Motion artefact removal in electroencephalography and electrocardiography by using multichannel inertial measurement units and adaptive filtering

FIGURE 1. Proposed overall design architecture. BUF = buffer amplifier, IA = instrumentation amplifier, all other acronyms are defined in the text

FIGURE 2. Photograph of the fabricated device. (Top) PCB design of DRL and active electrode compared to a UK 5p. (Bottom) The electronic circuit used to record EEG and ECG signals together with a standard self‐adhesive ECG electrode

FIGURE 3. Diagram of the overall adaptive filtering approach demonstrating the two adaptive filter processing stages

FIGURE 4. Example ECG recorded from V2 electrode and corresponding accelerometer and gyroscope data

FIGURE 5. Example EEG recorded from T5 and corresponding accelerometer and gyroscope data

FIGURE 6. Example comparing filtering of data using wavelet baseline removal (b) against the adaptive filtering method proposed in this paper (c). Red crosses indicate identified QRS complexes from the Pan–Tompkins algorithm

FIGURE 7. Example of a good case, filtering ECG using the accelerometer and gyroscope independently

FIGURE 8. Example of a poor case, filtering ECG using the accelerometer and gyroscope independently

FIGURE 9. Example filtering EEG using the accelerometer and gyroscope independently