Bidirectional brain-computer interfaces

Alzheimer's disease and other memory disorders in the age of AI: reflection and perspectives on the 120th anniversary of the birth of Dr. John von Neumann

Two themes are coming to the forefront in this decade: Cognitive impairment of an aging population and the quantum leap in developing artificial intelligence (AI). Both can be described as growing exponentially and presenting serious challenges. Although many questions have been addressed about the dangers of AI, we want to go beyond the fearful aspects of this topic and focus on the possible contribution of AI to solve the problem of chronic disorders of the elderly leading to cognitive impairment, like Alzheimer's disease, Parkinson's disease, and Lewy body dementia. Our second goal is to look at the ways in which modern neuroscience can influence the future design of computers and the development of AI. We wish to honor the memory of Dr. John von Neumann, who came up with many breakthrough details of the first electronic computer. Remarkably, Dr. von Neumann dedicated his last book to the comparison of the human brain and the computer as it stood in those years of the mid-1950s. We will point out how his ideas are more relevant than ever in the age of supercomputers, AI and brain implants.

Exploring Machine Learning Classification of Movement Phases in Hemiparetic Stroke Patients: A Controlled EEG-tDCS Study

BACKGROUND/OBJECTIVES

Noninvasive brain stimulation (NIBS) can boost motor recovery after a stroke. Certain movement phases are more responsive to NIBS, so a system that auto-detects these phases would optimize stimulation timing. This study assessed the effectiveness of various machine learning models in identifying movement phases in hemiparetic individuals undergoing simultaneous NIBS and EEG recordings. We hypothesized that transcranial direct current stimulation (tDCS), a form of NIBS, would enhance EEG signals related to movement phases and improve classification accuracy compared to sham stimulation.

METHODS

EEG data from 10 chronic stroke patients and 11 healthy controls were recorded before, during, and after tDCS. Eight machine learning algorithms and five ensemble methods were used to classify two movement phases (hold posture and reaching) during each of these periods. Data preprocessing included z-score normalization and frequency band power binning.

RESULTS

In chronic stroke participants who received active tDCS, the classification accuracy for hold vs. reach phases increased from pre-stimulation to the late intra-stimulation period (72.2% to 75.2%, p < 0.0001). Late active tDCS surpassed late sham tDCS classification (75.2% vs. 71.5%, p < 0.0001). Linear discriminant analysis was the most accurate (74.6%) algorithm with the shortest training time (0.9 s). Among ensemble methods, low gamma frequency (30-50 Hz) achieved the highest accuracy (74.5%), although this result did not achieve statistical significance for actively stimulated chronic stroke participants.

CONCLUSIONS

Machine learning algorithms showed enhanced movement phase classification during active tDCS in chronic stroke participants. These results suggest their feasibility for real-time movement detection in neurorehabilitation, including brain-computer interfaces for stroke recovery.

Brain-computer Interaction in the Smart Era

The brain-computer interface (BCI) system serves as a critical link between external output devices and the human brain. A monitored object's mental state, sensory cognition, and even higher cognition are reflected in its electroencephalography (EEG) signal. Nevertheless, unprocessed EEG signals are frequently contaminated with a variety of artifacts, rendering the analysis and elimination of impurities from the collected EEG data exceedingly challenging, not to mention the manual adjustment thereof. Over the last few decades, the rapid advancement of artificial intelligence (AI) technology has contributed to the development of BCI technology. Algorithms derived from AI and machine learning have significantly enhanced the ability to analyze and process EEG electrical signals, thereby expanding the range of potential interactions between the human brain and computers. As a result, the present BCI technology with the help of AI can assist physicians in gaining a more comprehensive understanding of their patients' physical and psychological status, thereby contributing to improvements in their health and quality of life.

Electroencephalography-Based Brain-Computer Interfaces in Rehabilitation: A Bibliometric Analysis (2013-2023)

EEG-based Brain-Computer Interfaces (BCIs) have gained significant attention in rehabilitation due to their non-invasive, accessible ability to capture brain activity and restore neurological functions in patients with conditions such as stroke and spinal cord injuries. This study offers a comprehensive bibliometric analysis of global EEG-based BCI research in rehabilitation from 2013 to 2023. It focuses on primary research and review articles addressing technological innovations, effectiveness, and system advancements in clinical rehabilitation. Data were sourced from databases like Web of Science, and bibliometric tools (bibliometrix R) were used to analyze publication trends, geographic distribution, keyword co-occurrences, and collaboration networks. The results reveal a rapid increase in EEG-BCI research, peaking in 2022, with a primary focus on motor and sensory rehabilitation. EEG remains the most commonly used method, with significant contributions from Asia, Europe, and North America. Additionally, there is growing interest in applying BCIs to mental health, as well as integrating artificial intelligence (AI), particularly machine learning, to enhance system accuracy and adaptability. However, challenges remain, such as system inefficiencies and slow learning curves. These could be addressed by incorporating multi-modal approaches and advanced neuroimaging technologies. Further research is needed to validate the applicability of EEG-BCI advancements in both cognitive and motor rehabilitation, especially considering the high global prevalence of cerebrovascular diseases. To advance the field, expanding global participation, particularly in underrepresented regions like Latin America, is essential. Improving system efficiency through multi-modal approaches and AI integration is also critical. Ethical considerations, including data privacy, transparency, and equitable access to BCI technologies, must be prioritized to ensure the inclusive development and use of these technologies across diverse socioeconomic groups.

Clinical Applications and Future Translation of Somatosensory Neuroprostheses

Somatosensory neuroprostheses restore, replace, or enhance tactile and proprioceptive feedback for people with sensory impairments due to neurological disorders or injury. Somatosensory neuroprostheses typically couple sensor inputs from a wearable device, prosthesis, robotic device, or virtual reality system with electrical stimulation applied to the somatosensory nervous system via noninvasive or implanted interfaces. While prior research has mainly focused on technology development and proof-of-concept studies, recent acceleration of clinical studies in this area demonstrates the translational potential of somatosensory neuroprosthetic systems. In this review, we provide an overview of neurostimulation approaches currently undergoing human testing and summarize recent clinical findings on the perceptual, functional, and psychological impact of somatosensory neuroprostheses. We also cover current work toward the development of advanced stimulation paradigms to produce more natural and informative sensory feedback. Finally, we provide our perspective on the remaining challenges that need to be addressed prior to translation of somatosensory neuroprostheses.

Neural interfaces: Bridging the brain to the world beyond healthcare

Neural interfaces, emerging at the intersection of neurotechnology and urban planning, promise to transform how we interact with our surroundings and communicate. By recording and decoding neural signals, these interfaces facilitate direct connections between the brain and external devices, enabling seamless information exchange and shared experiences. Nevertheless, their development is challenged by complexities in materials science, electrochemistry, and algorithmic design. Electrophysiological crosstalk and the mismatch between electrode rigidity and tissue flexibility further complicate signal fidelity and biocompatibility. Recent closed-loop brain-computer interfaces, while promising for mood regulation and cognitive enhancement, are limited by decoding accuracy and the adaptability of user interfaces. This perspective outlines these challenges and discusses the progress in neural interfaces, contrasting non-invasive and invasive approaches, and explores the dynamics between stimulation and direct interfacing. Emphasis is placed on applications beyond healthcare, highlighting the need for implantable interfaces with high-resolution recording and stimulation capabilities.

Noninvasive brain stimulation during EEG improves machine learning classication in chronic stroke

Background

In individuals with chronic stroke and hemiparesis, noninvasive brain stimulation (NIBS) may be used as an adjunct to therapy for improving motor recovery. Specific states of movement during motor recovery are more responsive to brain stimulation than others, thus a system that could auto-detect movement state would be useful in correctly identifying the most effective stimulation periods. The aim of this study was to compare the performance of different machine learning models in classifying movement periods during EEG recordings of hemiparetic individuals receiving noninvasive brain stimulation. We hypothesized that transcranial direct current stimulation, a form of NIBS, would modulate brain recordings correlating with movement state and improve classification accuracies above those receiving sham stimulation.

Methods

Electroencephalogram data were obtained from 10 participants with chronic stroke and 11 healthy individuals performing a motor task while undergoing transcranial direct current stimulation. Eight traditional machine learning algorithms and five ensemble methods were used to classify two movement states (a hold posture and an arm reaching movement) before, during and after stimulation. To minimize compute times, preprocessing and feature extraction were limited to z-score normalization and power binning into five frequency bands (delta through gamma).

Results

Classification of disease state produced significantly higher accuracies in the stimulation (versus sham) group at 78.9% (versus 55.6%, p < 0.000002). We observed significantly higher accuracies when classifying stimulation state in the chronic stroke group (77.6%) relative to healthy controls (64.1%, p < 0.0095). In the chronic stroke cohort, classification of hold versus reach was highest during the stimulation period (75.2%) as opposed to the pre- and post-stimulation periods. Linear discriminant analysis, logistic regression, and decision tree algorithms classified movement state most accurately in participants with chronic stroke during the stimulation period (76.1%). For the ensemble methods, the highest classification accuracy for hold versus reach was achieved using low gamma frequency (30-50 Hz) as a feature (74.5%), although this result did not achieve statistical significance.

Conclusions

Machine learning algorithms demonstrated sufficiently high movement state classification accuracy in participants with chronic stroke performing functional tasks during noninvasive brain stimulation. tDCS improved disease state and movement state classification in participants with chronic stroke.

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.

Considerations and discussions on the clear definition and definite scope of brain-computer interfaces

Brain-computer interface (BCI) is a revolutionizing human-computer interaction with potential applications in both medical and non-medical fields, emerging as a cutting-edge and trending research direction. Increasing numbers of groups are engaging in BCI research and development. However, in recent years, there has been some confusion regarding BCI, including misleading and hyped propaganda about BCI, and even non-BCI technologies being labeled as BCI. Therefore, a clear definition and a definite scope for BCI are thoroughly considered and discussed in the paper, based on the existing definitions of BCI, including the six key or essential components of BCI. In the review, different from previous definitions of BCI, BCI paradigms and neural coding are explicitly included in the clear definition of BCI provided, and the BCI user (the brain) is clearly identified as a key component of the BCI system. Different people may have different viewpoints on the definition and scope of BCI, as well as some related issues, which are discussed in the article. This review argues that a clear definition and definite scope of BCI will benefit future research and commercial applications. It is hoped that this review will reduce some of the confusion surrounding BCI and promote sustainable development in this field.

The evolution of neuromodulation for chronic stroke: From neuroplasticity mechanisms to brain-computer interfaces

Stroke is one of the most common and debilitating neurological conditions worldwide. Those who survive experience motor, sensory, speech, vision, and/or cognitive deficits that severely limit remaining quality of life. While rehabilitation programs can help improve patients' symptoms, recovery is often limited, and patients frequently continue to experience impairments in functional status. In this review, invasive neuromodulation techniques to augment the effects of conventional rehabilitation methods are described, including vagus nerve stimulation (VNS), deep brain stimulation (DBS) and brain-computer interfaces (BCIs). In addition, the evidence base for each of these techniques, pivotal trials, and future directions are explored. Finally, emerging technologies such as functional near-infrared spectroscopy (fNIRS) and the shift to artificial intelligence-enabled implants and wearables are examined. While the field of implantable devices for chronic stroke recovery is still in a nascent stage, the data reviewed are suggestive of immense potential for reducing the impact and impairment from this globally prevalent disorder.

Brain-Computer Interfaces with Intracortical Implants for Motor and Communication Functions Compensation: Review of Recent Developments

Brain-computer interfaces allow the exchange of data between the brain and an external device, bypassing the muscular system. Clinical studies of invasive brain-computer interface technologies have been conducted for over 20 years. During this time, there has been a continuous improvement of approaches to neuronal signal processing in order to improve the quality of control of external devices. Currently, brain-computer interfaces with intracortical implants allow completely paralyzed patients to control robotic limbs for self-service, use a computer or a tablet, type text, and reproduce speech at an optimal speed. Studies of invasive brain-computer interfaces regularly provide new fundamental data on functioning of the central nervous system. In recent years, breakthrough discoveries and achievements have been annually made in this sphere. This review analyzes the results of clinical experiments of brain-computer interfaces with intracortical implants, provides information on the stages of this technology development, its main discoveries and achievements.

Multifunctional Fiber-Based Optoacoustic Emitter as a Bidirectional Brain Interface

A bidirectional brain interface with both "write" and "read" functions can be an important tool for fundamental studies and potential clinical treatments for neurological diseases. Herein, a miniaturized multifunctional fiber-based optoacoustic emitter (mFOE) is reported thatintegrates simultaneous optoacoustic stimulation for "write" and electrophysiology recording of neural circuits for "read". Because of the intrinsic ability of neurons to respond to acoustic wave, there is no requirement of the viral transfection. The orthogonality between optoacoustic waves and electrical field provides a solution to avoid the interference between electrical stimulation and recording. The stimulation function of the mFOE is first validated in cultured ratcortical neurons using calcium imaging. In vivo application of mFOE for successful simultaneous optoacoustic stimulation and electrical recording of brain activities is confirmed in mouse hippocampus in both acute and chronical applications up to 1 month. Minor brain tissue damage is confirmed after these applications. The capability of simultaneous neural stimulation and recording enabled by mFOE opens up new possibilities for the investigation of neural circuits and brings new insights into the study of ultrasound neurostimulation.

Suppression of cortical electrostimulation artifacts using pre-whitening and null projection

Objective.Invasive brain-computer interfaces (BCIs) have shown promise in restoring motor function to those paralyzed by neurological injuries. These systems also have the ability to restore sensation via cortical electrostimulation. Cortical stimulation produces strong artifacts that can obscure neural signals or saturate recording amplifiers. While front-end hardware techniques can alleviate this problem, residual artifacts generally persist and must be suppressed by back-end methods.Approach.We have developed a technique based on pre-whitening and null projection (PWNP) and tested its ability to suppress stimulation artifacts in electroencephalogram (EEG), electrocorticogram (ECoG) and microelectrode array (MEA) signals from five human subjects.Main results.In EEG signals contaminated by narrow-band stimulation artifacts, the PWNP method achieved average artifact suppression between 32 and 34 dB, as measured by an increase in signal-to-interference ratio. In ECoG and MEA signals contaminated by broadband stimulation artifacts, our method suppressed artifacts by 78%-80% and 85%, respectively, as measured by a reduction in interference index. When compared to independent component analysis, which is considered the state-of-the-art technique for artifact suppression, our method achieved superior results, while being significantly easier to implement.Significance.PWNP can potentially act as an efficient method of artifact suppression to enable simultaneous stimulation and recording in bi-directional BCIs to biomimetically restore motor function.

A Real-Time Non-Implantation Bi-Directional Brain-Computer Interface Solution Without Stimulation Artifacts

The non-implantation bi-directional brain-computer interface (BCI) is a neural interface technology that enables direct two-way communication between the brain and the external world by both "reading" neural signals and "writing" stimulation patterns to the brain. This technology has vast potential applications, such as improving the quality of life for individuals with neurological and mental illnesses and even expanding the boundaries of human capabilities. Nonetheless, non-implantation bi-directional BCIs face challenges in generating real-time feedback and achieving compatibility between stimulation and recording. These issues arise due to the considerable overlap between electrical stimulation frequencies and electrophysiological recording frequencies, as well as the impediment caused by the skull to the interaction of external and internal currents. To address those challenges, this work proposes a novel solution that combines the temporal interference stimulation paradigm and minimally invasive skull modification. A longitudinal animal experiment has preliminarily validated the feasibility of the proposed method. In signal recording experiments, the average impedance of our scheme decreased by 4.59 kΩ , about 67%, compared to the conventional technique at 18 points. The peak-to-peak value of the Somatosensory Evoked Potential increased by 8%. Meanwhile, the signal-to-noise ratio of Steady-State Visual Evoked Potential increased by 5.13 dB, and its classification accuracy increased by 44%. The maximum bandwidth of the resting state rose by 63%. In electrical stimulation experiments, the signal-to-noise ratio of the low-frequency response evoked by our scheme rose by 8.04 dB, and no stimulation artifacts were generated. The experimental results show that signal quality in acquisition has significantly improved, and frequency-band isolation eliminates stimulation artifacts at the source. The acquisition and stimulation pathways are real-time compatible in this non-implantation bi-directional BCI solution, which can provide technical support and theoretical guidance for creating closed-loop adaptive systems coupled with particular application scenarios in the future.

Dynamic amplitude modulation of microstimulation evokes biomimetic onset and offset transients and reduces depression of evoked calcium responses in sensory cortices

BACKGROUND

Intracortical microstimulation (ICMS) is an emerging approach to restore sensation to people with neurological injury or disease. Biomimetic microstimulation, or stimulus trains that mimic neural activity in the brain through encoding of onset and offset transients, could improve the utility of ICMS for brain-computer interface (BCI) applications, but how biomimetic microstimulation affects neural activation is not understood. Current "biomimetic" ICMS trains aim to reproduce the strong onset and offset transients evoked in the brain by sensory input through dynamic modulation of stimulus parameters. Stimulus induced depression of neural activity (decreases in evoked intensity over time) is also a potential barrier to clinical implementation of sensory feedback, and dynamic microstimulation may reduce this effect.

OBJECTIVE

We evaluated how bio-inspired ICMS trains with dynamic modulation of amplitude and/or frequency change the calcium response, spatial distribution, and depression of neurons in the somatosensory and visual cortices.

METHODS

Calcium responses of neurons were measured in Layer 2/3 of visual and somatosensory cortices of anesthetized GCaMP6s mice in response to ICMS trains with fixed amplitude and frequency (Fixed) and three dynamic ICMS trains that increased the stimulation intensity during the onset and offset of stimulation by modulating the amplitude (DynAmp), frequency (DynFreq), or amplitude and frequency (DynBoth). ICMS was provided for either 1-s with 4-s breaks (Short) or for 30-s with 15-s breaks (Long).

RESULTS

DynAmp and DynBoth trains evoked distinct onset and offset transients in recruited neural populations, while DynFreq trains evoked population activity similar to Fixed trains. Individual neurons had heterogeneous responses primarily based on how quickly they depressed to ICMS, where neurons farther from the electrode depressed faster and a small subpopulation (1-5%) were modulated by DynFreq trains. Neurons that depressed to Short trains were also more likely to depress to Long trains, but Long trains induced more depression overall due to the increased stimulation length. Increasing the amplitude during the hold phase resulted in an increase in recruitment and intensity which resulted in more depression and reduced offset responses. Dynamic amplitude modulation reduced stimulation induced depression by 14.6 ± 0.3% for Short and 36.1 ± 0.6% for Long trains. Ideal observers were 0.031 ± 0.009 s faster for onset detection and 1.33 ± 0.21 s faster for offset detection with dynamic amplitude encoding.

CONCLUSIONS

Dynamic amplitude modulation evokes distinct onset and offset transients, reduces depression of neural calcium activity, and decreases total charge injection for sensory feedback in BCIs by lowering recruitment of neurons during long maintained periods of ICMS. In contrast, dynamic frequency modulation evokes distinct onset and offset transients in a small subpopulation of neurons but also reduces depression in recruited neurons by reducing the rate of activation.

Modulating Brain Activity with Invasive Brain-Computer Interface: A Narrative Review

Brain-computer interface (BCI) can be used as a real-time bidirectional information gateway between the brain and machines. In particular, rapid progress in invasive BCI, propelled by recent developments in electrode materials, miniature and power-efficient electronics, and neural signal decoding technologies has attracted wide attention. In this review, we first introduce the concepts of neuronal signal decoding and encoding that are fundamental for information exchanges in BCI. Then, we review the history and recent advances in invasive BCI, particularly through studies using neural signals for controlling external devices on one hand, and modulating brain activity on the other hand. Specifically, regarding modulating brain activity, we focus on two types of techniques, applying electrical stimulation to cortical and deep brain tissues, respectively. Finally, we discuss the related ethical issues concerning the clinical application of this emerging technology.

Application of a Brain-Computer Interface System with Visual and Motor Feedback in Limb and Brain Functional Rehabilitation after Stroke: Case Report

(1) Objective: To investigate the feasibility, safety, and effectiveness of a brain-computer interface (BCI) system with visual and motor feedback in limb and brain function rehabilitation after stroke. (2) Methods: First, we recruited three hemiplegic stroke patients to perform rehabilitation training using a BCI system with visual and motor feedback for two consecutive days (four sessions) to verify the feasibility and safety of the system. Then, we recruited five other hemiplegic stroke patients for rehabilitation training (6 days a week, lasting for 12-14 days) using the same BCI system to verify the effectiveness. The mean and Cohen's w were used to compare the changes in limb motor and brain functions before and after training. (3) Results: In the feasibility verification, the continuous motor state switching time (CMSST) of the three patients was 17.8 ± 21.0s, and the motor state percentages (MSPs) in the upper and lower limb training were 52.6 ± 25.7% and 72.4 ± 24.0%, respectively. The effective training revolutions (ETRs) per minute were 25.8 ± 13.0 for upper limb and 24.8 ± 6.4 for lower limb. There were no adverse events during the training process. Compared with the baseline, the motor function indices of the five patients were improved, including sitting balance ability, upper limb Fugel-Meyer assessment (FMA), lower limb FMA, 6 min walking distance, modified Barthel index, and root mean square (RMS) value of triceps surae, which increased by 0.4, 8.0, 5.4, 11.4, 7.0, and 0.9, respectively, and all had large effect sizes (Cohen's w ≥ 0.5). The brain function indices of the five patients, including the amplitudes of the motor evoked potentials (MEP) on the non-lesion side and lesion side, increased by 3.6 and 3.7, respectively; the latency of MEP on the non-lesion side was shortened by 2.6 ms, and all had large effect sizes (Cohen's w ≥ 0.5). (4) Conclusions: The BCI system with visual and motor feedback is applicable in active rehabilitation training of stroke patients with hemiplegia, and the pilot results show potential multidimensional benefits after a short course of treatment.

Explant Analysis of Utah Electrode Arrays Implanted in Human Cortex for Brain-Computer-Interfaces

Brain-computer interfaces are being developed to restore movement for people living with paralysis due to injury or disease. Although the therapeutic potential is great, long-term stability of the interface is critical for widespread clinical implementation. While many factors can affect recording and stimulation performance including electrode material stability and host tissue reaction, these factors have not been investigated in human implants. In this clinical study, we sought to characterize the material integrity and biological tissue encapsulation via explant analysis in an effort to identify factors that influence electrophysiological performance. We examined a total of six Utah arrays explanted from two human participants involved in intracortical BCI studies. Two platinum (Pt) arrays were implanted for 980 days in one participant (P1) and two Pt and two iridium oxide (IrOx) arrays were implanted for 182 days in the second participant (P2). We observed that the recording quality followed a similar trend in all six arrays with an initial increase in peak-to-peak voltage during the first 30–40 days and gradual decline thereafter in P1. Using optical and two-photon microscopy we observed a higher degree of tissue encapsulation on both arrays implanted for longer durations in participant P1. We then used scanning electron microscopy and energy dispersive X-ray spectroscopy to assess material degradation. All measures of material degradation for the Pt arrays were found to be more prominent in the participant with a longer implantation time. Two IrOx arrays were subjected to brief survey stimulations, and one of these arrays showed loss of iridium from most of the stimulated sites. Recording performance appeared to be unaffected by this loss of iridium, suggesting that the adhesion of IrOx coating may have been compromised by the stimulation, but the metal layer did not detach until or after array removal. In summary, both tissue encapsulation and material degradation were more pronounced in the arrays that were implanted for a longer duration. Additionally, these arrays also had lower signal amplitude and impedance. New biomaterial strategies that minimize fibrotic encapsulation and enhance material stability should be developed to achieve high quality recording and stimulation for longer implantation periods.

Solving musculoskeletal biomechanics with machine learning

Deep learning is a relatively new computational technique for the description of the musculoskeletal dynamics. The experimental relationships of muscle geometry in different postures are the high-dimensional spatial transformations that can be approximated by relatively simple functions, which opens the opportunity for machine learning (ML) applications. In this study, we challenged general ML algorithms with the problem of approximating the posture-dependent moment arm and muscle length relationships of the human arm and hand muscles. We used two types of algorithms, light gradient boosting machine (LGB) and fully connected artificial neural network (ANN) solving the wrapping kinematics of 33 muscles spanning up to six degrees of freedom (DOF) each for the arm and hand model with 18 DOFs. The input-output training and testing datasets, where joint angles were the input and the muscle length and moment arms were the output, were generated by our previous phenomenological model based on the autogenerated polynomial structures. Both models achieved a similar level of errors: ANN model errors were 0.08 ± 0.05% for muscle lengths and 0.53 ± 0.29% for moment arms, and LGB model made similar errors-0.18 ± 0.06% and 0.13 ± 0.07%, respectively. LGB model reached the training goal with only 103 samples, while ANN required 106 samples; however, LGB models were about 39 times slower than ANN models in the evaluation. The sufficient performance of developed models demonstrates the future applicability of ML for musculoskeletal transformations in a variety of applications, such as in advanced powered prosthetics.

Perception of microstimulation frequency in human somatosensory cortex

Microstimulation in the somatosensory cortex can evoke artificial tactile percepts and can be incorporated into bidirectional brain–computer interfaces (BCIs) to restore function after injury or disease. However, little is known about how stimulation parameters themselves affect perception. Here, we stimulated through microelectrode arrays implanted in the somatosensory cortex of two human participants with cervical spinal cord injury and varied the stimulus amplitude, frequency, and train duration. Increasing the amplitude and train duration increased the perceived intensity on all tested electrodes. Surprisingly, we found that increasing the frequency evoked more intense percepts on some electrodes but evoked less-intense percepts on other electrodes. These different frequency–intensity relationships were divided into three groups, which also evoked distinct percept qualities at different stimulus frequencies. Neighboring electrode sites were more likely to belong to the same group. These results support the idea that stimulation frequency directly controls tactile perception and that these different percepts may be related to the organization of somatosensory cortex, which will facilitate principled development of stimulation strategies for bidirectional BCIs.

Exploring the Use of Brain-Computer Interfaces in Stroke Neurorehabilitation

With the continuous development of artificial intelligence technology, "brain-computer interfaces" are gradually entering the field of medical rehabilitation. As a result, brain-computer interfaces (BCIs) have been included in many countries' strategic plans for innovating this field, and subsequently, major funding and talent have been invested in this technology. In neurological rehabilitation for stroke patients, the use of BCIs opens up a new chapter in "top-down" rehabilitation. In our study, we first reviewed the latest BCI technologies, then presented recent research advances and landmark findings in BCI-based neurorehabilitation for stroke patients. Neurorehabilitation was focused on the areas of motor, sensory, speech, cognitive, and environmental interactions. Finally, we summarized the shortcomings of BCI use in the field of stroke neurorehabilitation and the prospects for BCI technology development for rehabilitation.

Interface, interaction, and intelligence in generalized brain-computer interfaces

A brain-computer interface (BCI) establishes a direct communication channel between a brain and an external device. With recent advances in neurotechnology and artificial intelligence (AI), the brain signals in BCI communication have been advanced from sensation and perception to higher-level cognition activities. While the field of BCI has grown rapidly in the past decades, the core technologies and innovative ideas behind seemingly unrelated BCI systems have never been summarized from an evolutionary point of view. Here, we review various BCI paradigms and present an evolutionary model of generalized BCI technology which comprises three stages: interface, interaction, and intelligence (I3). We also highlight challenges, opportunities, and future perspectives in the development of new BCI technology.

A brain-computer interface that evokes tactile sensations improves robotic arm control

Prosthetic arms controlled by a brain-computer interface can enable people with tetraplegia to perform functional movements. However, vision provides limited feedback because information about grasping objects is best relayed through tactile feedback. We supplemented vision with tactile percepts evoked using a bidirectional brain-computer interface that records neural activity from the motor cortex and generates tactile sensations through intracortical microstimulation of the somatosensory cortex. This enabled a person with tetraplegia to substantially improve performance with a robotic limb; trial times on a clinical upper-limb assessment were reduced by half, from a median time of 20.9 to 10.2 seconds. Faster times were primarily due to less time spent attempting to grasp objects, revealing that mimicking known biological control principles results in task performance that is closer to able-bodied human abilities.

Injecting Information into the Mammalian Cortex: Progress, Challenges, and Promise

For 150 years artificial stimulation has been used to study the function of the nervous system. Such stimulation-whether electrical or optogenetic-eventually may be used in neuroprosthetic devices to replace lost sensory inputs and to otherwise introduce information into the nervous system. Efforts toward this goal can be classified broadly as either biomimetic or arbitrary. Biomimetic stimulation aims to mimic patterns of natural neural activity, so that the subject immediately experiences the artificial stimulation as if it were natural sensation. Arbitrary stimulation, in contrast, makes no attempt to mimic natural patterns of neural activity. Instead, different stimuli-at different locations and/or in different patterns-are assigned different meanings randomly. The subject's time and effort then are required to learn to interpret different stimuli, a process that engages the brain's inherent plasticity. Here we will examine progress in using artificial stimulation to inject information into the cerebral cortex and discuss the challenges for and the promise of future development.