All animal procedures were performed using C57BL/6J mice (JAX: 000664) according to the National Institutes of Health guidelines and approved by the Cincinnati Children’s Hospital Medical Center animal care and use committee’s regulations. Surgical lesions were performed on one lateral half of the spinal cord at C2 in adult male mice (60–176 days of age) under 1 L/min 1% isoflurane, as previously described (Jensen et al., 2024). A 2 cm incision was made at midline beginning at the back of the skull and extending to between the scapula of the animal, exposing the trapezius. The trapezius was then cut at midline caudal to rostral using microscissors to expose the paravertebral muscles. Paravertebral muscles were blunt dissected using cotton tipped applicators to expose the posterior skull and cervical vertebrae. Laminectomy of C1 and C2 vertebrae was performed using microscissors to expose the spinal cord. Once exposed, a durotomy of the spinal cord was performed using forceps, then a 30-gauge needle was inserted near the midline at the rostral edge of the C2 vertebra and passed laterally through the tissue to create a left hemisection injury. Up to 5 needle passes were performed until complete injury was achieved, which was determined by observation of lack of chest wall movement on the side ipsilateral to injury. Paravertebral and trapezius muscles were sutured back together following injury, and the incision was closed using dermal adhesive. Following surgery, 5 mg/kg*BW carprofen was administered subcutaneously for analgesia, nails were trimmed to prevent possible injury during grooming, and the animal was placed into a 32°C incubator overnight. Cages were supplied with nutritional gel and a water bottle. 5 mg/kg*BW carprofen was administered the morning after surgery and as needed for continued analgesic care. Diaphragm EMG recordings/ stimulation experiments were performed between 6 and 12 h after carprofen administration (for recordings performed 1 day post injury). Animals also received subcutaneous injections of 1.0 mL saline on the 1^st and 2^nd post-operative day. Animals which underwent 2 month post-injury recordings received weekly nail trimming to prevent injury when grooming.

Bilateral diaphragm EMG recording was performed under isoflurane in cohorts of uninjured animals, animals 1 day following C2Hx surgery, and 2 months following C2Hx surgery. Recordings were performed under 0.8–1% isoflurane, 1 L/min O2 flow anesthesia with body temperature maintained at 36°C using a heating pad with thermal probe, and animals placed in the prone position with forelimbs and tail secured using surgical tape to stabilize the animal. This anesthetic level was found to prevent any response to a foot pinch (prior to stimulation) and tail pinch (post sciatic nerve cut). Animals initially underwent exposure of the sciatic nerve bilaterally, with 2 cm incisions placed over the back legs of the animal and blunt dissection of the biceps femoris and gluteus maximus to expose the sciatic nerve. A cuff electrode (Microprobes microcuff electrode-consisting of 3 rings of 100um platinum wire spaced 1 mm apart, in an insulated cuff with a 0.5 mm inner diameter and 3 mm at each end outside of wires) was primed with sterile saline and placed around the exposed nerve. The nerve and cuff were covered with mineral oil to prevent drying out of the tissue. A 5 cm incision in the skin at the base of the rib cage exposed the oblique muscles, and 2 cm lateral incisions of the obliques on each side of the spine exposed the caudal surface of the diaphragm through the peritoneal cavity. Two sets of bipolar electrodes (each consisting of 2 NEE-3 needle electrodes (CWE Inc.) taped together with 1.5 mm spacing between electrode tips) connected to a BMA-400 AC/DC bioamplifier were inserted into the left and right side of the diaphragm, with grounded electrodes inserted into the cut external obliques. EMG signals were acquired and analyzed using Spike2 data analysis software (CED). A 30 Hz-10 kHz band pass filter was applied by the amplifier during signal acquisition.

Electrical stimulation was delivered using a WPI Model A365 constant current stimulus isolator. Stimulation was controlled using the Spike2 program, with key activated triggers. The threshold of stimulation required to elicit a motor response (T) was determined by stimulating over a range of currents (1, 5, 10, 15 and 20 μA) until a motor response was observed in the hind paw. The average current required to elicit activation of dorsiflexion/plantarflexion in the ipsilateral paw was found to be 9.3 +/− 1.9uA in healthy mice for both left and right sciatic nerve stimulation. For C2Hx animals at 1 day post injury, the average stimulation to elicit a motor response was 8.1uA+/−2.6 for the left (ipsilateral to injury) and 7.8uA+/− 2.7 for the right (contralateral to injury) sciatic nerve. For C2Hx animals at 2 months post injury, the average stimulation to elicit a motor response was 8.9uA+/− 2.2 for the left and 7.9uA+/−2.7 for the right sciatic nerve. We did not use currents above 200 μA in order to prevent potential activation of high threshold c-fiber nociceptive afferents (Sdrulla et al., 2015). Animals which did not exhibit motor response to stimulation (e.g., due to damage to the sciatic nerve during cuff placement) with both nerve cuffs were excluded from the study. T was assessed independently for both left and right sciatic nerves. The sciatic nerve was cut distal to the cuff following assessment of T to prevent limb movement on subsequent stimulations. Next, animals underwent a brief 15 s nasal occlusion to ensure that maximal chemosensory drive results in rhythmic activation of the paralyzed diaphragm, indicating that crossed phrenic pathways remain intact (Jensen et al., 2024; Mantilla et al., 2011). After nasal occlusion, a 10 min baseline was recorded. Following baseline, animals underwent sciatic nerve stimulation. Stimulation was delivered for 1 min in 20 Hz, 1 ms bipolar pulses, followed by a 1 min period of no stimulation. Except as noted, the same animal underwent consecutive stimulation on the left and right sciatic nerves at 1×, 2×, 5×, 10×, 15×, and 20× T. One cohort of animals underwent a different stimulation paradigm 1 day following C2Hx, with the goal of testing the effects of time under anesthesia on response to stimulation. For these experiments, stimulation was delivered at 10T current only at 1, 10, and 20 min following baseline. Out of a total of 40 animals, we excluded 2 due to incomplete paralysis of the ipsilateral diaphragm at 24 h post injury, 5 for failure to respond to nasal occlusion with rhythmic activity in the paralyzed diaphragm, 3 animals were excluded due to damage to the sciatic nerves during cuff, and 3 due to damage to the diaphragm from the leads during placement and/or animal movement. Additionally, 5 animals had stimulation delivered unilaterally to one nerve only, due to damage of the other sciatic nerve during cuff placement.

Acquired EMG signals were processed using a Finite Impulse Response (FIR) digital high-pass 60 Hz filter to remove DC components of the recording as well as movement artifacts. In channels where ECG activity was greater than inspiratory diaphragm activity, the ECG signal was reduced using “ECGDelete02” script (CED) in Spike2. This script subtracts an averaged QRS complex (ECG contamination) from the EMG signal based on a rolling average of 30 previous ECG complexes. Subsequently the root mean square (RMS) was calculated for the channel using a rolling 50 ms window (Jensen et al., 2024; Mantilla et al., 2014; Mantilla and Sieck, 2013). The RMS signal was used to measure the peak amplitude (highest RMS signal for each inspiratory period) and the tonic amplitude (the minimum RMS amplitude during non-inspiratory intervals) for selected 10 s intervals prior to stimulation, during peak stimulation, during the last 10 s of stimulation, and 10 s immediately after stimulation ceased. Two animals (one at 1 day and one at 2 months post-injury) contained a low frequency/high amplitude artifact (consistent with a movement artifact) during stimulation at 10 and 15T which was not removed by our previously described signal processing. For these animals, only regions of the trace that did not contain the low-frequency/high amplitude artifact were analyzed. Periods of inspiration were determined based on the right (contralateral to injury) diaphragm RMS signal. The inspiratory amplitude was calculated by subtracting the tonic amplitude from the inspiratory amplitude for each breath in the analyzed area. Inspiratory frequency was measured prior to (approximately 30s before stimulation) and during stimulation (the same 10s intervals used for determining peak amplitude) by counting the number of inspiratory peaks over a 10s interval and dividing by 10.

Upon completion of recordings, animals were terminally anesthetized with 0.2 ml pentobarbital, perfused with 6 ml cold phosphate buffered saline (1x PBS), followed by 1 ml*BW 4% PFA in 1xPB, the spinal cord and brainstem dissected and post-fixed in 4% PFA in 1x PB solution for 2 h. Tissue was then washed in 1x PBS overnight at 4°C to remove excess PFA, then cryoprotected in 30% sucrose for between 8 and 24 h. Cervical spinal cord (C1 to C4) was embedded in OCT and kept in a -80C freezer until sectioned (transverse) at 60 μM using a cryostat. Sections were collected sequentially on slides. Eriochrome cyanin stain was performed by processing tissue slides in the following solutions sequentially: 2× 30 min. Xylene, 3 min. Each of 100% ethanol, 95% ethanol, 70% ethanol, 50% ethanol, and 2 min. in dH20. Eriochrome stain was made by mixing 12 ml of 1% FeCl₃ + 2.7% HCl with 240 ml of 0.2% Eriochrome cyanin in 0.5% H₂SO₄ and bringing the total volume to 300 ml using dH₂O. Slides were stained with Eriochrome stain for 10 min, then excess stain was washed off using 2 dips in milliQ H₂O. After washing excess stain, a 30 s differentiation step was performed in a solution of 0.5% aqueous NH₄OH in dH₂O, followed by washing off excess solution by dipping in dH₂O twice. Slides was then air dried overnight in a chemical fume hood. The following day, slides were placed into xylene for 10 min, and allowed to dry for 20 min prior to cover slipping with Permount. Permount was allowed to harden overnight prior to further handling. Tissue was imaged using a Nikon Ni-E upright motorized microscope using a 4x objective lens. The section with maximal damage was assessed empirically, then used to calculate the extent of injury (Jensen et al., 2024; Mantilla et al., 2014). Using NIS-Elements, the spared white and grey matter of the cord were outlined and the area was calculated (TotalSparedTissueArea). To estimate the total area of the cord, the area of the uninjured half of the cord (HemicordArea) was doubled. Percent spared tissue area was calculated using the following calculation: % Spared = TotalSparedTissueArea 2 x HemicordArea x 100

Repeated measures statistics were used to analyze the changes in EMG amplitude and respiratory rate between stimulated and unstimulated conditions. ANOVAs were utilized where data collected was found to be normally distributed, while a Friedman’s F-test (for non-parametric data) was used when the distribution was not found to be normal (see Tables 1–3). For each group of animals, peak amplitude was compared at pre-stimulation baseline versus the 1, 2, 5, 10, and 15T stimulation conditions for each side of the diaphragm (left and right) and each sciatic nerve stimulated (left and right; see Tables 1–3). Tonic and inspiratory amplitude were analyzed similarly (see Tables 1–3). Dunn’s test (which controls for multiple comparisons) was used for post-hoc analysis to compare pre-stimulation baseline to each intensity of stimulation. A multivariate repeated measures ANOVA was used to analyze the changes in frequency during stimulation, comparing the frequency immediately prior to a given stimulation versus the frequency during stimulation for each of 1, 2, 5, 10 and 15T. Comparisons were made between the pre- and post-stimulated values for each stimulation tested, and the Bonferroni correction was used to control for multiple comparisons when using an ANOVA. In all tests, a p-value below 0.05 was considered statistically significant. P and F values for each test are given in Tables 1–3. Power analyses (α = 0.05, power = 80%) determined that we had a sufficient number of animals to detect a 49% change from baseline in uninjured animals, 63% change from baseline in the diaphragm ipsilateral to injury in the one-day post C2Hx group, and an 81% change from baseline in the diaphragm ipsilateral to injury in the two-months post C2Hx group.

Figures 1A, 4A were created with BioRender.com.

We tested whether electrical stimulation of hindlimb afferents could increase diaphragm activity in mice, as has previously been shown in other species (Haxhiu et al., 1984; Kanbar et al., 2016; Korsak et al., 2018; Mizumura and Kumazawa, 1976; Schiefer et al., 2018). In anesthetized, freely breathing adult mice, a cuff electrode was applied to the sciatic nerve on both the left and right hindlimbs. The sciatic nerve contains sensory and motor axons innervating a significant portion of the skin and muscles of the leg and foot. Needle electrodes were inserted into each side of the diaphragm to record electromyograph (EMG) activity (Figure 1A). Electrical pulses (1 ms pulse width, 20 Hz, 1 min train; Figure 1A) were delivered over a range of currents to determine the threshold sufficient to elicit a motor response (T) in the hind paw distal to each cuff electrode. The sciatic nerve distal to each cuff was cut to eliminate further hindlimb responses. We stimulated one sciatic nerve at a time (alternating between left and right), sequentially at currents corresponding to 1, 2, 5, 10, 15, and 20 times threshold (1–20T). However, since 20T often resulted in strong thoracic movements such that the diaphragm was occasionally damaged by the inserted electrodes, we eliminated this time point from our analyses. No movement of the head was noted during these high threshold stimulations. We elected to perform all the experiments in this study in non-ventilated, non-vagotomized mice to preserve their ability to respond to changes in blood gasses and simulate what might be observed in healthy or injured humans undergoing stimulation. As expected, electrical stimulation of the sciatic nerve at sufficient thresholds could increase the amplitude of inspiratory diaphragm activity during stimulation (Figure 1B). Diaphragm activity is increased both ipsilateral and contralateral to the stimulated nerve. Prolonged stimulation led to an attenuation of these effects by the end of the 60s period of stimulation (Figure 1B). To analyze in more detail the effects of sciatic nerve stimulation on diaphragm activity, we compared the EMG activity on both sides of the diaphragm during a 10s baseline period (prior to any stimulation) as well as a 10s period containing the peak activation of the diaphragm during stimulation (starting between 4 and 25 s after the onset of stimulation). The average duration of the experiment was 10 min (baseline) plus 28+/−9 min for stimulations. For each period, we analyzed the root mean square (RMS) of the diaphragm activity and used this to measure the peak amplitude (during inspiration) and tonic amplitude (during expiration) of the diaphragm EMG signal. To calculate the inspiratory amplitude, we subtracted the tonic from the peak amplitude for each inspiratory burst (Figure 1C). Examples of diaphragm EMG and RMS traces for the left and right diaphragm prior to and during stimulation are shown in Figures 1C–F. We measured peak diaphragm activity for each condition (left/right nerve stimulation and left/right diaphragm EMG; Figures 2A,B). We observed an increase in peak amplitude with left sciatic nerve stimulation that was statistically significant (Table 1). Tonic activity was also elevated significantly by stimulation of the left sciatic nerve (Figures 2C,D; Table 1). The inspiratory amplitude also showed a tendency towards increased amplitude upon stimulation but was only statistically significant for the left diaphragm following left sciatic nerve stimulation (Figures 2E,F; Table 1). Thus, stimulation of the sciatic nerve at high thresholds (10–15T) can elicit an increase in peak, tonic and inspiratory amplitude, but the degree of change from baseline (normal inspiratory activity) is variable in uninjured mice.

We next examined changes in respiratory rate following sciatic nerve stimulation. The number of inspiratory peaks in the RMS signal were counted for a 10s window just prior to stimulation and for a 10s window during stimulation at the time of peak diaphragm activation. There was a significant increase in breathing frequency during left and right sciatic nerve stimulation compared to baseline (Table 1). This increase was seen at 5–15T stimulation, but not at lower currents (Figures 3A,B). The elevated respiratory rate persisted through the last 10s of stimulation (Table 4). Thus, our results confirm that sciatic nerve stimulation can increase both the frequency and amplitude of diaphragm activity in uninjured anesthetized mice.

Having established that sciatic nerve stimulation increases diaphragm output in healthy mice, we tested the effects of stimulation on diaphragm function after a cervical spinal cord injury. We utilized a C2 hemisection injury (C2Hx) in which the left half of the spinal cord is surgically transected at C2, paralyzing the diaphragm ipsilateral to the injury while the diaphragm contralateral to injury continues to contract during inspiration and maintain ventilation (Figure 4A). Post-hoc evaluation of the extent of injury was performed by harvesting the cervical cords, sectioning, and staining with eriochrome cyanin to detect myelin. We quantified the extent of spared tissue in the section containing the most damage and expressed it as a percentage of the expected cross-sectional area of the intact cord (Figure 4B). The spared cord area averaged 57+/− 6% (mean +/− SD), with a range of 51 to 69%.

One day following injury, animals were anesthetized, nerve cuff electrodes applied to the left (ipsilateral to injury) and right (contralateral to injury) sciatic nerves, and EMG electrodes inserted into the left and right sides of the diaphragm. Previous studies demonstrated that at 1 day after injury there is no significant spontaneous recovery but one can reliably activate the paralyzed diaphragm using maximal chemosensory drive (i.e., the crossed phrenic phenomenon; Jensen et al., 2024; Minor et al., 2006) or by chemogenetic activation of propriospinal neurons (Jensen et al., 2024). Prior to any electrical stimulation, we measured inspiratory amplitude from the RMS of the diaphragm EMG as described for uninjured animals, using the contralateral diaphragm to assess periods of inspiration. As expected, minimal inspiratory amplitude was observed in the left (ipsilateral to injury) diaphragm of C2Hx animals (0.7 +/− 0.3 mV) compared to uninjured mice (3.5 +/− 1.1 mV). Further, the inspiratory amplitude of the contralateral diaphragm (11.0 +/− 6.6 mV) was increased in C2Hx mice compared to uninjured mice (3.2 +/− 1.6 mV), likely compensating for the paralyzed diaphragm as found in previous studies (Jensen et al., 2024; Minor et al., 2006).

To assess the effects of electrical stimulation of the sciatic nerve on diaphragm activity in C2Hx injured mice, we stimulated one sciatic nerve at a time (alternating between left and right), sequentially at currents corresponding to 1, 2, 5, 10, 15, and 20 times threshold (1–20T), as previously done for uninjured mice. The average duration of the experiment was 10 min (baseline) plus 22+/−4 min for stimulations. We excluded 20T stimulation from our analyses due to the potential for diaphragm damage in some animals. Stimulations above 5T typically resulted in an increase in diaphragm EMG peak amplitude ipsilateral to injury as well as an increase in respiratory frequency (Figure 4C). In contrast to healthy mice that all exhibited an attenuation of the increase in peak amplitude over 60s of stimulation, the majority of C2Hx mice showed a persistent increase in peak amplitude (6/8 mice) and frequency (6/8 mice) throughout the 60s of 10T stimulation. An example of the increase in diaphragm activity during stimulation at 10T compared to the unstimulated condition is shown in Figures 4D–G. Note that stimulation restores inspiratory bursting to the previously paralyzed diaphragm that is synchronous with inspiratory activity of the diaphragm contralateral to injury. We quantified the peak, tonic and inspiratory diaphragm amplitude prior to stimulation as well as during 10s of peak activity during stimulation for each condition (left/right sciatic nerve stimulation, diaphragm EMG ipsilateral/contralateral to injury, 0–15T stimulation). Table 2 describes the statistical tests performed and resulting F- and P-values. Peak amplitude was significantly elevated in the diaphragm ipsilateral to injury with both left and right sciatic nerve stimulation (10 and 15T; Figure 5A and Table 2). There was also a small (but statistically significant) increase in peak amplitude of the contralateral (not paralyzed) diaphragm during stimulation (Figure 5B). Tonic activity was significantly increased during stimulation of the left sciatic nerve, but not the right sciatic nerve (Figures 5C,D; Table 2). The inspiratory amplitude of the diaphragm ipsilateral to injury was significantly increased by electrical stimulation of either the left (ipsilateral to injury) or right (contralateral to injury) sciatic nerve at thresholds of 10T and 15T, but not at lower thresholds (Figure 5E). The inspiratory amplitude of the contralateral diaphragm was also significantly increased by stimulation (Figure 5F; Table 2). Our results show that electrical stimulation of either sciatic nerve can restore rhythmic inspiratory activity to the previously paralyzed diaphragm after a C2Hx injury.

The effects of electrical stimulation on respiratory frequency in C2Hx injured animals was measured by comparing the respiratory frequency (measured from the diaphragm RMS contralateral to injury) for a 10s window just prior to stimulation to a 10s window during the peak of stimulation, as described previously for uninjured mice. Stimulation at 5, 10, and 15T resulted in significant increases in respiratory rate for both left (Figure 6A) and right (Figure 6B) sciatic nerve stimulation. An elevated respiratory rate persisted through the last 10 s of stimulation (Table 4). Our results show that electrical stimulation of the sciatic nerve can increase both the frequency and amplitude of diaphragm activity after a C2Hx injury.

We observed a decrease in respiratory frequency prior to stimulation over the course of the experimental protocol (Figure 6), which we hypothesized could be due to a “run-down” effect of anesthesia. Thus, we assessed the potential effect that prolonged anesthesia could have on the ability to activate the diaphragm by sciatic nerve stimulation at a single current threshold (10T). We performed C2Hx injuries and the following day applied sciatic nerve stimulation and recorded diaphragm EMG as previously described. After an initial 10 min baseline period (similar to our previous experiments), animals were stimulated at 10T at 1 min, 10 min, and 20 min time points. The 20 min time point corresponds to the period during which 10T stimulations would have occurred in our previous experiments on healthy and C2Hx animals. Stimulation at 10T consistently produced an increase in inspiratory amplitude in the diaphragm ipsilateral to injury at 1, 10, and 20 min compared to pre-stimulation (Figure 7; Table 2). However, there was not a statistically significant difference in inspiratory amplitude between the 1, 10, and 20 min time points following either left or right sciatic nerve stimulation (Figure 7). Further, no significant difference was observed between inspiratory amplitude following 10T left sciatic nerve stimulation at 1 min in this group of animals (3.0 +/− 1.2 mV) and during the 10T stimulation in the group of animals shown in Figure 5 (2.0 +/− 1.3 mV, Student’s T-test, p = 0.1770). A similar lack of significance was noted following 10T stimulation of the right sciatic nerve at 1 min (1.8 +/− 0.69 mV) and during 10T stimulation in the group of animals shown in Figure 5 (2.6 +/− 1.3 mV, Student’s T-test p = 0.2820). Eriochrome staining showed that this group experienced a comparable amount of spared tissue (51 +/− 2% of the cord) following injury as the prior C2Hx injury group (Figure 4; p = 0.39). These results demonstrate that 10T stimulation elicits a consistent increase in diaphragm inspiratory amplitude within the time frame of our experiments.

We next tested the ability of sciatic nerve stimulation to restore diaphragm function in mice at chronic stages (2 months) after a C2Hx injury. A cohort of mice underwent a C2Hx injury followed by 2 months of recovery time. Using the same nerve stimulation and diaphragm EMG recording preparation described previously, we stimulated one sciatic nerve at a time (alternating between left and right), sequentially at currents corresponding to 1, 2, 5, 10, 15 and 20 times threshold (1 T-20 T). As with prior experiments, we excluded 20 T stimulations from our analyses due to the potential for diaphragm damage. Prior to stimulation, we observed significant spontaneous recovery in 1 out of 10 mice, so we did not include that animal in our analysis of the effects of nerve stimulation. The remaining 9 animals experienced little spontaneous recovery and the baseline (pre-stimulation) inspiratory amplitude was not significantly different from the 1 day post-C2Hx cohort of animals for the diaphragm ipsilateral to injury (1 day post injury 0.73 +/− 0.33 mV, 2 month post injury 0.79 +/− 0.49 mV, Student’s T-test p = 0.7863) or for the diaphragm contralateral to injury (1 day post injury 11.0 +/− 6.6 mV, 2 month post injury 9.0 +/− 5.3 mV, Student’s T-test, p = 0.5030). The extent of injury analysis showed comparable spared tissue area between groups (57 +/− 6% 1 day cohort; 55 +/− 9% 2 month cohort, p > 0.99). The average duration of the experiment was 10 min (baseline) plus 23+/−5 min for stimulations.

As for animals 1 day after injury, stimulation of the sciatic nerve at 10T could produce rhythmic inspiratory activity in the previously paralyzed diaphragm 2 months following a C2Hx injury (Figures 8A–D). Moreover, peak amplitude was significantly increased in the paralyzed diaphragm at stimulations between 5 and 15T (Figure 8E and Table 3). Thus, at chronic stages, sciatic nerve stimulation could increase diaphragm activity at even lower thresholds (5T) than was typically observed 1 day after injury. The effect of stimulation on the diaphragm contralateral to injury (not paralyzed) was only statistically significant for stimulation of the left sciatic nerve at 15T (Figure 8F). Similar to 1 day after injury, sciatic nerve stimulation could increase tonic diaphragm activity in both the paralyzed and contralateral diaphragm following left sciatic nerve stimulation (Figures 8G,H; Table 3).

Similar to the peak amplitude, we found that stimulation of the left or right sciatic nerve could increase the inspiratory amplitude of the diaphragm ipsilateral to injury at thresholds between 5 and 15T (Figure 8I; Table 3). In fact, stimulation at the highest thresholds was able to restore peak inspiratory diaphragm activity to levels comparable to uninjured mice. The inspiratory amplitude of the diaphragm contralateral to injury was not consistently increased by stimulation, although some individual animals showed a response (Figure 8J; Table 3). Following the withdrawal of stimulation (the 10s immediately following stimulation), inspiratory amplitude returned to pre-stimulation levels. For example, the inspiratory amplitude increased from 1.5 +/− 0.93 mV prior to left sciatic stimulation, to 3.4 +/− 2.4 mV during stimulation at 10T, and returned to 1.6 +/− 1.9 mV after stimulation (Friedman’s F test, F = 8.00, p = 0.0476 overall, pre to peak stimulation p = 0.0368, pre to post stimulation p > 0.9999). A similar trend was observed following right sciatic nerve stimulation where the inspiratory amplitude increased from 1.3 +/− 0.83 mV pre-stimulation to 2.1 +/− 1.0 mV during 10T stimulation, then returning to 1.3 +/− 0.70 mV following stimulation (Friedman’s F test, F = 6.22, p = 0.0476 overall, pre to peak stimulation p = 0.0151, pre to post stimulation p > 0.9999). As with animals 1 day following injury, stimulation at 5-15T increased respiratory frequency (Figure 9; Table 3). The increase in frequency declined near the end of the 60s stimulation period (Table 4). Thus, at chronic stages, even lower stimulation thresholds are required to elicit significant increases in inspiratory diaphragm activity ipsilateral to a C2Hx compared to 1 day after injury.