Bruna M. Carlos, B. Campos, M. Alvim, Manuel G. Patiño, F. Cendes, G. Castellano
{"title":"Epileptogenic zone classification with functional connectivity and graph measures","authors":"Bruna M. Carlos, B. Campos, M. Alvim, Manuel G. Patiño, F. Cendes, G. Castellano","doi":"10.1109/NER52421.2023.10123728","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123728","url":null,"abstract":"Epilepsy is one of the most common neurological diseases, and it has a great variety of possible diagnosis that can involve different treatments. The characterization of epileptogenic zones (EZs) is of extreme importance for the evaluation of patients, specially for pharmacoresistant epilepsy patients that are candidates for surgery. Many works have been showing that epilepsy interferes with brain network organization, in particular during seizures or epileptic discharges, although there is also indication that there are significant alterations during interictal periods. With most studies using intracranial electroencephalography (EEG), there are few results about the alterations present in regular EEG examinations that could be used to indicate the general location of the EZ in advance. This prior indication could be beneficial for a better understanding of network changes across the whole scalp. The aim of this work was to extract relevant information from patients with different EZs, using functional connectivity and graph measures from normal background EEG signals. Patients with EZs in the temporal and frontal lobe were included in this study. Our results reinforce that centrality graph measures from many areas of the scalp are useful in the distinction between patients with different EZs. We also found that connectivity in the alpha band can be used for EZ classification and might be of interest in future studies.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122291574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Privacy-Preserving Motor Intent Classification via Feature Disentanglement","authors":"Jiahao Fan, Xiaogang Hu","doi":"10.1109/NER52421.2023.10123842","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123842","url":null,"abstract":"Recent studies have revealed that sensitive and private attributes could be decoded from surface electromyogram (sEMG) signals, which can incur privacy threat to the users of sEMG-based pattern recognition applications. Most studies so far focus on improving the accuracy and reliability of sEMG classifiers, but much less attention has been paid to their privacy. To fill this gap, this study implemented and evaluated a framework to optimize sEMG-based data-sharing mechanism. Our primary goal was to remove sensitive attributes (i.e., identity-relevant information) in the sEMG features before sharing them with primary pattern recognition tasks. We disentangled the identity-insensitive, task-relevant representations from original sEMG features. We shared it with the downstream pattern recognition tasks to reduce the chance of sensitive attributes being inferred by potential attackers. The proposed method was evaluated on data from twenty subjects, with training and testing data acquired 3–25 days apart. Our results showed that the disentangled representations significantly reduced the success rate of identity inference attacks compared to the original feature and its sparse representations generated by the state-of-the-art feature projection methods. The disentangled representation was then evaluated in hand gesture recognition tasks. Our results revealed that the disentangled representations led to higher classification accuracy across classifiers compared with other feature implementations. This work shows that disentangled representations of sEMG signals are a promising solution for privacy-preserving motor intent recognition applications.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114201551","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Asymmetric Changes in Intersegmental Covariation Across Ambulation Levels and Prosthetic Devices for Transfemoral Amputee Gait","authors":"N. Krausz, T. Flash","doi":"10.1109/NER52421.2023.10123729","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123729","url":null,"abstract":"Resolution of kinematic redundancies during human locomotion has previously been described by the law of intersegmental coordination. This law describes how elevation angles of the thigh, shank, and foot covary during gait. Previous work showed that the covariation plane (CVP) orientation can provide insight into gait pathologies such as Parkinson's Disease (PD) and Cerebellar Ataxia (CA). We assess the intersegmental CVP for locomotion with an above knee amputation, as compared to other movement pathologies, showing that amputees do follow the law of intersegmental coordination. Specifically here we considered locomotion across different ambulation capacities (K level) and knee and ankle prostheses. Our results showed that there is an asymmetry in how the CVP varies with walking speed for the amputated leg versus the contralateral leg. Also, we showed that with certain ankle and knee prostheses the CVP of the contralateral leg varies with speed similarly to able-bodied subjects; this is not the case for the amputated leg. Importantly, this result may have implications for reduction of the energetic cost of walking with a prosthesis. Additionally, using the shape score, we present an important result showing that amputee gait seems to be more poorly coordinated than that of PD or CA subjects. Future work will be expanded to better understand how coordination differs in amputee subjects and whether powered prostheses improve this coordination.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"245 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121157497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Variability in Depolarization Sensitivity Underlies Differential Responses to High-frequency Stimulation of ON and OFF RGCs","authors":"Jae-Ik Lee, P. Werginz, S. Fried","doi":"10.1109/NER52421.2023.10123855","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123855","url":null,"abstract":"Retinal prostheses aim to restore vision to patients with outer retinal degeneration, but the quality of restored vision is still far below that of the healthy visual system. Although multiple factors are likely to contribute to this limited effectiveness, the inability to reproduce the complex neural coding patterns that arise naturally in the retina, e.g., the indiscriminate activation of diverse types of retinal ganglion cells (RGCs) by electric stimulation, is likely to play a significant role. Our previous research has shown that different types of RGCs can be preferentially activated by modulating the stimulus amplitude of high-frequency stimulation (HFS). As a step towards optimizing selectivity, our goal was to explore mechanisms underlying the sensitivity differences between cell types. By measuring responses of RGCs to 2 kHz stimulation with the whole-cell patch technique, we found that increasing the amplitude of stimulation led to greater depolarization of membrane potentials. While small increases in the depolarization level increased spike rates, excessive depolarization suppressed spiking activities. Subsequent comparison between ON and OFF a sustained RGCs revealed that ON cells are more depolarized by a given amplitude of HFS than OFF cells, resulting in stronger responses of ON cells at lower amplitudes but suppression of ON cell spiking at higher amplitudes. Our results suggest that a better understanding of the changes in membrane potential induced by HFS is essential to further optimize HFS-based strategies for retinal prosthesis.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116706119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"On the in vitro long-term stability of thin-film stimulation contacts in polyimide-based neural interfaces","authors":"P. C̆vanc̆ara, Inga Bartels, T. Stieglitz","doi":"10.1109/NER52421.2023.10123737","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123737","url":null,"abstract":"Micro-machined peripheral nerve interfaces have entered translational research successfully. Polyimide-based interface substrates showed chemical inertness, mechanical flexibility and low water uptake. They allow neural interface with a thicknesses of ten micrometer reducing the probability of severe inflammatory reactions. To realize device thicknesses in this range, photolithographic processes are used to deposit thin-film metallization in the range of a few hundred nanometers. In order to prove long-term stability within chemical safe stimulation limits, in vitro stimulation of sputtered iridium oxide film stimulation contact sites was performed. After 4.5 billion pulses, first electrochemical changes occurred indicated by a decrease in impedance magnitude and simultaneously an increase of the cathodic charge storage capacity. Both scenarios indicated a change in the effective surface area. These contact sites were used for an optical in depth analysis using white light interferometry and scanning electron microscopy in combination with a focused ion beam, which confirmed crack formation and delamination. However, for all other contact sites further stimulation up to 6.5 billion pulses was applied. This amount was estimated about 6.6 years of stimulation in chronic human application with a conservative approach assuming daily work with each stimulation contact. The outcome of the long-term stimulation indicates stability for chronic stimulation studies in humans with life-times reasonably long for first translational studies on larger subject cohorts.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122605749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Gabrielle Strandquist, Tanner C. Dixon, Tomasz Frączek, Shravanan Ravi, Alicia Zeng, Raphael Bechtold, Daryl Lawrence, S. Little, J. Gallant, Jeffrey A. Herron
{"title":"In-Home Video and IMU Kinematics of Self Guided Tasks Correlate with Clinical Bradykinesia Scores","authors":"Gabrielle Strandquist, Tanner C. Dixon, Tomasz Frączek, Shravanan Ravi, Alicia Zeng, Raphael Bechtold, Daryl Lawrence, S. Little, J. Gallant, Jeffrey A. Herron","doi":"10.1109/NER52421.2023.10123851","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123851","url":null,"abstract":"Deep brain stimulation (DBS) delivers electrical stimulation directly to brain tissue to treat neurological movement disorders such as Parkinson's Disease (PD). Adaptive DBS (aDBS) is an advancement on DBS that uses symptom-related biomarkers to adjust therapeutic stimulation parameters in real time to improve clinical outcomes and reduce side-effects. A significant challenge for the field of aDBS is developing automated methods to optimize stimulation parameters using remote assessments of symptom severity. To address this challenge, we designed a prototype at-home data collection platform that can remotely update aDBS algorithms and explore objective assessments of motor symptom severity. Our platform collects neural, inertial, and video data, and supports clinician validation of automated symptom assessments. We deployed the system to the home of an individual with PD and collected pilot data across six days. We evaluated motor symptom severity by recording data with stimulation amplitudes set to varying levels during self-guided clinical tasks and free behavior. We assessed movement features including frequency, speed, and peak angular velocity from video-derived pose estimates and inertial data during three clinical tasks. All features showed a reduction during periods of under-stimulation and were significantly correlated with video-based clinical scores of symptom severity (Spearman rank test, $p < 0.006)$. These results demonstrate that our prototype is capable of remote multimodal data collection and that these data can enhance aDBS research outside the clinic.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"11 4 Suppl 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131110202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
J. McLinden, D. Gemoets, Daniel Hahn, J. Brangaccio, Y. Shahriari, J. Wolpaw, James J. S. Norton
{"title":"Automating visual feedback in H-reflex operant conditioning studies: Feasibility and first steps","authors":"J. McLinden, D. Gemoets, Daniel Hahn, J. Brangaccio, Y. Shahriari, J. Wolpaw, James J. S. Norton","doi":"10.1109/NER52421.2023.10123817","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123817","url":null,"abstract":"H-reflex conditioning is a novel targeted-neuroplasticity-based method for the rehabilitation of movement that uses visual feedback to train participants to change the excitability of their reflexes over several months. Present H-reflex conditioning protocols require experimenters to manually control multiple covariates of the H-reflex to ensure the accuracy of the visual feedback. Manual control of these covariates is error prone, labor intensive, and reduces experimenter engagement with participants. Here, as a first step towards a system that automatically ensures the accuracy of visual feedback during H-reflex conditioning, we performed inference and prediction based on multivariate linear regression (MLR) to assess the effect of six covariates of the size of the H-reflex and whether this class of models can be used to predict those effects. Four participants completed experiments where we measured H-reflex size changes in response to changes in each of the six covariates. Inference from the MLR models show that background EMG activity and stimulation current affected H-reflex size in three of the four participants. In addition, our experiments show that MLR models can be used to predict H-reflex size. Compared to an intercept-only model, MLR reduced prediction error by more than 30% $(p < 0.05)$. Our results suggest that automatic adjustment (in response to changes in covariates of the H-reflex) of visual feedback during H-reflex conditioning is possible. With further development, this method could improve feedback during H-reflex operant conditioning, reduce the need for clinicians and researchers to manually control covariates of the H-reflex, and lead to improved H-reflex conditioning protocols for the rehabilitation of movement following neurological injury or illness.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131478659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Maurice Abou Jaoude, Aravind Ravi, Jiansheng Niu, Hubert J. Banville, Nicolas Florez Torres, Christopher Aimone
{"title":"Automated Sleep Staging on Wearable EEG Enables Sleep Analysis at Scale","authors":"Maurice Abou Jaoude, Aravind Ravi, Jiansheng Niu, Hubert J. Banville, Nicolas Florez Torres, Christopher Aimone","doi":"10.1109/NER52421.2023.10123829","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123829","url":null,"abstract":"This study presents automated sleep staging on a large number of sleep electroencephalography (EEG) recordings collected using the Muse S headband. Two recent deep learning models; a single-channel Deep Sleep Net (DSN) and a multi-channel Muse Net (MNet) were evaluated on a 5-class sleep stage classification task on 200 expert-labelled overnight sleep EEG recordings. The learned representations of the models were visualized using uniform manifold approximation projection (UMAP). Moreover, a large scale analysis of the relationship between sleep stage distribution of non-rapid eye movement (NREM) and rapid eye movement (REM) sleep with age was performed on 1020 unlabelled EEG recordings. The results showed that the proposed models achieved high accuracy (DSN: 85.2%, MNet: 86.3%) and Cohen's Kappa (DSN: 0.77, MNet: 0.79) indicating substantial agreement with human expert sleep scoring. Furthermore, the features learned by the deep neural networks showed a sleep continuum beyond the traditionally used sleep stages. Hypnogram analysis revealed a decrease in percentage of NREM 3 and REM sleep with increasing age, and an increase in percentage of NREM 2 sleep with increasing age. The results suggested that a 4-channel wearable EEG headband provides low-cost and powerful means to automatically score and analyze sleep at a large scale.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"133 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131580033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Laura A. Chicos, D. Rangaprakash, R. Barry, H. Herr
{"title":"Resting State Neurophysiology of Agonist Antagonist Myoneural Interface in Transtibial Amputees","authors":"Laura A. Chicos, D. Rangaprakash, R. Barry, H. Herr","doi":"10.1109/NER52421.2023.10123738","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123738","url":null,"abstract":"The agonist-antagonist myoneural interface (AMI) is a novel amputation surgery that preserves sensorimotor signaling mechanisms of the central-peripheral nervous systems. Our first neuroimaging study investigating AMI subjects focused on task-based neural signatures, and showed evidence of proprioceptive feedback to the central nervous system. The study of spontaneous neural activity in the brain at rest helps non-invasively characterize the spatially distributed networks that prime task response. In this first study on resting state fMRI in subjects with an AMI amputation, we compared resting state functional connectivity in patients with transtibial AMI amputation $(mathrm{n}=12)$ and transtibial traditional amputation $(mathrm{n}=7)$, as well as biologically intact control subjects $boldsymbol{(mathrm{n}=10)}$. We hypothesized that the AMI surgery will induce functional network reorganization that significantly differs from the traditional amputation surgery and also more closely resembles the neural configuration of controls. We found that the neurophysiology of AMI subjects reconfigured in such a way that implicated both a salience network and motor cortex region in relationships with significantly less coupling than in subjects with traditional amputation. We also found that the neuropathological signatures found in the spontaneous networks of traditional amputees decrease in the topology of AMI subjects, producing a connectivity pattern more closely resembling controls. These findings provide researchers and clinicians with a critical mechanistic understanding of the effects of the AMI surgery on the brain at rest, informing future research towards improved prosthetic control and embodiment.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126988602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Self-Supervised Task-Agnostic Embedding for EEG Signals","authors":"A. Partovi, A. Burkitt, D. Grayden","doi":"10.1109/NER52421.2023.10123767","DOIUrl":"https://doi.org/10.1109/NER52421.2023.10123767","url":null,"abstract":"Brain-Computer Interfaces (BCIs) have great potential for improving the lives of people with disabilities. The success of a BCI system is largely driven by the accuracy of the BCI decoder. This accuracy, in turn, may be limited by the amount of labelled training data available for supervised machine learning algorithms. The success of deep learning algorithms in other computer science areas has not reached the field of BCI decoding due to this lack of abundant labelled data. We use a novel deep learning architecture trained in a self-supervised manner to learn a common vector representation (embedding) of EEG signals that can be used in different BCI tasks. The vector representation is trained using EEG recordings without using any task labels. We validate our embedder using two separate BCI tasks: seizure detection and motor imagery, and assess its usefulness through distance similarity metrics in a clustering approach. The derived embeddings were successful in distinguishing binary classes in both tasks.","PeriodicalId":201841,"journal":{"name":"2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128015626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}