Altered muscle activation patterns (AMAP): an analytical tool to compare muscle activity patterns of hemiparetic gait with a normative profile

Abstract

Background: Stroke survivors often have lower extremity sensorimotor impairments, resulting in an inability to sufficiently recruit muscle activity at appropriate times in a gait cycle. Currently there is a lack of a standardized method that allows comparison of muscle activation in hemiparetic gait post-stroke to a normative profile.

Methods: We developed a new tool to quantify altered muscle activation patterns (AMAP). AMAP accounts for spatiotemporal asymmetries in stroke gait by evaluating the deviations of muscle activation specific to each gait sub-phase. It also recognizes the characteristic variability within the healthy population. The inter-individual variability of normal electromyography (EMG) patterns within some sub-phases of the gait cycle is larger compared to others, therefore AMAP penalizes more for deviations in a gait sub-phase with a constant profile (absolute active or inactive) vs variable profile. EMG data were collected during treadmill walking, from eight leg muscles of 34 stroke survivors at self-selected speeds and 20 healthy controls at four different speeds. Stroke survivors' AMAP scores, for timing and amplitude variations, were computed in comparison to healthy controls walking at speeds matched to the stroke survivors' self-selected speeds.

Results: Altered EMG patterns in the stroke population quantified using AMAP agree with the previously reported EMG alterations in stroke gait that were identified using qualitative methods. We defined scores ranging between ±2.57 as "normal". Only 9% of healthy controls were outside "normal" window for timing and amplitude. Percentages of stroke subjects outside the "normal" window for each muscle were, Soleus = 79%; 73%, Medial Gastrocnemius = 62%; 79%, Tibialis Anterior = 62%; 59%, and Gluteus Medius = 48%; 51% for amplitude and timing component respectively, alterations were relatively smaller for the other four muscles. Paretic-propulsion was negatively correlated to AMAP scores for the timing component of Soleus. Stroke survivors' self-selected walking speed was negatively correlated with AMAP scores for amplitude and timing of Soleus but only amplitude of Medial gastrocnemius (p < 0.05).

Conclusions: Our results validate the ability of AMAP to identify alterations in the EMG patterns within the stroke population and its potential to be used to identify the gait phases that may require more attention when developing an optimal gait training paradigm.

Trial registration: ClinicalTrials.gov NCT00712179 , Registered July 3rd 2008.

Keywords: EMG assessment; Electromyography; Locomotion; Muscle activation patterns; Stroke.

Conflict of interest statement

Ethics approval and consent to participate

Written informed consent approved by the institutional review board at the University of Florida was obtained from all the subjects for publication of this study.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

Representative data from stroke survivors’ muscle activity patterns demonstrating the sensitivity of k-means cluster analysis to identify “on/off” periods. Each row presents EMG signals for a) soleus, b) tibialis anterior, and c) medial gastrocnemius from stroke survivor’ single gait cycle. Left panels present original EMG data and the right panels present the same representative data after it were high pass filtered (20 Hz) with a zero lag fourth-order Butterworth filter, demeaned, rectified, and smoothed with a zero lag fourth-order low-pass (25 Hz) Butterworth filter. The red solid lines represent k-means clusters

Fig. 2

Left panel demonstrates the stroke survivors’ EMG activity amplitude during self-selected walking; the box plots indicate the range in the data, horizontal black line in center is the median, the upper and lower boundaries of the box indicating the upper and lower quartile respectively, and red markers represent extreme values. Right panel demonstrates AMAP scores of stroke survivors for the amplitude component of each muscle. The shaded gray area is the normal range of scores (±2.57). Each dot within a region of gait cycle represents score of a stroke survivor with solid red dots representing the subjects with scores outside the “normal” window of ±2.57. Positive scores, i.e. solid red dots above the “normal” window, represent EMG activation greater than “normal” and vice versa. For the clarity of data, we have adjusted the Y-axes scales on the right panel between ±5 and ± 11. The EMG data from each muscle were normalized to the averaged peak activation across all the steps taken by the subject

Fig. 3

Left panel demonstrates the stroke survivors’ EMG activity timing during self-selected walking; the box plots indicate the range in the data, horizontal black line in center is the median, the upper and lower boundaries of the box indicating the upper and lower quartile respectively, and red marker represent extreme values. Right panel demonstrates AMAP scores of stroke survivors for the timing component of each muscle. The shaded gray area is the normal range of scores (±2.57). Each dot within a region of gait cycle represents score of a stroke survivor with solid red dots representing the subjects with scores outside the “normal” window of ±2.57. Positive scores represent duration of EMG activation longer than “normal” and vice versa. For the clarity of data, we have adjusted the Y-axes scales on the right panel between ±5 and ± 11. The EMG data from each muscle were normalized to the averaged peak activation across all the steps taken by the subject

Fig. 4

Examples of altered EMG patterns in comparison to normal pattern are illustrated for EMG activity at self-selected walking speeds for all muscles. Black solid lines are the mean EMG activity of healthy controls and shaded grey area is the SD. Red solid lines are EMG activity of a representative stroke survivor with Table 3 representing the corresponding AMAP scores for the amplitude and timing components (values highlighted in red in the table represent scores outside the “normal” window of ±2.57). To clearly present the EMG activity, time is presented as percent of gait cycle

Fig. 5

Correlations between Pp and the total AMAP scores for SO averaged across all regions of the timing component. The total AMAP scores were negatively associated (p < 0.05) with Pp. The red dots represent stroke subjects that had AMAP scores outside of the “normal” window of ±2.57

Fig. 6

Correlations between self-selected walking speeds and total AMAP scores for the amplitude components (scores averaged across all regions of the amplitude component) of SO (left panel) and MG (right panel). The red dots represent stroke subjects that had AMAP scores outside of the “normal” window of ±2.57. The total AMAP scores of SO and MG for amplitude were negatively associated with the walking ability of stroke survivors (p