Objective: Development of an adaptive filtering algorithm to remove movement artefacts from EEG data acquired during a dual sensory-motor gait task when performed by People with Parkinson’s Disease (PwPD). Combined with wireless EEG measurement systems, adaptive filtering algorithms provide more quantifiable information on neuromotor ability during gait can be obtained.

Background: Assessing the risk of falling in PwPD is typically carried out by a combination of neuropsychological and dynamic movement tests. EEG-derived measures would allow clinicians to probe neural activation patterns, which may be indicative of falls risk. Wireless EEG systems allow for such neural activity to be simultaneously recorded along with gait-related movement data, allowing a better understanding of the origins of gait abnormalities in PwPD.

During the performance of these dynamic movement tasks, especially during gait, EEG measurements are often corrupted by movement artefacts. Conventional filtering methods can remove noise, but also large portion of the signal that can contain valuable information is simultaneously removed.  The challenge is to develop a filter that can efficiently filter the specific parts of the EEG signal containing gait-induced artefacts while maintaining a high signal-to-noise (SNR) ratio. Artefact Subspace Reconstruction (ASR) filtering is a method typically applied artefact removal.

Method: EEG data and movement data, obtained from inertial movement sensors (IMU) placed on the lower limbs, are recorded synchronously. However, applying ASR to the position of the EEG signal during known specific periods where the signal is corrupted by movement artefact (heel strikes, etc.) can prove to be more effective in artefact removal while maintaining the quality of neural information contained in the data. The performance of this artefact removal strategy is objectively quantified by metrics such as difference SNR. Testing is underway in a convenient sample of PwPD cohort (3) and aged-matched controls (3) as they follow a specific out-patient protocol to assess their balance and movement.

Results: The application of an optimal filtering algorithm to remove gait-induced movement artefacts will allow improved clinical interpretation of neuromotor ability of PwPD during different dynamic movement tasks.

Conclusion: .

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/determining-the-optimal-filtering-method-for-the-removal-of-gait-related-movement-artefacts-from-eeg-acquired-from-people-with-parkinsons-disease/