A Survey on Deep Learning-Based Short/Zero-Calibration Approaches for EEG-Based Brain–Computer Interfaces

Abstract

Brain–computer interfaces (BCIs) utilizing machine learning techniques are an emerging technology that enables a communication pathway between a user and an external system, such as a computer. Owing to its practicality, electroencephalography (EEG) is one of the most widely used measurements for BCI. However, EEG has complex patterns and EEG-based BCIs mostly involve a cost/time-consuming calibration phase; thus, acquiring sufficient EEG data is rarely possible. Recently, deep learning (DL) has had a theoretical/practical impact on BCI research because of its use in learning representations of complex patterns inherent in EEG. Moreover, algorithmic advances in DL facilitate short/zero-calibration in BCI, thereby suppressing the data acquisition phase. Those advancements include data augmentation (DA), increasing the number of training samples without acquiring additional data, and transfer learning (TL), taking advantage of representative knowledge obtained from one dataset to address the so-called data insufficiency problem in other datasets. In this study, we review DL-based short/zero-calibration methods for BCI. Further, we elaborate methodological/algorithmic trends, highlight intriguing approaches in the literature, and discuss directions for further research. In particular, we search for generative model-based and geometric manipulation-based DA methods. Additionally, we categorize TL techniques in DL-based BCIs into explicit and implicit methods. Our systematization reveals advances in the DA and TL methods. Among the studies reviewed herein, ~45% of DA studies used generative model-based techniques, whereas ~45% of TL studies used explicit knowledge transferring strategy. Moreover, based on our literature review, we recommend an appropriate DA strategy for DL-based BCIs and discuss trends of TLs used in DL-based BCIs.

Studies to Overcome Brain–Computer Interface Challenges

“…Over the past decades, there have been lots of BCI studies. They were usually focused on the methods to improve the prediction accuracy [5,8,23,[30][31][32], raise the number of commands [12,33], increase the information transfer rate (ITR) [34][35][36][37][38], or reduce the training efforts [7,30,34,39]. To enhance the prediction accuracy, new classification algorithms [30,40,41] or feature extraction methods have been proposed [31,32,42].…”

“…The accuracy and ITR of the hybrid study were 95.2% and 360.7 bpm, respectively [38]. To reduce the time measuring training data, data augmentation or transfer learning approaches were often used [39]. The prediction model can be trained with fewer data by augmenting the data.…”

2020 International brain–computer interface competition: A review

“…. Background Short calibration, which uses minimal training data in most BCI paradigms that aim to develop practical systems, is a major challenge (Benaroch et al, 2021;Ko et al, 2021a). The development of decoding models using less training data, described as few-shot learning, which allows a model to learn a method that enables fast adaptation to a new task or environment (Hospedales et al, 2020), is one of the challenges associated with machine learning and deep learning .…”

Semi-supervised generative and discriminative adversarial learning for motor imagery-based brain–computer interface

“…The (generative) adversarial learning has also been applied to BCI tasks for well generalization 39 . For instance, Tan et al 40 converted raw EEG signals to EEG optical flow images and obtained a general feature extractor for EEG optical flow images and ImageNet by adversarial learning to build a classification network capable of classifying category labels.…”