This thesis will be aimed at automated classification of intracranial EEG of pharmacoresistant epileptic patients who are performing a verbal memory task. The main goal of the thesis will be to utilize machine learning models to forecast whether presented words are going to be forgotten or recalled by the patient during the task performance. The secondary goal will be to identify the brain structures such recall prediction models should focus on.

The theoretical part of the thesis will provide the necessary medical/neuroscience background and investigate the state-of-the-art research in usage of machine learning models for verbal memory prediction. Methods suitable for achieving the thesis goal will be identified then. The practical part will implement chosen algorithms and train and evaluate them across multiple patients and clinical centers. The results will be written up in the form of a thesis.