Artifact-EOG Denoising Using FIR-Filtering In EEG Channel Selection For Monitoring and Rehabilitation of Stroke Patients

Electroencephalography (EEG) will produce a non-invasive recording of brain signals used to analyze brain performance activity for medical or clinical personnel in diagnosing brain-related diseases such as stroke. The recorded EEG signal is the original brain wave signal which still contains noise artefacts. Artefact noise, among others, occurs due to eye movement or Artifact Electrooculography EOG. The main problem of this paper is that there is still noise in interpreting the EEG signal, so it is necessary to denoise the artefacts of the EOG signal using the FIR filter method. In a previous study to determine the performance of the EEG signal FIR filter in the human brain in reducing EOG artefacts through a movement stimulus, it can reach less than 100 Hz and work very well on Delta signals with the Signal to Noise Ratio (SNR) value obtained after filtering is equal to 3.2198 dB. Theoretically, the Delta signal is a wave that appears when a person is deep asleep and without dreams so that no outside information can enter the brain. Usually, in a deep sleeping position, the brain waves generated by recording will experience very little noise because there is no motor activity. Because this research focuses on stroke patients who have to carry out motor activities in medical rehabilitation, motor activity is needed to find the proper signal to reduce EOG-artefacts so that the cleaning process can be carried out correctly. Therefore, this study will improve the determination of the most appropriate signal for motor activity, namely working on Gamma signals and the performance of the EEG signal FIR filter in the human brain in reducing EOG artefacts through movement stimuli that can reach less than 100 Hz and determining the obtained SNR value. After filtering 3.2284 dB, the higher the SNR value is expected to be less noise cleaned and faster in segmentation, data decomposition, and feature extraction.