VLSI Architecture for Modulating Nerve Activity in Parkinson’s disease Using STFT Transformation Methods

The progressive neurodegenerative condition known as Parkinson’s disease (PD) mostly impairs motor control and appears as bradykinesia, stiffness, and tremors. The abnormal neural activity that underlies these symptoms is frequently not sufficiently treated by existing medicines, despite tremendous advancements in therapy. It could render traditional signal processing methods unsuitable for continuous, long-term use in implanted or wearable medical devices. The research suggests a novel VLSI (Very Large-Scale Integration) design that uses Short-Time Fourier Transform (STFT) transformation techniques to modulate neural activity in PD. To analyze and modify nerve activity, it makes use of a dataset of electroencephalogram (EEG) signals that show neural activity in PD patients. The dataset includes time-series data and time-frequency characteristics. A band-pass filter was used to preprocess the EEG data in order to eliminate noise and unnecessary frequencies, maintaining the emphasis on the pertinent neural activity. The STFT was then used to extract features, which produced time-frequency representations that capture the EEG signals’ phase and amplitude information for additional examination. A Convolutional Neural Network (CNN) is used to determine if a person has PD or not. The CNN uses its capacity to identify intricate patterns to interpret the retrieved STFT characteristics from the EEG data and make precise predictions. The performance for the proposed method is analyzed by accuracy, throughput, latency, and power consumption. The technology can precisely detect and alter the oscillatory patterns linked to PD by applying STFT+CNN to neural signals, thereby providing a way to implement more individualized and successful treatment approaches.