Research Article

Deep Learning Approach for Automatic Classification of Ocular and Cardiac Artifacts in MEG Data

Table 1

In section A, the parameters to train the model are summarized. Section B shows the performance parameters of the DCNN model when applied to the validation dataset.

Aā€‰

Number of samples (training)4628
Number of samples (validation)2494
Minibatch size32
Learning rate0.001

Bā€‰

Epoch number (best model)32
True positive rate (sensitivity)91.8%
True negative rate (specificity)97.4%