Posts Tagged Neurotechnology

[ARTICLE] Brain-Machine Neurofeedback: Robotics or Electrical Stimulation? – Full Text

Neurotechnology such as brain-machine interfaces (BMI) are currently being investigated as training devices for neurorehabilitation, when active movements are no longer possible. When the hand is paralyzed following a stroke for example, a robotic orthosis, functional electrical stimulation (FES) or their combination may provide movement assistance; i.e., the corresponding sensory and proprioceptive neurofeedback is given contingent to the movement intention or imagination, thereby closing the sensorimotor loop. Controlling these devices may be challenging or even frustrating. Direct comparisons between these two feedback modalities (robotics vs. FES) with regard to the workload they pose for the user are, however, missing. Twenty healthy subjects controlled a BMI by kinesthetic motor imagery of finger extension. Motor imagery-related sensorimotor desynchronization in the EEG beta frequency-band (17–21 Hz) was turned into passive opening of the contralateral hand by a robotic orthosis or FES in a randomized, cross-over block design. Mental demand, physical demand, temporal demand, performance, effort, and frustration level were captured with the NASA Task Load Index (NASA-TLX) questionnaire by comparing these workload components to each other (weights), evaluating them individually (ratings), and estimating the respective combinations (adjusted workload ratings). The findings were compared to the task-related aspects of active hand movement with EMG feedback. Furthermore, both feedback modalities were compared with regard to their BMI performance. Robotic and FES feedback had similar workloads when weighting and rating the different components. For both robotics and FES, mental demand was the most relevant component, and higher than during active movement with EMG feedback. The FES task led to significantly more physical (p = 0.0368) and less temporal demand (p = 0.0403) than the robotic task in the adjusted workload ratings. Notably, the FES task showed a physical demand 2.67 times closer to the EMG task, but a mental demand 6.79 times closer to the robotic task. On average, significantly more onsets were reached during the robotic as compared to the FES task (17.22 onsets, SD = 3.02 vs. 16.46, SD = 2.94 out of 20 opportunities; p = 0.016), even though there were no significant differences between the BMI classification accuracies of the conditions (p = 0.806; CI = −0.027 to −0.034). These findings may inform the design of neurorehabilitation interfaces toward human-centered hardware for a more natural bidirectional interaction and acceptance by the user.


About half of all severely affected stroke survivors remain with persistent motor deficits in the chronic disease stage despite therapeutic interventions on the basis of the current standard of care (Winters et al., 2015). Since these patients cannot use the affected hand for activities of daily living, novel interventions investigate different neurotechnological devices to facilitate upper limb motor rehabilitation, such as brain-machine interfaces (BMI), robotic orthoses, neuromuscular functional electrical stimulation (FES), and brain stimulation (Coscia et al., 2019). BMI approaches, for example, aim at closing the impaired sensorimotor loop in severe chronic stroke patients. They use robotic orthoses (Ang et al., 2015Kasashima-Shindo et al., 2015Belardinelli et al., 2017), FES devices (Kim et al., 2016Biasiucci et al., 2018), and their combination (Grimm et al., 2016cResquín et al., 2017) to provide natural sensory and proprioceptive neurofeedback during movement intention or imagery. It is hypothesized that this approach will lead to reorganization of the corticospinal network through repetitive practice, and might ultimately restore the lost motor function (Naros and Gharabaghi, 20152017Belardinelli et al., 2017Guggenberger et al., 2018).

However, these novel approaches often result in no relevant clinical improvements in severe chronic stroke patients yet (Coscia et al., 2019). Therefore, recent research has taken a refined and rather mechanistic approach, e.g., by targeting physiologically grounded and clinically relevant biomarkers with BMI neurofeedback; this has led to the conceptional differentiation between restorative therapeutic BMIs on the one side (as those applied in this study) and classical assistive BMIs on the other side like those applied to control devices such as wheel-chairs (Gharabaghi, 2016): While assistive BMIs intend to maximize the decoding accuracy, restorative BMIs want to enhance behaviorally relevant biomarkers. Specifically, brain oscillations in the beta frequency band have been suggested as potential candidate markers and therapeutic targets for technology-assisted stroke rehabilitation with restorative BMIs (Naros and Gharabaghi, 20152017Belardinelli et al., 2017), since they are known to enhance signal propagation in the motor system and to determine the input-output ratio of corticospinal excitability in a frequency- and phase-specific way (Raco et al., 2016Khademi et al., 20182019Naros et al., 2019).

However, these restorative BMI devices differ from their predecessors, i.e., assistive BMIs, by an intentionally regularized and restricted feature space, e.g., by using the beta frequency band as a feedback signal for BMI control (Gharabaghi, 2016Bauer and Gharabaghi, 2017). Such a more specific approach is inherently different from previous more flexible algorithms that select and weight brain signal features to maximize the decoding accuracy of the applied technology; restorative BMIs like the those applied in this study have, therefore, relevantly less classification accuracy than classical assistive BMIs (Vidaurre et al., 2011Bryan et al., 2013). As the regularized and restricted feature space of such restorative BMI devices leads to a lower classification accuracy in comparison to more flexible approaches, it may be frustrating even for healthy participants (Fels et al., 2015). IN the context of the present study, we conjectured that such challenging tasks will increase the relevance of extraneous load aspects like the workload (Schnotz and Kürschner, 2007). Furthermore, the modulation range of the oscillatory beta frequency band is compromised in stroke patients, proportionally to their motor impairment level (Rossiter et al., 2014Shiner et al., 2015). That means that more severely affected patients show less oscillatory event-related desynchronization (ERD) and synchronization (ERS) during motor execution or imagery (Pfurtscheller and Lopes da Silva, 1999). To our understanding, this underlines the relevance of beta oscillations as a therapeutic target for post-stroke rehabilitation. At the same time, however, this poses a major challenge for the affected patients and may, thereby, compromise their therapeutic benefit (Gomez-Rodriguez et al., 2011a,bBrauchle et al., 2015).

To overcome these hurdles that are inherent to restorative BMI devices, we have investigated different approaches in the past: (i) Reducing the brain signal attenuation by the skull via the application of epidural interfaces (Gharabaghi et al., 2014b,cSpüler et al., 2014), (ii) Augmenting the afferent feedback of the robotic orthosis by providing concurrent virtual reality input (Grimm et al., 2016a,b), (iii) combining the orthosis-assisted movements with neuromuscular (Grimm and Gharabaghi, 2016Grimm et al., 2016c) or transcranial electrical stimulation (Naros et al., 2016a) to enhance the cortical modulation range (Reynolds et al., 2015), and (iv) optimizing the mental workload related to the use of BMI devices.

In this study, we focus on the latter approach, i.e., optimizing the mental workload related to the use of BMI devices. For the latter approach it would be necessary to better understand the workloads related to different technologies applied in the context of BMI feedback (robotics vs. FES). We, therefore, investigated the mental demand, physical demand, temporal demand, performance, effort, and frustration of healthy subjects controlling a BMI by motor imagery of finger extension. Motor imagery-related sensorimotor desynchronization in the beta frequency-band was turned into passive opening of the contralateral hand by a robotic exoskeleton or FES in a randomized, cross-over block design. The respective workloads were compared to the task-related aspects of active hand movement with EMG feedback. We conjectured a feedback-specific workload profile that would be informative for more personalized future BMI approaches.



We recruited 20 healthy subjects (age = 23.5 ± 1.08 yeas [mean ± SD], range 19–27, 15 female) for this study. Subjects were not naive to the tasks. All were right-handed and reached a score equal or above 60 in the Edinburgh Handedness Inventory (Oldfield, 1971). The subjects gave their written informed consent before participation and the study protocol was approved by the Ethics Committee of the Medical Faculty of the University of Tübingen. They received monetary compensation.

Subject Preparation

We used Ag/AgCl electrodes in a 32 channel setup according to the international 10-20 system (Fp1, Fp2, F3, Fz, F4, FC5, FC3, FC1, FCz, FC2, FC4, FC6, C5, C3, C1, Cz, C2, C4, C6, TP9, CP5, CP3, CP1, CPz, CP2, CP4, CP6, P3, Pz, P4, O1, O2 with TP10 as Reference and AFz as Ground) to examine the cortical activation pattern during the training session. Electrode impedances were set below 10 kΩ. All signals are digitalized at a sampling frequency of 1,000 Hz (robotic block) or 5,000 Hz (FES block) using Brain Products Amplifiers and transmitted online to BCI2000 software. BCI2000 controlled in combination with a custom-made software the respective feedback device, i.e., either the robotic orthosis or the functional electrical stimulation. Depending on the task, one of the following preparations was performed. Either the robotic hand orthosis (Amadeo, Tyromotion) was attached to the subject’s left hand (Figure 1A), fixated with Velcro strips across the forearm and with magnetic pads on the fingertips (Gharabaghi et al., 2014aNaros et al., 2016b); or functional electrical stimulation (FES, Figure 1B) was applied to the M. extensor digitorum communis (EDC) by the RehaMove2 (Hasomed GmbH, Magdeburg) with two self-adhering electrodes (50 mm, HAN-SEN Trading & Consulting GmbH, Hamburg). First an electrode was fixed to the distal end of the EDC’s muscle belly serving as ground. Then a rectangular electrode prepared with contact gel was used to find the optimal place for the second electrode where maximal extension of the left hand could be achieved. Here a custom written Matlab script was executed to detect the current threshold needed for the extension. Starting at 1 mA, the current was increased in steps of 0.5–1 mA. During each trial, FES was applied for 3 s with a pulse width of 1,000 μs and a frequency of 100 Hz. At the beginning of stimulation, a ramping protocol was implemented for 500 ms. Once, the correct position and threshold of stimulation were found, the temporary electrode was replaced by the second stimulation electrode and both were fixed with tape. A mean stimulation intensity of 6.5 mA (SD = 2.27) was required to cause the desired contraction in this study.

Figure 1. Experimental set-up. (Left) Robotic hand orthosis as feedback device (Amadeo, Tyromotion GmbH, Graz). (Middle) Neuromuscular forearm stimulation as feedback device (RehaMove 2, Hasomed GmbH, Magdeburg). In both cases, a brain-machine interface (BMI) detected motor imagery-related oscillations in the beta frequency band by an electroencephalogram (EEG) and provided via a BCI2000-system contingent feedback by moving the hand with either the robot or the electrical stimulation. (Right) The EEG montage used in this study.


Continue —-> Frontiers | Brain-Machine Neurofeedback: Robotics or Electrical Stimulation? | Bioengineering and Biotechnology

, , , , , , , ,

Leave a comment

[WEB SITE] Research work in Neurotechnology directs efforts towards treating chronic stroke

Scientific work undertaken at Wyss Center for Bio and Neuroengineering in Geneva, Switzerland has developed a rehabilitation arm in order to improve recovery during severe chronic strokes in patients.

Stroke is regarded as one of the major health problems among people today. A common symptom observed among cases of stroke is the long-term impairment of upper arm function. This results in complications in daily life chores and hampers the quality of life.

The Neurotechnology includes a host of therapies, like robotics, brain stimulation, brain-machine interfaces, etc. According to experts, these will in return be fruitful in treating patients, centering on their individual needs. Moreover, the new study also sheds light on longitudinal clinical studies in order to understand the rehabilitation benefits of individual therapies. Furthermore, the study also focuses on various combinations of complementary therapies used over a period of time.

“Our findings show that neurotechnology-aided upper limb rehabilitation is promising for severe chronic stroke patients. However, we also found that the ‘one size fits all’ approach doesn’t lead to the best outcome. We suggest a move towards a personalized combination of neurotechnology-based stroke rehabilitation therapies, ideally in a home-based environment where prolonged therapy is more feasible than in a clinic. We believe that by sequentially introducing stroke therapies according to individual progress, we could allow patients to continue their recovery beyond what is possible today,” says Dr. Martina Coscia, lead author and Staff Engineer at Wyss Center.

As per experts, rehabilitation therapies show the best results within the first three months after the incidence of stroke. After the first three months, the scope of natural recovery is limited and patients are considered chronic, commonly observed scenario, especially among patients who are severely affected.

For the study, authors reportedly compared data from 64 cases of clinical studies based on upper limb neurotechnology treatments among stroke patients. The findings mainly centered on brain stimulation, electrical stimulation of muscles, and brain-computer interfaces, in addition to a combination of these.

Further reports suggest the team is directing efforts towards undertaking clinical traits in order to test the results. For the trial, experimental design such as robotics, functional electrical stimulation, brain-computer interfaces is used to monitor the after-effects of treatment in individual patients. Scientists believe to use a combination of neurotechnological and new personalized therapies in order to improve recovery among patients. The study published in the journal Brain alleges that the trial will begin in Switzerland in summer 2019.

via Research work in Neurotechnology directs efforts towards treating chronic stroke – Xaralite

, , , , , , , , , , , , ,

Leave a comment

[WEB SITE] Neurotechnology-Aided Rehab Holds Promise for Chronic Stroke Patients

Published on 


Personalized neurotechnology-aided rehabilitation of the arm could improve recovery in severe chronic stroke patients, according to a study published recently in the journal Brain.

Neurotechnology-based therapies, including brain-machine interfaces, robotics, and brain stimulation among others, will lead to largest treatment effects and success if they are tailored to the needs of individual patients, and used in combination, according to the authors from the Wyss Center for Bio and Neuroengineering, Swiss Federal Institute of Technology Lausanne (EPFL), Scuola Superiore Sant’Anna, University of Geneva Faculty of Medicine and Clinique Romande de Réadaptation.

In their study, they call for longitudinal clinical studies to show the rehabilitation benefits of individual therapies as well as the use of multiple complementary therapies used in combination over long time periods.

“Our findings show that neurotechnology-aided upper limb rehabilitation is promising for severe chronic stroke patients,” says lead author Dr. Martina Coscia, Staff Engineer at the Wyss Center, in a media release.

“However, we also found that the ‘one size fits all’ approach doesn’t lead to the best outcome. We suggest a move towards a personalized combination of neurotechnology-based stroke rehabilitation therapies, ideally in a home-based environment where prolonged therapy is more feasible than in a clinic.

“We believe that by sequentially introducing stroke therapies according to individual progress, we could allow patients to continue their recovery beyond what is possible today.”

One of the most common consequences of stroke is impaired upper arm function, which has a direct impact on daily tasks and quality of life. Rehabilitation therapies generally have the largest effect in the first three months after stroke. After this time, patients are considered chronic and the likelihood of further natural recovery is limited; this is especially true for those most severely affected.

“What we would like to see in the future are long-term trials in which multiple neurotechnology-based therapies are used within the same patient,”  Professor Friedhelm Hummel from EPFL (Director, Defitech Chair of Clinical Neuroengineering) and the University of Geneva Medical School, shares in the release.

“We believe that this synergistic approach could uncover previously undiscovered treatment pathways for chronic stroke patients.”

In their study, the authors compared effectiveness data from 64 clinical studies on upper limb neurotechnology-aided treatments in chronic stroke patients. The interventions analyzed in the paper included robotics, functional electrical stimulation of muscles, brain stimulation, and brain-computer interfaces as well as their use in combination.

The interdisciplinary research team is now starting a clinical trial to test these ideas. The trial uses a new experimental design with a personalized therapy approach using brain-computer interfaces, robotics, functional electrical stimulation, and brain stimulation specifically chosen to maximize treatment effects in each individual patient. The goal is to keep incrementally improving recovery by using new personalized, neurotechnology-based therapies in combination. The trial will start in Switzerland in summer 2019.

[Source(s): Wyss Center for Bio and Neuroengineering, Science Daily]


via Neurotechnology-Aided Rehab Holds Promise for Chronic Stroke Patients – Rehab Managment

, , , , , , ,

Leave a comment

[WEB SITE] Electrically stimulating the brain may restore movement after stroke

June 18, 2018, University of California, San Francisco

Micrograph showing cortical pseudolaminar necrosis, a finding seen in strokes on medical imaging and at autopsy. H&E-LFB stain. Credit: Nephron/Wikipedia

UC San Francisco scientists have improved mobility in rats that had experienced debilitating strokes by using electrical stimulation to restore a distinctive pattern of brain cell activity associated with efficient movement. The researchers say they plan to use the new findings to help develop brain implants that might one day restore motor function in human stroke patients.

After a , roughly one-third of  recover fully, one-third have significant lingering  problems, and one-third remain virtually paralyzed, said senior author Karunesh Ganguly, MD, Ph.D., associate professor of neurology and a member of the UCSF Weill Institute for Neurosciences. Even patients who experience partial recovery often continue to struggle with “goal-directed” movements of the arms and hands, such as reaching and manipulating objects, which can be crucial in the workplace and in daily living.

“Our main impetus was to understand how we can develop implantable neurotechnology to help stroke patients,” said Ganguly, who conducts research at the San Francisco VA Health Care System. “There’s an enormous field growing around the idea of neural implants that can help neural circuits recover and improve function. We were interested in trying to understand the circuit properties of an injured brain relative to a healthy brain and to use this information to tailor neural implants to improve  after stroke.”

Over the past 20 years, neuroscientists have presented evidence that coordinated patterns of neural activity known as oscillations are important for efficient brain function. More recently, low-frequency oscillations (LFOs)—which were first identified in studies of sleep—have been specifically found to help organize the firing of neurons in the brain’s primary motor cortex. The motor cortex controls voluntary movement, and LFOs chunk the cells’ activity together to ensure that goal-directed movements are smooth and efficient.

In the new study, published in the June 18, 2018 issue of Nature Medicine, the researchers first measured neural activity in rats while the animals reached out to grab a small food pellet, a task designed to emulate human goal-directed movements. They detected LFOs immediately before and during the action, which inspired the researchers to investigate how these activity patterns might change after stroke and during recovery.

To explore these questions, they caused a stroke in the rats that impaired the animals’ movement ability, and found that LFOs diminished. In rats that were able to recover, gradually making faster and more precise movements, the LFOs also returned. There was a strong correlation between recovery of function and the reemergence of LFOs. Animals that fully recovered had stronger low-frequency activity than those that partially recovered, and those that didn’t recover had virtually no low-frequency activity.

To try to boost recovery, the researchers used electrodes to both record activity and deliver a mild electrical current to the rats’ brains, stimulating the area immediately surrounding the center of the . This stimulation consistently enhanced LFOs in the damaged area and appeared to improve motor function: when the researchers delivered a burst of electricity right before a rat made a movement, the rat was up to 60 percent more accurate at reaching and grasping for a food pellet.

“Interestingly, we observed this augmentation of LFOs only on the trials where stimulation was applied,” said Tanuj Gulati, Ph.D., a postdoctoral researcher in the Ganguly lab who is co-first author of the study, along with Dhakshin Ramanathan, MD, Ph.D., now assistant professor of psychiatry at UC San Diego, and Ling Guo, a neuroscience graduate student at UCSF.

“We are not creating a new frequency, we are amplifying the existing frequency,” added Ganguly. “By amplifying the weak low-frequency oscillations, we are able to help organize the task-related . When we delivered the electrical current in step with their intended actions, motor control actually got better.”

The researchers wanted to know whether their findings might also apply to humans, so they analyzed recordings made from the surface of the brain of an epilepsy patient who had suffered a stroke that had impaired the patient’s arm and hand movements. The recordings revealed significantly fewer LFOs than recordings made in two epilepsy patients who hadn’t had a stroke. These findings suggest that, just as in rats, the stroke had caused a loss of low-frequency  that impaired the patient’s movement.

Physical therapy is the only treatment currently available to aid stroke patients in their recovery. It can help people who are able to recover neurologically get back to being fully functional more quickly, but not those whose stroke damage is too extensive. Ganguly hopes that electrical brain stimulation can offer a much-needed alternative for these latter patients, helping their brain circuits to gain better control of motor neurons that are still functional. Electrical  stimulation is already widely used to help patients with Parkinson’s disease and epilepsy, and Ganguly believes stroke patients may be the next to benefit.

 Explore further: Electrical nerve stimulation could help patients regain motor functions sooner

More information: Low-frequency cortical activity is a neuromodulatory target that tracks recovery after stroke, Nature Medicine (2018).

via Electrically stimulating the brain may restore movement after stroke

, , , , , , , , ,

Leave a comment

[WEB SITE] Building a better brain-computer interface

Building a better brain-computer interface

Photo of a dummy BrainGate interface. Credit: Paul Wick/Wikimedia Commons

October 2, 2018 by Matt Miles, Medical Xpress

Brain-computer interfaces, or BCIs, represent relatively recent advances in neurotechnology that allow computer systems to interact directly with human or animal brains. This technology is particularly promising for use in cases of spinal cord injury or paralysis. In these situations, patients may be able to use neural decoders that access part of their brain to operate a prosthetic limb or even to re-animate a paralyzed limb through functional electrical stimulation (FES).

Michael A. Schwemmer and colleagues, in a recent Nature Medicine article, detail their research on BCIs using  decoders with a participant with tetraplegia due to spinal cord injury. Their research focuses on addressing several key needs identified by end-users of BCI systems, namely: high , minimal daily setup, rapid response time, and multifunctionality—all of which are characteristics heavily influenced by a BCI’s particular neural decoding algorithm.

Schwemmer’s group describes several different approaches to training and testing three variations on neural network decoders (NN-BCI) in comparison with each other and a benchmark support vector machine (SVM) decoder. The four BCI decoder paradigms were developed and tested over the course of several years in association with a 27-year-old male participant with tetraplegia. The participant had a 96-channel microelectrode array implanted in the area of his left primary motor cortex corresponding to the hand and arm. Using intracortical data collected from 80 sessions over 865 days, the investigators trained and evaluated these BCI decoders. These sessions consisted of two 104-second blocks of a four-movement task: index extension, index flexion, wrist extension, and wrist flexion.

The initial neural network (NN)  was developed and calibrated using data from the first 40 sessions (80 blocks); it was not updated over the second half of the training/testing period, and is referred to here as the fixed neural network (fNN) model. From the fNN, two other neural network models were created: a supervised updating (sNN) model and an unsupervised updating (uNN) model. Both models used data from the first block of the second 40-session (updating/testing) period. The sNN model’s algorithm relies on explicit training labels, that is, known timing and type of movement, whereas the uNN model relies on undifferentiated or unknown direct input in relation to intended action of the limb. The second block of the second 40-session period was used for accuracy testing of all models—fNN, sNN, uNN, and SVM.

The purpose of using four separate models here was to test and demonstrate various aspects of the three neural network models in relation to each other and the benchmark SVM model. For instance, the supervised neural network (sNN) model was updated daily (during the first block of the second 40-session period) and compared directly with the daily-retrained SVM model. The fixed neural network (fNN) model was provided to demonstrate that a BCI could sustain accuracy for over a year with no updates.

The unsupervised neural network (uNN) was perhaps the most interesting comparator, as we shall see, because it attempted to combine the improved accuracy gained from daily updates but without the consequent daily setup time required by the sNN model. Accuracy was the key performance measure in all tests, defined here as a percentage of correctly predicted time-bins in the second block of the second 40 sessions; the criterion of greater than 90% accuracy was one of the four end-user requirements originally articulated at the outset of the study.

The sNN consistently outperformed the daily-retrained SVM: in 37 out of 40 sessions, its accuracy was > 90%, whereas the SVM only achieved > 90% accuracy in 12 sessions. The fNN also outperformed the SVM in 36 of 40 sessions; it achieved > 90% accuracy in 32 sessions. The fNN accuracy was, not surprisingly, lower than the accuracy of the sNN, and both fixed decoders, fNN and SVM, declined in accuracy over the course of the study period, in contrast to the daily-updated decoders.

Perhaps the most interesting finding of this research however, is the performance of the unsupervised neural network (uNN), which outperformed both fixed models in terms of accuracy, while also meeting the end-user requirement of minimal daily set-up. Where the sNN model required explicit daily training, the uNN incorporated data from general use in its update schema, which required no such daily set-up. In comparison with the fNN, a performance gap emerged over time, and the benefits of the uNN distinguished themselves. The uNN also outperformed the SVM in terms of response time, another key end-user requirement.

Another important aspect of this study with regard to NNs focused on transfer learning, whereby new movements can be added to the existing repertoire with minimal additional training and data. In this case, “hand open” and “hand close” were added to the previous four movements, and all decoders were rebuilt. Here too, unsupervised updating was used to build an unsupervised transfer  (utNN), which, after only one session of training oupterformed the SVM model.

Finally, the previous research—all of which was conducted in an “offline” setting—was applied, via the participant’s FES-controlled hand and forearm, to show that a transfer learning uNN trained on the original four-movement task could be used to quickly create a new decoder to control, in real time, an open hand and three grips (can, fork, and peg). In a test of the system, the participant was able to perform all three hand movement grip tasks, with no failures, in 45 attempts. Previously, he was only able to perform one grip task successfully.

In summarizing how the results of their study relate to the main end-user expectations previously described, the investigators cite the following achievements: “(i) using deep NNs to create robust neural decoders that sustain high fidelity BCI control for more than a year without retraining; (ii) introducing a new updating procedure that can improve performance using data obtained through regular system use; (iii) extension of functionality through transfer learning using minimal additional data; and (iv) introducing a decoding framework that simultaneously addresses these four competing aspects of BCI performance (accuracy, speed, longevity, and multifunctionality). In addition, we provide a clinical demonstration that a decoder calibrated using historical data of imagined hand movements with no feedback can be successfully used in real-time to control FES-evoked grasp function for object manipulation.”

Schwemmer and colleagues go on to offer a more in-depth discussion of their results amidst the broader landscape of BCI research, and offer commentary on some of the specific challenges and limitations of their experiment. While noting that the median response time for uNN decoders (0.9 s) is still faster than that of SVM decoders (1.1 s), they acknowledge that a target of 750 ms or less is probably closer to realistic end-user expectations.

Ultimately they conclude: “We have demonstrated that decoders based on NNs may be superior to other implementations because new functions can be easily added after the initial decoder calibration using transfer learning. Crucially, we show that this secondary update to add more movements requires a minimal amount of additional data.” And “insights gained from offline data and analyses can carry over to a realistic online BCI scenario with minimal additional data collection.”

 Explore further: Using multi-task learning for low-latency speech translation

More information: Michael A. Schwemmer et al. Meeting brain–computer interface user performance expectations using a deep neural network decoding framework, Nature Medicine(2018). DOI: 10.1038/s41591-018-0171-y

via Building a better brain-computer interface

, , , , , ,

Leave a comment

Advances in Neurotechnology, Electronics and Informatics: Revised Selected … – Google Books

This book is a timely report on current neurotechnology research. It presents a snapshot of the state of the art in the field, discusses current challenges and identifies new directions. The book includes a selection of extended and revised contributions presented at the 2nd International Congress on Neurotechnology, Electronics and Informatics (NEUROTECHNIX 2014), held October 25-26 in Rome, Italy.

The chapters are varied: some report on novel theoretical methods for studying neuronal connectivity or neural system behaviour; others report on advanced technologies developed for similar purposes; while further contributions concern new engineering methods and technological tools supporting medical diagnosis and neurorehabilitation.

All in all, this book provides graduate students, researchers and practitioners dealing with different aspects of neurotechnologies with a unified view of the field, thus fostering new ideas and research collaborations among groups from different disciplines.

Source: Advances in Neurotechnology, Electronics and Informatics: Revised Selected … – Google Books

, , ,

Leave a comment

%d bloggers like this: