Music-based intervention drives paretic limb acceleration into intentional movement frequencies in chronic stroke rehabilitation

Abstract

This study presented a novel kinematic assessment of paretic limb function "online" during the actual therapeutic exercisers rooted within the acceleration domain. Twenty-eight patients at chronic stroke stages participated in an auditory-motor intervention mapping reaching movements of the paretic arm unto surfaces of large digital musical instruments and sound tablets that provided rhythmic entrainment cues and augmented auditory feedback. Patients also wore a tri-axial accelerometer on the paretic limb during the nine-session intervention. The resulting acceleration profiles were extracted and quantified within the frequency domain. Measures of peak power and peak width were leveraged to estimate volitional control and temporal consistency of paretic limb movements, respectively. Clinical assessments included the Wolf Motor Function Test and Fugl-Meyer - Upper Extremity subtest. The results showed that peak power increased significantly from Session 1 to Session 9 within oscillatory frequency ranges associated with intentional movement execution (i.e., 4.5 Hz). Decreases in peak width over time provided additional evidence for improved paretic arm control from a temporal perspective. In addition, Peak width values obtained in Session 1 was significantly correlated with pre-test Fugl-Meyer - Upper Extremity scores. These results highlighted improvements in paretic limb acceleration as an underlying mechanism in stroke motor recovery and shed further light on the utility of accelerometry-based measures of paretic limb control in stroke rehabilitation. The data reported here was obtained from a larger clinical trial: https://ichgcp.net/clinical-trials-registry/NCT03246217 ClinicalTrials.gov Identifier: NCT03246217.

Keywords: limb acceleration; motor control; rehabilitation; sonification; stroke.

Conflict of interest statement

Author NS was employed by company BeSB GmbH Berlin Sound Engineering. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

© 2022 Loria, de Grosbois, Haire, Vuong, Schaffert, Tremblay and Thaut.

Figures

Figure 1

Summary of the acceleration analysis. On the left side are the raw acceleration profiles for Session 1 (magenta) and 9 (teal) along the resultant movement axis. On the right side is the resulting power spectral density function showing peak power and peak width for the same participant.

Figure 2

The findings for peak power. Peak power was greater in Session 9 compared to Session 1, which supports the effectiveness of the intervention. Note: Error bars denote standard error of the mean, * = significance at <0.05.

Figure 3

The results for peak width. Although not significant, peak width numerically decreased from Sessions 1 to 9, localized to frequency ranges commonly associated with voluntary motor control.

Figure 4

The results for relationship between measures. The relationship between WMFT and FM-UE scores (i.e., Panel A) as well as between Δ peak power and Δ peak width (i.e., Panel B) are shown. Δ peak power and Δ FM-UE and Δ WMFT are shown in panels C and D, respectively. Lastly, Panels E and F show the relationship between Δ peak width and Δ FM-UE and Δ WMFT.