Real time cascaded moving average filter for detrending of electroencephalogram signals

Electroencephalogram signals are widely used in computer aided diagnosis of patients suffering from seizures and in detection of emotions. The signals are measured by placing electrodes on the scalp. The major sources of baseline wander in EEG signals are because of the poor contact of the electrodes and change in impedance of the electrode which can be attributed to perspiration. For effective analysis of such signals, denoising methods have to be developed. This paper proposes a low complexity method to remove baseline wander and detrend EEG signals. Dedicated handheld devices require a computationally simpler algorithm to pre-process EEG signals. This paper proposes a cascaded moving average filter as a simple and computationally efficient methodology for detrending of EEG signals. To indicate the simplicity, this work has concentrated on real-time implementation by implementing it on Xilinx XC7Z020-CLG484-1. Parameters have been derived indicating the efficiency of the proposed methodology for detrending of EEG signals. The filter discussed in this work has its uses in preprocessing of EEG signals for applications such as Brain-Computer Interface and epileptic seizure detection.