/brain_mas_classifier

Mental Attention State Classification for the Human Brain using EEG-based BCI data

Primary LanguageJupyter Notebook

GA-ANFIS and CRNN for Mental Attention State Classification using Passive BCI-based EEG data

By using electroencephalography (EEG) based BCI intrinsic or passive activity data self-generated by specified individuals under simulation or obtained live [3], we aim to detect and classify the current mental attention state of an individual into several categories of states, the most prominent being - focused, unfocused (may also be described as lost-in-thought or mind-wandering) and drowsy (or sleep). Our approach uses a hybrid neuro-genetic fuzzy system optimized and trained on a select number of features and channels extracted from the data as well as a shallow convolutional and convolutional recurrent neural network trained on the raw EEG signal data to predict human mental attention state with signals of trial lengths as short as 6 seconds.

View the complete project report here.

Dataset

Collaborators