015. 562 20. Obayashi S, and Takahashi R. Repetitive peripheral magnetic stimulation improves 563 severe upper limb paresis in early acute phase stroke survivors. NeuroRehabilitation 46: 564 569–575, 2020. 565 21. Struppler A, Binkofski F, Angerer B, Bernhardt M, Spiegel S, Drzezga A, and 566 Bartenstein P. A fronto-parietal network is mediating improvement of motor function 567 related to repetitive peripheral magnetic stimulation: A PET-H2O15 study. Neuroimage 36 568 Suppl 2: T174–186, 2007. 569 22. Werner C, Schrader M, Wernicke S, Bryl B, and Hesse S. Repetitive Peripheral 570 Magnetic Stimulation (rpMS) in Combination with Muscle Stretch Decreased the Wrist 571 and Finger Flexor Muscle Spasticity in Chronic Patients after CNS Lesion. International 572 Journal of Physical Medicine & Rehabilitation 4: 2016. 573 23. Kandel ER, Schwartz JH, Jessell TM, Siegelbaum S, Hudspeth AJ, and Mack S. 574 Principles of Neural Science. McGraw-Hill Medical, 2013. 575 24. Pierrot-Deseilligny E, and Burke D. The Circuitry of the Human Spinal Cord: Spinal 576 and Corticospinal Mechanisms of Movement. Cambridge: Cambridge University Press, 577 2012. 578 25. Rothwell JC. Control of human voluntary movement. Croom Helm, 1987. 579 26. Sasaki R, Kotan S, Nakagawa M, Miyaguchi S, Kojima S, Saito K, Inukai Y, and 580 Onishi H. Presence and Absence of Muscle Contraction Elicited by Peripheral Nerve 581 Electrical Stimulation Differentially Modulate Primary Motor Cortex Excitability. Front 582 Hum Neurosci 11: 146, 2017. 583 27. Schabrun SM, Ridding MC, Galea MP, Hodges PW, and Chipchase LS. Primary 584 sensory and motor cortex excitability are co-modulated in response to peripheral electrical 585 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 22 nerve stimulation. PLoS One 7: e51298, 2012. 586 28. Tinazzi M, Zarattini S, Valeriani M, Romito S, Farina S, Moretto G, Smania N, 587 Fiaschi A, and Abbruzzese G. Long-lasting modulation of human motor cortex 588 following prolonged transcutaneous electrical nerve stimulation (TENS) of forearm 589 muscles: evidence of reciprocal inhibition and facilitation. Exp Brain Res 161: 457–464, 590 2005. 591 29. Panizza M, Nilsson J, Roth BJ, Basser PJ, and Hallett M. Relevance of stimulus 592 duration for activation of motor and sensory fibers: implications for the study of 593 H-reflexes and magnetic stimulation. Electroencephalogr Clin Neurophysiol 85: 22–29, 594 1992. 595 30. Chapman LJ, and Chapman JP. The measurement of handedness. Brain Cogn 6: 596 175–183, 1987. 597 31. Kakuda N, and Nagaoka M. Dynamic response of human muscle spindle afferents to 598 stretch during voluntary contraction. J Physiol 513 ( Pt 2): 621–628, 1998. 599 32. Magladery JW, McDougal DBJ, and Stoll J. Electrophysiological studies of nerve and 600 reflex activity in normal man. II. The effects of peripheral ischemia. Bull Johns Hopkins 601 Hosp 86: 291–312, 1950. 602 33. Fournier E, Meunier S, Pierrot-Deseilligny E, and Shindo M. Evidence for 603 interneuronally mediated Ia excitatory effects to human quadriceps motoneurones. J 604 Physiol 377: 143–169, 1986. 605 34. Mendell LM, and Henneman E. Terminals of single Ia fibers: location, density, and 606 distribution within a pool of 300 homonymous motoneurons. J Neurophysiol 34: 171–187, 607 1971. 608 35. Ellaway PH. Cumulative sum technique and its application to the analysis of peristimulus 609 time histograms. Electroencephalogr Clin Neurophysiol 45: 302–304, 1978. 610 36. Shinozaki K, Nito M, Kobayashi S, Hayashi M, Miyasaka T, Hashizume W, Shindo 611 M, and Naito A. Monosynaptic facilitation of group I afferents between brachioradialis 612 and extensor carpi radialis in humans. Neurosci Res 114: 30–34, 2017. 613 37. Birnbaum A, and Ashby P. Postsynaptic potentials in individual soleus motoneurons in 614 man produced by achilles tendon taps and electrical stimulation of tibial nerve. 615 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 23 Electroencephalogr Clin Neurophysiol 54: 469–471, 1982. 616 38. Burke D, Gandevia SC, and McKeon B. The afferent volleys responsible for spinal 617 proprioceptive reflexes in man. J Physiol 339: 535–552, 1983. 618 39. Ashby P, and Zilm D. Characteristics of postsynaptic potentials produced in single 619 human motoneurons by homonymous group 1 volleys. Exp Brain Res 47: 41–48, 1982. 620 40. Fetz EE, and Gustafsson B. Relation between shapes of post-synaptic potentials and 621 changes in firing probability of cat motoneurones. J Physiol 341: 387–410, 1983. 622 41. Macefield G, Gandevia SC, and Burke D. Conduction velocities of muscle and 623 cutaneous afferents in the upper and lower limbs of human subjects. Brain 112 ( Pt 6): 624 1519–1532, 1989. 625 42. Taylor JL, and Gandevia SC. Noninvasive stimulation of the human corticospinal tract. 626 J Appl Physiol (1985) 96: 1496–1503, 2004. 627 43. Ugawa Y, Rothwell JC, Day BL, Thompson PD, and Marsden CD. Percutaneous 628 electrical stimulation of corticospinal pathways at the level of the pyramidal decussation 629 in humans. Ann Neurol 29: 418–427, 1991. 630 631 FIGURE LEGENDS 632 Figure 1. Effects of different intensities of repetitive peripheral magnetic stimulation 633 (PMS) on motor-evoked potentials (MEPs). 634 Changes in MEP amplitudes of the extensor carpi radialis (ECR; A) and flexor carpi radialis 635 (FCR; B) muscles were recorded from 19 participants. The repetitive PMS was delivered at 636 three intensities: below motor threshold (0.9×MT; white circle), above motor threshold 637 (1.2×MT; light gray circle), and the intensity inducing maximum wrist dorsiflexion 638 (high-intensity, dark gray circle). The ordinate shows MEP amplitude normalized to the 639 baseline value for each participant, and the abscissa shows the time point of measurements: 640 before repetitive PMS (T0), after one session (T3), two sessions (T6), three sessions (T9), four 641 sessions (T12), and five sessions of repetitive PMS (T15). Each circle and error bar represent 642 the mean value and SD, respectively. Asterisks indicate significant differences compared with 643 “T0,” and the daggers indicate significant differences compared with the other intensities (p < 644 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 24 0.05). 645 646 Figure 2. Lasting effects of repetitive PMS on MEPs. 647 Time courses of MEP changes in the ECR (A) and FCR (B) muscles were recorded from 17 648 participants. The ordinates show MEP amplitude normalized to baseline amplitude for each 649 participant, and the abscissae show the time point of measurements: before (Pre) and after the 650 repetitive PMS (P0–P50). Each circle and error bar represent the mean value and SD, 651 respectively. Asterisks indicate significant differences compared with “Pre.” 652 653 Figure 3. Effects of ischemia on MEP facilitation induced by repetitive PMS. 654 The time courses of MEP changes in the ECR were recorded from five participants without (A) 655 and with ischemia (B). Individual MEP changes and their mean values are represented as gray 656 lines and circles, respectively. The ordinate and abscissa are the same as those in Figure 2. 657 658 Figure 4. Changes in the firing probability of an ECR motor unit following different 659 intensities of PMS. 660 A, Left: Time histograms were obtained without PMS (white bar) or with PMS to the radial 661 nerve innervating the ECR (black bar) within a sequence. Each histogram was constructed 662 from 600 triggers. Middle: Each column (dark gray bar) represents the difference between the 663 situations with and without PMS. The analysis time window (light gray area) was determined 664 from several consecutive bins of the time histograms obtained during 1.8×MT PMS. Right: 665 Cumulative sums are obtained from each subtracted histogram. The ordinates represent the 666 number of counts as a percentage of the number of triggers, and the abscissae represent the 667 latency after stimulation. Bin width: 0.5 ms. B, Stacked bar graph showing the proportion of 668 motor units that exhibited significant peaks in the time histograms under each stimulation 669 intensity. The numbers in parentheses indicate the number of motor units. C, Changes in the 670 peak amplitude obtained by cumulative sums within the analysis time window of the difference 671 histograms under different stimulation intensities. Individual changes and their mean values 672 are represented as gray lines and circles, respectively. Asterisks indicate significant differences 673 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 25 revealed by post-hoc tests (p < 0.05). 674 675 Figure 5. Changes in the firing probability of an ECR motor unit following electrical and 676 magnetic stimulation. 677 Time histograms were obtained by subtracting the situation without stimulation from that with 678 stimulation. Each histogram was constructed from 400 triggers. The ordinate and abscissa are 679 the same as in Figure 4A. Bin width: 0.2 ms. The difference between the latencies of peaks 680 induced by electrical stimulation (20.4 ms) and magnetic stimulation (22.0 ms) was 1.6 ms. 681 The distance between the two stimulation sites was 10 cm. 682 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 26 683 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 27 684 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 28 685 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 29 686 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint 30 687 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted December 29, 2025. ; https://doi.org/10.64898/2025.12.29.696815doi: bioRxiv preprint