More Like this

1. A Comparison of Classification Techniques to Predict Brain-Computer Interfaces Accuracy Using Classifier-Based Latency Estimation

   https://doi.org/10.3390/brainsci10100734  

   Mowla, Md Rakibul; Gonzalez-Morales, Jesus D.; Rico-Martinez, Jacob; Ulichnie, Daniel A.; Thompson, David E. (October 2020, Brain Sciences)

   P300-based Brain-Computer Interface (BCI) performance is vulnerable to latency jitter. To investigate the role of latency jitter on BCI system performance, we proposed the classifier-based latency estimation (CBLE) method. In our previous study, CBLE was based on least-squares (LS) and stepwise linear discriminant analysis (SWLDA) classifiers. Here, we aim to extend the CBLE method using sparse autoencoders (SAE) to compare the SAE-based CBLE method with LS- and SWLDA-based CBLE. The newly-developed SAE-based CBLE and previously used methods are also applied to a newly-collected dataset to reduce the possibility of spurious correlations. Our results showed a significant (p<0.001) negative correlation between BCI accuracy and estimated latency jitter. Furthermore, we also examined the effect of the number of electrodes on each classification technique. Our results showed that on the whole, CBLE worked regardless of the classification method and electrode count; by contrast the effect of the number of electrodes on BCI performance was classifier dependent. 

   more » « less
2. Development of an Adaptive Artifact Subspace Reconstruction Based on Hebbian/Anti-Hebbian Learning Networks for Enhancing BCI Performance

   https://doi.org/10.1109/TNNLS.2022.3174528  

   Tsai, Bo-Yu; Diddi, Sandeep Vara; Ko, Li-Wei; Wang, Shuu-Jiun; Chang, Chi-Yuan; Jung, Tzyy-Ping (June 2022, IEEE Transactions on Neural Networks and Learning Systems)

   Brain-computer interface (BCI) actively translates the brain signals into executable actions by establishing direct communication between the human brain and external devices. Recording brain activity through electroencephalography (EEG) is generally contaminated with both physiological and nonphysiological artifacts, which significantly hinders the BCI performance. Artifact subspace reconstruction (ASR) is a well-known statistical technique that automatically removes artifact components by determining the rejection threshold based on the initial reference EEG segment in multichannel EEG recordings. In real-world applications, the fixed threshold may limit the efficacy of the artifact correction, especially when the quality of the reference data is poor. This study proposes an adaptive online ASR technique by integrating the Hebbian/anti-Hebbian neural networks into the ASR algorithm, namely, principle subspace projection ASR (PSP-ASR) and principal subspace whitening ASR (PSW-ASR) that segmentwise self-organize the artifact subspace by updating the synaptic weights according to the Hebbian and anti-Hebbian learning rules. The effectiveness of the proposed algorithm is compared to the conventional ASR approaches on benchmark EEG dataset and three BCI frameworks, including steady-state visual evoked potential (SSVEP), rapid serial visual presentation (RSVP), and motor imagery (MI) by evaluating the root-mean-square error (RMSE), the signal-to-noise ratio (SNR), the Pearson correlation, and classification accuracy. The results demonstrated that the PSW-ASR algorithm effectively removed the EEG artifacts and retained the activity-specific brain signals compared to the PSP-ASR, standard ASR (Init-ASR), and moving-window ASR (MW-ASR) methods, thereby enhancing the SSVEP, RSVP, and MI BCI performances. Finally, our empirical results from the PSW-ASR algorithm suggested the choice of an aggressive cutoff range of c = 1-10 for activity-specific BCI applications and a moderate range of for the benchmark dataset and general BCI applications. 

   more » « less
3. Benchtop and bedside validation of a low-cost programmable cortical stimulator in a testbed for bi-directional brain-computer-interface research

   https://doi.org/10.3389/fnins.2022.1075971  

   Sohn, Won Joon; Lim, Jeffrey; Wang, Po T.; Pu, Haoran; Malekzadeh-Arasteh, Omid; Shaw, Susan J.; Armacost, Michelle; Gong, Hui; Kellis, Spencer; Andersen, Richard A.; et al (January 2023, Frontiers in Neuroscience)

   Introduction Bi-directional brain-computer interfaces (BD-BCI) to restore movement and sensation must achieve concurrent operation of recording and decoding of motor commands from the brain and stimulating the brain with somatosensory feedback. Methods A custom programmable direct cortical stimulator (DCS) capable of eliciting artificial sensorimotor response was integrated into an embedded BCI system to form a safe, independent, wireless, and battery powered testbed to explore BD-BCI concepts at a low cost. The BD-BCI stimulator output was tested in phantom brain tissue by assessing its ability to deliver electrical stimulation equivalent to an FDA-approved commercial electrical cortical stimulator. Subsequently, the stimulator was tested in an epilepsy patient with subcortical electrocorticographic (ECoG) implants covering the sensorimotor cortex to assess its ability to elicit equivalent responses as the FDA-approved counterpart. Additional safety features (impedance monitoring, artifact mitigation, and passive and active charge balancing mechanisms) were also implemeneted and tested in phantom brain tissue. Finally, concurrent operation with interleaved stimulation and BCI decoding was tested in a phantom brain as a proof-of-concept operation of BD-BCI system. Results The benchtop prototype BD-BCI stimulator's basic output features (current amplitude, pulse frequency, pulse width, train duration) were validated by demonstrating the output-equivalency to an FDA-approved commercial cortical electrical stimulator ( R 2 > 0.99). Charge-neutral stimulation was demonstrated with pulse-width modulation-based correction algorithm preventing steady state voltage deviation. Artifact mitigation achieved a 64.5% peak voltage reduction. Highly accurate impedance monitoring was achieved with R 2 > 0.99 between measured and actual impedance, which in-turn enabled accurate charge density monitoring. An online BCI decoding accuracy of 93.2% between instructional cues and decoded states was achieved while delivering interleaved stimulation. The brain stimulation mapping via ECoG grids in an epilepsy patient showed that the two stimulators elicit equivalent responses. Significance This study demonstrates clinical validation of a fully-programmable electrical stimulator, integrated into an embedded BCI system. This low-cost BD-BCI system is safe and readily applicable as a testbed for BD-BCI research. In particular, it provides an all-inclusive hardware platform that approximates the limitations in a near-future implantable BD-BCI. This successful benchtop/human validation of the programmable electrical stimulator in a BD-BCI system is a critical milestone toward fully-implantable BD-BCI systems. 

   more » « less
4. Role of Scalp EEG Brain Connectivity in Motor Imagery Decoding for BCI Applications

   https://doi.org/10.1109/EMBC53108.2024.10781532  

   Delavari, Fatemeh; Santaniello, Sabato (July 2024, IEEE)

   Brain Connectivity (BC) features of multichannel EEG have been proposed for Motor Imagery (MI) decoding in Brain-Computer Interface applications, but the advantages of BC features vs. single-channel features are unclear. Here, we consider three BC features, i.e., Phase Locking Value (PLV), Granger Causality, and weighted Phase Lag Index, and investigate the relationship between the most central nodes in BC-based networks and the most influential EEG channels in single-channel classification based on common spatial pattern filtering. Then, we compare the accuracy of MI decoders that use BC features in source vs. sensor space. Applied to the BCI Competition VI Dataset 2a (left- vs. right-hand MI decoding), our study found that PLV in sensor space achieves the highest classification accuracy among BC features and has similar performance compared to single-channel features, while the transition from sensor to source space reduces the average accuracy of BC features. Across all BC measures, the network topology is similar in left- vs. right-hand MI tasks, and the most central nodes in BC-based networks rarely overlap with the most influential channels in single-channel classification. 

   more » « less
5. Context-aware Multimodal Auditory BCI Classification through Graph Neural Networks

   https://doi.org/10.1109/EMBC40787.2023.10339984  

   Kumar, Chetan; Rahimi, Neela; Gonjari, Rohan; McLinden, John; Hosni, Sarah Ismail; Shahriari, Yalda; Shao, Ming (July 2023, IEEE)

   The prospect of electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) in the presence of topological information of participants is often left unexplored in most of the brain-computer interface (BCI) systems. Additionally, the usage of these modalities together in the field of multimodality analysis to support multiple brain signals toward improving BCI performance is not fully examined. This study first presents a multimodal data fusion framework to exploit and decode the complementary synergistic properties in multimodal neural signals. Moreover, the relations among different subjects and their observations also play critical roles in classifying unknown subjects. We developed a context-aware graph neural network (GNN) model utilizing the pairwise relationship among participants to investigate the performance on an auditory task classification. We explored standard and deviant auditory EEG and fNIRS data where each subject was asked to perform an auditory oddball task and has multiple trials regarded as context-aware nodes in our graph construction. In experiments, our multimodal data fusion strategy showed an improvement up to 8.40% via SVM and 2.02% via GNN, compared to the single-modal EEG or fNIRS. In addition, our context-aware GNN achieved 5.3%, 4.07% and 4.53% higher accuracy for EEG, fNIRS and multimodal data based experiments, compared to the baseline models. 

   more » « less