Gait event prediction from surface electromyography in parkinsonian patients

Abstract

Gait disturbances are common manifestations of Parkinson's disease (PD), with unmet therapeutic needs. Inertial measurement units (IMU) are capable of monitoring gait, but they lack neu-rophysiological information that may be crucial for studying gait disturbances in these patients. Here, we present a machine-learning approach to approximate IMU angular velocity profiles, and subsequently gait events from electromyographic (EMG) channels. We recorded six parkinsonian patients while walking for at least three minutes. Patient-agnostic regression models were trained on temporally-embedded EMG time series of different combinations of up to five leg muscles bi-laterally (i.e., tibialis anterior, soleus, gastrocnemius medialis, gastrocnemius lateralis, and vastus lateralis). Gait events could be detected with high temporal precision (median displace-ment <50 msec), low numbers of missed events (<2%), and next to no false positive event detec-tions (<0.1%). Swing and stance phases could thus be determined with high fidelity (median F1 score ~0.9). Interestingly, the best performance was obtained using as few as two EMG probes placed on the left and right vastus lateralis. Our results demonstrate the practical utility of the proposed EMG-based system for gait event prediction while allowing the simultaneous acquisition of an electromyographic signal. This gait analysis approach has the potential to make additional measurement devices such as IMU and force plates less essential, and thereby to reduce financial and preparation overheads and discomfort factors in gait studies.

Cortical networks of parkinsonian gait: a metabolic and functional connectivity study

“…IMU, EMG and EEG recordings were synchronized with a transistor-transistor logic (TTL) reference signal and an electrical artifact provided by a transcutaneous electrical nerve stimulation (TENS) device at the beginning and end of each recording session 10 . We detected heel strike and toe-off events from the angular velocity profiles measured with respect to the IMU medio-lateral axis as previously described 13 and computed: i) swing duration (as the duration between successive toe-off and heel strike events), ii) stride duration (as the time difference between two successive heel strikes), and the swing/stride duration (as ratio of the swing and stride duration).…”

“…Artifactual channels and segments were identified using automated artifact rejection 16 . We manually detected remaining channels with artifacts by inspecting topographic plots of the alpha power (8)(9)(10)(11)(12)(13). Artifactual channels were replaced with spherical spline interpolation on the scalp.…”

“…In three subjects, we could not acquire IMU recordings due to technical issues. In two of them, we defined gait events with surface electromyography (EMG) as described in 13 . The second patient (N2) showed artifacts in the EMG signals and was discarded for kinematic analyses.…”

Pallidal Recordings in Chronically Implanted Dystonic Patients: Mitigation of Tremor-Related Artifacts

“…The -dimensional vector of regression coefficients was estimated using ordinary least-squares (OLS) regression, where , , and T denoted the number of available paired measurements of EMG/IMU and LFP activity. Using the fitted model, the part of the LFP signal that could be predicted from EMG/IMU was obtained as [ 62 ]. The cleaned LFP signal was then obtained as the residual .…”