This is a EEG dataset that records brain activity during the overnight driving on a car simulator. The recorded observations contain microsleep patterns which can be used to train the model to forecast brain activity and predict when will be the next microsleep event happen.

The dataset is split across multiple files which you can iteratively load into a dataframe, preprocess it and pass it to the model by using a generator function.

The 1st column is time-domain start from -12 to 2 which you can use to plot EEG signals. The rest of the 36 columns are EEG signals to train your model.