Brain-machine interface based on deep learning to control asynchronously a lower-limb robotic exoskeleton: a case-of-study.

BACKGROUND: This research focused on the development of a motor imagery (MI) based brain-machine interface (BMI) using deep learning algorithms to control a lower-limb robotic exoskeleton. The study aimed to overcome the limitations of traditional BMI approaches by leveraging the advantages of deep learning, such as automated feature extraction and transfer learning. The experimental protocol to evaluate the BMI was designed as asynchronous, allowing subjects to perform mental tasks at their own will. METHODS: A total of five healthy able-bodied subjects were enrolled in this study to participate in a series of experimental sessions. The brain signals from two of these sessions were used to develop a generic deep learning model through transfer learning. Subsequently, this model was fine-tuned during the remaining sessions and subjected to evaluation. Three distinct deep learning approaches were compared: one that did not undergo fine-tuning, another that fine-tuned all layers of the model, and a third one that fine-tuned only the last three layers. The evaluation phase involved the exclusive closed-loop control of the exoskeleton device by the participants' neural activity using the second deep learning approach for the decoding. RESULTS: The three deep learning approaches were assessed in comparison to an approach based on spatial features that was trained for each subject and experimental session, demonstrating their superior performance. Interestingly, the deep learning approach without fine-tuning achieved comparable performance to the features-based approach, indicating that a generic model trained on data from different individuals and previous sessions can yield similar efficacy. Among the three deep learning approaches compared, fine-tuning all layer weights demonstrated the highest performance. CONCLUSION: This research represents an initial stride toward future calibration-free methods. Despite the efforts to diminish calibration time by leveraging data from other subjects, complete elimination proved unattainable. The study's discoveries hold notable significance for advancing calibration-free approaches, offering the promise of minimizing the need for training trials. Furthermore, the experimental evaluation protocol employed in this study aimed to replicate real-life scenarios, granting participants a higher degree of autonomy in decision-making regarding actions such as walking or stopping gait.

How Integration of a Brain-Machine Interface and Obstacle Detection System Can Improve Wheelchair Control via Movement Imagery.

This study presents a human-computer interaction combined with a brain-machine interface (BMI) and obstacle detection system for remote control of a wheeled robot through movement imagery, providing a potential solution for individuals facing challenges with conventional vehicle operation. The primary focus of this work is the classification of surface EEG signals related to mental activity when envisioning movement and deep relaxation states. Additionally, this work presents a system for obstacle detection based on image processing. The implemented system constitutes a complementary part of the interface. The main contributions of this work include the proposal of a modified 10-20-electrode setup suitable for motor imagery classification, the design of two convolutional neural network (CNNs) models employed to classify signals acquired from sixteen EEG channels, and the implementation of an obstacle detection system based on computer vision integrated with a brain-machine interface. The models developed in this study achieved an accuracy of 83% in classifying EEG signals. The resulting classification outcomes were subsequently utilized to control the movement of a mobile robot. Experimental trials conducted on a designated test track demonstrated real-time control of the robot. The findings indicate the feasibility of integration of the obstacle detection system for collision avoidance with the classification of motor imagery for the purpose of brain-machine interface control of vehicles. The elaborated solution could help paralyzed patients to safely control a wheelchair through EEG and effectively prevent unintended vehicle movements.

Short-Term Effects of Gamma Stimulation on Neuroinflammation at the Tissue-Electrode Interface in Motor Cortex.

OBJECTIVES: The reliability of long-term neural recordings as therapeutic interventions for motor and sensory disorders is hampered by the brain tissue response. Previous work showed that flickering light at gamma frequencies (ie, 20-50 Hz) causes enhanced microglial recruitment in the visual cortex. The effects of gamma stimulation on glial cells surrounding implanted neural electrodes are not well understood. We hypothesized that invasive stimulation in the gamma frequency band increases microglial recruitment in the short term and reduces astrogliosis at the tissue-electrode interface. MATERIALS AND METHODS: Male Long Evans rats were implanted with dual-shank silicon microelectrode arrays into the motor cortex. After implantation, rats received one hour of 40-Hz stimulation at a constant current of 10 µA using charge-balanced, biphasic pulses on one shank, and the other shank served as the nonstimulated control. Postmortem, tissue sections were stained with ectodermal dysplasia 1 (ED1) for activated microglia, glial fibrillary acidic protein (GFAP) for astrocytes, and 4',6-diamidino-2-phenylindole (DAPI) for nonspecific nuclei. Fluorescent intensity and cell number as a function of distance from the tissue-electrode interface were used to quantify all stained sections. RESULTS: Fluorescent intensity for ED1 was nearly 40% lower for control than for stimulated sites (0-500 µm away from the implant), indicating increased microglial recruitment to the stimulated site (p < 0.05). Fluorescent intensity for GFAP was >67% higher for control than for stimulated sites (0-500 µm away from the implant), indicating reduced astrogliosis at the stimulated site (p < 0.05). No differences were observed in DAPI-stained sections between conditions. CONCLUSIONS: These results suggest that short-term gamma stimulation modulates glial recruitment in the immediate vicinity of the microelectrode. Future studies will investigate the long-term effects of gamma stimulation on glial recruitment at the tissue-electrode interface as a strategy to improve long-term recording reliability.

Broad Therapeutic Time Window for Driving Motor Recovery After TBI Using Activity-Dependent Stimulation.

BACKGROUND: After an acquired injury to the motor cortex, the ability to generate skilled movements is impaired, leading to long-term motor impairment and disability. While rehabilitative therapy can improve outcomes in some individuals, there are no treatments currently available that are able to fully restore lost function. OBJECTIVE: We previously used activity-dependent stimulation (ADS), initiated immediately after an injury, to drive motor recovery. The objective of this study was to determine if delayed application of ADS would still lead to recovery and if the recovery would persist after treatment was stopped. METHODS: Rats received a controlled cortical impact over primary motor cortex, microelectrode arrays were implanted in ipsilesional premotor and somatosensory areas, and a custom brain-machine interface was attached to perform the ADS. Stimulation was initiated either 1, 2, or 3 weeks after injury and delivered constantly over a 4-week period. An additional group was monitored for 8 weeks after terminating ADS to assess persistence of effect. Results were compared to rats receiving no stimulation. RESULTS: ADS was delayed up to 3 weeks from injury onset and still resulted in significant motor recovery, with maximal recovery occurring in the 1-week delay group. The improvements in motor performance persisted for at least 8 weeks following the end of treatment. CONCLUSIONS: ADS is an effective method to treat motor impairments following acquired brain injury in rats. This study demonstrates the clinical relevance of this technique as it could be initiated in the post-acute period and could be explanted/ceased once recovery has occurred.

Deep Convolutional Neural Network for EEG-Based Motor Decoding.

Brain-machine interfaces (BMIs) have been applied as a pattern recognition system for neuromodulation and neurorehabilitation. Decoding brain signals (e.g., EEG) with high accuracy is a prerequisite to building a reliable and practical BMI. This study presents a deep convolutional neural network (CNN) for EEG-based motor decoding. Both upper-limb and lower-limb motor imagery were detected from this end-to-end learning with four datasets. An average classification accuracy of 93.36 ± 1.68% was yielded on the four datasets. We compared the proposed approach with two other models, i.e., multilayer perceptron and the state-of-the-art framework with common spatial patterns and support vector machine. We observed that the performance of the CNN-based framework was significantly better than the other two models. Feature visualization was further conducted to evaluate the discriminative channels employed for the decoding. We showed the feasibility of the proposed architecture to decode motor imagery from raw EEG data without manually designed features. With the advances in the fields of computer vision and speech recognition, deep learning can not only boost the EEG decoding performance but also help us gain more insight from the data, which may further broaden the knowledge of neuroscience for brain mapping.

Volitional Control of Brain Motor Activity and Its Therapeutic Potential.

BACKGROUND: Neurofeedback training is a closed-loop neuromodulatory technique in which real-time feedback of brain activity and connectivity is provided to the participant for the purpose of volitional neural control. Through practice and reinforcement, such learning has been shown to facilitate measurable changes in brain function and behavior. OBJECTIVES: In this review, we examine how neurofeedback, coupled with motor imagery training, has the potential to improve or normalize motor function in neurological diseases such as Parkinson disease and chronic stroke. We will also explore neurofeedback in the context of brain-machine interfaces (BMIs), discussing both noninvasive and invasive methods which have been used to power external devices (eg, robot hand orthosis or exoskeleton) in the context of motor neurorehabilitation. CONCLUSIONS: The published literature provides mounting high-quality evidence that neurofeedback and BMI control may lead to clinically relevant changes in brain function and behavior.

Cloud-Based Functional Magnetic Resonance Imaging Neurofeedback to Reduce the Negative Attentional Bias in Depression: A Proof-of-Concept Study.

Individuals with depression show an attentional bias toward negatively valenced stimuli and thoughts. In this proof-of-concept study, we present a novel closed-loop neurofeedback procedure intended to remediate this bias. Internal attentional states were detected in real time by applying machine learning techniques to functional magnetic resonance imaging data on a cloud server; these attentional states were externalized using a visual stimulus that the participant could learn to control. We trained 15 participants with major depressive disorder and 12 healthy control participants over 3 functional magnetic resonance imaging sessions. Exploratory analysis showed that participants with major depressive disorder were initially more likely than healthy control participants to get stuck in negative attentional states, but this diminished with neurofeedback training relative to controls. Depression severity also decreased from pre- to posttraining. These results demonstrate that our method is sensitive to the negative attentional bias in major depressive disorder and showcase the potential of this novel technique as a treatment that can be evaluated in future clinical trials.

Sensory Integration in Human Movement: A New Brain-Machine Interface Based on Gamma Band and Attention Level for Controlling a Lower-Limb Exoskeleton.

Brain-machine interfaces (BMIs) can improve the control of assistance mobility devices making its use more intuitive and natural. In the case of an exoskeleton, they can also help rehabilitation therapies due to the reinforcement of neuro-plasticity through repetitive motor actions and cognitive engagement of the subject. Therefore, the cognitive implication of the user is a key aspect in BMI applications, and it is important to assure that the mental task correlates with the actual motor action. However, the process of walking is usually an autonomous mental task that requires a minimal conscious effort. Consequently, a brain-machine interface focused on the attention to gait could facilitate sensory integration in individuals with neurological impairment through the analysis of voluntary gait will and its repetitive use. This way the combined use of BMI+exoskeleton turns from assistance to restoration. This paper presents a new brain-machine interface based on the decoding of gamma band activity and attention level during motor imagery mental tasks. This work also shows a case study tested in able-bodied subjects prior to a future clinical study, demonstrating that a BMI based on gamma band and attention-level paradigm allows real-time closed-loop control of a Rex exoskeleton.

Pain Control by Co-adaptive Learning in a Brain-Machine Interface.

Innovation in the field of brain-machine interfacing offers a new approach to managing human pain. In principle, it should be possible to use brain activity to directly control a therapeutic intervention in an interactive, closed-loop manner. But this raises the question as to whether the brain activity changes as a function of this interaction. Here, we used real-time decoded functional MRI responses from the insula cortex as input into a closed-loop control system aimed at reducing pain and looked for co-adaptive neural and behavioral changes. As subjects engaged in active cognitive strategies orientated toward the control system, such as trying to enhance their brain activity, pain encoding in the insula was paradoxically degraded. From a mechanistic perspective, we found that cognitive engagement was accompanied by activation of the endogenous pain modulation system, manifested by the attentional modulation of pain ratings and enhanced pain responses in pregenual anterior cingulate cortex and periaqueductal gray. Further behavioral evidence of endogenous modulation was confirmed in a second experiment using an EEG-based closed-loop system. Overall, the results show that implementing brain-machine control systems for pain induces a parallel set of co-adaptive changes in the brain, and this can interfere with the brain signals and behavior under control. More generally, this illustrates a fundamental challenge of brain decoding applications-that the brain inherently adapts to being decoded, especially as a result of cognitive processes related to learning and cooperation. Understanding the nature of these co-adaptive processes informs strategies to mitigate or exploit them.

Intra-cortical brain-machine interfaces for controlling upper-limb powered muscle and robotic systems in spinal cord injury.

OBJECTIVE: Intracortical brain-machine interface (iBMI) is an assistive strategy to restore lost sensorimotor function by bridging the disrupted neural pathways to reanimate paralyzed limbs. However, to date, none of the studies explored the trade-offs between the performance criteria of different iBMI systems that decode discrete upper limb movements from intracortical neural recordings. METHODS: A systematic review of electronic databases using different MeSH terms from January 1990 to December 2019 was conducted. IBM® SPSS statistics version 25 (Released 2017, Armonk, NY: IBM) was used to evaluate for differences between groups using independent sample t-tests. RESULTS: A total of 18 patients from 15 studies were included in our analysis. The included studies involved iBMI controlled 5-robotic and 10-neuromuscular stimulated orthotics to perform skillful and coordinated movements that resulted in a clinically significant gain in tests of upper-limb functions. Pooled analysis revealed that the mean response time to execute 3-D reach and grasp task by the robotic-assisted limb was relatively longer (46.8 +/-101.5 s) compared to the neuro-muscular stimulated orthotics (15.8 +/-15.2 s); however, statistically insignificant [Mean difference (MD): 30.9, 95 % Confidence Interval (CI): -40.4-102.3, p = 0.35]. Furthermore, the accuracy in performing 3-D reach and grasp tasks after repetitive trials were better among patients with neuro-muscular stimulated orthotics (83.5 +/-12.7 %) compared to those with robotic-assisted prosthetic limb (69.1 +/- 23.6 %) with statistically significant difference (MD: 15.9, 95 % CI: 1.65-32.5, p = 0.05). CONCLUSION: Our study demonstrates that iBMI-assisted prosthetic limbs showed better accuracy and shorter response time among patients with neuro-muscular stimulated orthotics compared to robotic neuro-prosthetics.