Myoelectric Signals During Upper-Limb Movements in a Person with Elbow Disarticulation
Description
This dataset contains myoelectric signals acquired from a 37-year-old participant with right elbow disarticulation, 13 years after upper-limb amputation. The data were collected as part of a study focused on the detection of muscle activation onset and offset during upper-limb movement. Myoelectric signals were acquired using H124SG conductive electrodes connected to MyoWare sensors (AT-04-001). Four electrodes were positioned over the short head of the biceps brachii, middle head of the deltoid, lateral head of the triceps brachii, and long head of the triceps brachii. Signals were digitized using a BeagleBone board at a sampling frequency of 1 kHz, with a voltage range from 0 to 1.8 V. The experimental protocol included six movement classes: elbow flexion, elbow extension, forearm pronation, forearm supination, hand opening, and hand closing, together with resting periods. Data acquisition was conducted in three sessions on different days. The protocol incorporated alternating periods of muscle activation and relaxation to allow the identification of contraction onset and offset. The data were generated to evaluate methods for automatic detection of muscle activation onset and offset, particularly the Root Mean Square (RMS) index and the Hilbert-Huang Transform. In the associated study, the Hilbert-Huang Transform showed higher onset and offset detection success rates for this participant than the RMS-based method. These data can be used for research on myoelectric signal processing, muscle activation detection, upper-limb prosthetic control, feature extraction, and the development or validation of algorithms for onset and offset detection in individuals with upper-limb amputation.
Files
Steps to reproduce
Download the main compressed dataset file and extract its contents using any standard ZIP-compatible decompression software. The extracted folder contains seven compressed files, corresponding to the six movements and the resting condition: M1 – elbow flexion, M2 – elbow extension, M3 – hand opening, M4 – hand closing, M5 – forearm pronation, M6 – forearm supination, and M7 – rest. Extract the compressed file corresponding to the movement or condition of interest. Each extracted folder contains plain-text files with the acquired myoelectric signals. Identify each recording using the filename convention described in the README file. The identifiers indicate participant (P), acquisition set (S), day (D), movement (M), and signal channel (s1–s4). Only recordings from participant P3, the participant with elbow disarticulation, are included in this dataset. Import the plain-text files into signal-processing software. The signals were acquired at a sampling frequency of 1 kHz and digitized within an input voltage range of 0–1.8 V. To reproduce the preprocessing used in the associated study, apply a second-order zero-phase high-pass Butterworth filter with a 5 Hz cutoff frequency, a tenth-order low-pass Butterworth filter with a 300 Hz cutoff frequency, and a notch filter to reduce 60 Hz power-line interference. The associated study used MATLAB (MathWorks, proprietary software) for signal processing. RMS-based processing was performed using a 200 ms window with 100 ms overlap. The Hilbert-Huang Transform was also applied for muscle activation onset and offset detection. Results can be validated by comparing the automatically detected muscle activation onset and offset instants with visual identification of the contractions, following the procedure reported in the associated publication.
Institutions
- Universidad Santiago de CaliValle del Cauca Department, Santiago de Cali