Brain-machine Interface Applications Research Articles

E-FNet: A EEG-fNIRS dual-stream model for Brain-Computer Interfaces

With the rapid advancements in Brain-Computer Interfaces (BCI) and multimodal technology, researchers are actively exploring the intricate domain of multimodal BCI. The integration of electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) in BCI systems has garnered significant attention due to the combination of strengths from both modalities and the overcoming of limitations inherent in standalone systems. In this study, the E-FNet architecture is introduced, a (2+1)D dual-stream model explicitly designed to process multimodal EEG and fNIRS data for brain-machine interface applications. By leveraging the (2+1)D architecture and adopting a bias-free design, E-FNet efficiently processes multimodal data. A unified 4D tensor representation enables the achievement of both temporal and spatial alignment of EEG and fNIRS signals. For Motor Imagery (MI) and Mental Arithmetic (MA) tasks, the early integration of EEG and fNIRS signals yields remarkable accuracy rates of 95.86% and 95.80% respectively, outperforming the accuracy obtained with EEG signals alone. Moreover, the minimal variability in results across subjects underscores the complementary effects achieved through the integration of different signals.

Whole-Head Noninvasive Brain Signal Measurement System with High Temporal and Spatial Resolution Using Static Magnetic Field Bias to the Brain.

Noninvasive brain signal measurement techniques are crucial for understanding human brain function and brain-machine interface applications. Conventionally, noninvasive brain signal measurement techniques, such as electroencephalography, magnetoencephalography, functional magnetic resonance imaging, and near-infrared spectroscopy, have been developed. However, currently, there is no practical noninvasive technique to measure brain function with high temporal and spatial resolution using one instrument. We developed a novel noninvasive brain signal measurement technique with high temporal and spatial resolution by biasing a static magnetic field emitted from a coil on the head to the brain. In this study, we applied this technique to develop a groundbreaking system for noninvasive whole-head brain function measurement with high spatiotemporal resolution across the entire head. We validated this system by measuring movement-related brain signals evoked by a right index finger extension movement and demonstrated that the proposed system can measure the dynamic activity of brain regions involved in finger movement with high spatiotemporal accuracy over the whole brain.

Interface-Mediated Neurogenic Signaling: The Impact of Surface Geometry and Chemistry on Neural Cell Behavior for Regenerative and Brain-Machine Interfacing Applications.

Nanomaterial advancements have driven progress in central and peripheral nervous system applications such as tissue regeneration and brain-machine interfacing. Ideally, neural interfaces with native tissue shall seamlessly integrate, a process that is often mediated by the interfacial material properties. Surface topography and material chemistry are significant extracellular stimuli that can influence neural cell behavior to facilitate tissue integration and augment therapeutic outcomes. This review characterizes topographical modifications, including micropillars, microchannels, surface roughness, and porosity, implemented on regenerative scaffolding and brain-machine interfaces. Their impact on neural cell response is summarized through neurogenic outcome and mechanistic analysis. The effects of surface chemistry on neural cell signaling with common interfacing compounds like carbon-based nanomaterials, conductive polymers, and biologically inspired matrices are also reviewed. Finally, the impact of these extracellular mediated neural cues on intracellular signaling cascades is discussed to provide perspective on the manipulation of neuron and neuroglia cell microenvironments to drive therapeutic outcomes.

Advancements in brain-machine interfaces for application in the metaverse.

In recent years, with the shift of focus in metaverse research toward content exchange and social interaction, breaking through the current bottleneck of audio-visual media interaction has become an urgent issue. The use of brain-machine interfaces for sensory simulation is one of the proposed solutions. Currently, brain-machine interfaces have demonstrated irreplaceable potential as physiological signal acquisition tools in various fields within the metaverse. This study explores three application scenarios: generative art in the metaverse, serious gaming for healthcare in metaverse medicine, and brain-machine interface applications for facial expression synthesis in the virtual society of the metaverse. It investigates existing commercial products and patents (such as MindWave Mobile, GVS, and Galea), draws analogies with the development processes of network security and neurosecurity, bioethics and neuroethics, and discusses the challenges and potential issues that may arise when brain-machine interfaces mature and are widely applied. Furthermore, it looks ahead to the diverse possibilities of deep and varied applications of brain-machine interfaces in the metaverse in the future.

A low-latency and energy-efficient 4-bit absolute value detector for brain-machine interface applications

This research article aims to develop a 4-bit absolute value detector, balancing speed and power efficiency, with potential applications in Brain-Machine Interface (BMI) systems. The detector outputs a binary signal, indicating whether the absolute value of the input surpasses a predefined threshold. The design integrates two primary modules: an absolute value calculator and a comparator. Initially, the study focuses on enhancing the architecture of a multiplexer-based adder for absolute value calculation and selecting an efficient comparator structure, emphasizing least significant bit comparison. Further, the implementation of logic gates using Complementary Metal-Oxide-Semiconductor (CMOS) technology is elaborated. The research concludes by assessing the minimum delay achievable in the critical path, quantified at 74.22 units, and investigating strategies to minimize energy consumption. This is achieved by adjusting gate dimensions and supply voltage, aiming for a delay 1.5 times the minimum. The energy expenditure of the critical path is extrapolated to estimate the overall circuit consumption. The findings demonstrate that, at 1.5 times the minimal delay, the circuit achieves a maximum energy savings of 62.8% with a supply voltage of 0.815V.

Depletion of complement factor 3 delays the neuroinflammatory response to intracortical microelectrodes

The neuroinflammatory response to intracortical microelectrodes (IMEs) used with brain-machine interfacing (BMI) applications is regarded as the primary contributor to poor chronic performance. Recent developments in high-plex gene expression technologies have allowed for an evolution in the investigation of individual proteins or genes to be able to identify specific pathways of upregulated genes that may contribute to the neuroinflammatory response. Several key pathways that are upregulated following IME implantation are involved with the complement system. The complement system is part of the innate immune system involved in recognizing and eliminating pathogens – a significant contributor to the foreign body response against biomaterials. Specifically, we have identified Complement 3 (C3) as a gene of interest because it is the intersection of several key complement pathways. In this study, we investigated the role of C3 in the IME inflammatory response by comparing the neuroinflammatory gene expression at the microelectrode implant site between C3 knockout (C3–/-) and wild-type (WT) mice. We have found that, like in WT mice, implantation of intracortical microelectrodes in C3–/- mice yields a dramatic increase in the neuroinflammatory gene expression at all post-surgery time points investigated. However, compared to WT mice, C3 depletion showed reduced expression of many neuroinflammatory genes pre-surgery and 4 weeks post-surgery. Conversely, depletion of C3 increased the expression of many neuroinflammatory genes at 8 weeks and 16 weeks post-surgery, compared to WT mice. Our results suggest that C3 depletion may be a promising therapeutic target for acute, but not chronic, relief of the neuroinflammatory response to IME implantation. Additional compensatory targets may also be required for comprehensive long-term reduction of the neuroinflammatory response for improved intracortical microelectrode performance.

An Ultraflexible Electrode Array for Large-Scale Chronic Recording in the Nonhuman Primate Brain.

Single-unit (SU) recording in nonhuman primates (NHPs) is indispensible in the quest of how the brain works, yet electrodes currently used for the NHP brain are limited in signal longevity, stability, and spatial coverage. Using new structural materials, microfabrication, and penetration techniques, we develop a mechanically robust ultraflexible, 1 µm thin electrode array (MERF) that enables pial penetration and high-density, large-scale, and chronic recording of neurons along both vertical and horizontal cortical axes in the nonhuman primate brain. Recording from three monkeys yields 2,913 SUs from 1,065 functional recording channels (up to 240 days), with some SUs tracked for up to 2 months. Recording from the primary visual cortex (V1) reveals that neurons with similar orientation preferences for visual stimuli exhibited higher spike correlation. Furthermore, simultaneously recorded neurons in different cortical layers of the primary motor cortex (M1) show preferential firing for hand movements of different directions. Finally, it is shown that a linear decoder trained with neuronal spiking activity across M1 layers during monkey's hand movements can be used to achieve on-line control of cursor movement. Thus, the MERF electrode array offers a new tool for basic neuroscience studies and brain-machine interface (BMI) applications in the primate brain.

EEG-based Mouse Cursor Control

The last decade has seen an increase in the use of artificial intelligence (AI) and machine learning. Recent advances in the field of BC have led to renewed interest in the use of electroencephalography (EEG) for different fields. EEG is used in medical and biomedical applications such as analyzing mental workload and fatigue, diagnosing brain tumors, and rehabilitation of central nervous system disorders; EEG-based movement analysis and classification is widely used in many areas, from clinical applications to brain-machine interface and robotic applications. This article reviews applications of several BC algorithms used in EEG signal processing, introducing commonly used algorithms, typical application scenarios, key advances, and current problems. The study explored current ML applications in EEG, including brain-computer interfaces, cognitive neuroscience, diagnosis of brain disorders. First, the basic principles of ML algorithms used in EEG signal processing, including convolutional neural networks, support vector machines, K-nearest neighbor, and omnidirectional convolutional neural networks, are briefly described. Additionally, a general survey of BC applications used in EEG analysis is presented. As a result, it was determined that SVM methods were used most in the studies, and the study topics were mainly on epilepsy, BCI, and Emotion, and least on Sleep States and Perception.

A Low Noise Neural Recording Frontend IC With Power Management for Closed-Loop Brain-Machine Interface Application.

Brain-machine Interface (BMI) with implantable bioelectronics systems can provide an alternative way to cure neural diseases, while a power management system plays an important role in providing a stable voltage supply for the implanted chip. a prototype system of power management integrated circuit (PMIC) with heavy load capability supplying artifacts tolerable neural recording integrated circuit (ATNR-IC) is presented in this work. A reverse nested miller compensation (RNMC) low dropout regulator (LDO) with a transient enhancer is proposed for the PMIC. The power consumption is 0.55 mW and 22.5 mW at standby (SB) and full stimulation (ST) load, respectively. For a full load transition, the overshoot and downshoot of the LDO are 110 mV and 71 mV, respectively, which help improve the load transient response during neural stimulation. With the load current peak-to-peak range is about 560 μA supplied by a 4-channel stimulator, the whole PMIC can output a stable 3.3 V supply voltage, which indicates that this PMIC can be extended for more stimulating channels' scenarios. When the ATNR-IC is supplied for presented PMIC through a voltage divider network, it can amplify the signal consisting of 1 mVpp simulated neural signal and 20 mVpp simulated artifact by 28 dB with no saturation.

Deployment of an electrocorticography system with a soft robotic actuator.

Electrocorticography (ECoG) is a minimally invasive approach frequently used clinically to map epileptogenic regions of the brain and facilitate lesion resection surgery and increasingly explored in brain-machine interface applications. Current devices display limitations that require trade-offs among cortical surface coverage, spatial electrode resolution, aesthetic, and risk consequences and often limit the use of the mapping technology to the operating room. In this work, we report on a scalable technique for the fabrication of large-area soft robotic electrode arrays and their deployment on the cortex through a square-centimeter burr hole using a pressure-driven actuation mechanism called eversion. The deployable system consists of up to six prefolded soft legs, and it is placed subdurally on the cortex using an aqueous pressurized solution and secured to the pedestal on the rim of the small craniotomy. Each leg contains soft, microfabricated electrodes and strain sensors for real-time deployment monitoring. In a proof-of-concept acute surgery, a soft robotic electrode array was successfully deployed on the cortex of a minipig to record sensory cortical activity. This soft robotic neurotechnology opens promising avenues for minimally invasive cortical surgery and applications related to neurological disorders such as motor and sensory deficits.

Intraarterial encephalography from an acutely implanted aneurysm embolization device in awake humans.

Endovascular electroencephalography (evEEG) uses the cerebrovascular system to record electrical activity from adjacent neural structures. The safety, feasibility, and efficacy of using the Woven EndoBridge Aneurysm Embolization System (WEB) for evEEG has not been investigated. Seventeen participants undergoing awake WEB endovascular treatment of unruptured cerebral aneurysms were included. After WEB deployment and before detachment, its distal deployment wire was connected to an EEG receiver, and participants performed a decision-making task for 10 minutes. WEB and scalp recordings were captured. All patients underwent successful embolization and evEEG with no complications. Event-related potentials were detected on scalp EEG in 9/17 (53%) patients. Of these 9 patients, a task-related low-gamma (30-70 Hz) response on WEB channels was captured in 8/9 (89%) cases. In these 8 patients, the WEB was deployed in 2 middle cerebral arteries, 3 anterior communicating arteries, the terminal internal carotid artery, and 2 basilar tip aneurysms. Electrocardiogram artifact on WEB channels was present in 12/17 cases. The WEB implanted within cerebral aneurysms of awake patients is capable of capturing task-specific brain electrical activities. Future studies are warranted to establish the efficacy of and support for evEEG as a tool for brain recording, brain stimulation, and brain-machine interface applications.

Thin flexible arrays for long-term multi-electrode recordings in macaque primary visual cortex

Objective. Basic, translational and clinical neuroscience are increasingly focusing on large-scale invasive recordings of neuronal activity. However, in large animals such as nonhuman primates and humans—in which the larger brain size with sulci and gyri imposes additional challenges compared to rodents, there is a huge unmet need to record from hundreds of neurons simultaneously anywhere in the brain for long periods of time. Here, we tested the electrical and mechanical properties of thin, flexible multi-electrode arrays (MEAs) inserted into the primary visual cortex of two macaque monkeys, and assessed their magnetic resonance imaging (MRI) compatibility and their capacity to record extracellular activity over a period of 1 year. Approach. To allow insertion of the floating arrays into the visual cortex, the 20 by 100 µm2 shafts were temporarily strengthened by means of a resorbable poly(lactic-co-glycolic acid) coating. Main results. After manual insertion of the arrays, the ex vivo and in vivo MRI compatibility of the arrays proved to be excellent. We recorded clear single-unit activity from up to 50% of the electrodes, and multi-unit activity (MUA) on 60%–100% of the electrodes, which allowed detailed measurements of the receptive fields and the orientation selectivity of the neurons. Even 1 year after insertion, we obtained significant MUA responses on 70%–100% of the electrodes, while the receptive fields remained remarkably stable over the entire recording period. Significance. Thus, the thin and flexible MEAs we tested offer several crucial advantages compared to existing arrays, most notably in terms of brain tissue compliance, scalability, and brain coverage. Future brain-machine interface applications in humans may strongly benefit from this new generation of chronically implanted MEAs.

High Classification Accuracy of Touch Locations from S1 LFPs Using CNNs and Fastai.

The primary somatosensory cortex (S1) is a region often targeted for input via somatosensory neuroprosthesis as tactile and proprioception are represented in S1. How this information is represented is an ongoing area of research. Neural signals are high-dimensional, making accurate models for decoding a significant challenge. Artificial neural networks (ANNs) have proven efficient at classification tasks in multiple fields. Moreover, ANNs allow for transfer learning, which exploits feature extraction trained on a large and more general dataset than may be available for a particular problem. In this work, convolutional neural networks (CNN), used for image recognition, were fine-tuned with somatosensory cortical recordings during experiments with naturalistic touch stimuli. We created a highly accurate ( correct) classifier for cutaneous stimulation locations as part of a somatosensory neuroprosthesis pipeline. Here we present the classifier results. Clinical Relevance- Our work provides a method for classifying cortical activity in brain-machine interface applications, specifically towards somatosensory neuroprosthetics.

EEG Sinyallerini İşlemek İçin Makine Öğreniminin Kullanıldığı Konular Üzerine Bir İnceleme

Son on yılda, yapay zekâ (YZ) ve makine öğrenimi (MÖ) kullanımlarında bir artış görülmüştür. MÖ alanındaki son gelişmeler, farklı alanlar için elektroensefalografinin (EEG) kullanımına yeniden ilgi duyulmasına yol açmıştır. EEG, zihinsel iş yükünü ve yorgunluğu analiz etmek, beyin tümörlerini teşhis etmek ve merkezi sinir sistemi bozukluklarının rehabilitasyonu gibi tıbbi ve biyomedikal uygulamalarda; EEG tabanlı hareket analizi ve sınıflandırması ise, klinik uygulamalardan beyin-makine ara yüzüne ve robotik uygulamalara kadar birçok alanda yaygın olarak kullanılmaktadır. Bu makale, EEG sinyal işlemede kullanılan birçok MÖ algoritmalarının uygulamalarını gözden geçirmekte, yaygın olarak kullanılan algoritmaları, tipik uygulama senaryolarını, önemli ilerlemeleri ve mevcut sorunları tanıtmaktadır. Çalışmada, beyin-bilgisayar arayüzleri, bilişsel sinirbilim, beyin bozukluklarının teşhisi ve daha farklı konular dahil olmak üzere, EEG'deki mevcut MÖ uygulamaları araştırılmıştır. İlk olarak, evrişimli sinir ağı, destek vektör makineleri, K-en yakın komşu ve çok yönlü evrişim sinir ağı dahil olmak üzere EEG sinyal işlemede kullanılan MÖ algoritmalarının temel ilkeleri kısaca açıklanmıştır. Ayrıca EEG analizinde kullanılan MÖ uygulamalarına dair genel bir araştırma sunulmuştur. Sonuç olarak çalışmalarda en fazla DVM ve CNN yöntemlerinin kullanıldığı, çalışma başlıklarının ise ağırlıklı olarak epilepsi, BCI ve Duygu konularında en az ise Alkol, Uyku Durumları, Algı konularında yapıldığı belirlenmiştir.

Bridging the gap between patient-specific and patient-independent seizure prediction via knowledge distillation

Objective. Deep neural networks (DNNs) have shown unprecedented success in various brain-machine interface applications such as epileptic seizure prediction. However, existing approaches typically train models in a patient-specific fashion due to the highly personalized characteristics of epileptic signals. Therefore, only a limited number of labeled recordings from each subject can be used for training. As a consequence, current DNN based methods demonstrate poor generalization ability to some extent due to the insufficiency of training data. On the other hand, patient-independent models attempt to utilize more patient data to train a universal model for all patients by pooling patient data together. Despite different techniques applied, results show that patient-independent models perform worse than patient-specific models due to high individual variation across patients. A substantial gap thus exists between patient-specific and patient-independent models. Approach. In this paper, we propose a novel training scheme based on knowledge distillation which makes use of a large amount of data from multiple subjects. It first distills informative features from signals of all available subjects with a pre-trained general model. A patient-specific model can then be obtained with the help of distilled knowledge and additional personalized data. Main results. Four state-of-the-art seizure prediction methods are trained on the Children’s Hospital of Boston-MIT sEEG database with our proposed scheme. The resulting accuracy, sensitivity, and false prediction rate show that our proposed training scheme consistently improves the prediction performance of state-of-the-art methods by a large margin. Significance. The proposed training scheme significantly improves the performance of patient-specific seizure predictors and bridges the gap between patient-specific and patient-independent predictors.

Getting a grasp on BMIs: Decoding prehension and speech signals

Wandelt etal. (2022) show that different grasps can be decoded from neural activity in the human supramarginal gyrus (SMG), ventral premotor cortex, and somatosensory cortex during motor imagery and speech, highlighting the attractiveness of higher-level areas such as the SMG for brain-machine interface applications.

Perovskite nickelate ionotronics for AI and brain-machine interfaces

Human brain is the ultimate computing machine in nature. Creating brain-like devices that emulate how the brain works and can communicate with the brain is crucial for fabricating highly efficient computing circuits, monitoring the onset of diseases at early stages, and transferring information across brain-machine interfaces. Simultaneous transduction of ionic-electronic signals would be of particular interest in this context since ionic transmitters are the means of information transfer in human brain while traditional electronics utilize electrons or holes. In this perspective, we propose strongly correlated oxides (mainly focused on perovskite nickelates) as potential candidates for this purpose. The capability of reversibly accepting small ions and converting ionic signal to electrical signals renders perovskite nickelates strong candidates for neuromorphic computing and bioelectrical applications. We will discuss the mechanism behind the interplay between ionic doping and the resistivity modulation in perovskite nickelates. We will also present case studies of using the perovskite nickelates in neuromorphic computing and brain-machine interface applications. We then conclude by pointing out the challenges in this field and provide our perspectives. We hope the utilization of strong electron correlation in the perovskite nickelates will provide exciting new opportunities for future computation devices and brain-machine interfaces.

Investigation of the Feasibility of Ventricular Delivery of Resveratrol to the Microelectrode Tissue Interface.

(1) Background: Intracortical microelectrodes (IMEs) are essential to basic brain research and clinical brain–machine interfacing applications. However, the foreign body response to IMEs results in chronic inflammation and an increase in levels of reactive oxygen and nitrogen species (ROS/RNS). The current study builds on our previous work, by testing a new delivery method of a promising antioxidant as a means of extending intracortical microelectrodes performance. While resveratrol has shown efficacy in improving tissue response, chronic delivery has proven difficult because of its low solubility in water and low bioavailability due to extensive first pass metabolism. (2) Methods: Investigation of an intraventricular delivery of resveratrol in rats was performed herein to circumvent bioavailability hurdles of resveratrol delivery to the brain. (3) Results: Intraventricular delivery of resveratrol in rats delivered resveratrol to the electrode interface. However, intraventricular delivery did not have a significant impact on electrophysiological recordings over the six-week study. Histological findings indicated that rats receiving intraventricular delivery of resveratrol had a decrease of oxidative stress, yet other biomarkers of inflammation were found to be not significantly different from control groups. However, investigation of the bioavailability of resveratrol indicated a decrease in resveratrol accumulation in the brain with time coupled with inconsistent drug elution from the cannulas. Further inspection showed that there may be tissue or cellular debris clogging the cannulas, resulting in variable elution, which may have impacted the results of the study. (4) Conclusions: These results indicate that the intraventricular delivery approach described herein needs further optimization, or may not be well suited for this application.

Investigation of dual-band antenna with low-SAR characteristics for bidirectional brain-machine interface applications

In this paper, we propose an innovative dual-band printed monopole antenna (PMA) for bidirectional brain-machine interface (Bi-BMI) application. The proposed antenna configuration contains of a square radiator with a pair of rectangular slots and an inverted U-shaped slot and a truncated ground plane with a pair of L-shaped slots, which provides two usable radiating bands for down-link data transferring around 2.3–2.5 GHz and for up-link data transferring 3.21–9.03 GHz. By inserting two rectangular slots surrounded by a modified inverted U-shaped slot with modified dimensions on the radiating square patch and also by cutting two L-shaped slots, new resonances are created and therefore much wider impedance bandwidth can be generated at both lower and higher bands. Despite recently reported antenna designs in the literature, the proposed antenna demonstrates wideband radiative characteristics even at higher frequencies, and also the specific absorption rate (SAR) level in 1 mW input power case is satisfactory less than 0.4 W/kg in the whole radiating band. The designed antenna has a compact configuration and small size. Simulated and measured S-parameter and radiation patterns results are demonstrated to confirm the utility of the designed antenna configuration for brain machine interface application.

Impact of referencing scheme on decoding performance of LFP-based brain-machine interface

Objective. There has recently been an increasing interest in local field potential (LFP) for brain-machine interface (BMI) applications due to its desirable properties (signal stability and low bandwidth). LFP is typically recorded with respect to a single unipolar reference which is susceptible to common noise. Several referencing schemes have been proposed to eliminate the common noise, such as bipolar reference, current source density (CSD), and common average reference (CAR). However, to date, there have not been any studies to investigate the impact of these referencing schemes on decoding performance of LFP-based BMIs. Approach. To address this issue, we comprehensively examined the impact of different referencing schemes and LFP features on the performance of hand kinematic decoding using a deep learning method. We used LFPs chronically recorded from the motor cortex area of a monkey while performing reaching tasks. Main results. Experimental results revealed that local motor potential (LMP) emerged as the most informative feature regardless of the referencing schemes. Using LMP as the feature, CAR was found to yield consistently better decoding performance than other referencing schemes over long-term recording sessions. Significance. Overall, our results suggest the potential use of LMP coupled with CAR for enhancing the decoding performance of LFP-based BMIs.