Stretch-reflex threshold modulation during active elbow movements in post-stroke survivors with spasticity

Accepted Manuscript Stretch-reflex threshold modulation during active elbow movements in poststroke survivors with spasticity Nicolas A. Turpin, Anatol G. Feldman, Mindy F. Levin PII: DOI: Reference:

S1388-2457(17)30899-4 http://dx.doi.org/10.1016/j.clinph.2017.07.411 CLINPH 2008224

To appear in:

Clinical Neurophysiology

Received Date: Revised Date: Accepted Date:

8 May 2017 24 June 2017 17 July 2017

Please cite this article as: Turpin, N.A., Feldman, A.G., Levin, M.F., Stretch-reflex threshold modulation during active elbow movements in post-stroke survivors with spasticity, Clinical Neurophysiology (2017), doi: http:// dx.doi.org/10.1016/j.clinph.2017.07.411

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Stretch-reflex threshold modulation during active elbow movements in post-stroke survivors with spasticity Nicolas A. Turpin, PhD1, 2, 4, Anatol G. Feldman, PhD 1, 2, 4, Mindy F. Levin, PhD 3, 4 1

Department of Neuroscience and 2Institute of Biomedical Engineering, University of Montreal; School of Physical and Occupational Therapy, McGill University; 4Center for Interdisciplinary Research in Rehabilitation of Greater Montreal (CRIR), Montreal, QC, Canada 3

Corresponding author: Dr. Mindy F. Levin School of Physical and Occupational Therapy McGill University 3654 Promenade Sir William Osler Montreal, Quebec. Canada H3G-1Y5 Tel.: +1-514-398-3994 Fax: +1-514-398-6360 E-mail: [email protected]

2

Abstract Objectives. Voluntary movements post-stroke are affected by abnormal antagonist activation due to exaggerated stretch reflexes (SRs). We examined the ability of post-stroke subjects to regulate SRs in spastic muscles. Methods. Elbow flexor and extensor EMGs and angle were recorded in 13 subjects with chronic post-stroke spasticity. Muscles were either stretched passively (relaxed arm) or actively (antagonist contraction) at different velocities. Velocity-dependent SR thresholds were defined as angles where stretched muscle EMG exceeded 3SDs of baseline. Sensitivity of SRs to stretch velocity was defined as µ. The regression through thresholds was interpolated to zero velocity to obtain the tonic SR threshold (TSRT) angle. Results. Compared to passive stretches, TSRTs during active motion occurred at longer muscle lengths (i.e., increased in flexors and decreased in extensors by 10-40°). Values of µ increased by 1.5-4.0. Changes in flexor TSRTs during active compared to passive stretches were correlated with clinical spasticity (r=-0.68) and arm motor impairment (r=0.81). Conclusions. Spasticity thresholds measured at rest were modulated during active movement. Arm motor impairments were related to the ability to modulate SR thresholds between the two states rather than to passive-state values. Significance. Relationship between spasticity and movement disorders may be explained by deficits in SR threshold range of regulation and modifiability, representing a measure of strokerelated sensorimotor deficits.

Highlights •

Tonic stretch-reflex thresholds in post-stroke spasticity occurred within the joint range at rest.

•

Threshold modulation during active movements was related to clinical spasticity and motor impairment.

•

Characteristics of threshold modulation provide information about post-stroke sensorimotor deficits.

Keywords: muscle spasticity; reflex, stretch; stroke; rehabilitation; voluntary movement.

3

1.0 Introduction Spasticity is a common complication of stroke, occurring in ~20-50% of patients in the first year (Wissel et al., 2013) and often associated with other sensory and motor impairments (e.g., muscle weakness, loss of dexterity). Spasticity is generally assessed by resistance or EMG responses to passive muscle stretches and has been attributed to exaggerated spinal stretch reflexes (SRs) and alterations in intrinsic muscle properties (Dietz and Sinkjaer, 2007). For example, motor units of spastic muscles often have an impaired ability to relax (Lewek et al., 2007), prolonged spontaneous firing (Mottram et al., 2010) and low firing rates (Young and Mayer, 1982; Gemperline et al., 1995). Neural mechanisms underlying spasticity include deficits in the regulation of inhibitory reflex pathways (e.g., reciprocal Ia inhibition, presynaptic inhibition) and hyper-excitability of α-motoneurons (MNs; Nielsen et al., 2007). A causal relationship may exist between spasticity and limitations in performance of daily activities (Pandyan et al., 2005). However, clear understanding of the relationship between spasticity and movement deficits remains elusive (O'Dwyer et al., 1996; Mirbagheri et al., 2001; Dietz and Sinkjaer, 2007). This situation might be changed by considering spasticity and movement production within the same conceptual framework such as the threshold position control theory (an extension of the Equilibrium-point hypothesis). The theory is based on findings that voluntary muscle activation originates from central shifts in the spatial stretch-reflex threshold (SRT), defined as the muscle length (or corresponding joint angle) at which muscle activity emerges in response to stretch (Feldman, 2015; Raptis et al., 2010; Fig. 1). The SRT is velocity-dependent (dynamic SRT or DSRT; e.g., Powers et al., 1989), such that the length at which the muscle begins to be activated decreases with increasing stretch velocity (see equation 1). DSRT has a velocity-independent component called the tonic stretch reflex threshold (TSRT; e.g., Matthews, 1959). It has been shown that central changes in SRTs underlie voluntary movements in humans (Asatryan and Feldman, 1965). By shifting TSRT, the nervous system predetermines the spatial (angular) range in which the muscle can generate active forces (Fig. 1A). Shifts in TSRT can be accomplished by different descending systems (vestibulo-, reticulocortico- and rubro-spinal) that mediate influences on α-MNs mono- and/or poly-synaptically as well as pre-synaptically or via γ-MNs (Feldman and Orlovsky, 1972; Capaday, 1995). Within this framework, spasticity can be understood as an impaired ability to increase SRTs in affected muscles to prevent active responses to passive muscle stretching (Powers et al., 1988, 1989; Levin and Feldman, 1994; Musampa et al., 2007; Fig. 1B, right panel), as in healthy subjects. However, the framework also implies that in addition to muscle relaxation, SRT modulation also underlies the control of voluntary movements. In particular, the muscle that is inactive at a given length, can be activated by shifting the threshold below this length. Therefore, a reduced capacity to modulate SRTs might contribute to movement disorders in spastic patients (Jones and Yang, 1994; Sinkjær et al., 1996; Faist, 1999; Morita, 2001; Burne, 2005). In contrast, procedures facilitating SR modulation can improve motor function, as observed in chronic spinal cord injury (Manella et al., 2013). The possibility remains that SRTs are also modulated during movements in spastic muscles and limitations in this capacity may be related to movement disorders.

4

We measured DSRTs from which we computed TSRTs in spastic muscles at rest (i.e. when patients were instructed to completely relax arm muscles) and determined whether and how the thresholds (TSRTs) were modified during active movement. Since movement production depends on the capacity to modulate TSRTs, we assumed that in patients who can produce elbow movements, some residual ability to regulate TSRTs in spastic muscles would be retained (Morita, 2001). Three hypotheses were tested: 1) TSRTs determined at rest identify the joint angle at which spasticity begins to be manifested; 2) active movement influences TSRTs at which spasticity is manifested, and 3) differences in TSRTs between passive and active movement will be related to clinical movement deficits. 2.0 Materials and Methods 2.1 Subjects Thirteen adults with stroke following a cerebrovascular accident participated (age = 39-71 yrs; 10 males; Table 1) after signing informed consent forms approved by the Ethics Committee of CRIR according to the Declaration of Helsinki. Subjects were included if they had a single unilateral stroke >1 mo previously, >5/16 on the Composite Spasticity Index (CSI) in elbow flexors and/or extensors (Levin and Hui-Chan, 1993) and some residual function in the affected upper limb (>2/7 on the Chedoke-McMaster Arm Stroke Assessment (Gowland et al., 1993). Patients were excluded if they had pain in the upper limb or could not understand simple commands. Patients usually took cholesterol-lowering and anti-diabetic medications, which did not prevent them from participating. Two patients (S6 and S7) were taking spasticity-lowering medications (Baclofen). 2.2 Procedures Spasticity was measured clinically with the CSI, which assesses phasic reflex excitability (0-4), resistance to passive stretch on a doubly-weighted 4-pt scale (0-8) and the presence of wrist clonus (0-4) for a maximal score of 16. In addition, upper limb motor and sensory functions were assessed with the Fugl-Meyer Arm Assessment (FMA, Fugl-Meyer et al., 1975) on scales of 0-66 and 0-12 respectively. To determine SRTs, subjects sat in a comfortable chair with a back support. The semi-supinated forearm of the affected arm was placed in a custom-made rigid cast attached to a horizontal manipulandum that could be rotated about a vertical axis aligned with that of the elbow. The arm was placed in 70° shoulder flexion and abduction. Stretches of elbow flexor and extensor muscles were made either passively by the experimenter by extending or flexing the joint respectively, or actively by the subject through the full joint range in both directions. The order of conditions was randomized across participants. Spasticity thresholds were measured at rest, i.e. when patients were instructed to fully relax arm muscles. Despite the instruction to relax, it has been shown that patients cannot prevent activation of spastic muscles stretched beyond a certain muscle length, even if the stretch velocity is low (Levin et al., 2000). In contrast, healthy subjects do not have stretch responses in elbow muscles at rest unless stretch velocities are quite high (>300 deg/s; Jobin and Levin, 2000). Thus, a phasic response to a tendon tap in a healthy subject is a manifestation of a DSRT evoked at a high velocity. However, the fact that the response can only be evoked at high velocity supports the assumption that the TSRT lies beyond the biomechanical

5

range of the joint. Because of the absence of TSRT responses in this group, they were not included as controls. 2.3 Data recording and analysis An optical encoder on the torque motor (Parker iBE342G) connected to the axis of the manipulandum measured the elbow angle. Bipolar surface EMG was recorded from two elbow flexors (Biceps Brachii, BB; Brachioradialis, BR) and two elbow extensors (Triceps BrachiiMedial head, TM; Triceps Brachii-lateral head, TL) with 1 cm Ag-AgCl disk electrodes placed over motor points (inter-electrode distance, 1 cm). Signal quality was visually checked in realtime on an oscilloscope. EMG signals were amplified (×2000; Noraxon telemetric system, Telemyo 16; USA) and band-pass filtered (20-400 Hz; Butterworth, 4th order). A customized program (LabView, National Instruments, USA) recorded the angular position and EMG signals (sampling rate: 2 kHz). To obtain the EMG envelope (iEMG), EMG data were rectified and integrated over 25-ms windows (trapezoid method) that shifted with each sample. iEMG was normalized by the maximal EMG value across all trials (Turpin et al., 2016). Velocity was obtained by differentiation of the angle trace (5th order polynomial method). Angle and velocity were low-pass filtered (cut-off frequency 5 Hz, 2nd order Butterworth). Mean velocities were computed between passive or active movement onsets and offsets. Movement onset was identified as the time when the velocity increased and remained above 5% of the maximal velocity for at least 200 ms. Offset was defined as the time when the velocity returned to this value for more than 150 ms. Dynamic stretch-reflex thresholds (DSRTs) were defined as the joint angles at which EMG activity started to increase in the stretched muscles for each velocity of stretch (Musampa et al., 2007). To a first approximation, DSRT = TSRT - µ × V

(1)

where V is the angular velocity (positive for elbow extension). TSRT is the tonic, velocityindependent (static) component of the SRT and is obtained by extrapolating a linear regression line through the DSRTs to zero velocity (Levin and Feldman, 1994; Musampa et al., 2007). Parameter µ is a positive time-dimensional (s) scalar representing the sensitivity of the DSRT to velocity, represented by the slope of the regression line (see Figs. 3 and 4). Tonic stretch-reflex threshold (TSRT, i.e. DSRT for zero velocity) and parameter µ were determined for elbow flexors and extensors. In the passive condition, the forearm was flexed or extended respectively, by the examiner through an arc of ~110° from ~50° to ~160° (180°= full extension) for flexors or vice versa for extensors. Muscles were stretched at slow, medium and high velocities (movement durations ~10, 5 and 1-2 s respectively; peak velocities ranging from 10 to 300°/s). For each stretch, the elbow was initially positioned in maximal flexion or extension. Subjects were instructed to fully relax at these positions (verified by monitoring EMG activity on an oscilloscope) and not to intentionally activate their muscle during passive stretching. Stretches were alternated between flexion and

6

extension (five in each direction for each velocity, 2-3 s between stretches, 1-2 min rest between randomized velocity blocks. To determine TSRT and µ during active motion, subjects were instructed to fully extend or flex the elbow at slow, moderate or high velocities (movement durations ~10, 5 and 1-2 s, respectively). For each trial, the elbow was initially placed in maximal flexion or extension, and subjects performed 3-5 flexions or extensions in each trial (6 trials in total; 1-2 min rest between trials). EMG onset was identified when iEMG of the stretched muscle exceeded 3 SDs of the mean baseline value in the 300 ms period prior to stretch. A graphical user interface was used to verify muscle activation onset determined in each trial. DSRTs were identified in each trial in the passively stretched muscles and then the same muscles were stretched as antagonists during active movements made by the subjects (e.g., flexor DSRTs were determined during passive and voluntary elbow extensions). In some subjects, TSRTs could be computed during passive stretches but not during active movement when the stretched antagonist muscle was not activated. In these cases, the TSRT during active motion was considered to be beyond the biomechanical range and the flexor and extensor TSRTs were set to 180° and 30° (the minimal biomechanical angle) respectively. For correlational analyses, TSRT and µ values for each condition (passive, active) were averaged for the two flexors and the two extensors to obtain global measures for each muscle group (Musampa et al., 2007). ∆TSRT and ∆µ were defined as the differences between values obtained during voluntary and passive stretching for each muscle group. 2.4 Statistical analysis For all tests, data normality was verified using Shapiro-Wilk tests. Movement amplitude during active flexion and extension were compared with paired t-tests. Mean peak velocities during active movement were compared using ANOVA with two repeated measures (i.e., movement velocity = slow, medium and fast, and direction = extension and flexion). For Hypothesis 1 and 3, clinical spasticity and motor impairment scores were correlated with TSRT, µ, ∆TSRT and ∆µ using Pearson r statistics. For Hypothesis 2, TSRT and µ values between passive stretches and active motion were compared with paired t-tests, separately for each muscle. Then, correlations between TSRT and µ values in passive and active conditions were assessed using Pearson’s r. Effect sizes were determined with Cohen’s d statistics for paired t-tests and partial eta squared (np²) for ANOVA. Z-scores were computed as (x – m)/σ, where x, m and σ were the subject’s value, the group mean and the group standard deviation respectively. A |Z-score| >1.96 indicates that the subject’s value falls outside of the 95% confidence-interval (CI) of the group. Significance level was set at p<0.05.

7

3.0 Results 3.1 Spasticity thresholds in different conditions Amplitudes and peak stretch velocities were similar for active and passive movements for flexors and extensors. For example, for passive flexor stretches, stretch amplitudes were 106.7±12.4° with mean peak velocities of 22.9±11.7, 114.5±60.2 and 264.6±49.4°/s for slow, medium and high stretch velocities respectively and were not different from those during active stretches (t12=1.60; p=0.136; d=0.44 for amplitudes, F1,12=4.27; p=0.061; np²=0.26 for velocities). Examples of responses in elbow flexors and extensors during passive stretches and active movements stretching flexor muscles are shown in Fig. 2. BR activity emerged due to the stretchreflex at ~110° in both conditions but at a higher velocity during active stretch (i.e., 50°/s vs. 80°/s) indicating a change in the reflex behavior. Typical examples of the determination of TSRT and µ values for passive and active movement are shown in Fig. 3 (for extension movements; S8) and Fig. 4 (for flexion movements; S7). Figure 5 shows TSRT and µ values (group data) during passive and voluntary motion when flexor or extensor muscles were stretched. TSRTs significantly increased during active stretches in BB (i.e., +32.5±17.7°, t11=6.4; p<0.001; d=1.8) and BR (i.e., +32.2±25.8°; t11=4.3; p=0.001; d=1.2). Conversely, TSRTs significantly decreased in TL (-44.9±30.5°; t7=-4.2; p=0.004; d=-1.5) and TM (-49.9±28.4°; t7= -5.0; p=0.002; d=-1.8). Thus, in both cases, muscle activity emerged at greater muscle lengths during active compared to passive stretches. Parameter µ significantly increased in BB from an initial value of 0.197±0.188 s by 0.569±0.530 s (i.e., by a factor 3.9; t10=3.6; p=0.005, d=1.1), from 0.221±0.164 s in BR by 0.490±0.402 s (i.e., factor 3.3; t9=3.8; p=0.004; d=1.2) and from 0.273±0.075 s in TM by 0.430±0.346 s (i.e., factor 1.6; t4=2.8; p=0.049; d=1.24) in voluntary compared to passive motion but not in TL. TSRT and µ were positively correlated only for flexors during passive stretches (r=0.91; p<0.001; n=12). For one of two participants who took Baclofen (S7), Z-scores fell outside of the 95% CIs. For passive stretches, Z-scores were 2.4 (98th percentile) for TSRT and 2.7 (96th percentile) for µ in BB and 2.1 (96th percentile) for TSRT in BR. Values for the other participant taking Baclofen (S6) had stretch-reflex parameters that were within the 95% CIs. 3.2 Relationship between SR characteristics and clinical scores Clinical CSI scores of each muscle group were not related to TSRT or µ parameters in extensor or flexor muscles for passive stretches or active motion alone. However, motor impairment (i.e., FMA) was correlated with flexor TSRTs (r=0.69; p=0.022; n=12) and with extensor µ (r=0.70; p<0.001; n=8) during active motion. In contrast, the ∆TSRT for flexors was negatively correlated with CSI scores in flexors (r=-0.68, p=0.016, n=12) and with clinical motor impairment (i.e., FMA, r=0.81, p=0.004, n=12). Sensory status was not correlated with any stretch-reflex parameter. 4.0 Discussion

8

We evaluated the position and velocity at which activity of elbow muscles emerged when the spastic muscles were stretched passively by the examiner, or actively during voluntary motion. We found that TSRTs could be obtained in spastic muscles at rest in all subjects, thus not rejecting Hypothesis 1. We found that the maximal muscle length at which EMG activity emerged (i.e., the static SR threshold, TSRT) as well as the sensitivity of the SRT to velocity (µ) were greater during active motion compared to passive stretch in both flexors and extensors. These findings suggest that the patients retained the capacity to modulate SR characteristics, thus not rejecting Hypothesis 2 Moreover, active movements reduced the spatial range in which muscles manifested spasticity. 4. 1 Contribution to the understanding of spasticity There was no relationship between TSRT measured at rest and mechanical resistance (gain) to passive stretch measured clinically (e.g, by CSI), suggesting that the TSRT measures a different aspect of spasticity not captured in clinical scales. This also confirmed previous findings (Powers et al., 1988; Katz and Rymer, 1989; Levin et al., 2000; Musampa et al., 2007) of the independence between stretch-reflex threshold and gain. The finding is not entirely surprising since the clinical measure assesses resistance that accumulates in the stretched muscle after it has been activated by stretch above the centrally pre-set SRT. This suggests that TSRT and SR gain provide different information about stretch-reflex characteristics and central inputs in spastic muscles. Two patients (S6 and S7) took Baclofen, an anti-spastic medication that acts primarily on presynaptic afferent terminals (Dario and Tomei, 2004). Compared to the other subjects, only one of them (S7) had a greater TSRT angle (i.e., lower spasticity range) and lower µ (decreased sensitivity of SRT to velocity) values for flexors during passive stretches. During active stretches, the TSRT and µ values for extensors were within the range of the rest of the group (|Z-score| <1.64). The effects of Baclofen on TSRT and µ values are therefore not clear and deserve further investigation. We propose that additional information about the neurological origin of spasticity can be gained by characterizing where, in joint space, increased muscle activity begins. This spatial measure of SR excitability includes two parameters: TSRT and µ. In patients with neurological lesions, the EMG response increases when the muscle is stretched beyond the SR threshold (Powers et al., 1989; Levin and Feldman, 1994). As mentioned in Section 1.0, this threshold can be shifted by various central descending and spinal inhibitory and facilitatory, direct or indirect influences on α-MNs, via spinal interneurons or γ-MNs. These influences can also be mediated by interneurons responsible for reflex intermuscular interactions such as reciprocal Ia inhibition between flexors and extensors, crossed-extensor and cutaneous reflexes (Matthews, 1959; Feldman and Orlovsky, 1972; Faist et al., 1999; Morita et al., 2001). The role of intermuscular interactions in the regulation of TSRT can be accounted for by adding ρ to equation 1 for DSRT (Feldman, 2015): DSRT = TSRT - µ×V + ρ (2) Thus, one can assume that changes in TSRT during active stretches found in the present study result in part from changes in the state of intermuscular interactions in spasticity.

9

Characterizing stretch-reflex behavior in terms of TSRT angle may provide a more reliable evaluation of spasticity (Levin et al., 2000; Mullick et al., 2013) than current clinical scales. With some reservations, this method of spasticity evaluation is similar to the Tardieu scale that determines “catch” angles at which muscle resistance is felt during slow and fast stretches (Ansari et al., 2008), resembling TSRT and DSRT, respectively. However, the Tardieu scale relies on the subjective feeling of muscle resistance without distinguishing between its passive and active components. Our method is more robust since it identifies SR reflex threshold based on electromyographic measurements to determine the angle at which active muscle resistance emerges. The use of EMG data ensures that the extracted SR characteristics, including their central regulation (changes in TSRT and µ) are linked to the underlying neurophysiology rather than to only mechanical factors (e.g., applied force, intrinsic muscle resistance). Moreover, our method can be used during passive and active muscle stretches to determine the ability of patients to modulate TSRTs, which may potentially be a biomarker of motor recovery. 4.2 Changes in spasticity characteristics during active movement Changes in TSRT angle between active and passive stretching of flexors (∆TSRT) were correlated with clinical scores of phasic reflex excitability and resistance to passive stretch (CSI) and upper limb motor impairment (FMA). Thus, Hypotheses 3 was not rejected since compared to the spasticity threshold measured at rest, active movements in which spastic muscles play the role of antagonists can partly decrease the spatial range in which spasticity occurs, and this effect is related to residual motor ability (Morita, 2001). The increase in TSRT during active stretches may involve inhibitory circuits responsible for reflex reciprocal inhibition (Morita, 2001), central inhibition of MNs or other elements of the SR loop (Nielsen et al., 2007). The higher TSRT and µ during active motion could also result from changes in muscle spindle sensitivity (e.g. from a decrease in static and an increase in dynamic fusimotor sensitivity, respectively). A similar pattern (decrease in static and simultaneous increase in dynamic spindle afferent responses to stretch) was observed in decerebrated cats during midbrain stimulation (Ellaway et al., 2015). The relationship between TSRT and µ at rest is also consistent with this pattern. This suggests the involvement of midbrain structures in the regulation of TSRT and µ or in their abnormal values at rest (Burke et al., 1972). Previous studies have not found evidence of alterations in fusimotor activity in spastic muscles in patients compared to healthy controls (Wilson, 1999), but these studies did not test the ability to modulate fusimotor drive in passive compared to active states. Exaggerated reflexes in spastic muscles at rest may be caused by a loss of control of the mechanisms that inhibit reflexes and spontaneous motoneuronal activity (Nielsen et al., 2007). Mechanisms causing the reduction of TSRT at rest are not clear but the present findings suggest that the SRT at rest was independent of the ability to modulate it (i.e., independent of ∆TSRT). In other words, the residual capacity to produce inhibition during active motion might not involve the same mechanisms as those responsible for the exaggerated reflexes at rest. Changes in SRTs and µ in spastic muscles during voluntary motion were not observed in Musampa et al. (2007), likely because of very low velocities of motion (<5°/s). In the present study the velocity of stretch was varied and taken into account to define DSRTs and TSRT. The

10

present procedure allowed us also to define µ (the sensitivity to velocity) in both conditions and likely provided a more accurate estimation of SRTs. We found a significant negative correlation between CSI and the change in TSRT between active and passive conditions, meaning that the greater the clinical impairment, the lower the ability to modulate the SR threshold. This result provides insight and support to the concept that deficits in threshold regulation may be one of the mechanisms underlying disordered motor control after stroke. An interesting implication of our results that can be further tested is that repetitions of active movements in which spastic muscles play the role of antagonists may result in a sustained decrease of spasticity at rest. 4.3 Limitations Repetition of stretches may result in a decrease in stretch-reflex amplitude due to i) mechanical alterations in muscle structure (Hagbarth et al., 1985) and ii) stretch-reflex habituation (Rothwell et al., 1986; Turpin et al., 2016). Mechanical influences may have had an effect on the size of responses (Hagbarth et al., 1985). However, they cannot account for the changes between passive and active movement as the order of conditions was randomized. Habituation has been attributed to supra-spinal rather than spinal mechanisms (Rothwell et al., 1986) and is likely a learned phenomenon associated with predictable stretch perturbations (Turpin et al., 2016). During passive stretching, velocities occurred randomly which limited their predictability, making it unlikely that reflex habituation played a major role in the present results. It is possible that, contrary to the instructions given to the subjects, muscles were activated voluntarily during the passive stretches. If this were the case, a relatively constant reaction time would have been expected, such that the greater the velocity, the greater the muscle length at which EMG would occur during passive stretches. However the opposite was observed (Fig. 3 and 4), suggesting that effects of voluntary interventions were minimal. 5.0 Conclusion Subjects with post-stroke spasticity retained some ability to modulate stretch-reflex behavior in elbow muscles and this ability was correlated with the level of upper limb motor impairment. The study also suggests that i) clinical spasticity, although evaluated at rest, may provide some information about the ability of subjects to regulate the SR threshold during active motion and ii) an increase in TSRTs during active motion may result from the involvement of inhibitory mechanisms that are suppressed at rest. Finally, the study highlights the notion of the threshold position control theory that TSRT regulation, spasticity and movement disorders are coupled in a unified framework. Acknowledgements Thanks to Aditi Mullick and Rhona Guberek. Supported by National Science and Engineering Research Council and Heart and Stroke Foundation, Canada. MFL holds a Canada Research Chair in Motor Recovery and Rehabilitation. NAT was supported by Mentor (REPAR/FRSQ). Disclosures None.

11

References Ansari NN, Naghdi S, Hasson S, Azarsa MH, Azarnia S. The Modified Tardieu Scale for the measurement of elbow flexor spasticity in adult patients with hemiplegia. Brain Inj 2008;22:1007-12. Asatryan DG, Feldman AG. Functional tuning of the nervous system with control of movement or maintenance of a steady posture: I. Mechanographic analysis of the work of the joint or execution of a postural task. Biophys 1965;10:925-34. Burke D, Knowles L, Andrews C, Ashby P. Spasticity, decerebrate rigidity and the clasp-knife phenomenon: an experimental study in the cat. Brain 1972;95:31-48. Burne JA. The spasticity paradox: movement disorder or disorder of resting limbs? J Neurol Neurosurg Psychiatry 2005;76:47-54. Capaday C. The effects of baclofen on the stretch reflex parameters of the cat. Exp Brain Res 1995;104:287-96. Dario A, Tomei G. A benefit-risk assessment of baclofen in severe spinal spasticity. Drug Safety 2004;27:799–818. Dietz V, Sinkjaer T. Spastic movement disorder: impaired reflex function and altered muscle mechanics. Lancet Neurol 2007;6:725-33. Ellaway PH, Taylor A, Durbaba R. Muscle spindle and fusimotor activity in locomotion. J Anat 2015;227:157-66. Faist M. Impaired modulation of quadriceps tendon jerk reflex during spastic gait: differences between spinal and cerebral lesions. Brain 1999;122:567-79. Feldman AG. Referent control of action and perception: Challenging conventional theories in behavioral neuroscience. New York: Springer; 2015. Feldman AG, Orlovsky GN. The influence of different descending systems on the tonic stretch reflex in the cat. Exp Neurol 1972;37:481-94. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. a method for evaluation of physical performance. Scand J Rehabil Med 1975;7:13-31. Gemperline JJ, Allen S, Walk D, Rymer WZ. Characteristics of motor unit discharge in subjects with hemiparesis. Muscle Nerve 1995;18:1101–14. Gowland C, Stratford P, Ward M, Moreland J, Torresin W, Van Hullenaar S, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke 1993;24:58-63. Hagbarth KE, Hägglund JV, Nordin M, Wallin EU. Thixotropic behaviour of human finger flexor muscles with accompanying changes in spindle and reflex responses to stretch. J Physiol 1985;368:323-42. Jobin A, Levin MF. Regulation of stretch reflex threshold in elbow flexors in children with cerebral palsy: a new measure of spasticity. Dev Med Child Neurol 2000;42:531-50. Jones CA, Yang JF. Reflex behavior during walking in incomplete spinal-cord-injured subjects. Exp Neurol 1994;128:239-48. Katz RT, Rymer WZ. Spastic hypertonia: mechanisms and measurement. Arch Phys Med Rehabil 1989;70:144-55.

12

Levin MF, Feldman AG. The role of stretch reflex threshold regulation in normal and impaired motor control. Brain Res1994;657:23-30. Levin MF, Hui-Chan C. Are H and stretch reflexes in hemiparesis reproducible and correlated with spasticity? J Neurol 1993;240:63-71. Levin MF, Selles RW, Verheul MHG, Meijer OG. Deficits in the coordination of agonist and antagonist muscles in stroke patients: implications for normal motor control. Brain Res 2000;853:352-69. Lewek MD, Hornby TG, Dhaher YY, Schmit BD. Prolonged quadriceps activity following imposed hip extension: a neurophysiological mechanism for stiff-knee gait? J Neurophysiol 2007;98: 3153–62. Manella KJ, Roach KE, Field-Fote EC. Operant conditioning to increase ankle control or decrease reflex excitability improves reflex modulation and walking function in chronic spinal cord injury. J Neurophysiol 2013;109:2666-79. Matthews PBC. A study of certain factors influencing the stretch reflex of the decerebrate cat. J Physiol 1959;147:547-64. Mirbagheri MM, Barbeau H, Ladouceur M, Kearney RE. Intrinsic and reflex stiffness in normal and spastic, spinal cord injured subjects. Exp Brain Res 2001;141:446-59. Morita H. Modulation of presynaptic inhibition and disynaptic reciprocal Ia inhibition during voluntary movement in spasticity. Brain 2001;124:826-37. Mottram CJ, Wallace CL, Chikando CN, Rymer WZ. Origins of spontaneous firing of motor units in the spastic-paretic biceps brachii muscle of stroke survivors. J Neurophysiol 2010;104:3168–79. Mullick AA, Musampa NK, Feldman AG, Levin MF. Stretch reflex spatial threshold measure discriminates between spasticity and rigidity. Clin Neurophysiol 2013;124:740-51. Musampa NK, Mathieu PA, Levin MF. Relationship between stretch reflex thresholds and voluntary arm muscle activation in patients with spasticity. Exp Brain Res 2007;181:579-93. Nielsen JB, Crone C, Hultborn H. The spinal pathophysiology of spasticity from a basic science point of view. Acta Physiol 2007;189:171-80. O'Dwyer NJ, Ada L, Neilson PD. Spasticity and muscle contracture following stroke. Brain 1996;119:1737-49. Pandyan AD, Gregoric M, Barnes MP, Wood D, Wijck FV, Burridge J, et al. Spasticity: Clinical perceptions, neurological realities and meaningful measurement. Disabil Rehabil 2005;27:2-6. Powers RK, Campbell DL, Rymer WZ. Stretch reflex dynamics in spastic elbow flexor muscles. Ann Neurol 1989;25:32-42. Powers RK, Marder-Meyer J, Rymer WZ. Quantitative relations between hypertonia and stretch reflex threshold in spastic hemiparesis. Ann Neurol 1988;23:115-24. Raptis H, Burtet L, Forget R, Feldman AG. Control of wrist position and muscle relaxation by shifting spatial frames of reference for motoneuronal recruitment: possible involvement of corticospinal pathways. J Physiol 2010;588:1551-70. Rothwell JC, Day BL, Berardelli A, Marsden CD. Habituation and conditioning of the human long latency stretch reflex. Exp Brain Res 1986;63:197-204.

13

Sinkjær T, Andersen JB, Nielsen JF, Nielsen JF. Impaired stretch reflex and joint torque modulation during spastic gait in multiple sclerosis patients. J Neurol 1996;243:566-74. Turpin NA, Levin MF, Feldman AG. Implicit learning and generalization of stretch response modulation in humans. J Neurophysiol 2016;115:3186-94. Wilson LR. Muscle spindle activity in the affected upper limb after a unilateral stroke. Brain 1999;122:2079-88. Wissel J, Manack A, Brainin M. Toward an epidemiology of poststroke spasticity. Neurol 2013;80:S13-S9. Young JL, Mayer RF. Physiological alterations of motor units in hemiplegia. J Neurological Sci 1982;54:401–12.

14

Figure legends Figure 1. Threshold position control (schematic). A) Activity and torque is generated when the muscle length exceeds the velocity-dependent threshold muscle length also called the dynamic spatial stretch-reflex threshold (DSRT). Static (tonic) muscle activity and torque increase with the increasing difference between the actual and the tonic stretch reflex threshold (TSRT). Panel A shows that, in healthy individuals, the range of central TSRT regulation exceeds the biomechanical limits of the joint. B) Left: To prevent flexor activation during passive stretching (full muscle relaxation during elbow extension), healthy subjects shift TSRT to the right, beyond the upper biomechanical joint limit. Right: In contrast, due to a deficit in central inhibitory mechanisms, TSRT is not shifted far enough and occurs abnormally within the biomechanical range, resulting in spasticity beyond the TSRT angle (spasticity zone, shaded area), i.e., patients cannot relax (de-activate) spastic muscles stretched beyond certain muscle lengths, even at low velocities. Left: Active stretching of flexors is produced when subjects intentionally extend the joint. This is achieved by facilitating extensor and de-facilitating/inhibiting flexor motoneurons, shifting flexors and extensors TSRTs to the right. Right: Extensor muscles are activated to overcome spastic flexor resistance during active extension into the spasticity zone. Figure 2. Measuring dynamic stretch-reflex threshold in elbow flexors. A). Passive stretching of spastic Brachioradialis (BR) produced by the experimenter. The upper panel shows elbow angle (solid line) and velocity (dashed line). B) Active stretching of BR produced by the subject activating elbow extensors, including Triceps Brachii-Lateralis (TL). Note flexor and extensor muscle coactivation in B. Lower traces show integrated EMG (iEMG) normalized to maximal EMG level of the respective muscle in all trials. Full elbow extension is 180°. Figure 3. Stretch-reflex characteristics of spastic flexor muscles in a representative subject. A) Each point represents the joint angle and velocity at which muscle activity emerged when muscle was passively stretched. The tonic stretch-reflex threshold (TSRT) and sensitivity of the dynamic stretch thresholds (DSRT) to velocity (µ) were determined from linear orthogonal regressions (dashed lines) using equation 1 in text. B) Similar procedures were used to determine TSRT and µ during active stretching of spastic elbow flexors, BB (Biceps Brachii), BR (Brachoradialis). Figure 4. Stretch-reflex characteristics of spastic extensor muscles in a different subject. Stretching of extensors produced by flexing the elbow joint either passively in A or actively in B. TL: Triceps Brachii-Lateralis; TM: Triceps Brachii-Medialis. Figure 5. Mean±SD values of TSRT and µ for each spastic muscle (group data). Asterisks (*) indicate significant differences (p<0.05) between values observed during passive and active stretches. A) For both flexor (BB, BR) and extensor muscles (TL, TM), the muscle length at which EMG emerged during active stretches was greater than during passive stretches, resulting in a greater threshold angle for flexors and lower joint angles for extensors (180°=full elbow extension). B) Except for TL, the sensitivity to the velocity of stretch (µ) was higher during active than passive stretches.

15

Table 1. Study participants. Fugl-Meyer

CSI

Subject

Age (yr)

Time Since Stroke (mo)

Sex/ Dominance/ Hemiparetic side

1 2 3

46 71 54

1 21 96

M/R/R M/R/L F/Ambi/R

MCA Fronto-parietal, Pons MCA, BG

I I I

5 4 4

55 54 51

12 3 7

5 6 7

7 6 5

4 5 6

56 39 65

52 10 3

M/R/L M/R/L M/R/L

Pons MCA, IC Brainstem

I I I

3 3 4

47 46 44

12 11 12

8 7 11

5 7 8

7 8 9 10 11 12 13

50 68 58 46 53 59 64

72 252 17 104 13 110 60

M/R/R M/R/R M/R/R M/R/L M/R/R M/R/R M/R/R

BG Frontal BG Temporal, Lentiform, IC, BG Temporal-parietal, Precentral gyrus Occipital, IC Fronto-parietal, Putamen, CR, IC

H H I I I I I

3 3 3 3 3 3 2

36 32 29 26 25 16 15

3 5 12 3 3 12 12

10 12 13 12 11 8 8

11 9 11 9 10 8 10

Location

Type

CMM (/7)

Motor Function (/66)

Sensory Function (/12)

Flexors (/16)

Extensors (/16)

Ambi: Ambidextrous; BG: Basal Ganglia; CR: Corona Radiata; CSI: Composite Spasticity Score; F: Female; H: Hemorrhagic; I: Ischemic; IC: Internal Capsule; L: Left; M: Male; MCA: Middle Cerebral Artery; R: Right.

16

Active flexor torque

A

No spasticity

0

Spasticity zone

Threshold shift

0

0

Threshold range Biomechanical range

180o

40

Threshold range Biomechanical range

180o

40

Active muscle stretch

Threshold range Biomechanical range

40

Spasticity

B

SR threshold

C

180o Active torque

Active flexor torque

Passive muscle stretch

Flexors Coactivation

Threshold shift

0

0

Threshold range

40

Extensors

180o

Threshold range

40

Joint angle

Figure 1

180o

17

B Active stretch

150 100 50

100 50 0 -50 -100

velocity angle 0.5 s

BR

10 mV

TL 5 mV

Figure 2

velocity (°/s)

angle (°)

A Passive stretch

18

A Passive stretch 160 140

BB

velocity (°/s)

120 100

160 140

BR

120 100

80

80

60

60 TSRT = 114.7° μ = 0.271

40 20

TSRT = 93.1° μ = 0.247

40 20 0

0 0

20 40 60 80 100 120 140 160 180

0

20 40 60 80 100 120 140 160 180

joint angle (°)

B Active stretch

100 90 80 70 60 50 40 30 20 10 0

100 90 80 70 60 50 40 TSRT = 151.7° 30 μ = 0.989 20 10 0 20 40 60 80 100 120 140 160 180 0

BB

0

BR

TSRT = 147.7° μ = 1.005

20 40 60 80 100 120 140 160 180

joint angle (°)

Figure 3

19

A Passive stretch 0

joint angle (°)

20 40 60 80 100 120 140 160 180

0

-20

TL

-40

-40

-60

-60

-80

-120

-140

-140

B Active stretch

joint angle (°)

20 40 60 80 100 120 140 160 180

TL

TSRT = 100.2° μ = 0.224

-100

-120

0

TM

-80

TSRT = 112.7° μ = 0.147

-100

0 -20 -40 -60 -80 -100 -120 -140 -160 -180

20 40 60 80 100 120 140 160 180

0

-20

velocity (°/s)

0

TSRT = 42.2° μ = 0.384

0 -20 -40 -60 -80 -100 -120 -140 -160 -180

0

20 40 60 80 100 120 140 160 180

TM

TSRT = 39.9° μ = 0.412

Figure 4

20

A 180

BB

120

*

BR

*

B

Passive Active stretch

TL

Passive Active stretch

TM

0

*

0 Passive Active stretch

1.5

TL

Passive Active stretch

TM

*

1

120 60

BR

*

0.5

µ (s)

TSRT (°)

180

BB

1

60 0

1.5

*

*

0.5 0

Passive Active stretch

Passive Active stretch

Passive Active stretch

Passive Active stretch

Figure 5