Transfer learning of an ensemble of DNNs for SSVEP BCI spellers without user-specific training

Abstract

Objective: Steady-state visually evoked potentials (SSVEPs), measured with EEG (electroencephalogram), yield decent information transfer rates (ITR) in brain-computer interface (BCI) spellers. However, the current high performing SSVEP BCI spellers in the literature require an initial lengthy and tiring user-specific training for each new user for system adaptation, including data collection with EEG experiments, algorithm training and calibration (all are before the actual use of the system). This impedes the widespread use of BCIs. To ensure practicality, we propose a highly novel target identification method based on an ensemble of deep neural networks (DNNs), which does not require any sort of user-specific training. Approach: We exploit already-existing literature datasets from participants of previously conducted EEG experiments to train a global target identifier DNN first, which is then fine-tuned to each participant. We transfer this ensemble of fine-tuned DNNs to the new user instance, determine the k most representative DNNs according to the participants’ statistical similarities to the new user, and predict the target character through a weighted combination of the ensemble predictions. Main results: The proposed method significantly outperforms all the state-of-the-art alternatives for all stimulation durations in [0.2 − 1.0] seconds on two large-scale benchmark and BETA datasets, and achieves impressive 155.51 bits/min and 114.64 bits/min ITRs. Code is available for reproducibility: https://github.com/osmanberke/Ensemble-of-DNNs Significance: Our Ensemble-DNN method has the potential to promote the practical widespread deployment of BCI spellers in daily lives as we provide the highest performance while enabling the immediate system use without any user-specific training.

Self-supervised contrastive learning for EEG-based cross-subject motor imagery recognition

“…Electroencephalography (EEG) is widespread in acquiring BCI signals owing to its noninvasive nature, high temporal resolution, and costeffectiveness. The EEG signals commonly utilized in studies associated with BCI encompass P300 [8], steady-state visual evoked potential [9], and motor imagery (MI) [10]. The most typical application of EEG-based MI-BCI is stroke rehabilitation.…”

Dataset Evaluation Method and Application for Performance Testing of SSVEP-BCI Decoding Algorithm

“…However, the majority of current algorithmic research utilizes one or two datasets to verify their performance [ 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 ], which did not make full use of public data resources, and the results were limited by the distribution of data samples in individual datasets, so it was not conducive to judge the application effect of the algorithm in the actual scene through the result. This issue has two underlying causes.…”

Exploring the Potential of Attention Mechanism-Based Deep Learning for Robust Subject-Independent Motor-Imagery Based BCIs

“…B RAIN-Computer interfaces (BCIs) have shown considerable promise in redefining the interaction between humans and external devices. Numerous applications have been explored, from aiding individuals with severe disabilities such as amyotrophic lateral sclerosis (ALS) and locked-in syndrome [1], [2], to the early identification of epileptic seizures [3]- [5]. Further applications include the use of advanced prosthetics [6]- [8], engagement in gaming and virtual reality [9], [10], as well as advancements in scientific research [6], [11], [12].…”