Somatosensory motion-dependent feedback is a critical component of the locomotor control system, and terrestrial legged animals from insects to humans have evolved similar sensory modalities and control mechanisms that provide for stable and efficient locomotion; see recent reviews (Edwards & Prilutsky, 2017; Frigon et al., 2021; Tuthill & Azim, 2018). Specifically, length- and load-dependent sensory feedback entrains locomotor rhythm (Kriellaars et al., 1994; Pearson & Collins, 1993) and regulates the level of muscle or motoneuronal activity (af Klint et al., 2010; Akay et al., 2004; Chung et al., 2015; Guertin et al., 1995; Lam & Pearson, 2002) and timing of the transitions between the swing (flexor) and stance (extensor) phases of the locomotor cycle (Schomburg et al., 1998; Stecina et al., 2005). The lack of somatosensory feedback due to genetic mutations, viral infection, or pharmacological or surgical interventions causes severe locomotor impairments (Akay et al., 2014; Chesler et al., 2016; Cole & Paillar, 1998; Goldberger, 1988; Housley et al., 2021).
Complete loss of somatosensory feedback in people is rare. Rather, specific somatosensory modalities from selected parts of the body may be affected to a different extent by aging (Goble et al., 2009; Henry & Baudry, 2019), disease (Crowell & Gwathmey, 2017; Dyck et al., 2013), or limb loss (Bensmaia et al., 2020). The specific somatosensory pathologies, for example, compromised tactile sensations from the feet, still negatively affect locomotor function and balance control. For example, several studies in humans and cats have demonstrated that this type of sensory feedback is critical for maintaining balance during postural tasks (Honeycutt & Nichols, 2010; Meyer et al., 2004), for recovery of balance after postural perturbations during locomotion (Bolton & Misiaszek, 2009; Hohne et al., 2011), and for the successful performance of demanding locomotor motor behaviors, for example, walking on a horizontal ladder or slopped surface (Bouyer & Rossignol, 2003). Cutaneous paw pad afferents are also involved in regulation of the stance phase duration during locomotion. A low-intensity electrical stimulation of the foot plantar skin or its cutaneous afferents in the sural or distal tibial nerves during the stance (or extensor) phase of walking enhances the activity of extensor muscles and increases the duration of the stance (extensor) phase (Duysens & Pearson, 1976; Guertin et al., 1995; Loeb, 1993; Ollivier-Lanvin et al., 2011), whereas the same stimulation of these nerves during flexor activity may trigger the transition from the flexor to extensor activity (Guertin et al., 1995) or enhance flexor activity (Duysens & Pearson, 1976; Ollivier-Lanvin et al., 2011).
In addition to modulating locomotor activity in quadrupedal animals, electrical stimulation of intact or residual cutaneous nerves could also improve locomotor performance in people with spinal cord injury (Knikou, 2010) or evoke natural somatosensation in people with limb loss (Christie et al., 2020; George et al., 2020; Graczyk et al., 2022) and improve their motor performance (Christie et al., 2020; Raspopovic et al., 2014; Tan et al., 2014). Quality of tactile perception strongly depends on stimulation parameters of cutaneous or mixed nerves, that is, pulse shape and duration, stimulation frequency, and magnitude (Charkhkar et al., 2018; Tan et al., 2014).
In recent studies, we have found additional evidence of the important role of cutaneous sensory feedback in the regulation of frontal balance and interlimb coordination during quadrupedal locomotion (Latash et al., 2020; Park et al., 2019). Unilateral paw pad anesthesia in cats walking on a split-belt treadmill with different speed ratios caused a shift of the center of mass (COM) toward the anesthetized side, a decrease in the relative duration of two-leg and three-leg support periods, an increase in the four-leg support time, an increase in the hindlimb step width, and improvements in dynamic stability. Most of these changes were consistent with locomotor adaptations aimed to improve lateral balance during walking in quadrupeds (Bolton & Misiaszek, 2009; Farrell, Bulgakova, et al., 2014; Galvez-Lopez et al., 2011) and in humans (Dunlap et al., 2012; Onushko et al., 2019). These changes, however, occurred typically at speed ratios above 1.0, supporting the previous suggestions that cutaneous sensory feedback from paw pads is important for challenging locomotor tasks (Bouyer & Rossignol, 2003).
Given similar changes in frontal plane COM displacements and measures of dynamic stability with increasing speed ratios of split-belt treadmill walking between cats (Latash et al., 2020; Park et al., 2019) and humans (Buurke et al., 2018, 2019) and the ability to easily modulate locomotor mechanics by manipulating cutaneous feedback by paw pad anesthesia and electrical stimulation of cutaneous afferents, the cat appears to be a convenient animal model to investigate locomotor compensations to compromised tactile sensations and their potential restoration.
The goal of this study was to examine whether electrical stimulation of the distal tibial nerve during the stance phase could compensate for the effects of unilateral anesthesia of the hindpaw and forepaw during split-belt treadmill walking in cats. We tested the hypothesis that locomotor changes caused by paw anesthesia could be reduced or reversed by electrical stimulation of cutaneous and proprioceptive afferents in the distal tibial nerve during stance. In addition, we investigated the effects of similar electrical stimulation of the residual distal tibial nerve on walking mechanics of one cat with a unilateral, transtibial, bone-anchored prosthesis (Park et al., 2018), lacking sensory feedback from the foot. We hypothesized that the nerve stimulation, in this case, would cause locomotor changes comparable with those in cats with anesthetized paws.
Methods
All surgical and experimental procedures involving animals followed the Guide for the Care and Use of Laboratory Animals. Eighth Edition (National Research Council of the National Academies, 2011) and were approved by the Institutional Animal Care and Use Committees of the Georgia Institute Technology (Protocols 16089 and 100013) and T3 Laboratories (Protocol GT27F). Cats of both sexes were included in the study. Female and male cats were housed separately in a single room of animal facilities at Georgia Tech and T3 Laboratories, respectively. Most of the surgical and experimental procedures have been described elsewhere (Jarrell et al., 2018; Klishko et al., 2021; Ollivier-Lanvin et al., 2011; Park et al., 2019; Pitkin et al., 2021; Prilutsky et al., 2011) and, thus, only briefly outlined here.
Surgical Procedures for Experiments With Cutaneous Anesthesia and Stimulation
Prior to surgery, four adult female cats (mass 2.55–4.11 kg; Table 1) underwent locomotor training on a split-belt treadmill with the speed ratios between the right and left belts of 0.4:0.8 m/s (speed ratio 0.5), 0.4:0.4 m/s (ratio 1.0), 0.6:0.4 m/s (ratio 1.5), and 0.8:0.4 m/s (ratio 2.0). Subsequently, during a survival surgery in aseptic conditions and under general isoflurane anesthesia, a bipolar nerve cuff stimulating electrode was implanted on the right distal tibial nerve, and a tripolar recording nerve cuff electrode was placed on the right sciatic nerve. The nerve cuff electrodes were made of platinum foil (thickness 0.025 mm and width 3 mm) embedded in silicon (Sigma-Aldrich) as described in Ollivier-Lanvin et al. (2011). In addition, Teflon-insulated multistranded stainless steel fine wires (CW5402, Cooner Wire) were implanted in selected muscles of the right hindlimb for electromyography (EMG) recordings. Electrode leads were directed under the skin along the back to two multipin Amphenol connectors attached to the skull with screws and dental cement. The animal’s physiological variables (temperature, blood pressure, heart rate, etc.) were continuously monitored during the surgery. After implantation, the skin was closed in layers using Vicryl 4-0 and 5-0 sutures. The animals recovered after surgery for 2 weeks with the administration of pain medication for 3 days (fentanyl transdermal patch, 12–25 μg/hr and/or buprenorphine, s.c., 0.01 mg/kg; or ketoprofen, 2 mg/kg, s.c.) and antibiotics for 10 days (Cefovecin 8 mg/kg, s.c. or Ceftiofur 4 mg/kg, s.c.).
Number of Analyzed Cycles for Each Animal and Condition in Experiments With Paw Pad Anesthesia and Distal Tibial Nerve Stimulation
| Cat ID | Mass (kg) | Speed ratio | Intact walking | Paw anesthesia | Paw anesthesia @ stimulation |
|---|---|---|---|---|---|
| MO | 4.11 | 0.5 | 76 | 80 | 38 |
| 1.0 | 47 | 41 | 37 | ||
| 1.5 | 31 | 41 | 37 | ||
| 2.0 | 38 | 74 | 53 | ||
| NO | 3.91 | 0.5 | 60 | 13 | 3 |
| 1.0 | 31 | 14 | 30 | ||
| 1.5 | 24 | 22 | 35 | ||
| 2.0 | 15 | 14 | 7 | ||
| TA | 3.00 | 0.5 | — | — | — |
| 1.0 | 51 | 49 | 61 | ||
| 1.5 | 69 | 65 | 28 | ||
| 2.0 | 23 | 19 | — | ||
| WE | 2.55 | 0.5 | 31 | 47 | 42 |
| 1.0 | 33 | 30 | 31 | ||
| 1.5 | 75 | 74 | 30 | ||
| 2.0 | 34 | 25 | 43 | ||
| Mean mass ± SD: 3.39 ± 0.74 | Total: 638 | Total: 608 | Total: 475 | ||
Determining Stimulation Threshold
Following recovery, we established the stimulation threshold (T) for activating afferents in the distal tibial nerve in each awake cat by recording the compound action potential in the sciatic cuff electrode in response to distal tibial nerve stimulations of different intensities (Ollivier-Lanvin et al., 2011). We selected nerve stimulation parameters similar to those that evoked natural touch perception in the phantom hand or foot of human subjects with limb loss whose residual cutaneous and mixed nerves were electrically stimulated (Charkhkar et al., 2018; Tan et al., 2014). We stimulated the distal tibial nerve by trains of 200-μs biphasic rectangular pulses at 100 Hz for 500 ms duration (Figure 1a1,2). The duration of stimulation was selected based on the approximate mean duration of the stance phase during walking across all experimental conditions. Stimulation trains of a given strength were delivered to the distal tibial nerve for 500 ms, and the responses were recorded in the sciatic nerve while the awake animal was eating and being petted by the researcher. Each stimulation train was separated by a period of 2–3 min. The stimulation strength varied from 5 μA to 1.5 mA with a logarithmic scale of steps from 2.5 to 250 μA. An example of sciatic nerve activity in response to distal tibial nerve stimulation is shown in Figure 1a2. The smallest magnitude of stimulation that evoked a detectable response in the sciatic nerve was ±10–20 μA and was considered an activation threshold of sensory afferents. These afferents include low-threshold Groups I and II afferents innervating the spindles and Golgi tendon organs of foot intrinsic muscles and cutaneous mechanoreceptors in the plantar skin of the foot and paw. The conduction velocities of the activated afferents (estimated as the ratio of the distance between the tibial and sciatic cuff electrodes and the latency of the evoked response) were in the range of Groups I and II muscle spindle and Golgi tendon afferents as well as cutaneous afferents of hair and touch and pressure receptors (approximate range of conduction velocities 25–90 m/s and fiber diameters 5–15 μm; Boyd & Kalu, 1979; Burgess et al., 1968; Hunt & McIntyre, 1960). Muscle twitches in response to distal tibial nerve stimulation typically occurred at stimulation strength of 0.2–0.5 mA in all four cats, which was consistent with a previously reported threshold for recruiting efferent fibers in a cat mixed nerve by biphasic stimulation trains through a cuff nerve electrode (Gorman & Mortimer, 1983). The threshold determinations were performed periodically over the course of the study and were found to be stable.
—Illustration of experimental methods. (a) Characteristics of electrical stimulation of the distal tibial nerve. (1) Profile of train of 200-μs biphasic rectangular pulses. (2) Compound action potential recorded in the sciatic nerve evoked in the awake cat (cat WE) by stimulation of the distal tibial nerve at threshold (T) 1 × T and 2 × T for 500 ms. (b) Illustrations of the implanted pylon with nerve cuff electrodes and the passive and powered transtibial prostheses. (b1) X-ray image of the porous titanium implant (ii) inside the residual tibia. Leads from implanted electromyography (EMG) electrodes and nerve cuff electrodes on the residual tibial nerve (iv) and sciatic nerve (v) are passed through the tibia and the channel inside the pylon to the connector secured outside the pylon in a metal container (iii); a J-shaped passive prosthesis (i) attached to the pylon. (b2) Powered sensing transtibial prosthesis. (i) Prosthetic foot with a pressure transducer glued to the foot plantar surface. (ii) Aluminum frame to hold prosthesis components and to attach to the external part of the pylon. (iii) Attachment point of the prosthesis linear actuator. (iv) “Ankle” joint. (v) Connector for connecting leads from implanted electrodes. (c1) Examples of experimental recordings of soleus (SO) EMG activity and foot pressure of the cat during level walking (see text for explanations). (c2) Image of longitudinal section through the right tibia with implanted pylon stained with H&E. Black areas below and above white area are porous titanium with channel inside, respectively. The image demonstrates skin and bone ingrowth inside the pylon pores with good apposition of a porous implant to the surrounding bone and skin surface (see areas marked by the three rectangles).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Illustration of experimental methods. (a) Characteristics of electrical stimulation of the distal tibial nerve. (1) Profile of train of 200-μs biphasic rectangular pulses. (2) Compound action potential recorded in the sciatic nerve evoked in the awake cat (cat WE) by stimulation of the distal tibial nerve at threshold (T) 1 × T and 2 × T for 500 ms. (b) Illustrations of the implanted pylon with nerve cuff electrodes and the passive and powered transtibial prostheses. (b1) X-ray image of the porous titanium implant (ii) inside the residual tibia. Leads from implanted electromyography (EMG) electrodes and nerve cuff electrodes on the residual tibial nerve (iv) and sciatic nerve (v) are passed through the tibia and the channel inside the pylon to the connector secured outside the pylon in a metal container (iii); a J-shaped passive prosthesis (i) attached to the pylon. (b2) Powered sensing transtibial prosthesis. (i) Prosthetic foot with a pressure transducer glued to the foot plantar surface. (ii) Aluminum frame to hold prosthesis components and to attach to the external part of the pylon. (iii) Attachment point of the prosthesis linear actuator. (iv) “Ankle” joint. (v) Connector for connecting leads from implanted electrodes. (c1) Examples of experimental recordings of soleus (SO) EMG activity and foot pressure of the cat during level walking (see text for explanations). (c2) Image of longitudinal section through the right tibia with implanted pylon stained with H&E. Black areas below and above white area are porous titanium with channel inside, respectively. The image demonstrates skin and bone ingrowth inside the pylon pores with good apposition of a porous implant to the surrounding bone and skin surface (see areas marked by the three rectangles).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Illustration of experimental methods. (a) Characteristics of electrical stimulation of the distal tibial nerve. (1) Profile of train of 200-μs biphasic rectangular pulses. (2) Compound action potential recorded in the sciatic nerve evoked in the awake cat (cat WE) by stimulation of the distal tibial nerve at threshold (T) 1 × T and 2 × T for 500 ms. (b) Illustrations of the implanted pylon with nerve cuff electrodes and the passive and powered transtibial prostheses. (b1) X-ray image of the porous titanium implant (ii) inside the residual tibia. Leads from implanted electromyography (EMG) electrodes and nerve cuff electrodes on the residual tibial nerve (iv) and sciatic nerve (v) are passed through the tibia and the channel inside the pylon to the connector secured outside the pylon in a metal container (iii); a J-shaped passive prosthesis (i) attached to the pylon. (b2) Powered sensing transtibial prosthesis. (i) Prosthetic foot with a pressure transducer glued to the foot plantar surface. (ii) Aluminum frame to hold prosthesis components and to attach to the external part of the pylon. (iii) Attachment point of the prosthesis linear actuator. (iv) “Ankle” joint. (v) Connector for connecting leads from implanted electrodes. (c1) Examples of experimental recordings of soleus (SO) EMG activity and foot pressure of the cat during level walking (see text for explanations). (c2) Image of longitudinal section through the right tibia with implanted pylon stained with H&E. Black areas below and above white area are porous titanium with channel inside, respectively. The image demonstrates skin and bone ingrowth inside the pylon pores with good apposition of a porous implant to the surrounding bone and skin surface (see areas marked by the three rectangles).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Effects of electrical stimulation of the distal tibial nerve on the kinematic variables of split-belt treadmill walking with different speed ratios while the ipsilateral (right) paw pads of the fore- and hindlimb are anesthetized. Presented mean (±95% confidence interval) values of kinematic variables were computed across all animals and cycles; see Table 1. (a) Center of mass (COM) bias, that is, the mean COM displacement during the cycle in the lateral direction with respect to lateral positions of the paws on the ground. (b) Durations of the cycle, stance, and swing phases of the right forelimb. (c) Durations of the cycle, stance, and swing phases of the right hindlimb. (d) Duty factor of the right forelimb. (e) Duty factor of the right hindlimb. (f) Pacing of the right (ipsilateral) limbs. (g) Step length of the right forelimb. (h) Step length of the right hindlimb.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Effects of electrical stimulation of the distal tibial nerve on the kinematic variables of split-belt treadmill walking with different speed ratios while the ipsilateral (right) paw pads of the fore- and hindlimb are anesthetized. Presented mean (±95% confidence interval) values of kinematic variables were computed across all animals and cycles; see Table 1. (a) Center of mass (COM) bias, that is, the mean COM displacement during the cycle in the lateral direction with respect to lateral positions of the paws on the ground. (b) Durations of the cycle, stance, and swing phases of the right forelimb. (c) Durations of the cycle, stance, and swing phases of the right hindlimb. (d) Duty factor of the right forelimb. (e) Duty factor of the right hindlimb. (f) Pacing of the right (ipsilateral) limbs. (g) Step length of the right forelimb. (h) Step length of the right hindlimb.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Effects of electrical stimulation of the distal tibial nerve on the kinematic variables of split-belt treadmill walking with different speed ratios while the ipsilateral (right) paw pads of the fore- and hindlimb are anesthetized. Presented mean (±95% confidence interval) values of kinematic variables were computed across all animals and cycles; see Table 1. (a) Center of mass (COM) bias, that is, the mean COM displacement during the cycle in the lateral direction with respect to lateral positions of the paws on the ground. (b) Durations of the cycle, stance, and swing phases of the right forelimb. (c) Durations of the cycle, stance, and swing phases of the right hindlimb. (d) Duty factor of the right forelimb. (e) Duty factor of the right hindlimb. (f) Pacing of the right (ipsilateral) limbs. (g) Step length of the right forelimb. (h) Step length of the right hindlimb.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
These four animals participated in locomotor experiments described later as well as in additional studies, including muscle self-reinnervation (Gregor et al., 2018) and stretch reflexes (Lyle et al., 2016). At completion of these studies, the animals were euthanized with either an overdose of concentrated pentobarbital (Euthasol) or potassium chloride as recommended by guidelines from the American Veterinary Medical Association.
Surgical Procedures for Implanting Bone-Anchored Transtibial Prosthesis
As part of another study on developing a powered sensing transtibial prosthesis integrated with the bone, skin, muscles, and peripheral nerves (Farrell, Prilutsky, Kistenberg, et al., 2014; Farrell, Prilutsky, Ritter, et al., 2014; Park et al., 2018; Pitkin et al., 2021), we also investigated one male cat (mass 4.7 kg). This animal was trained to walk overground inside a horizontal Plexiglass enclosed walkway with three force plates (Bertec) embedded in the walkway floor. After locomotor training and baseline locomotor recording, a two-stage implantation procedure was performed in aseptic conditions and under general isoflurane anesthesia. In the first surgery, we implanted in the distal part of the right tibia the porous titanium pylon (pore size ∼100 μm, pore volume fraction >50%) with a channel inside (SBIP-PNI, Poly-Orth International) for interfacing a prosthesis with residual muscles and peripheral nerves (Pitkin & Raykhtsaum, 2012; Pitkin et al., 2012). The details of the distal tibia amputation and pylon implantation have been described previously (Farrell, Prilutsky, Kistenberg, et al., 2014). During this first surgery, the distal tibial nerve was transected distally and resutured to the skin on the proximal shank, and the passage for electrode leads inside the residual tibia was prepared. We placed a cast on the residual limb to secure the implanted pylon for 70 days to allow integration of the pylon with the residuum (Farrell, Prilutsky, Kistenberg, et al., 2014). During the second surgery, we implanted fine wires in the residual SO for the recording of EMG signals and nerve cuff electrodes on the residual distal tibial nerve and the sciatic nerve (Figure 1b1iv,v). The leads from these electrodes were passed through the bone and pylon channel (Figure 1b1ii), attached to the connector, which was secured in a small aluminum container inside an aluminum frame fixed on the external end of the pylon (Figure 1b1iii). A J-shaped passive prosthesis (Figure 1b1i) was then attached to the aluminum frame. Following a 14-day recovery period after the second surgery (with administration of the same pain medication and antibiotics as after the first surgery), the animal was trained to stand and walk on the passive prosthesis as described previously (Farrell, Prilutsky, Kistenberg, et al., 2014; Jarrell et al., 2018). After the animal started using the passive prosthesis for walking confidently (i.e., loading the prosthetic limb in each walking cycle for body support; Jarrell et al., 2018; rather than keeping the limb in a flexed position), we substituted the passive prosthesis for a powered sensing prosthesis (Figure 1b2) while the cat was trained and recorded in the lab (described later). After completion of locomotor experiments 29 months after initial implantation surgery, the residual shank with the implant was surgically removed for histological analysis and the animal was adopted by an approved owner. The shank with implant was fixed in 10% neutral buffered formalin and sent to Alizée Pathology, LLC for histological analysis. At Alizée Pathology, the shank with the implant was trimmed and embedded in methyl methacrylate, sectioned, ground, and micropolished to an approximate 100 μm thickness using Exakt grinding system (EXAKT Technologies, Inc.) and stained with hematoxylin and eosin (H&E). An example of histological images is shown in Figure 1c2.
Experimental Protocol for Paw Pad Anesthesia and Stimulation of Distal Tibial Nerve
We recorded 3D kinematics by a six-camera motion capture system (Vicon-612) at a sampling rate of 250 Hz and EMG activity of selected hindlimb muscles (sampling rate 3,000 Hz) in four adult female cats during split-belt treadmill walking with four speed ratios between the right and left belts (described earlier). Each speed ratio condition was recorded three times: (a) without paw pad anesthesia or nerve stimulation (control condition), (b) with paw anesthesia, and (c) with paw anesthesia and nerve stimulation. Each of the conditions was tested on separate days, and the order of conditions was randomized across animals. Prior to each recording session, 26 light reflective markers were attached on shaved skin over the major limb joints and the lateral aspects of the head using a double-sided tape (Farrell, Bulgakova, et al., 2014; Park et al., 2019). Prior to each session involving distal nerve stimulation, we tested a twitch response of the toes (motor threshold) to stimulation of the distal tibial nerve in the awake animal. In recording sessions with paw pad anesthesia, we sedated the cat with dexmedetomidine (40–60 μg/kg, i.m.) and injected each digital pad and the three parts of the metacarpal (metatarsal) pad of the right forelimb and right hindlimb with 1 ml of 1% lidocaine solution. Then the animal was awakened by administration of atipamezole (40–60 μg/kg, i.m.), and the recording session started within ∼5 min. The anesthesia lasted for at least 30 min as verified by pinpricks or attaching a piece of adhesive tape to the paw to evoke limb withdrawal or paw shake response. In each experimental condition, the cat first walked at a tied-belt condition (left and right belt speeds were 0.4 m/s) for 15 s, then we changed the speed of one belt to the desired speed ratio within 1 s and maintained it for 60 s, and after that, we changed the speed to the initial tied-belt condition within 1 s and maintained this speed for another 15 s. During several minutes of rest between experimental conditions, the animal received dry food as a reward.
In conditions with stimulation of the distal tibial nerve, stimulation was triggered by the contact of the right hindpaw with the treadmill belt. To detect a hindpaw ground contact, a pressure monitoring system was installed below the right treadmill belt. The system consisted of a two-channel 64-bit pressure sensor array (SEN-09375, SparkFun Electronics) with 12-bit analog-to-digital converters at 100 Hz sampling (Park et al., 2016). The pressure sensor triggered the stimulation if the applied force was larger than 5 N. The distal tibial nerve was stimulated with the following parameters: trains of 200-μs biphasic rectangular pulses at 100 Hz (Figure 1a1) for 500 ms duration, stimulation strength 1.2 × T.
Experimental Protocol for Recordings of Prosthetic Walking With Stimulation of Residual Distal Tibial Nerve
Overground level walking was recorded in one adult male cat before surgery and after the attachment of a unilateral, powered, sensing transtibial prosthesis; prosthetic walking was recorded over the period of 20–29 months following the initial implantation. We did not test the cat with the prosthesis on the treadmill for safety reasons to avoid the prosthetic foot getting accidently under the moving belts and potentially causing injury. The powered sensing transtibial prosthesis (Figure 1b2) was developed and benchtop characterized in Park et al. (2018). Several modes of prosthetic operations were used in this study. In Mode 0, the powered prosthesis performs ankle extension if there is contact between the foot and the ground. If there is no contact with the ground, the foot flexes. The foot contact was detected by a pressure transducer ThinPot (Spectra Symbol) on the bottom surface of the prosthetic foot. In Mode 2, EMG activity in the residual SO muscle was also required for ankle extension. The custom LabView program (LabView 2018, National Instruments) was used to control the prosthesis electronics wirelessly and to set the EMG band-pass (30–1,000 Hz) and low-pass (cutoff frequency 25 Hz) filtering parameters and a threshold above the background EMG noise level. EMG signals were processed by the prosthesis electronic circuits and then sent wirelessly to a PC and saved (Park et al., 2018). Figure 1c1 shows an example of EMG activity in the residual SO and the foot pressure signal during level walking with prosthesis operating in Mode 2. In addition, the program set parameters of the distal tibial nerve stimulation (e.g., intensity, frequency) that were triggered by the onset of foot–ground contact and lasted for 500 ms. Additional details about the prosthesis can be found in Park et al. (2018). Each mode of prosthesis operation was investigated with and without stimulation of the residual distal tibial nerve.
Prior to each experimental session, the passive J-shaped prosthesis was substituted with the powered prosthesis, and the female connector with the electrode leads (Figure 1b1iii) was plugged into the male connector (Figure 1b2v) on the powered prosthesis. Major joints and the head were marked with 28 reflective markers, and the animal was placed in the walkway. We recorded full-body kinematics using the Vicon-612 motion capture system at 120 Hz and ground reaction forces at 360 Hz. Detailed descriptions of overground locomotor recordings can be found in Farrell, Bulgakova, et al. (2014) and Gregor et al. (2018).
Data Analysis
We selected for kinematic and EMG analyses episodes of steady-state walking with adequate quality of recordings; the number of analyzed cycles for each cat and condition are shown in Tables 1 and 2. Marker coordinates recorded during treadmill and overground walking were low-pass filtered (fourth-order Butterworth zero-lag filter, cutoff frequency 6 Hz), individual walking cycles identified based on forelimb and hindlimb orientations (Pantall et al., 2012), and basic kinematic parameters of walking computed. These parameters included COM lateral bias (mean COM lateral position with respect to the left and right hindpaw positions in stance; Park et al., 2019); step length (the distance between paws of the leading and contralateral trailing limb at stance onset of the leading limb); cycle, stance phase, and swing phase durations; duty factor; and pacing. The duty factor was defined as the ratio of stance phase duration to cycle duration. Pacing was defined as the phase difference between the ipsilateral hindlimb and forelimb footfalls (Hildebrand, 1965). We did not compute the COM lateral bias for overground walking because of low recording quality of head markers in the majority of trials.
Number of Analyzed Cycles for Each Condition in Experiments in the Cat With Unilateral Transtibial Prosthesis
| Distal tibial nerve stimulation | |||
|---|---|---|---|
| Limb | Prosthetic condition | Off | On |
| RH | RH intact | 21 | — |
| RH prosthetic Mode 0 | 28 | 102 | |
| RH prosthetic Mode 2 | 38 | 61 | |
| Left hindlimb | RH intact | 20 | — |
| RH prosthetic Mode 0 | 33 | 125 | |
| RH prosthetic Mode 2 | 53 | 83 | |
| Right forelimb | RH intact | 21 | — |
| RH prosthetic Mode 0 | 28 | 113 | |
| RH prosthetic Mode 2 | 46 | 66 | |
| Left forelimb | RH intact | 24 | — |
| RH prosthetic Mode 0 | 34 | 125 | |
| RH prosthetic Mode 2 | 52 | 83 | |
| Total | 398 | 761 | |
Note. RH = right hindlimb.
The EMG activity recorded during treadmill walking from SO, medial gastrocnemius (MG), and iliopsoas (IL) muscles was checked for signal artifacts caused by nerve stimulation during the stance phase (see examples in Figure 3c). Cycles with large magnitude artifacts overlapping with the extensor bursts of SO and MG (exceeding ∼5% of the mean EMG burst magnitude) were removed from the analysis. The stimulation artifacts did not interfere with the flexor bursts of IP and were ignored. Recorded EMG signals were band-pass filtered (30–1,000 Hz, Butterworth filter) and full-wave rectified. The rectified raw EMG was used to identify EMG burst onset and offset using a two-standard deviation threshold above the baseline. The identified bursts were time integrated to determine the mean burst activity. The mean EMG activity was normalized to the maximum mean activity found for a given muscle across all walking trials of a given cat.
—Examples of EMG recordings of right (ipsilateral) SO, MG, and IL at three speed ratios of split-belt treadmill walking in control condition (a), in condition of ipsilateral anesthesia of fore- and hindpaw pads (b), and in condition of anesthesia and electrical stimulation of the distal tibial nerve during the stance phase (c). Cat WE. (d–f) Mean (± 95% confidence interval) magnitude of normalized EMG activity averaged across all cats and cycles (Table 1). (d) Magnitude of normalized EMG activity of SO. (e) Magnitude of normalized EMG activity of MG. (f) Magnitude of normalized EMG activity of IL. EMG = electromyography; LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb; MG = medial gastrocnemius; IL = iliopsoas; SO = soleus.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Examples of EMG recordings of right (ipsilateral) SO, MG, and IL at three speed ratios of split-belt treadmill walking in control condition (a), in condition of ipsilateral anesthesia of fore- and hindpaw pads (b), and in condition of anesthesia and electrical stimulation of the distal tibial nerve during the stance phase (c). Cat WE. (d–f) Mean (± 95% confidence interval) magnitude of normalized EMG activity averaged across all cats and cycles (Table 1). (d) Magnitude of normalized EMG activity of SO. (e) Magnitude of normalized EMG activity of MG. (f) Magnitude of normalized EMG activity of IL. EMG = electromyography; LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb; MG = medial gastrocnemius; IL = iliopsoas; SO = soleus.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Examples of EMG recordings of right (ipsilateral) SO, MG, and IL at three speed ratios of split-belt treadmill walking in control condition (a), in condition of ipsilateral anesthesia of fore- and hindpaw pads (b), and in condition of anesthesia and electrical stimulation of the distal tibial nerve during the stance phase (c). Cat WE. (d–f) Mean (± 95% confidence interval) magnitude of normalized EMG activity averaged across all cats and cycles (Table 1). (d) Magnitude of normalized EMG activity of SO. (e) Magnitude of normalized EMG activity of MG. (f) Magnitude of normalized EMG activity of IL. EMG = electromyography; LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb; MG = medial gastrocnemius; IL = iliopsoas; SO = soleus.
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
Statistics
We used a linear mixed effect model analysis based on restricted maximum likelihood method (MIXED, SPSS 19, IBM SPSS) to determine the significance of effects of fixed, within-animal, independent factors on the dependent kinematic and EMG variables obtained during treadmill walking. We selected this analysis because it allowed unequal or missing data points in repeated measurements within subjects, permitted inclusion of fixed and random effect factors, and was typically robust to violations of the assumptions of constant variance, independence, and distribution of residual and random effects (Schielzeth et al., 2020; Smith, 2012; West et al., 2015). Nevertheless, we tested the assumption of normal distribution for each dependent variable at different combinations of fixed factors using the Kolmogorov–Smirnov test (significance level .05) and found that this assumption could not be rejected for the majority of dependent variables. The fixed factors in our analysis were anesthesia (off, on), nerve stimulation (off, on), and speed ratio (0.5, 1.0, 1.5, and 2.0). The random factors were animals and cycles (see Table 1). The dependent variables included the COM bias; the cycle, stance phase, and swing phase durations of the right hindlimbs; right fore- and hindlimb step lengths; duty factor; and pacing of the right limbs. For each dependent variable, effects of the independent fixed factors were determined. If the effects were found to be significant, a post hoc paired comparison was performed using the Games-Howell test with Bonferroni adjustments. Significance level was set at .05.
As, for safety reasons, we tested prosthetic walking overground rather than on the treadmill and because prosthetic walking is inherently asymmetric (Farrell, Prilutsky, Kistenberg, et al., 2014; Jarrell et al., 2018), we compared dependent kinematic variables of the ipsilateral and contralateral limbs between intact walking (control) and prosthetic walking with and without the residual distal tibial nerve stimulation for each prosthetic operation mode. We used a similar linear mixed effect model analysis to test the significance of effects of independent fixed factors on kinematic variables of prosthetic walking. Independent fixed factors were prosthetic condition (intact, prosthetic Mode 0, prosthetic Mode 2), limb (right hindlimb, right forelimb, left hindlimb, and left forelimb), and nerve stimulation (off, on). Walking cycles were considered a random factor. Dependent variables were the cycle, stance phase, and swing phase durations, step length, duty factor, pacing, and the mean speed of each limb (computed as the ratio of the limb stride length and cycle time). Pairwise comparisons were performed using the Games-Howell test with Bonferroni adjustments. Significance level was set at .05.
Results
Effects of Unilateral Paw Pad Anesthesia and Distal Tibial Nerve Stimulation
The COM lateral bias was significantly affected by speed ratio, anesthesia of the right paws, and nerve stimulation (p < .05; Figure 2a). As described previously (Latash et al., 2020; Park et al., 2019), the COM shifted toward the slow belt without anesthesia, and anesthesia caused the COM shift to the anesthetized side regardless of the speed ratio, except the ratio of 1.0 at which no COM shift was noticed (p > .05). Stimulation of the right distal tibial nerve during the stance phase caused a significant reduction of the anesthesia effect (COM shifted toward the control condition, p < .05) but only for speed ratios of 1.5 and 2.0, that is, when the anesthetized paws were moving on a faster right belt. Stimulation effects were not significant for speed ratios of 0.5 and 1.0 (Figure 2a).
Speed ratio, anesthesia, and nerve stimulation all had significant effects on the duration of the cycle and swing and stance phases of the right fore- and hindlimb (p < .05; Figure 2b and 2c). Specifically, the cycle duration in the control conditions was longest (near 0.9 s) for all limbs at speed ratio 1.0 and decreased when one of the belts moved at a faster speed. Anesthesia of right paw pads decreased the cycle duration of the ipsilateral fore- and hindlimb at all speed ratios but ratio 1.0 for the right forelimb and ratio 0.5 for the right hindlimb (p < .05; Figure 2b and 2c). The stance phase duration of the ipsilateral right limbs decreased (p < .05) only at speed ratio of 1.0 for the forelimb and 1.0 and 1.5 for the hindlimb. Paw anesthesia decreased the swing duration of the right fore- and hindlimbs at speed ratios of 1.5 and 2.0 (p < .05). Stimulation of the right distal tibial nerve during stance while the right hind- and forepaw pads were anesthetized reversed the effects of anesthesia on durations of the walking cycle and its phases at many speed ratios. Specifically, stimulation increased the duration of the cycle, stance, and swing phases compared with the anesthetized condition for the right fore- and hindlimbs typically for speed ratios of 1.0, 1.5, and 2.0 (Figure 2b and 2c). One noticeable exception was the stance duration of the right hindlimb at ratios of 1.5 and 2.0—stimulation did not change the stance duration. In most cases, stimulation increased the cycle duration of the fore- and hindlimbs beyond values of control conditions (p < .05) but did not affect the stance duration (with one exception, it increased beyond control values for the forelimb at speed ratio of 1.5). Stimulation increased the swing duration beyond control values only in the right hindlimb and only at speed ratios 1.5 and 2.0.
Duty factor of the right forelimb and right hindlimb in the control conditions decreased as a function of the speed ratio from 0.768 and 0.746 at ratio 0.5 to 0.637 and 0.602 at ratio 2.0 (p < .05; Figure 2d and 2e). The duty factor was also significantly lower for the hindlimbs than forelimbs (p < .05). Unilateral paw anesthesia increased the duty factor of the right fore- and hindlimbs at speed ratios 1.5 and 2.0 (p < .05; Figure 2d and 2e). Stimulation of the right distal tibial nerve during stance reversed the anesthesia effects on duty factor and reduced duty factor beyond the control values (p < .05).
Pacing of right limbs was the highest (the phase difference between hindlimb and forelimb footfalls the lowest, 0.16) at speed ratio 0.5; pacing was lower and essentially constant (0.19–0.20) at the other speed ratios (Figure 2f). Anesthesia of right paws increased pacing of the ipsilateral limbs at speed ratios 1.0, 1.5, and 2.0 (p < .05) but not at ratio 0.5 (p > .05; Figure 2f). Stimulation of the right distal tibial nerve significantly increased the pacing of the right limbs at speed ratio 0.5 and decreased it at speed ratio 1.5 (p < .05; Figure 2f).
In control conditions, step length of the right fore- and hindlimb was the longest at speed ratio 0.5 (22.5 and 23.5 cm, respectively; the difference was significant, p < .05). Step length drastically shortened (p < .05) at speed ratio 1.0–16.5 cm and 17.4 cm, respectively—and remained essentially constant at speed ratios 1.5 and 2.0; forelimb step length was greater than hindlimb length at speed ratio 2.0 (p < .05; Figure 2g and 2h). Anesthesia of right fore- and hindpaw did not change the right forelimb and hindlimb step length (p > .05; Figure 2g and 2h). Stimulation of the right distal tibial nerve increased step length of the right fore- and hindlimb at speed ratios 1.0, 1.5, and 2.0 compared with intact and anesthetized conditions.
Figure 3a–c demonstrates examples of EMG recordings of right ankle extensors SO and MG and a hip flexor IP in control, anesthesia, and anesthesia-with-stimulation conditions at speed ratios 0.5, 1.0, and 2.0. The duration and magnitude of extensor bursts were typically greater at speed ratio 0.5, at which the COM was shifted to the right (Figure 2a), and thus, load on the right hindlimb was higher. With increasing speed ratio, the MG EMG bursts decreased in magnitude and duration dramatically, whereas flexor EMG bursts increased in size. In the control condition, the mean EMG burst magnitude of SO and MG averaged across all cats and cycles decreased when speed ratio changed from 0.5 to 1.0 (p < .05); the EMG magnitude was relatively constant with a further increase of speed ratio (Figure 3d and 3e). The mean burst magnitude of IP also decreased with increasing speed ratio from 0.5 to 1.0 but increased with changing speed ratio from 1.5 to 2.0 in control conditions (p < .05; Figure 3f). Anesthesia of the right fore- and hindpaw increased the EMG magnitude in SO at speed ratio 1.0, MG at ratios 1.5 and 2.0, and IP at ratios 1.0–2.0 (p < .05; Figure 3d-f). Stimulation of the right distal tibial nerve reduced EMG magnitude with respect to the anesthesia condition in SO (speed ratio 1.0) and IP (ratios 1.0 and 1.5); it increased EMG magnitude in SO at speed ratio 1.5.
Effects of Distal Tibial Nerve Stimulation on Prosthetic Walking
Independent fixed factors of prosthetic condition (intact, prosthetic Mode 0, prosthetic Mode 2), limb, and stimulation all had significant effects on the analyzed kinematic variables (p < .05). The mean speed of each limb during intact walking cycle was between 0.54 m/s for the left hindlimb and 0.56 m/s for the right forelimb (Figure 4a). The speed of each limb during prosthetic walking in Modes 1 and 2 without stimulation was significantly lower (p < .05). Stimulation of the right residual distal tibial nerve significantly increased speed for each limb from the minimal speed of 0.47 m/s in the nonstimulation condition to the maximum speed of 0.57 m/s during stimulation (p < .05; Figure 4a).
—Mean (± 95% confidence interval) of kinematic characteristics of overground level walking before surgery (intact) and with the right transtibial prosthesis operating in Mode 0 and Mode 2 (see text). The number of analyzed cycles for each limb and experimental condition is presented in Table 2. (a) Mean movement speed of each limb computed as the ratio of the stride length over the cycle time. (b) Mean step length. (c) Mean cycle time. (d) Mean stance time. (e) Mean swing time. (f) Mean duty factor. (g) Mean pacing. LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb (prosthetic limb).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Mean (± 95% confidence interval) of kinematic characteristics of overground level walking before surgery (intact) and with the right transtibial prosthesis operating in Mode 0 and Mode 2 (see text). The number of analyzed cycles for each limb and experimental condition is presented in Table 2. (a) Mean movement speed of each limb computed as the ratio of the stride length over the cycle time. (b) Mean step length. (c) Mean cycle time. (d) Mean stance time. (e) Mean swing time. (f) Mean duty factor. (g) Mean pacing. LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb (prosthetic limb).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
—Mean (± 95% confidence interval) of kinematic characteristics of overground level walking before surgery (intact) and with the right transtibial prosthesis operating in Mode 0 and Mode 2 (see text). The number of analyzed cycles for each limb and experimental condition is presented in Table 2. (a) Mean movement speed of each limb computed as the ratio of the stride length over the cycle time. (b) Mean step length. (c) Mean cycle time. (d) Mean stance time. (e) Mean swing time. (f) Mean duty factor. (g) Mean pacing. LF = left forelimb; LH = left hindlimb; RF = right forelimb; RH = right hindlimb (prosthetic limb).
Citation: Motor Control 27, 1; 10.1123/mc.2022-0096
The step length determined for each limb during intact walking was between 0.258 m for the right hindlimb and 0.301 m for the left hindlimb. The step length of the prosthetic right hindlimb significantly increased in Modes 0 and 2 without stimulation compared with the intact right hindlimb (p < .05). The step length of the sound limbs (right forelimb, left fore- and hindlimbs) was typically lower during prosthetic walking without stimulation than in the intact walking (p < .05). Stimulation of the residual nerve during prosthetic walking increased the step length only in the contralateral left hindlimb in both prosthetic modes and in left forelimb of Mode 2 (p < .05; Figure 4b).
The cycle duration of the sound limbs during prosthetic walking in Mode 0 without stimulation increased significantly compared with intact walking (p < .05); however, there were no changes in the cycle duration of all limbs in Mode 2 compared with intact walking (p > .05; Figure 4c). Stimulation of the residual nerve during stance of prosthetic walking in Modes 0 and 2 decreased the cycle duration for each limb (p < .05; Figure 4c); the longest cycle time without stimulation was 1.051 s for the left hindlimb, and the shortest cycle time with stimulation was 0.880 s for the right hindlimb.
The stance time and duty factor of the contralateral limbs during prosthetic walking in Modes 0 and 2 without stimulation were greater than those of intact walking, whereas the stance time and duty factor of the prosthetic limb were significantly smaller compared with intact walking (p < .05; Figure 4d and 4f). The decrease in the duty factor of the prosthetic limb was especially noticeable, that is, 0.58 and 0.57 in Modes 0 and 2 versus 0.64 in intact walking. Nerve stimulation decreased the stance phase duration in all limbs and prosthetic modes (p < .05; Figure 4d), whereas it only slightly, although significantly, decreased the duty factor of the contralateral fore- and hindlimbs in Mode 0 and the contralateral forelimb in Mode 2.
The swing phase duration of the prosthetic right hindlimb was much longer during prosthetic walking without stimulation as opposed to the swing duration of the same hindlimb in intact walking (0.443 and 0.380 s in prosthetic Modes 0 and 2 vs. 0.352 s in intact walking; p < .05; Figure 4e). The swing duration of the contralateral forelimb and hindlimb significantly decreased from 0.295 and 0.349 s in intact walking to 0.272 and 0.265 s in prosthetic Mode 0 and 0.271 and 0.260 s in Mode 2, respectively. Stimulation of the residual nerve during stance of prosthetic walking consistently reduced the swing duration in all limbs and prosthetic modes (p < .05) with greater changes observed in the ipsilateral limbs, especially the prosthetic limb (Figure 4e).
Pacing during intact walking (defined as the phase difference between the hindlimb and forelimb footfalls for each side) was 0.17 and 0.20 for the contralateral (left) and ipsilateral (right) sides, respectively. Walking with the prosthesis without stimulation in Modes 0 and 2 increased pacing (decreased the phase difference) for the ipsilateral side to 0.14 and 0.15, respectively (p < .05). The contralateral side decreased pacing (increased the phase difference) for Mode 2 (p < .05) but not for Mode 0 (Figure 4g). Stimulation only affected pacing of the contralateral side in both prosthetic modes (pacing decreased, p < .05) but had no effect on the ipsilateral side.
Discussion
The goal of this study was to test whether electrical stimulation of the distal tibial nerve during the stance phase of split-belt walking could reduce or reverse effects of ipsilateral anesthesia of the hind- and forepaws. As the distal tibial nerve innervates skin on the plantar surface of the paw and foot, in addition to mechanoreceptors in the foot muscles, we expected that activation of cutaneous pressure afferents by electrical stimulation during stance might lead to tactile sensations and evoke appropriate spinal postural reflexes to improve locomotor function. In addition, we investigated the effects of stimulation of the residual distal tibial nerve during stance phase of walking with a powered sensing transtibial prosthesis. Amputation of the foot and distal shank completely removed somatosensory feedback from these parts of the hindlimb. Activation of cutaneous and muscle afferents during the stance phase by electrical stimulation could potentially recover some sensory motion-dependent feedback.
Effects of Speed Ratio and Unilateral Paw Pad Anesthesia on Extensor and Flexor EMG Activity
We have already discussed the effects of unilateral anesthesia of paw pads on walking kinematics during split-belt treadmill walking (Latash et al., 2020; Park et al., 2019). Here, we presented new results on how paw pad anesthesia and speed ratio influence EMG activity of ipsilateral SO, MG, and IP. The higher EMG activity of right ipsilateral extensors SO and MG during walking with speed ratio 0.5 when the ipsilateral hindlimb moved slower than the contralateral one (0.4 m/s vs. 0.8 m/s) can be explained by the COM shift toward the slower ipsilateral limb and loading it more (Figure 2a). A similar increase in EMG activity of extensors on the slow belt was also reported in Frigon et al. (2015). The smaller magnitude and duration of extensor bursts at speed ratios 1.0, 1.5, and 2.0 is likely caused by unloading of the limb due to the COM shift in the opposite direction (Figure 2a) and the corresponding decrease in the stance duration (Figure 2c) and duty factor (Figure 2e). The decrease in pacing at higher speed ratios (Figure 2f) could contribute to increased loading of the ipsilateral hindlimb as the ipsilateral forelimb is on the ground for a shorter fraction of the cycle and contributes less to weight support. As demonstrated previously (Latash et al., 2020; Park et al., 2019), unilateral paw pad anesthesia offsets effects of increasing speed ratio on the COM shift (see also Figure 2a), that is, ipsilateral anesthesia reduces ipsilateral unloading of the hind- and forelimbs at larger speed ratios. This, then, can explain the increase in MG EMG activity during anesthesia at speed ratios 1.5 and 2.0 (Figure 3e). SO EMG activity increased after application of anesthesia only at speed ratio 1.0 when the ipsilateral belt moved at the slowest speed of 0.4 m/s. The differential change in EMG activity of SO and MG in response to anesthesia at slow (0.4 m/s) and faster speeds (0.6 and 0.8 m/s), respectively, could be related to the division of labor between the slow-twitch SO and fast-twitch MG reported previously (Prilutsky et al., 1996; Walmsley et al., 1978). The increase in IP EMG activity with ipsilateral anesthesia at speed ratios 1.0, 1.5, and 2.0 appears consistent with greater demands for faster hindlimb swing at these speed ratios. Ipsilateral anesthesia causes shortening of the hindlimb swing phase at speed ratios 1.5 and 2.0 (Figure 2c) without changing the hindlimb step length (Figure 2h).
Distal Tibial Nerve Stimulation Reduces or Reverses Some Effects of Paw Pad Anesthesia
Electrical stimulation of the distal tibial nerve at strength 1.2 × T presumably activated low threshold Groups I and II muscle as well as cutaneous afferents from hindpaw pads based on estimated ranges of conduction velocities in the activated afferents, that is, 25–90 m/s (Boyd & Kalu, 1979; Burgess et al., 1968; Hunt & McIntyre, 1960). The nerve stimulation was triggered by hindpaw contact with the treadmill and lasted 500 ms. This stimulation period was shorter than the stance phase at speed ratios 0.5 and 1.0, it occupied almost the entire stance duration for speed ratio 1.5, and it was typically longer than the stance phase at speed ratio 2.0 (Figure 2c). The difference in stimulation phase could cause differential responses, for example, enhancing extensor or flexor actions (Duysens & Pearson, 1976). According to our results, the distal tibial nerve stimulation at 1.2 × T during the stance phase while the ipsilateral paw pads were anesthetized resulted in a partial or complete cancelation of anesthesia effects for many kinematic characteristics. First, the COM shift toward the anesthetized side at speed ratios 1.5 and 2.0 was completely eliminated by stimulation (Figure 2a). This result could be interpreted as if the animal sensed contact with the ground during stance due to electrical activation of paw cutaneous afferents (as well as foot muscle load-sensitive afferents) and, thus, no longer needed to press the supported limbs harder to improve tactile perception degraded by anesthesia. Second, the swing duration and duty factor of both the ipsilateral fore- and hindlimb returned to the control values or slightly overshot them during stimulation (Figure 2b–e). In addition, we also noticed effects of nerve stimulation that did not seem related to anesthesia. For example, nerve stimulation in the anesthetized condition dramatically increased pacing at speed ratio 0.5, whereas anesthesia did not affect pacing at this speed ratio; on the other hand, increased pacing by anesthesia at speed ratio 1.5 was significantly reduced by nerve stimulation. Also, although paw pad anesthesia had no effect on ipsilateral forelimb and hindlimb step lengths, nerve stimulation substantially increased them, especially for speed ratios of 1.5 and 2.0. This effect of stimulation was seen at speed ratios corresponding to relatively short stance phase (Figure 2c, speed ratios 1.5 and 2.0), that is, stimulation was continuing at the end of stance or early swing and, thus, was expected to elicit enhanced hindlimb flexion action (Duysens & Pearson, 1976; Guertin et al., 1995; Ollivier-Lanvin et al., 2011). This expectation was only partially supported. Although stimulation increased the relative duration of the swing phase (reduced the duty factor, Figure 2e), the absolute swing duration decreased (Figure 4c), and EMG activity of a hip flexor IP also decreased during nerve stimulation at speed ratios 1.0 and 1.5 (Figure 3f). It is possible that the substantial increase in the step length of the ipsilateral fore- and hindlimbs by nerve stimulation at high speed ratios is caused by complex interactions between extension and flexion actions evoked by the distal nerve stimulation in the early-mid and late stance and in the early swing (Duysens & Pearson, 1976; Guertin et al., 1995; Ollivier-Lanvin et al., 2011).
Effects of Residual Distal Tibial Nerve Stimulation on Kinematics of Prosthetic Walking
The cat that participated in this portion of the study had an implant with the attached prosthesis for 29 months. Although the animal observations and histological analysis (partially shown in Figure 1c2) indicated excellent pylon integration with the bone and skin (this is the first report of successful integration of a transcutaneous implant over such an extended period; Jeyapalina et al., 2017), the structure and function of the residual distal tibial nerve as well as the quality of implanted electrodes could have degraded. During the recording period, we could not reliably record responses to stimulation of the residual distal tibial nerve in the sciatic nerve. On the other hand, axotomized nerves in the cat can still demonstrate rhythmic activity during locomotion (Gordon et al., 1980), and electrical stimulation of cutaneous nerves in human amputees evokes natural tactile perceptions several years after amputation (Charkhkar et al., 2018; Christie et al., 2020; George et al., 2020; Graczyk et al., 2022).
As demonstrated in this study (Figure 4), stimulation of the residual distal tibial nerve during stance of prosthetic walking had robust and consistent effects on many kinematic variables. We hypothesized that effects of this stimulation would be similar to those observed in intact animals with anesthetized paws. Comparing stimulation results between overground prosthetic walking at self-selected speeds between ∼0.45 and 0.55 m/s (Figure 4a) and treadmill walking with anesthetized paws at speeds 0.4 m/s for both sides (speed ratio 1.0, the only symmetric condition), we conclude that the hypothesis was not supported. In prosthetic walking, stimulation decreased the duration of the cycle, stance phase, and swing phase of all limbs (Figure 4c–4e) and did not affect the ipsilateral hindlimb and forelimb step length (Figure 4b); these results were opposite to those observed during treadmill walking (Figure 2b, 2c, 2g, and 2h, speed ratio 1.0). Nerve stimulation did not cause significant changes in the duty cycle and pacing of the ipsilateral side, which was consistent with the results of treadmill walking at speed ratio 1.0 (Figures 2d–f and 4f,g).
There are several possible explanations for the different effects of nerve stimulation in the two cases. First, removal of somatosensory feedback by transtibial amputation is much more extensive than local paw anesthesia. In addition, the lack of ankle muscles in the prosthetic hindlimb reduces limb support and propulsion even with the powered prosthesis, whose power appeared to be limited (Pitkin et al., 2021). As a result, the contralateral limbs, as well as the ipsilateral forelimb, increased their contribution to weight support and propulsion as evident from the longer stance time and higher duty factor of the contralateral limbs (Figure 4d and 4f) and the greater pacing on the ipsilateral side (Figure 4g) as compared with the intact walking before surgery; see also (Jarrell et al., 2018). Furthermore, the distal tibial nerve transection and its resuture to the skin on the proximal shank likely led to imperfect reinnervation of skin receptors, and reinnervation of muscle proprioceptors was impossible (de Ruiter et al., 2014; Gordon, 2016); that affected responses to nerve stimulation. Even if the natural spinal reflex responses to activation of the stimulated nerve were compromised, it would also be possible that the animal learned to associate the perceived nerve stimulation with the stance phase and consciously adjusted locomotor strategy based on this information.
Taken together, the results of this study suggest that low-intensity stimulation of cutaneous and muscle afferents innervating the plantar foot may provide functionally meaningful motion-dependent sensory feedback and could potentially improve locomotor function. Responses to distal tibial nerve stimulation depend on the limb and stimulated nerve conditions. Future studies should address the long-term reliability of implanted electrodes for nerve stimulation and muscle activity recordings and how the nervous system modulates responses to peripheral nerve stimulation in different limb conditions.
Acknowledgments
Park and Klishko contributed equally to this manuscript. The authors wish to thank the veterinary staff of animal facilities at Georgia Institute of Technology and T3 Laboratories for excellent animal care. This study was supported by Grants HD057492, HD032571, HD090768, and R01 NS110550 from the U.S. National Institutes of Health and by Grant W81XWH-16-1-0791 from the U.S. Department of Defense.
References
af Klint, R., Mazzaro, N., Nielsen, J.B., Sinkjaer, T., & Grey, M.J. (2010). Load rather than length sensitive feedback contributes to soleus muscle activity during human treadmill walking. Journal of Neurophysiology, 103(5), 2747–2756. https://doi.org/10.1152/jn.00547.2009
Akay, T., Haehn, S., Schmitz, J., & Buschges, A. (2004). Signals from load sensors underlie interjoint coordination during stepping movements of the stick insect leg. Journal of Neurophysiology, 92(1), 42–51. https://doi.org/10.1152/jn.01271.2003
Akay, T., Tourtellotte, W.G., Arber, S., & Jessell, T.M. (2014). Degradation of mouse locomotor pattern in the absence of proprioceptive sensory feedback. Proceedings of the National Academy of Sciences of the United States of America, 111(47), 16877–16882. https://doi.org/10.1073/pnas.1419045111
Bensmaia, S.J., Tyler, D.J., & Micera, S. (2020). Restoration of sensory information via bionic hands. Nature Biomedical Engineering. https://doi.org/10.1038/s41551-020-00630-8
Bolton, D.A., & Misiaszek, J.E. (2009). Contribution of hindpaw cutaneous inputs to the control of lateral stability during walking in the cat [Research Support, Non-U.S. Gov’t]. Journal of Neurophysiology, 102(3), 1711–1724. https://doi.org/10.1152/jn.00445.2009
Bouyer, L.J., & Rossignol, S. (2003). Contribution of cutaneous inputs from the hindpaw to the control of locomotion. I. Intact cats. Journal of Neurophysiology, 90(6), 3625–3639. https://doi.org/10.1152/jn.00496.2003
Boyd, I.A., & Kalu, K.U. (1979). Scaling factor relating conduction velocity and diameter for myelinated afferent nerve fibres in the cat hind limb. Journal of Physiology, 289(1), 277–297. https://doi.org/10.1113/jphysiol.1979.sp012737
Burgess, P.R., Petit, D., & Warren, R.M. (1968). Receptor types in cat hairy skin supplied by myelinated fibers. Journal of Neurophysiology, 31(6), 833–848. https://doi.org/10.1152/jn.1968.31.6.833
Buurke, T.J.W., Lamoth, C.J.C., van der Woude, L.H.V., Hof, A.L., & den Otter, R. (2019). Bilateral temporal control determines mediolateral margins of stability in symmetric and asymmetric human walking. Scientific Reports, 9(1), 12494. https://doi.org/10.1038/s41598-019-49033-z
Buurke, T.J.W., Lamoth, C.J.C., Vervoort, D., van der Woude, L.H.V., & den Otter, R. (2018). Adaptive control of dynamic balance in human gait on a split-belt treadmill. Journal of Experimental Biology, 221(13), Article jeb174896. https://doi.org/10.1242/jeb.174896
Charkhkar, H., Shell, C.E., Marasco, P.D., Pinault, G.J., Tyler, D.J., & Triolo, R.J. (2018). High-density peripheral nerve cuffs restore natural sensation to individuals with lower-limb amputations. Journal of Neural Engineering, 15(5), 056002. https://doi.org/10.1088/1741-2552/aac964
Chesler, A.T., Szczot, M., Bharucha-Goebel, D., Ceko, M., Donkervoort, S., Laubacher, C., Hayes, L.H., Alter, K., Zampieri, C., Stanley, C., Innes, A.M., Mah, J.K., Grosmann, C.M., Bradley, N., Nguyen, D., Foley, A.R., Le Pichon, C.E., & Bönnemann, C.G. (2016). The role of PIEZO2 in human mechanosensation. New England Journal of Medicine, 375(14), 1355–1364. https://doi.org/10.1056/NEJMoa1602812
Christie, B.P., Charkhkar, H., Shell, C.E., Burant, C.J., Tyler, D.J., & Triolo, R.J. (2020). Ambulatory searching task reveals importance of somatosensation for lower-limb amputees. Scientific Reports, 10(1), 10216. https://doi.org/10.1038/s41598-020-67032-3
Chung, B., Bacque-Cazenave, J., Cofer, D.W., Cattaert, D., & Edwards, D.H. (2015). The effect of sensory feedback on crayfish posture and locomotion: I. Experimental analysis of closing the loop. Journal of Neurophysiology, 113(6), 1763–1771. https://doi.org/10.1152/jn.00248.2014
Cole, J., & Paillar, J. (1998). Living without touch and peripheral information about body position and movement: Studies with deafferented subjects. In J. Bermúdez (Ed.), The body and the self. MIT Press.
Crowell, A., & Gwathmey, K.G. (2017). Sensory neuronopathies. Current Neurology and Neuroscience Reports, 17(10), Article 79. https://doi.org/10.1007/s11910-017-0784-4
de Ruiter, G.C., Spinner, R.J., Verhaagen, J., & Malessy, M.J. (2014). Misdirection and guidance of regenerating axons after experimental nerve injury and repair. Journal of Neurosurgery, 120(2), 493–501. https://doi.org/10.3171/2013.8.JNS122300
Dunlap, P., Perera, S., VanSwearingen, J.M., Wert, D., & Brach, J.S. (2012). Transitioning to a narrow path: The impact of fear of falling in older adults [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t]. Gait & Posture, 35(1), 92–95. https://doi.org/10.1016/j.gaitpost.2011.08.013
Duysens, J., & Pearson, K.G. (1976). The role of cutaneous afferents from the distal hindlimb in the regulation of the step cycle of thalamic cats. Experimental Brain Research, 24(3), 245–255. https://doi.org/10.1007/BF00235013
Dyck, P.J., Herrmann, D.N., Staff, N.P., & Dyck, P.J. (2013). Assessing decreased sensation and increased sensory phenomena in diabetic polyneuropathies. Diabetes, 62(11), 3677–3686. https://doi.org/10.2337/db13-0352
Edwards, D.H., & Prilutsky, B.I. (2017). Sensory feedback in the control of posture and locomotion. In S.L. Hooper & A. Büschges (Eds.), Neurobiology of motor control: Fundamental concepts and new directions (pp. 263–304). Wiley.
Farrell, B.J., Bulgakova, M.A., Beloozerova, I.N., Sirota, M.G., & Prilutsky, B.I. (2014). Body stability and muscle and motor cortex activity during walking with wide stance. Journal of Neurophysiology, 112(3), 504–524. https://doi.org/10.1152/jn.00064.2014
Farrell, B.J., Prilutsky, B.I., Kistenberg, R.S., Dalton, J.F., & Pitkin, M. (2014). An animal model to evaluate skin-implant-bone integration and gait with a prosthesis directly attached to the residual limb. Clinical Biomechanics, 29(3), 336–349. https://doi.org/10.1016/j.clinbiomech.2013.12.014
Farrell, B.J., Prilutsky, B.I., Ritter, J.M., Kelley, S., Popat, K., & Pitkin, M. (2014). Effects of pore size, implantation time, and nano-surface properties on rat skin ingrowth into percutaneous porous titanium implants. Journal of Biomedical Materials Research Part A, 102(5), 1305–1315. https://doi.org/10.1002/jbm.a.34807
Frigon, A., Akay, T., & Prilutsky, B.I. (2021). Control of mammalian locomotion by somatosensory feedback. Comprehensive Physiology, 12(1), 2877–2947. https://doi.org/10.1002/cphy.c210020
Frigon, A., Thibaudier, Y., & Hurteau, M.F. (2015). Modulation of forelimb and hindlimb muscle activity during quadrupedal tied-belt and split-belt locomotion in intact cats. Neuroscience, 290, 266–278. https://doi.org/10.1016/j.neuroscience.2014.12.084
Galvez-Lopez, E., Maes, L.D., & Abourachid, A. (2011). The search for stability on narrow supports: An experimental study in cats and dogs [Research Support, Non-U.S. Gov’t]. Zoology, 114(4), 224–232. https://doi.org/10.1016/j.zool.2011.03.001
George, J.A., Page, D.M., Davis, T.S., Duncan, C.C., Hutchinson, D.T., Rieth, L.W., & Clark, G.A. (2020). Long-term performance of Utah slanted electrode arrays and intramuscular electromyographic leads implanted chronically in human arm nerves and muscles. Journal of Neural Engineering, 17(5), 056042. https://doi.org/10.1088/1741-2552/abc025
Goble, D.J., Coxon, J.P., Wenderoth, N., Van Impe, A., & Swinnen, S.P. (2009). Proprioceptive sensibility in the elderly: Degeneration, functional consequences and plastic-adaptive processes. Neuroscience & Biobehavioral Reviews, 33(3), 271–278. https://doi.org/10.1016/j.neubiorev.2008.08.012
Goldberger, M.E. (1988). Partial and complete deafferentation of cat hindlimb: The contribution of behavioral substitution to recovery of motor function. Experimental Brain Research, 73(2), 343–353. https://doi.org/10.1007/BF00248226
Gordon, T. (2016). Nerve regeneration: Understanding biology and its influence on return of function after nerve transfers. Hand Clinics, 32(2), 103–117. https://doi.org/10.1016/j.hcl.2015.12.001
Gordon, T., Hoffer, J.A., Jhamandas, J., & Stein, R.B. (1980). Long-term effects of axotomy on neural activity during cat locomotion. Journal of Physiology, 303(1), 243–263. https://doi.org/10.1113/jphysiol.1980.sp013283
Gorman, P.H., & Mortimer, J.T. (1983). The effect of stimulus parameters on the recruitment characteristics of direct nerve stimulation. IEEE Transactions on Biomedical Engineering, BME-30(7), 407–414. https://doi.org/10.1109/TBME.1983.325041
Graczyk, E.L., Christie, B.P., He, Q., Tyler, D.J., & Bensmaia, S.J. (2022). Frequency shapes the quality of tactile percepts evoked through electrical stimulation of the nerves. Journal of Neuroscience, 42(10), 2052–2064. https://doi.org/10.1523/JNEUROSCI.1494-21.2021
Gregor, R.J., Maas, H., Bulgakova, M.A., Oliver, A., English, A.W., & Prilutsky, B.I. (2018). Time course of functional recovery during the first 3 mo after surgical transection and repair of nerves to the feline soleus and lateral gastrocnemius muscles. Journal of Neurophysiology, 119(3), 1166–1185. https://doi.org/10.1152/jn.00661.2017
Guertin, P., Angel, M.J., Perreault, M.C., & McCrea, D.A. (1995). Ankle extensor group I afferents excite extensors throughout the hindlimb during fictive locomotion in the cat. Journal of Physiology, 487(1), 197–209. https://doi.org/10.1113/jphysiol.1995.sp020871
Henry, M., & Baudry, S. (2019). Age-related changes in leg proprioception: Implications for postural control. Journal of Neurophysiology, 122(2), 525–538. https://doi.org/10.1152/jn.00067.2019
Hildebrand, M. (1965). Symmetrical gaits of horses. Science, 150(3697), 701–708. https://doi.org/10.1126/science.150.3697.701
Hohne, A., Stark, C., Bruggemann, G.P., & Arampatzis, A. (2011). Effects of reduced plantar cutaneous afferent feedback on locomotor adjustments in dynamic stability during perturbed walking. Journal of Biomechanics, 44(12), 2194–2200. https://doi.org/10.1016/j.jbiomech.2011.06.012
Honeycutt, C.F., & Nichols, T.R. (2010). Disruption of cutaneous feedback alters magnitude but not direction of muscle responses to postural perturbations in the decerebrate cat. Experimental Brain Research, 203(4), 765–771. https://doi.org/10.1007/s00221-010-2281-8
Housley, S.N., Nardelli, P., Rotterman, T.M., & Cope, T.C. (2021). Neural circuit mechanisms of sensorimotor disability in cancer treatment. Proceedings of the National Academy of Sciences of the United States of America, 118(51), Article e2100428118. https://doi.org/10.1073/pnas.2100428118
Hunt, C.C., & McIntyre, A.K. (1960). An analysis of fibre diameter and receptor characteristics of myelinated cutaneous afferent fibres in cat. Journal of Physiology, 153(1), 99–112. https://doi.org/10.1113/jphysiol.1960.sp006521
Jarrell, J.R., Farrell, B.J., Kistenberg, R.S., Dalton, J.F.t., Pitkin, M., & Prilutsky, B.I. (2018). Kinetics of individual limbs during level and slope walking with a unilateral transtibial bone-anchored prosthesis in the cat. Journal of Biomechanics, 76, 74–83. https://doi.org/10.1016/j.jbiomech.2018.05.021
Jeyapalina, S., Beck, J.P., Agarwal, J., & Bachus, K.N. (2017). A 24-month evaluation of a percutaneous osseointegrated limb-skin interface in an ovine amputation model. Journal of Materials Science: Materials in Medicine, 28(11), 179. https://doi.org/10.1007/s10856-017-5980-x
Klishko, A.N., Akyildiz, A., Mehta-Desai, R., & Prilutsky, B.I. (2021). Common and distinct muscle synergies during level and slope locomotion in the cat. Journal of Neurophysiology, 126(2), 493–515. https://doi.org/10.1152/jn.00310.2020
Knikou, M. (2010). Plantar cutaneous afferents normalize the reflex modulation patterns during stepping in chronic human spinal cord injury. Journal of Neurophysiology, 103(3), 1304–1314. https://doi.org/10.1152/jn.00880.2009
Kriellaars, D.J., Brownstone, R.M., Noga, B.R., & Jordan, L.M. (1994). Mechanical entrainment of fictive locomotion in the decerebrate cat. Journal of Neurophysiology, 71(6), 2074–2086.https://doi.org/10.1152/jn.1994.71.6.2074
Lam, T., & Pearson, K.G. (2002). Sartorius muscle afferents influence the amplitude and timing of flexor activity in walking decerebrate cats. Experimental Brain Research, 147(2), 175–185. https://doi.org/10.1007/s00221-002-1236-0
Latash, E.M., Barnett, W.H., Park, H., Rider, J.M., Klishko, A.N., Prilutsky, B.I., & Molkov, Y.I. (2020). Frontal plane dynamics of the centre of mass during quadrupedal locomotion on a split-belt treadmill. Journal of the Royal Society Interface, 17(170), Article 20200547. https://doi.org/10.1098/rsif.2020.0547.
Loeb, G.E. (1993). The distal hindlimb musculature of the cat: Interanimal variability of locomotor activity and cutaneous reflexes. Experimental Brain Research, 96(1), 125–140. https://doi.org/10.1007/BF00230446
Lyle, M.A., Prilutsky, B.I., Gregor, R.J., Abelew, T.A., & Nichols, T.R. (2016). Self-reinnervated muscles lose autogenic length feedback, but intermuscular feedback can recover functional connectivity. Journal of Neurophysiology, 116(3), 1055–1067. https://doi.org/10.1152/jn.00335.2016
Meyer, P.F., Oddsson, L.I., & De Luca, C.J. (2004). Reduced plantar sensitivity alters postural responses to lateral perturbations of balance. Experimental Brain Research, 157(4), 526–536. https://doi.org/10.1007/s00221-004-1868-3
National Research Council of the National Academies. (2011). Guide for the care and use of laboratory animals. Washington, D.C.: National Academies Press.
Ollivier-Lanvin, K., Krupka, A.J., AuYong, N., Miller, K., Prilutsky, B.I., & Lemay, M.A. (2011). Electrical stimulation of the sural cutaneous afferent nerve controls the amplitude and onset of the swing phase of locomotion in the spinal cat [Comparative Study Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t]. Journal of Neurophysiology, 105(5), 2297–2308. https://doi.org/10.1152/jn.00385.2010
Onushko, T., Boerger, T., Van Dehy, J., & Schmit, B.D. (2019). Dynamic stability and stepping strategies of young healthy adults walking on an oscillating treadmill. PLoS One, 14(2), Article e0212207. https://doi.org/10.1371/journal.pone.0212207
Pantall, A., Gregor, R.J., & Prilutsky, B.I. (2012). Stance and swing phase detection during level and slope walking in the cat: Effects of slope, injury, subject and kinematic detection method. Journal of Biomechanics, 45(8), 1529–1533. https://doi.org/10.1016/j.jbiomech.2012.03.013
Park, H., Islam, M.S., Grover, M.A., Klishko, A.N., Prilutsky, B.I., & DeWeerth, S.P. (2018). A prototype of a neural, powered, transtibial prosthesis for the cat: Benchtop characterization. Frontiers in Neuroscience, 12, Article 471. https://doi.org/10.3389/fnins.2018.00471
Park, H., Latash, E.M., Molkov, Y.I., Klishko, A.N., Frigon, A., DeWeerth, S.P., & Prilutsky, B.I. (2019). Cutaneous sensory feedback from paw pads affects lateral balance control during split-belt locomotion in the cat. Journal of Experimental Biology, 222(14), Article jeb.198648. https://doi.org/10.1242/jeb.198648
Park, H., Oh, K., Prilutsky, B.I., & DeWeerth, S.P. (2016). A real-time closed-loop control system for modulating gait characteristics via electrical stimulation of peripheral nerves. Paper presented at the IEEE Biomedical Circuits and Systems Conference (BioCAS) (pp. 95–98).
Pearson, K.G., & Collins, D.F. (1993). Reversal of the influence of group Ib afferents from plantaris on activity in medial gastrocnemius muscle during locomotor activity. Journal of Neurophysiology, 70(3), 1009–1017. https://doi.org/10.1152/jn.1993.70.3.1009
Pitkin, M., Cassidy, C., Muppavarapu, R., & Edell, D. (2012). Recording of electric signal passing through a pylon in direct skeletal attachment of leg prostheses with neuromuscular control. IEEE Transactions on Biomedical Engineering, 59(5), 1349–1353. https://doi.org/10.1109/TBME.2012.2187784
Pitkin, M., Cassidy, C., Shevtsov, M.A., Jarrell, J.R., Park, H., Farrell, B.J., . . . Prilutsky, B.I. (2021). Recent progress in animal studies of the skin- and bone-integrated pylon with deep porosity for bone-anchored limb prosthetics with and without neural interface. Military Medicine, 186, 688–695. https://doi.org/10.1093/milmed/usaa445
Pitkin, M., & Raykhtsaum, G. (2012). Skin integrated device (U.S. Patent No. 8257435). http://www.google.com/patents/US8257435
Prilutsky, B.I., Herzog, W., & Allinger, T.L. (1996). Mechanical power and work of cat soleus, gastrocnemius and plantaris muscles during locomotion: Possible functional significance of muscle design and force patterns. Journal of Experimental Biology, 199(4), 801–814. https://doi.org/10.1242/jeb.199.4.801
Prilutsky, B.I., Maas, H., Bulgakova, M., Hodson-Tole, E.F., & Gregor, R.J. (2011). Short-term motor compensations to denervation of feline soleus and lateral gastrocnemius result in preservation of ankle mechanical output during locomotion [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t]. Cells Tissues Organs, 193(5), 310–324. https://doi.org/10.1159/000323678
Raspopovic, S., Capogrosso, M., Petrini, F.M., Bonizzato, M., Rigosa, J., Di Pino, G., . . . Micera, S. (2014). Restoring natural sensory feedback in real-time bidirectional hand prostheses. Science Translational Medicine, 6(222), 222ra219. https://doi.org/10.1126/scitranslmed.3006820
Schielzeth, H., Dingemanse, N.J., Nakagawa, S., Westneat, D.F., Allegue, H., Teplitsky, C., . . . Araya-Ajoy, Y.G. (2020). Robustness of linear mixed-effects models to violations of distributional assumptions. Methods in Ecology and Evolution, 11(9), 1141–1152. https://doi.org/10.1111/2041-210X.13434
Schomburg, E.D., Petersen, N., Barajon, I., & Hultborn, H. (1998). Flexor reflex afferents reset the step cycle during fictive locomotion in the cat. Experimental Brain Research, 122(3), 339–350. https://doi.org/10.1007/s002210050522
Smith, P.F. (2012). A note on the advantages of using linear mixed model analysis with maximal likelihood estimation over repeated measures ANOVAs in psychopharmacology: Comment on Clark et al. (2012). Journal of Psychopharmacology, 26(12), 1605–1607. https://doi.org/10.1177/0269881112463471
Stecina, K., Quevedo, J., & McCrea, D.A. (2005). Parallel reflex pathways from flexor muscle afferents evoking resetting and flexion enhancement during fictive locomotion and scratch in the cat. Journal of Physiology, 569(1), 275–290. https://doi.org/10.1113/jphysiol.2005.095505
Tan, D.W., Schiefer, M.A., Keith, M.W., Anderson, J.R., Tyler, J., & Tyler, D.J. (2014). A neural interface provides long-term stable natural touch perception. Science Translational Medicine, 6(257), 257ra138. https://doi.org/10.1126/scitranslmed.3008669
Tuthill, J.C., & Azim, E. (2018). Proprioception. Current Biology, 28(5), R194–R203. https://doi.org/10.1016/j.cub.2018.01.064
Walmsley, B., Hodgson, J.A., & Burke, R.E. (1978). Forces produced by medial gastrocnemius and soleus muscles during locomotion in freely moving cats. Journal of Neurophysiology, 41(5), 1203–1216. https://doi.org/10.1152/jn.1978.41.5.1203
West, B.T., Welch, K.B., & Gatecki, A.T. (2015). Linear mixed models. A practical guide using statistical software (2nd ed.). CRC Press.
