/EEGMMI-Deep-ConvNET-

Deep Learning with Convolutional Neural Network Predicts Imagery Tasks Through EEG

Primary LanguageMATLABMIT LicenseMIT

Deep Learning with Convolutional Neural Network Predicts Imagery Tasks Through EEG

Train Deep ConvNET with raw EEG data

downloadEEGMMI

Downloads and saves EEGMMI dataset from Pyhsionet with corresponding events event

           X .................vector includes subject ids to be downloaded (1-109)
           n..................tasks to be downloaded (1-14)

USAGE

               subject=1:15
	   task=[2,5,7]
               downloadEEGMMI(subject, task)

FILES

	 S001R08.mat.......data file id:1, task: 8
         ann_S001R08.mat...event file id:1, task: 8

isegment

[data, label]=isegment(id)

        Segments data for any subject
            id ..............subject number (choices: 1-62)
            data.............EEG data
            label............label related to data(with considering
                             annotation)
            requires ann.mat file that includes annotations

USAGE

            [data2, label2]=isegment(2);

tempimag

feature=tempimag(data, channelnumber)

Constructs temporal images 

            data ..............EEG data (each row corresponds to a channel X trials)
            channelnumber......Number of channles 
            feature............Images related to each trial with
                               struct  data type

USAGE

            data=randn(2880, 656);
            feature=tempimag(data, 64);
            figure, imshow(feature.images{1})

sendimage2folders

sendimage2folders(feature, label)

Sends images to labeled folders

ReadyforClassifyConvNet

Makes data ready for deep ConvNet

implementConvNET

Scripts to train and test with deep ConvNet

Multi-Layer Perceptron with Hand-Crafted Spectral Features

ReadyforClassifyMLP

Makes data ready for MLP including Welch periodogram

implementMLP

Scripts to train and test with MLP

Apdullah Yayık, 2018 ayayik@kho.edu.tr