Create a machine/deep learning based model which will sucessfully predict if a 20 minute baseline EEG session is that of a schizophrenic.
To enhance quick diagnosis of patients that land in the emergency room to get to know if they're schizophrenic and need specializied care.
Research is a team effort, 3-4 people on board at the moment.
Not sure about sharing all code (privacy etc) but i will aim to share the most of it.
There are obvious issues that can arise from the contextual setting of landing on ER that will have to be accounted for if we manage to make a working model based on the dataset.
This notebook is meant for exploration and notes for the TUH EEG dataset, alongside with ideas on what to extract from it.
The next step will be to extract the necessary samples. around 300-500 of control and maybe 100 of diagnosed schizophrenics (not on meds) to enhance our initial dataset.
Find the best way to traverse through the TUH EEG dataset.Get a sample that will be used for the prototype models.- Create a prototype model.