Median Nerve Stimulation Based BCI: A New Approach to Detect Intraoperative Awareness During General Anesthesia

Brain-computer interfaces: the innovative key to unlocking neurological conditions

Neurological disorders such as Parkinson's disease, stroke, and spinal cord injury can pose significant threats to human mortality, morbidity, and functional independence. Brain-Computer Interface (BCI) technology, which facilitates direct communication between the brain and external devices, emerges as an innovative key to unlocking neurological conditions, demonstrating significant promise in this context. This comprehensive review uniquely synthesizes the latest advancements in BCI research across multiple neurological disorders, offering an interdisciplinary perspective on both clinical applications and emerging technologies. We explore the progress in BCI research and its applications in addressing various neurological conditions, with a particular focus on recent clinical studies and prospective developments. Initially, the review provides an up-to-date overview of BCI technology, encompassing its classification, operational principles, and prevalent paradigms. It then critically examines specific BCI applications in movement disorders, disorders of consciousness, cognitive and mental disorders, as well as sensory disorders, highlighting novel approaches and their potential impact on patient care. This review reveals emerging trends in BCI applications, such as the integration of artificial intelligence and the development of closed-loop systems, which represent significant advancements over previous technologies. The review concludes by discussing the prospects and directions of BCI technology, underscoring the need for interdisciplinary collaboration and ethical considerations. It emphasizes the importance of prioritizing bidirectional and high-performance BCIs, areas that have been underexplored in previous reviews. Additionally, we identify crucial gaps in current research, particularly in long-term clinical efficacy and the need for standardized protocols. The role of neurosurgery in spearheading the clinical translation of BCI research is highlighted. Our comprehensive analysis presents BCI technology as an innovative key to unlocking neurological disorders, offering a transformative approach to diagnosing, treating, and rehabilitating neurological conditions, with substantial potential to enhance patients' quality of life and advance the field of neurotechnology.

A cross-dataset adaptive domain selection transfer learning framework for motor imagery-based brain-computer interfaces

Objective. In brain-computer interfaces (BCIs) that utilize motor imagery (MI), minimizing calibration time has become increasingly critical for real-world applications. Recently, transfer learning (TL) has been shown to effectively reduce the calibration time in MI-BCIs. However, variations in data distribution among subjects can significantly influence the performance of TL in MI-BCIs.Approach.We propose a cross-dataset adaptive domain selection transfer learning framework that integrates domain selection, data alignment, and an enhanced common spatial pattern (CSP) algorithm. Our approach uses a huge dataset of 109 subjects as the source domain. We begin by identifying non-BCI illiterate subjects from this huge dataset, then determine the source domain subjects most closely aligned with the target subjects using maximum mean discrepancy. After undergoing Euclidean alignment processing, features are extracted by multiple composite CSP. The final classification is carried out using the support vector machine.Main results.Our findings indicate that the proposed technique outperforms existing methods, achieving classification accuracies of 75.05% and 76.82% in two cross-dataset experiments, respectively.Significance.By reducing the need for extensive training data, yet maintaining high accuracy, our method optimizes the practical implementation of MI-BCIs.

Modeling Complex EEG Data Distribution on the Riemannian Manifold Toward Outlier Detection and Multimodal Classification

OBJECTIVE

The usage of Riemannian geometry for Brain-computer interfaces (BCIs) has gained momentum in recent years. Most of the machine learning techniques proposed for Riemannian BCIs consider the data distribution on a manifold to be unimodal. However, the distribution is likely to be multimodal rather than unimodal since high-data variability is a crucial limitation of electroencephalography (EEG). In this paper, we propose a novel data modeling method for considering complex data distributions on a Riemannian manifold of EEG covariance matrices, aiming to improve BCI reliability.

METHODS

Our method, Riemannian spectral clustering (RiSC), represents EEG covariance matrix distribution on a manifold using a graph with proposed similarity measurement based on geodesic distances, then clusters the graph nodes through spectral clustering. This allows flexibility to model both a unimodal and a multimodal distribution on a manifold. RiSC can be used as a basis to design an outlier detector named outlier detection Riemannian spectral clustering (odenRiSC) and a multimodal classifier named multimodal classifier Riemannian spectral clustering (mcRiSC). All required parameters of odenRiSC/mcRiSC are selected in data-driven manner. Moreover, there is no need to pre-set a threshold for outlier detection and the number of modes for multimodal classification.

RESULTS

The experimental evaluation revealed odenRiSC can detect EEG outliers more accurately than existing methods and mcRiSC outperformed the standard unimodal classifier, especially on high-variability datasets.

CONCLUSION

odenRiSC/mcRiSC are anticipated to contribute to making real-life BCIs outside labs and neuroergonomics applications more robust.

SIGNIFICANCE

RiSC can work as a robust EEG outlier detector and multimodal classifier.

EEG-based emergency braking intention detection during simulated driving

BACKGROUND

Current research related to electroencephalogram (EEG)-based driver's emergency braking intention detection focuses on recognizing emergency braking from normal driving, with little attention to differentiating emergency braking from normal braking. Moreover, the classification algorithms used are mainly traditional machine learning methods, and the inputs to the algorithms are manually extracted features.

METHODS

To this end, a novel EEG-based driver's emergency braking intention detection strategy is proposed in this paper. The experiment was conducted on a simulated driving platform with three different scenarios: normal driving, normal braking and emergency braking. We compared and analyzed the EEG feature maps of the two braking modes, and explored the use of traditional methods, Riemannian geometry-based methods, and deep learning-based methods to predict the emergency braking intention, all using the raw EEG signals rather than manually extracted features as input.

RESULTS

We recruited 10 subjects for the experiment and used the area under the receiver operating characteristic curve (AUC) and F1 score as evaluation metrics. The results showed that both the Riemannian geometry-based method and the deep learning-based method outperform the traditional method. At 200 ms before the start of real braking, the AUC and F1 score of the deep learning-based EEGNet algorithm were 0.94 and 0.65 for emergency braking vs. normal driving, and 0.91 and 0.85 for emergency braking vs. normal braking, respectively. The EEG feature maps also showed a significant difference between emergency braking and normal braking. Overall, based on EEG signals, it was feasible to detect emergency braking from normal driving and normal braking.

CONCLUSIONS

The study provides a user-centered framework for human-vehicle co-driving. If the driver's intention to brake in an emergency can be accurately identified, the vehicle's automatic braking system can be activated hundreds of milliseconds earlier than the driver's real braking action, potentially avoiding some serious collisions.

Detection of Motor Cerebral Activity After Median Nerve Stimulation During General Anesthesia (STIM-MOTANA): Protocol for a Prospective Interventional Study

BACKGROUND

Accidental awareness during general anesthesia (AAGA) is defined as an unexpected awareness of the patient during general anesthesia. This phenomenon occurs in 1%-2% of high-risk practice patients and can cause physical suffering and psychological after-effects, called posttraumatic stress disorder. In fact, no monitoring techniques are satisfactory enough to effectively prevent AAGA; therefore, new alternatives are needed. Because the first reflex for a patient during an AAGA is to move, but cannot do so because of the neuromuscular blockers, we believe that it is possible to design a brain-computer interface (BCI) based on the detection of movement intention to warn the anesthetist. To do this, we propose to describe and detect the changes in terms of motor cortex oscillations during general anesthesia with propofol, while a median nerve stimulation is performed. We believe that our results could enable the design of a BCI based on median nerve stimulation, which could prevent AAGA.

OBJECTIVE

To our knowledge, no published studies have investigated the detection of electroencephalographic (EEG) patterns in relation to peripheral nerve stimulation over the sensorimotor cortex during general anesthesia. The main objective of this study is to describe the changes in terms of event-related desynchronization and event-related synchronization modulations, in the EEG signal over the motor cortex during general anesthesia with propofol while a median nerve stimulation is performed.

METHODS

STIM-MOTANA is an interventional and prospective study conducted with patients scheduled for surgery under general anesthesia, involving EEG measurements and median nerve stimulation at two different times: (1) when the patient is awake before surgery (2) and under general anesthesia. A total of 30 patients will receive surgery under complete intravenous anesthesia with a target-controlled infusion pump of propofol.

RESULTS

The changes in event-related desynchronization and event-related synchronization during median nerve stimulation according to the various propofol concentrations for 30 patients will be analyzed. In addition, we will apply 4 different offline machine learning algorithms to detect the median nerve stimulation at the cerebral level. Recruitment began in December 2022. Data collection is expected to conclude in June 2024.

CONCLUSIONS

STIM-MOTANA will be the first protocol to investigate median nerve stimulation cerebral motor effect during general anesthesia for the detection of intraoperative awareness. Based on strong practical and theoretical scientific reasoning from our previous studies, our innovative median nerve stimulation-based BCI would provide a way to detect intraoperative awareness during general anesthesia.

TRIAL REGISTRATION

Clinicaltrials.gov NCT05272202; https://clinicaltrials.gov/ct2/show/NCT05272202.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

PRR1-10.2196/43870.

EEG-Based Spectral Analysis Showing Brainwave Changes Related to Modulating Progressive Fatigue During a Prolonged Intermittent Motor Task

Repeatedly performing a submaximal motor task for a prolonged period of time leads to muscle fatigue comprising a central and peripheral component, which demands a gradually increasing effort. However, the brain contribution to the enhancement of effort to cope with progressing fatigue lacks a complete understanding. The intermittent motor tasks (IMTs) closely resemble many activities of daily living (ADL), thus remaining physiologically relevant to study fatigue. The scope of this study is therefore to investigate the EEG-based brain activation patterns in healthy subjects performing IMT until self-perceived exhaustion. Fourteen participants (median age 51.5 years; age range 26−72 years; 6 males) repeated elbow flexion contractions at 40% maximum voluntary contraction by following visual cues displayed on an oscilloscope screen until subjective exhaustion. Each contraction lasted ≈5 s with a 2-s rest between trials. The force, EEG, and surface EMG (from elbow joint muscles) data were simultaneously collected. After preprocessing, we selected a subset of trials at the beginning, middle, and end of the study session representing brain activities germane to mild, moderate, and severe fatigue conditions, respectively, to compare and contrast the changes in the EEG time-frequency (TF) characteristics across the conditions. The outcome of channel- and source-level TF analyses reveals that the theta, alpha, and beta power spectral densities vary in proportion to fatigue levels in cortical motor areas. We observed a statistically significant change in the band-specific spectral power in relation to the graded fatigue from both the steady- and post-contraction EEG data. The findings would enhance our understanding on the etiology and physiology of voluntary motor-action-related fatigue and provide pointers to counteract the perception of muscle weakness and lack of motor endurance associated with ADL. The study outcome would help rationalize why certain patients experience exacerbated fatigue while carrying out mundane tasks, evaluate how clinical conditions such as neurological disorders and cancer treatment alter neural mechanisms underlying fatigue in future studies, and develop therapeutic strategies for restoring the patients' ability to participate in ADL by mitigating the central and muscle fatigue.

The Differences Between Motor Attempt and Motor Imagery in Brain-Computer Interface Accuracy and Event-Related Desynchronization of Patients With Hemiplegia

Background: Motor attempt and motor imagery (MI) are two common motor tasks used in brain-computer interface (BCI). They are widely researched for motor rehabilitation in patients with hemiplegia. The differences between the motor attempt (MA) and MI tasks of patients with hemiplegia can be used to promote BCI application. This study aimed to explore the accuracy of BCI and event-related desynchronization (ERD) between the two tasks.

Materials and Methods: We recruited 13 patients with stroke and 3 patients with traumatic brain injury, to perform MA and MI tasks in a self-control design. The BCI accuracies from the bilateral, ipsilesional, and contralesional hemispheres were analyzed and compared between different tasks. The cortical activation patterns were evaluated with ERD and laterality index (LI).

Results: The study showed that the BCI accuracies of MA were significantly (p < 0.05) higher than MI in the bilateral, ipsilesional, and contralesional hemispheres in the alpha-beta (8–30 Hz) frequency bands. There was no significant difference in ERD and LI between the MA and MI tasks in the 8–30 Hz frequency bands. However, in the MA task, there was a negative correlation between the ERD values in the channel CP1 and ipsilesional hemispheric BCI accuracies (r = −0.552, p = 0.041, n = 14) and a negative correlation between the ERD values in channel CP2 and bilateral hemispheric BCI accuracies (r = −0.543, p = 0.045, n = 14). While in the MI task, there were negative correlations between the ERD values in channel C4 and bilateral hemispheric BCI accuracies (r = −0.582, p = 0.029, n = 14) as well as the contralesional hemispheric BCI accuracies (r = −0.657, p = 0.011, n = 14). As for motor dysfunction, there was a significant positive correlation between the ipsilesional BCI accuracies and FMA scores of the hand part in 8–13 Hz (r = 0.565, p = 0.035, n = 14) in the MA task and a significant positive correlation between the ipsilesional BCI accuracies and FMA scores of the hand part in 13–30 Hz (r = 0.558, p = 0.038, n = 14) in the MI task.

Conclusion: The MA task may achieve better BCI accuracy but have similar cortical activations with the MI task. Cortical activation (ERD) may influence the BCI accuracy, which should be carefully considered in the BCI motor rehabilitation of patients with hemiplegia.

Influence of the visuo-proprioceptive illusion of movement and motor imagery of the wrist on EEG cortical excitability among healthy participants

INTRODUCTION

Motor Imagery (MI) is a powerful tool to stimulate sensorimotor brain areas and is currently used in motor rehabilitation after a stroke. The aim of our study was to evaluate whether an illusion of movement induced by visuo-proprioceptive immersion (VPI) including tendon vibration (TV) and Virtual moving hand (VR) combined with MI tasks could be more efficient than VPI alone or MI alone on cortical excitability assessed using Electroencephalography (EEG).

METHODS

We recorded EEG signals in 20 healthy participants in 3 different conditions: MI tasks involving their non-dominant wrist (MI condition); VPI condition; and VPI with MI tasks (combined condition). Each condition lasted 3 minutes, and was repeated 3 times in randomized order. Our main judgment criterion was the Event-Related De-synchronization (ERD) threshold in sensori-motor areas in each condition in the brain motor area.

RESULTS

The combined condition induced a greater change in the ERD percentage than the MI condition alone, but no significant difference was found between the combined and the VPI condition (p = 0.07) and between the VPI and MI condition (p = 0.20).

CONCLUSION

This study demonstrated the interest of using a visuo-proprioceptive immersion with MI rather than MI alone in order to increase excitability in motor areas of the brain. Further studies could test this hypothesis among patients with stroke to provide new perspectives for motor rehabilitation in this population.

Optimizing Motor Intention Detection With Deep Learning: Towards Management of Intraoperative Awareness

OBJECTIVE

This article shows the interest in deep learning techniques to detect motor imagery (MI) from raw electroencephalographic (EEG) signals when a functional electrical stimulation is added or not. Impacts of electrode montages and bandwidth are also reported. The perspective of this work is to improve the detection of intraoperative awareness during general anesthesia.

METHODS

Various architectures of EEGNet were investigated to optimize MI detection. They have been compared to the state-of-the-art classifiers in Brain-Computer Interfaces (based on Riemannian geometry, linear discriminant analysis), and other deep learning architectures (deep convolution network, shallow convolutional network). EEG data were measured from 22 participants performing motor imagery with and without median nerve stimulation.

RESULTS

The proposed architecture of EEGNet reaches the best classification accuracy (83.2%) and false-positive rate (FPR 19.0%) for a setup with only six electrodes over the motor cortex and frontal lobe and for an extended 4-38 Hz EEG frequency range while the subject is being stimulated via a median nerve. Configurations with a larger number of electrodes result in higher accuracy (94.5%) and FPR (6.1%) for 128 electrodes (and respectively 88.0% and 12.9% for 13 electrodes).

CONCLUSION

The present work demonstrates that using an extended EEG frequency band and a modified EEGNet deep neural network increases the accuracy of MI detection when used with as few as 6 electrodes which include frontal channels.

SIGNIFICANCE

The proposed method contributes to the development of Brain-Computer Interface systems based on MI detection from EEG.

Deep Learning Techniques to Improve Intraoperative Awareness Detection from Electroencephalographic Signals

Every year, millions of patients regain conscious- ness during surgery and can potentially suffer from post-traumatic disorders. We recently showed that the detection of motor activity during a median nerve stimulation from electroencephalographic (EEG) signals could be used to alert the medical staff that a patient is waking up and trying to move under general anesthesia [1], [2]. In this work, we measure the accuracy and false positive rate in detecting motor imagery of several deep learning models (EEGNet, deep convolutional network and shallow convolutional network) directly trained on filtered EEG data. We compare them with efficient non-deep approaches, namely, a linear discriminant analysis based on common spatial patterns, the minimum distance to Riemannian mean algorithm applied to covariance matrices, a logistic regression based on a tangent space projection of covariance matrices (TS+LR). The EEGNet improves significantly the classification performance comparing to other classifiers (p- value <; 0.01); moreover it outperforms the best non-deep classifier (TS+LR) for 7.2% of accuracy. This approach promises to improve intraoperative awareness detection during general anesthesia.

Bispectrum-Based Channel Selection for Motor Imagery Based Brain-Computer Interfacing

The performance of motor imagery (MI) based Brain-computer interfacing (BCI) is easily affected by noise and redundant information that exists in the multi-channel electroencephalogram (EEG). To solve this problem, many temporal and spatial feature based channel selection methods have been proposed. However, temporal and spatial features do not accurately reflect changes in the power of the oscillatory EEG. Thus, spectral features of MI-related EEG signals may be useful for channel selection. Bispectrum analysis is a technique developed for extracting non-linear and non-Gaussian information from non-linear and non-Gaussian signals. The features extracted from bispectrum analysis can provide frequency domain information about the EEG. Therefore, in this study, we propose a bispectrum-based channel selection (BCS) method for MI-based BCI. The proposed method uses the sum of logarithmic amplitudes (SLA) and the first order spectral moment (FOSM) features extracted from bispectrum analysis to select EEG channels without redundant information. Three public BCI competition datasets (BCI competition IV dataset 1, BCI competition III dataset IVa and BCI competition III dataset IIIa) were used to validate the effectiveness of our proposed method. The results indicate that our BCS method outperforms use of all channels (83.8% vs 69.4%, 86.3% vs 82.9% and 77.8% vs 68.2%, respectively). Furthermore, compared to the other state-of-the-art methods, our BCS method also can achieve significantly better classification accuracies for MI-based BCI (Wilcoxon signed test, p < 0.05).

Hypnotic State Modulates Sensorimotor Beta Rhythms During Real Movement and Motor Imagery

Hypnosis techniques are currently used in the medical field and directly influences the patient's state of relaxation, perception of the body, and its visual imagination. There is evidence to suggest that a hypnotic state may help patients to better achieve tasks of motor imagination, which is central in the rehabilitation protocols after a stroke. However, the hypnosis techniques could also alter activity in the motor cortex. To the best of our knowledge, the impact of hypnosis on the EEG signal during a movement or an imagined movement is poorly investigated. In particular, how event-related desynchronization (ERD) and event-related synchronization (ERS) patterns would be modulated for different motor tasks may provide a better understanding of the potential benefits of hypnosis for stroke rehabilitation. To investigate this purpose, we recorded EEG signals from 23 healthy volunteers who performed real movements and motor imageries in a closed eye condition. Our results suggest that the state of hypnosis changes the sensorimotor beta rhythm during the ERD phase but maintains the ERS phase in the mu and beta frequency band, suggesting a different activation of the motor cortex in a hypnotized state.

Accidental awareness while under general anaesthesia

Accidental awareness during general anaesthesia may cause many intraoperative discomforts and bring further moderate to severe long-term symptoms including flashbacks, nightmares, hyperarousal or post-traumatic stress disorder. The incidence of awareness varied from 0.017% to 4% among studies. The relatively reliable incidence of intraoperative awareness with postoperative recall is 0.02%. The reason causing awareness was unclear. Insufficient anaesthetic dosing was thought as the principal cause. Even awareness was not comprehensively understood, some endeavors have been raised to prevent or reduce it, including i) Reducing the insufficient anaesthetic dosing induced by negligence; ii) Providing close clinical observation and clinical parameters from the monitor such as bispectral index or electroencephalogram, as well as isolated forearm technique and passive brain-computer interface may bring some effects sometimes. Because current studies still have some flaws, further trials with new detecting approach, superior methodology and underlying aetiology are needed to unfasten the possible factors causing awareness.

MOTANA: study protocol to investigate motor cerebral activity during a propofol sedation

Background

Accidental Accidental awareness during general anesthesia (AAGA) occurs in 1–2% of high-risk practice patients and is a cause of severe psychological trauma, termed post-traumatic stress disorder (PTSD). However, no monitoring techniques can accurately predict or detect AAGA. Since the first reflex for a patient during AAGA is to move, a passive brain-computer interface (BCI) based on the detection of an intention of movement would be conceivable to alert the anesthetist. However, the way in which propofol (i.e., an anesthetic commonly used for the general anesthesia induction) affects motor brain activity within the electroencephalographic (EEG) signal has been poorly investigated and is not clearly understood. For this reason, a detailed study of the motor activity behavior with a step-wise increasing dose of propofol is required and would provide a proof of concept for such an innovative BCI. The main goal of this study is to highlight the occurrence of movement attempt patterns, mainly changes in oscillations called event-related desynchronization (ERD) and event-related synchronization (ERS), in the EEG signal over the motor cortex, in healthy subjects, without and under propofol sedation, during four different motor tasks.

Methods

MOTANA is an interventional, prospective, exploratory, physiological, monocentric, and randomized study conducted in healthy volunteers under light anesthesia, involving EEG measurements before and after target-controlled infusion of propofol at three different effect-site concentrations (0 μg.ml ⁻¹, 0.5 μg.ml ⁻¹, and 1.0 μg.ml ⁻¹). In this exploratory study, 30 healthy volunteers will perform 50 trials for the four motor tasks (real movement, motor imagery, motor imagery with median nerve stimulation, and median nerve stimulation alone) in a randomized sequence. In each conditions and for each trial, we will observe changes in terms of ERD and ERS according to the three propofol concentrations. Pre- and post-injection comparisons of propofol will be performed by paired series tests.

Discussion

MOTANA is an exploratory study aimed at designing an innovative BCI based on EEG-motor brain activity that would detect an attempt to move by a patient under anesthesia. This would be of interest in the prevention of AAGA.

Trial registration

Agence Nationale de Sécurité du Médicament (EUDRACT 2017-004198-1), NCT03362775. Registered on 29 August 2018. https://clinicaltrials.gov/ct2/show/NCT03362775?term=03362775&rank=1

Electronic supplementary material

The online version of this article (10.1186/s13063-019-3596-9) contains supplementary material, which is available to authorized users.