Abstract:Objective. Recently, deep learning models have been successfully applied in functional magnetic resonance imaging (fMRI) modeling and associated applications. However, there still exist at least two challenges. Firstly, due to the lack of sufficient data, deep learning models tend to suffer from overfitting in the training process. Secondly, it is still challenging to model the temporal dynamics from fMRI, due to that the brain state is continuously changing over scan time. In addition, existing methods rarely studied and applied fMRI data augmentation.Approach. In this work, we construct a deep recurrent variational auto-encoder (DRVAE) that combined variational auto-encoder and recurrent neural network, aiming to address all of the above mentioned challenges. The encoder of DRVAE can extract more generalized temporal features from assumed Gaussian distribution of input data, and the decoder of DRVAE can generate new data to increase training samples and thus partially relieve the overfitting issue. The recurrent layers in DRVAE are designed to effectively model the temporal dynamics of functional brain activities. LASSO (least absolute shrinkage and selection operator) regression is applied on the temporal features and input fMRI data to estimate the corresponding spatial networks.Main results. Extensive experimental results on seven tasks from HCP dataset showed that the DRVAE and LASSO framework can learn meaningful temporal patterns and spatial networks from both real data and generated data. The results on group-wise data and single subject suggest that the brain activities may follow certain distribution. Moreover, we applied DRVAE on four resting state fMRI datasets from ADHD-200 for data augmentation, and the results showed that the classification performances on augmented datasets have been considerably improved.Significance. The proposed method can not only derive meaningful temporal features and spatial networks from fMRI, but also generate high-quality new data for fMRI data augmentation and associated applications.

iVAE - An Improved Deep Learning Structure for EEG Signal Characterization and Reconstruction.

VAEEG: Variational Auto-encoder for Extracting EEG Representation

vEEGNet: learning latent representations to reconstruct EEG raw data via variational autoencoders

hvEEGNet: exploiting hierarchical VAEs on EEG data for neuroscience applications

Exploiting Asymmetric EEG Signals with EFD in Deep Learning Domain for Robust BCI

A hierarchical Bayesian approach for learning sparse spatio-temporal decompositions of multichannel EEG.

Deep Convolution Neural Network and Autoencoders-Based Unsupervised Feature Learning of EEG Signals.

SincVAE: a New Approach to Improve Anomaly Detection on EEG Data Using SincNet and Variational Autoencoder

Investigating Self-Supervised Deep Representations for EEG-based Auditory Attention Decoding

Modeling and augmenting of fMRI data using deep recurrent variational auto-encoder

DCAE: A dual conditional autoencoder framework for the reconstruction from EEG into image

Variational Autoencoder Learns Better Feature Representations for EEG-based Obesity Classification

Learning identifiable and interpretable latent models of high-dimensional neural activity using pi-VAE

A novel AI-driven EEG generalized classification model for cross-subject and cross-scene analysis

Deep Learning in Automatic Sleep Staging With a Single Channel Electroencephalography

Structure invariance-driven collaborative contrastive network for EEG decoding

EEG Classification of Motor Imagery Using a Novel Deep Learning Framework

Images Structure Reconstruction from Fmri by Unsupervised Learning Based on VAE

EEGDiR: Electroencephalogram denoising network for temporal information storage and global modeling through Retentive Network

A Deep Knowledge Distillation framework for EEG assisted enhancement of single-lead ECG based sleep staging

A Novel Data Augmentation Approach Using Mask Encoding for Deep Learning-Based Asynchronous SSVEP-BCI