Absence of a TMS-induced lower limb corticomotor response does not affect walking speed in chronic stroke survivors

Abstract

Background and purpose

Transcranial magnetic stimulation (TMS) is used to measure the functional integrity of the corticomotor system via motor evoked potentials (MEPs) in stroke. The association between corticomotor mechanisms and walking recovery is still not completely understood. This study determined the association between TMS-induced MEPs and walking outcomes, and examined the contribution of the contralesional hemisphere to walking recovery.

Methods

Contralateral and ipsilateral TMS responses from the contralesional and ipsilesional hemispheres were collected from 61 chronic stroke survivors. Clinical assessments included gait speeds, six-minute walk distance, timed up and go test, lower limb Fugl Meyer (FMLE) scale and strength measurements.

Results

Stroke participants were classified based on the presence (MEP+ (n = 28)) or absence (MEP- (n = 33)) of MEPs in the paretic tibialis anterior (TA) and rectus femoris (RF) muscles. A between-group analyses showed no significant differences for any gait variable. MEP+ group showed significantly higher FMLE and ankle dorsiflexor strength. Ipsilateral conductivity was not significantly different between groups. Finally, in the MEP+ group, MEP parameters did not predict gait recovery.

Conclusion

Our study investigated the association between walking outcomes and neurophysiological parameters of lower limb function in a large cohort of stroke survivors. We did not find an associations between TMS-induced TA and RF MEPs and walking speeds. Further work is required to develop more comprehensive models in stroke for predicting walking recovery.

Introduction

Regaining the ability to walk independently is an important functional goal for stroke survivors. Gait speed is an important determinant of walking recovery, and descending corticomotor control is a significant contributor to gait recovery post stroke.¹ Several studies have shown that the presence or absence of a transcranial magnetic stimulation (TMS)-induced motor evoked potential (MEP) is related to upper limb (UL) functional recovery in acute and chronic stroke.² For the lower limb (LL), few studies suggest that absent MEP responses may be associated with greater walking difficulty.^{3, 4} However, the relationship of the MEP to gait speed and other measures of LL function still needs to be elucidated. In addition, there remains a large gap in our understanding of the adaptive or maladaptive nature of the contralesional hemisphere and its contribution to walking recovery. Few LL stroke studies have shown that greater ipsilateral drive from the contralesional hemisphere is associated with greater LL impairment and reduced performance in a skilled motor task.^5,6 Whether this increased ipsilateral drive also affects walking speed is still unknown.

Identification of MEP as a neurophysiological biomarker for walking recovery has the potential to effectively tailor neuromodulation related treatments and other therapies. In this study, our primary aim was to determine if LL functional corticospinal tract (CST) integrity, determined by the presence or absence of TA and RF MEPs, was associated with gait speeds in chronic stroke. We also examined the relationship between ipsilateral connectivity from the contralesional M1 to the paretic LL muscles, and its association to walking recovery.

Methods

De-identified data that support the findings of this study will be available on reasonable request from the corresponding author (SM) after the completion of the ongoing RCT. Please see online-only Data Supplement for detailed methodology. Briefly, subjects with a first-ever mono-hemispheric stroke > 6 months since onset, residual hemiparetic gait deficits (abnormal gait pattern or 10-meter walk time exceeding age-related time by 2s), and ability to walk without an ankle orthotic for 5 minutes at self-paced speed were included in this study.⁷ Subjects with contraindications to TMS, brainstem or cerebellar lesions, presence of cognitive and cardiorespiratory impairments were excluded. A written informed consent was obtained from everyone, and the study was approved by the Institutional Review Board.

A physical therapist assessed gait speed using the 10-meter walk test (2 trials each of self-selected and fast speed), endurance with the 6-minute walk test (6MWT), dynamic balance using the timed-up and go test (TUG), LL impairment with the Fugl Meyer (FMLE) scale and muscle strength using maximum voluntary contractions (MVC).

For TMS, a double cone coil with a posterior-anterior current orientation was used to determine corticomotor excitability (CME) for the tibialis anterior (TA) and rectus femoris (RF) muscles. For contralateral responses the coil was placed over the hemisphere contralateral to the muscle, and for ipsilateral responses the coil was placed ipsilateral to the muscle (Figure 1). TMS-induced responses were collected from the paretic TA (PTA) and RF (PRF) and the non-paretic TA (NPTA) and RF (NPRF). MEP area were considered as the primary outcome of CME. To determine the relative magnitude of ipsilateral contributions, we calculated a physiological index of CME (ICE) for the PTA and NPTA.⁶

Fig. 1.

Schematic showing an example of a contralateral MEP (green) and an ipsilateral MEP (blue). Ideally, when the TMS coil is positioned contralateral to the target muscle, the MEP is larger (green) and when placed ipsilateral to the target muscle, the MEP (blue) is smaller. Ipsilateral conductivity is assumed when the ipsilateral MEP slope is higher than the contralateral MEP slope, suggestive of a lower index of corticomotor excitability (ICE).

We classified our participants into MEP+ (present) and MEP– (absent) groups based on the presence of MEPs from the ipsilesional hemisphere for both the contralateral PTA and PRF. Between group comparisons were performed for all clinical and neurophysiological parameters. For the MEP+ group, multiple regression models were used to investigate relationships between MEP parameters, MVC (PTA and PRF), FMLE, age, time since stroke and gait speeds.

Results

Data from 61 participants (age range 41-76 years) were analyzed (Table I in the supplement). MEPs were elicitable from the paretic TA and RF in 28 participants. There were no significant differences between the MEP+ and MEP- groups for self-selected and fast walking speeds (Figure I in the supplement), 6 MWT, or TUG. The FMLE and PTA MVC values were significantly higher in the MEP+ group (Table 1). The MEP- group showed significantly higher active motor threshold and lower contralateral recruitment curve slopes for the NPTA and NPRF compared to the MEP+ group. No significant differences were noted for the ICE values between the MEP+ and MEP– groups (Table 2). For the MEP+ group, the multiple regression models were unable to significantly predict self-selected or fast gait speeds. (Table II in the supplement).

Table 1.

Comparison of clinical parameters in both groups

	MEP+ (n = 28)	MEP– (n = 33)	95% CI Lower Upper		P value
Gait Speed (m/s)
Self-selected	0.74 (0.2)	0.74 (0.2)	−0.1	0.11	0.97
Fast velocity	0.99 (0.3)	0.97 (0.3)	−0.1	0.14	0.87
TUG (seconds)a	15.8 (6)	15.01 (5)			0.65
6MWD (m)	278.66 (94.08)	285.4 (86.1)	−38	54.5	0.72
FMLE (paretic)	24.14 (3.34)	18.7 (3.8)	−7.2	−3.4	< 0.001
MVC
NPTA	0.19 (0.09)	0.15 (0.07)	−0.08	0.002	0.06
PTAa	0.09 (0.05)	0.05 (0.03)			0.004
NPRF	0.08 (0.06)	0.08 (0.05)	−0.03	0.02	0.7
PRF	0.04 (0.02)	0.03 (0.01)	−0.08	0.002	0.18

Values are means (standard deviations).

^aResults for the TUG and PTA variables are from Mann Whitney U tests and 95% CI are not reported for these variables. Abbreviations: TUG, Timed Up and Go test; 6MWD, 6 Minute Walk Distance; FMLE, Fugl Meyer Lower Extremity scale; MVC; Maximum Voluntary Contraction; NPTA, Non-Paretic TA; PTA, Paretic TA; NPRF, Non-Paretic RF; PRF, Paretic RF. Bold type P value indicates statistical significance

Table 2.

Comparison of neurophysiological parameters in both groups

	MEP+ (n = 28)	MEP– (n = 33)	P value
Active motor threshold (% MSO)
NPTAcontra	41.67 (9.3)	50.39 (8.7)	<0.001
NPRFcontra	45.1 (10.1)	52.73 (9.7)	0.007
PTAcontra	52.1 (10.2)	N/A
PRFcontra	56.14 (12.2)	N/A
RC slope
NPTAcontra	0.072 (0.004)	0.041(0.003)	0.005
NPRFcontra	0.076 (0.005)	0.069 (0.009)	0.04
PTAcontra	0.043 (0.003)	N/A
PRFcontra	0.038 (0.004)	N/A
Index of corticomotor excitability (ICE)
NPTA	0.23	0.24	0.68
PTA	0.02	−0.07	0.3

Values are means (standard deviations). Abbreviations: RC, Recruitment Curve, NPTA_contra, Non-paretic tibialis anterior; NPRF_contra, Non-paretic rectus femoris; PTA_contra, Paretic tibialis anterior; PRF_contra, Paretic rectus femoris. N/A, not applicable. Bold type P value indicates statistical significance.

Discussion

Our results showed that the presence or absence of a LL MEP from the ipsilesional hemisphere does not affect gait outcomes in chronic stroke. These findings may imply that functional CST integrity may not be a useful biomarker for explaining walking recovery in chronic stroke survivors. Even though TMS has been shown to reliably predict UL motor recovery in acute and subacute stroke,^{8, 9} its role in the explanation of walking recovery in chronic stroke may be more complex. Our findings concur with studies that did not show an association between LL MEPs and independent ambulation.^10-12 Cho et al. (2013) reported that chronic stroke survivors without TA MEPs and reduced CST tract integrity could walk independently,¹⁰ and Smith et al. (2017) reported that even in the acute-subacute stages, presence of a MEP was not predictive of independent ambulation.¹¹ Our finding that participants with MEPs demonstrate lesser motor impairment (higher FMLE scores) is in line with other studies that prospectively evaluated MEPs from the acute to chronic stages and showed that the presence of a MEP was associated with better clinical recovery.^{3, 4, 13, 14}

A plausible explanation for the absence of differences in gait speeds in stroke survivors with and without MEPs could be the possible recruitment of redundant pathways, such as the reticulospinal tract, in those without MEPs,¹⁵ or it could be a reflection of motor compensation such as increased swing amplitudes on the non-paretic side.

We did not find any differences in ipsilateral activity from the contralesional hemisphere between groups. These results suggest that in chronic stroke survivors, the contralesional hemisphere may not be upregulated in those with reduced ipsilesional drive or it is possible that TMS may not be very sensitive to capture ipsilateral activity in the LL M1. Interestingly, the contralesional hemisphere in the MEP- group showed reduced CME. This may be clinically relevant as these individuals may benefit from facilitatory bi-hemispheric non-invasive brain stimulation, compared to suppression of the contralesional hemisphere which is standard for neuromodulation for UL recovery.

Our study is limited by the lack of gait kinematic and kinetic measures and quantification of MEPs for other LL muscles such as plantar flexors which may provide further explanation of our results. The proximity of the LL motor cortices may have accounted for inadvertent stimulation of both hemispheres during TMS, thus confounding some of our TMS measures. Finally, our participants were community ambulators who walked with higher speeds thus limiting the generalizability of our findings.

Conclusions

The results of this study suggest that the absence of a TMS-induced MEP of the TA and RF does not affect gait speed in chronic stroke survivors. Our study is the first to investigate the association between different gait outcomes and neurophysiological parameters for both the TA and RF muscles, and quantify ipsilateral connectivity to the paretic TA in a large cohort of stroke survivors. Future research with a larger, heterogeneous sample and comprehensive predictive models is warranted to identify the factors influencing gait recovery.