This project aims to address the critical need for assisting patients with neurological disabilities or limb amputations in regaining the ability to perform fundamental hand functions. To achieve this, we leverage EEG (Electroencephalogram) data analysis to identify and understand simple hand movements.
There are 12 subjects in total, 10 series of trials for each subject, and approximately 30 trials within each series. The number of trials varies for each series. The training set contains the first 8 series for each subject. The test set contains the 9th and 10th series. For each GAL, the project is aimed to detect 6 events always occuring in the same order:
The development of a predictive model has involved the implementation of the following procedural steps:
To enhance the accuracy of the models and gain a deeper understanding of the EEG signals, several avenues for future work are proposed. Feature engineering is a crucial next step to extract more relevant information from the data. Furthermore, the treatment of EEG signals as raw data can be revised to incorporate a temporal unit into the code, potentially leading to increased accuracy and improved insights into neurological disability-related hand movements.
GitHub repositoryRaw data
Processed data
Dimentionality reduction (principal component analysis)
Random forest evaluation