Persistence of locomotor-related interlimb reflex networks during walking after stroke

https://doi.org/10.1016/j.clinph.2011.07.049Get rights and content

Abstract

Objective

Cutaneous nerve stimulation evokes coordinated and phase-modulated reflex output widely distributed to muscles of all four limbs during walking. Accessibility to this distributed network after stroke offers insight into the pathological changes and suggests utility for therapeutic applications. Here we examined muscles in both the more (MA) and less affected (LA) legs evoked by stimulation at the ankle and wrist during walking in chronic (>6 months post CVA) stroke.

Methods

Stroke and control participants walked on a treadmill with a harness support system. Reflexes were evoked with trains of electrical stimuli delivered separately to the cutaneous superficial peroneal (SP; at the ankle) and superficial radial (SR; at the wrist) nerves. Background locomotor and reflex EMG were phase-averaged across the gait cycle and analyzed off line.

Results

Locomotor background muscle activation patterns were altered bilaterally in stroke, as compared with control. Phase-dependent modulation of interlimb cutaneous reflexes was found in both stroke and control subjects with stimulation of each nerve, but responses were blunted in stroke. Reflex reversal in tibialis anterior (TA) at heel strike with SP nerve stimulation was present in both groups. Notably, SR nerve stimulation produced facilitation during the swing-to-stance transition in the TA and suppression of MG in the MA leg during stance.

Conclusions

Interlimb cutaneous inputs may access coordinated reflex pathways in the MA limb during walking after stroke. Importantly activation in these pathways could provoke responses to counter foot drop during swing phase of walking. Additionally, our data support the perspective that there is no “unaffected” side after stroke and that caution should be used when interpreting the LA side as “control” after stroke.

Significance

The presence of functionally-relevant interlimb cutaneous reflexes in the MA leg presents a substrate that may be strengthened by rehabilitation.

Highlights

► Somatosensory networks assisting in locomotor coordination between the arms and legs remain accessible after stroke. ► Cutaneous input from the hand evokes coordinated reflexes in leg muscles that could assist with foot drop during swing and spasticity during stance. ► There is no completely “unaffected” side during walking after stroke since responses on both sides differed from control.

Introduction

Regulation of rhythmic arm and leg movement during human locomotion is supported by somatosensory linkages in the form of “interlimb reflexes” (Dietz et al., 2001, Haridas and Zehr, 2003, Lamont and Zehr, 2007) and neural coupling between lumbar and cervical spinal cord networks (Mezzarane et al., 2011, Zehr et al., 2007a, Zehr et al., 2007b). The extensive distribution across many muscles (Zehr et al., 2001) and modulation of cutaneous reflexes during walking (Haridas and Zehr, 2003, Zehr and Duysens, 2004) suggests a widespread interlimb network in neurologically intact (NI) humans. Phase-dependent reflexes are also typical such as suppression of the middle latency reflex bilaterally in tibialis anterior (TA) muscle at heel strike (Dietz et al., 1994, Duysens et al., 1990, Duysens et al., 1992, Duysens and Tax, 1994, Haridas and Zehr, 2003, Yang and Stein, 1990, Zehr and Haridas, 2003). In addition, phase-related responses are found in muscles of all four limbs during gait regardless of which limb is directly stimulated (Haridas and Zehr, 2003). This finely regulated control depends upon the integrity of the nervous system and excitability of interneuronal networks for useful and natural function. Interruption of connectivity occurring as a result of spinal cord injury or stroke, contributes to deficits in the neural regulation of walking.

Previously, stimulation applied to the cutaneous nerve innervating the top of the foot (SP nerve) of the more affected MA leg in stroke participants with mild spasticity revealed a partially-preserved stumble-correction (Zehr et al., 1998). Despite that, significant differences were found such as a dominance of suppression in extensor muscles of stroke participants during stance while in NI subjects, ankle extensor soleus (Sol) was unaffected and knee extensor vastus lateralis (VL) was facilitated. Thus, cutaneous input can evoke inhibitory responses in some leg muscles on the more affected (MA) sided during walking after stroke. Available data on cutaneous reflexes evoked in the opposite leg (crossed reflexes) in neurologically intact subjects suggests smaller amplitude responses (Van Wezel et al., 1997, Haridas and Zehr, 2003). In people who are spinal cord injured, tibial (Jones and Yang, 1994) and sural (Knikou et al., 2009) cutaneous nerve reflexes are modulated during walking implying the presence of spinal regulation below the level of injury.

Measuring reflexes in the MA limb from stimulation of the less affected (LA) limb gives information on crossed spinal responses. Crossed responses to stimulation of the cutaneous SP nerve provides an index of the participation of the contralateral leg muscles in corrections for the ipsilateral stumble-like perturbation (Zehr and Stein, 1999). Reflexes evoked in the leg muscles by cutaneous superficial radialis (SR) nerve stimulation at the wrist indicate interlimb responses to afferent input from the hand (Haridas and Zehr, 2003, Zehr and Haridas, 2003). Currently there are no comparable data for interlimb reflexes after stroke. Such data would be useful to assess the integrity of the widespread and integrated locomotor network after stroke. Additionally, this information could be usefully applied for guiding therapeutic intervention involving clinically applied neurostimulation (e.g. TENS) (Zehr, 2006, Dewald and Given, 1994, Seib et al., 1994).

The purpose of this study was to examine the integrity of the interlimb reflex network after stroke as assessed by responses in the MA muscles evoked by SP and SR nerve stimulation of the LA ankle and wrist, respectively. Given our previous work showing liminal preservation of ipsilateral responses in the MA leg (Zehr et al., 1998) as well as arm and leg coupling during rhythmic arm movement (Barzi and Zehr, 2008) after stroke, we hypothesized that interlimb cutaneous reflexes would be present in stroke subjects. However, the amplitude of the evoked responses and extensive network of responses would be “blunted” when compared with those seen in control, neurologically intact participants.

Much of the general methodology for nerve stimulation and muscle recording is similar to that previously described (Zehr et al., 1998, Haridas and Zehr, 2003), and is only briefly described here.

Stroke subjects were identified with a more (MA; contralesional) and a less affected (LA; ipsilesional) side. Additionally, muscles are described as ipsilateral or contralateral with respect to the site of nerve stimulation.

Twenty-three subjects with chronic stroke between the ages of 39 and 81 years (average 60, median 57) and 22 neurologically and metabolically intact control subjects between the ages of 37 and 88 years (average 64, median 69) participated with informed, written consent according to a protocol approved by the Human Research Ethics board at the University of Victoria. The stroke subjects comprised 12 with left and 11 with right-sided lesions. All participants were at least 6 months (average 59.61 ± 52 months, median 45 months) post-infarct (see Table 1). For four subjects, brain imaging was not available and the extent of infarction is not noted. Clinical outcomes were measured by a registered physiotherapist. Muscle tone and the level of motor recovery in the arms and legs of each subject were found using the Modified Ashworth Scale (Ashworth, 1964) and the Brunnstrom Stroke Scale (Bohannon and Smith, 1987), respectively. Ten subjects were tested for and exhibited nonsustaining clonus at the ankle. Cutaneous sensation was measured using the 5 piece Semmes Weinstein calibrated filaments (Sammons Preston Roylan, Cedarburg WI). Of the 22 subjects tested, 2 demonstrated diminished light touch (DLT), 2 demonstrated diminished protective sensation, 15 exhibited a loss of protective sensation (LPS) and 2 were untestable (i.e. no response to largest filament). A functional measure of gait was derived using the 6 point Functional Ambulation Categories Scale (FAC) (Holden et al., 1986).

Subjects performed a maximum of three experimental trials; one static trial and one walking trial for each nerve (SP and SR) examined. In some participants insufficient functional capacity was present to record two walking trials. In those cases only the SP nerve stimulation walking condition was examined.

Static trials: All subjects stood in place on the treadmill to generate background electromyographic (EMG) levels in arm and leg muscles and to determine if the stimulus intensity (see below) was adequate to elicit a reflex.

Walking trials: Stroke (see Table 2) and control subjects (see Table 3) walked at self-selected (“comfortable”) speeds (stroke X 1.75 ± 1.21 km/h; control average 4.09 ± 0.93 km/h) on a motorized treadmill (Woodway USA, Waukesha, WI) with an overhead safety harness (Pneu-weight, Pneumex Inc., Sandpoint, ID) in place. Three of the stroke subjects required support in the harness by a percentage (average 23% ± 11.30) of their body weight. Otherwise all stroke and control subjects wore the safety harness in the safety mode without body weight support. 19 of the 24 stroke subjects held either the side or the frontrail of the treadmill with their LA (ipsilateral) hand during both walking trials (7 of these held with both hands on front or side rails). Five of the control subjects held a railing either in front or at the side of the treadmill (1 with the ipsilateral hand, 2 with the contralateral hand, 2 with both hands). Five of the stroke subjects used an ankle-foot-orthosis (AFO) as per their usual walking method.

Cutaneous reflexes were evoked with trains (5 × 1.0 ms pulses at 300 Hz) of isolated constant current stimulation (Grass S88 stimulator with SIU5 and CCU1 units, AstroMed-Grass Inc.) applied with surface electrodes (Thought Technologies, Ltd.) to the SP nerve at the LA ankle of the stroke subjects and the right side in control subjects. In a separate trial, cutaneous reflexes were evoked in the same manner to the SR nerve at the LA wrist of 19 of the stroke subjects and to the right wrist of all of the control subjects. Stimulation trains (n = 160) were applied pseudorandomly across the gait cycle (yielding ∼15–20 stimulations per bin) in all walking trials with an intensity set as a multiple of the threshold at which a clear radiating paresthesia (radiating threshold (RT)) was reported into the innervation area of the nerve. That is, into the dorsum of the foot for the SP nerve and the dorsum of the hand towards the thumb and second digit for the SR nerve. RT was determined in a standing (static) position. Stimulation intensities were set to evoke a strong cutaneous sensation not deemed painful by the subjects and a minimal reflex response (SP nerve: stroke 2.90x RT ± 1.22, control 2.20x RT ± 0.65; SR nerve: stroke 2.54x RT ± 0 .87, control 2.13x RT ± 0.52). In two stroke subjects with sensory loss we were not able to determine the RT; therefore stimulation intensities in those subjects were determined by the presence of a minimal observable reflex response in the static position (average 2.55x RT ± .05).

After abrading and cleansing the skin with alcohol swabs, EMG was collected using disposable 1 cm surface EMG electrodes (Thought Technologies Ltd.) in a bipolar configuration with a 2 cm inter-electrode distance. EMG was recorded bilaterally from: Vastus Lateralis (VL), Biceps Femoris (BF), Tibialis Anterior (TA) and Medial Gastrocnemius (MG). On the stimulated side EMG was additionally collected from Anterior (AD) and Posterior Deltoid (PD) to ensure SR stimulation intensity was above threshold. Ground electrodes were placed over electrically neutral tissue. EMG signals were preamplified (at 5000x) and bandpass filtered at 100–300 Hz (P511 Grass Instruments, AstroMed, Inc.).

Step-cycle parameters (e.g. heel contact, toe-off) for walking were obtained with the use of custom-made force sensing resistors, located in the heel of the right shoe insole of the control subjects’ and in the heel of the LA shoe insole of the stroke subject. Use of the LA side in stroke subjects ensured a step cycle was determined in the absence of clear heel strike of the MA limb. Offline we separated the step cycle into 8 bins (see (Lamont and Zehr, 2006, Lamont and Zehr, 2007) with bins 1–5 for stance and 6–8 as swing with transitions for stance to swing in bin 5 and the swing to stance transition occurring in bin 1. Three static trials (heel strike, toe off and standing with feet together) were used to collect data for goniometer calibration.

All data were sampled at 1000 Hz with a 12 bit A/D converter connected to a microcomputer running custom-written LabView software (National Instruments, Austin, TX). Offline, custom-written software programs (Matlab, The Mathworks, Natick, MA) were used to separate the step cycles into eight equal parts or phases beginning with the onset of stance phase. During the offline analysis, EMG signals were full-wave rectified and filtered (see below) prior to averaging. At each phase of movement, the average trace from the non-stimulated steps was subtracted from the corresponding stimulated average trace to produce a subtracted EMG “reflex” trace for each subject. The stimulus artifact was removed from the subtracted reflex trace and it was then filtered at 40 Hz using a dual-pass 4th order Butterworth low-pass filter.

Cutaneous reflex excitability was determined based on analysis of the middle latency (80–120 ms to peak) response. Reflexes for a given muscle were only analyzed if at least one response at any latency exceeded a 2-SD band (centered about the mean prestimulus EMG level) for any task. Reflex amplitudes were calculated as an average value from a 10 ms time window centered about the peak for each response. All EMG values for each subject were normalized to the averaged peak control (non-stimulated) EMG amplitude value. Further, a modulation index (MI) of overall change in EMG parameters over the step cycle was determined by subtracting the maximum EMG amplitude in the step cycle from the minimum EMG amplitude in the step cycle and expressed as percentages. This measure provides an index of overall amplitude modulation independent of the pattern of modulation across the walking cycle. These data were then normalized to the maximum undisturbed EMG and averaged for all stroke and control subjects.

Data from SP and SR nerve stimulation were analyzed separately for EMG amplitude, modulation index and kinematics. For all EMG measures analysis was performed using the averaged normalized values for each subject from each phase of the step cycle. Student’s t tests for independent samples (with Lavene’s F-test corrections) were performed at each phase of the cycle to determine significant differences between stroke and control subjects in background EMG levels and cutaneous reflex amplitudes. Analyses on the modulation indices were performed using the averaged normalized values of all subjects for each muscle. Student’s t-tests for independent samples (with Lavene’s F-test corrections) were performed to determine significant differences between stroke and control subjects in background EMG and cutaneous reflex amplitudes. As well, to assess any interactions between subject type and muscle the modulation index, a repeated measures analysis of variance was used with a post hoc Tukey’s test.

Descriptive statistics included means ± standard error of the means (SE). The level of significance was set at an alpha level of p < =0.05.

Section snippets

Background locomotor EMG patterns

Single subject traces from all muscles in stroke are shown in Fig. 1 and for control in Fig. 2. For both figures, the averaged single subject full step cycle EMG records are shown as thin grey lines and the grand average of all subjects as the thick black lines. Quantified levels of background EMG for control and stroke subjects during walking are shown in Fig. 3. Muscles for the LA leg are in the top panels (A) and MA in the bottom panels (B). For the LA muscles, many differences were seen for

Discussion

The main result of this experiment demonstrates the persistence and modulation of interlimb reflexes during walking after stroke despite the interruption of some descending regulation of interneuronal excitability arising from the supraspinal lesion. There are three main outcomes of note. First, somatosensory input coming from the hand (i.e. SR nerve reflexes) generated significant reflexes in all of the MA leg muscles during walking. Some of these responses occur at critical gait phases and

Acknowledgments

This work was supported by a Grant-in-Aid of research from the Heart and Stroke Foundation of Canada (BC and Yukon) to EPZ.

References (48)

  • B. Ashworth

    Preliminary trial of carisoprodal in multiple sclerosis

    Practitioner

    (1964)
  • J.E. Balter et al.

    Neural coupling between the arms and legs during rhythmic locomotor-like cycling movement

    J Neurophysiol

    (2007)
  • R.W. Bohannon et al.

    Interrater Reliablilty of a Modified Ashworth Scale of Muscle Spasticity

    Phys Ther

    (1987)
  • D.A. Brown et al.

    Increased workload enhances force output during pedaling exercise in persons with poststroke hemiplegia

    Stroke

    (1998)
  • D.A. Brown et al.

    Muscle activity adapts to anti-gravity posture during pedalling in persons with post-stroke hemiplegia

    Brain

    (1997)
  • J.P.A. Dewald et al.

    Electrical Stimulation and spasticity reduction: fact or fiction?

    Phys Med and Rehab: State of the Art Rev

    (1994)
  • V. Dietz et al.

    Neuronal coordination of arm and leg movements during human locomotion

    Eur J Neurosci

    (2001)
  • V. Dietz et al.

    Human neuronal interlimb coordination during split-belt locomotion

    Exp Brain Res

    (1994)
  • B.H. Dobkin

    Rehabilitation after stroke

    NEJM

    (2005)
  • J. Duysens et al.

    Gait acts as a gate for reflexes from the foot

    Can J Physiol Pharmacol

    (2004)
  • J. Duysens et al.

    Phase-dependent reversal of reflexly induced movements during human gait

    Exp Brain Res

    (1992)
  • J. Duysens et al.

    Gating and reversal of reflexes in ankle muscles during human walking

    Exp Brain Res

    (1990)
  • D.P. Ferris et al.

    Moving the arms to activate the legs

    Exerc Sport Sci Rev

    (2006)
  • C. Haridas et al.

    Coordinated interlimb compensatory responses to electrical stimulation of cutaneous nerves in the hand and foot during walking

    J Neurophysiol

    (2003)
  • Cited by (0)

    View full text