Transfer Learning in Brain-Computer Interfaces with Adversarial\n Variational Autoencoders

2018/12/17 by Ozan Özdenizci, Ozdenizci, Ozan, Ye Wang +5
Neuroscience · #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #FOS: Electrical engineering #Functional Brain Connectivity Studies #Human-Computer Interaction (cs.HC) #Machine Learning (cs.LG) #Neural dynamics and brain function #Signal Processing (eess.SP) #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.1812.06857

openalex publication_date 2018/12/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We introduce adversarial neural networks for representation learning as a\nnovel approach to transfer learning in brain-computer interfaces (BCIs). The\nproposed approach aims to learn subject-invariant representations by\nsimultaneously training a conditional variational autoencoder (cVAE) and an\nadversarial network. We use shallow convolutional architectures to realize the\ncVAE, and the learned encoder is transferred to extract subject-invariant\nfeatures from unseen BCI users' data for decoding. We demonstrate a\nproof-of-concept of our approach based on analyses of electroencephalographic\n(EEG) data recorded during a motor imagery BCI experiment.\n

Citations

• BCI2000: A General-Purpose Brain-Computer Interface (BCI) System
• PhysioBank, PhysioToolkit, and PhysioNet
• Deep learning with convolutional neural networks for EEG decoding and visualization
• Invariant Representations from Adversarially Censored Autoencoders
• Adversarial Discriminative Domain Adaptation
• Learning Representations from EEG with Deep Recurrent-Convolutional Neural Networks

Related

• Mutual Information-driven Subject-invariant and Class-relevant Deep\n Representation Learning in BCI
• Confidence-Aware Subject-to-Subject Transfer Learning for Brain-Computer Interface
• Leveraging Transfer Learning and User-Specific Updates for Rapid Training of BCI Decoders
• Deep Transfer Learning for EEG-based Brain Computer Interface
• Reconstructing Perceived Images from Brain Activity by Visually-guided Cognitive Representation and Adversarial Learning