Browsing by Author "Contreras-Vidal, Jose L."
Now showing 1 - 20 of 45
- Results Per Page
- Sort Options
Item 18-Month Mobile Brain-Body Imaging (MoBI) Data Correlating with Daily Tasks: Findings in Alpha-band Frequencies(2019) Alarcon, Christian Bernard; Bellman, Devon E.Current neuroscience studies have failed to capture the progressive, long-term nature of the creative process, limiting the intricate system into single-session controlled experiments. Through the advancement of MoBI technology, we utilized context-aware documentation to monitor and record EEG data from a multimedia installation artist as she undergoes the creative process. This dataset propels brain-computer interfaces closer to real-world applications by answering the question: can EEG data from natural settings be analyzed using MoBI technology? In this 18-month longitudinal study, using a dry-electrode wireless headset a home-security camera, and a personal journaling phone app, EEG data is collected from real-world settings -- the comfort of an artist's home as she creates an art installation. Then, the data was separated by task-specific labels based off video and journal annotations. EEG and video were simultaneously recorded, resulting in over 400 hours of data. To determine the validity of the datasets, we have explored EEG findings in the alpha-band region (8-12 Hz). After scalp mapping the average EEG of the tasks, we notice a difference in alpha power from the prefrontal cortex (PFC) to the parietal region. Also, when comparing alpha power through potential baseline activities, a shift toward the parietal regions is also evident. We are working to open-source the multimodal dataset to allow others to verify findings and discover potential uses. We hope for the public EEG data to help create advances in merging brain-machine interfaces closer to the real world as wireless, wearable, non-invasive systems. This project was completed with contributions from Jesus G. Cruz-Garza from Corner University.Item A LONGITUDINAL BRAIN-MACHINE INTERFACE TRAINING PARADIGM WITH A LOWER-LIMB EXOSKELETON & ITS INDUCED CORTICAL CHANGES(2021-05) Nathan, Kevin C.; Contreras-Vidal, Jose L.; Grossman, Robert G.; Faghih, Rose T.; Mayerich, David; Zhang, YingchunIntroduction: Brain-machine interfaces (BMIs) have been developed to enable cognitive control of computers and robotic devices. Such technology might potentially lead to restoring movement for persons with motor disabilities by allowing them to control robotic prostheses or orthoses naturally with their mind. However, BMIs are still in their infancy, and long-term usage with closed-loop systems has not been thoroughly studied, nor the subsequent changes in the brain induced by cortical plasticity. Methods: Seven able-bodied subjects were recruited for a longitudinal BMI training paradigm with the Rex lower-limb exoskeleton. Participants developed their ability to use motor imagery over nine sessions to initiate the Rex’s walking and stopping as a Go-No Go task. The BMI consisted of active EEG processed through a Localized Fisher Discriminant Analysis dimensionality reduction and a Gaussian Mixture Model classifier on time-lagged δ band amplitudes. Training data were accumulated to update the decoding model over the first five sessions, after which model parameters were fixed for subjects to adapt to their personalized model. Subjects underwent a final session with simultaneous EEG-fMRI recording while watching video playback of themselves walking in the Rex performing the same motor imagery. Discussion: BMI decoding for control of the Rex’s gait varied among the subjects, with at least some achieving significantly above chance classification performance by the end of training. The fMRI scans showed contrasts in activation between the Walk and Stop conditions localized in the precentral gyrus among other areas associated with motor imagery. Offline EEG analysis identified ERPs corresponding to the walk cue, but these may not have been reliably detected by the classifier. Significance: The novelty in this study is the extended use of a subject pool continuously for many sessions of BMI training to control a walking exoskeleton. The longitudinal aspect provides insights into how much training subjects may need to achieve reliable classification, what factors separate good BMI operators from poor ones, and what other features may be more relevant in future BMI applications.Item A Non-invasive Brain Computer Interface Decoder for Gait(2020-05) Nakagome, Sho; Contreras-Vidal, Jose L.; Prasad, Saurabh; Mayerich, David; Nguyen, Hien Van; Pollonini, LucaBrain Computer Interface (BCI) systems enable control of machines and computers using signals extracted from the brain, such as data recorded using electroencephalography (EEG). Naturally, this technology is expected to help people with disabilities, such as lost speech or motor impairment, by providing an alternative approach to interact with the world. Being able to walk is one of the most fundamental human functions, and BCIs could help those with walking impairment by providing direct control of an exoskeleton directly from brain signals. The most crucial part of building such a system is the neural decoding–i.e., the specific algorithm that trans- lates neural signals into movement signals. Developing an effective neural decoding model does not only provide accurate control of the device, but could also open a new path towards understanding the neural representation of gait. A wide variety of algorithms have been proposed for neural decodings, such as linear regression, kalman filters, and artificial neural networks. However, there is a lack of rigorous comparisons of different decoding models and parameter choices. Furthermore, it is unclear how well each of these models will generalize to new data from either new environments or different subjects. This dissertation thesis aims to investigate those issues by: 1) Benchmarking the proposed models and understanding the representation of the brain during gait and 2) Study ways to generalize the model. In the first specific aim, we showed that neural networks not only performed better than conventional methods when trained within a specific walking environment, but resulted in models that were robust to external disturbances such as channel distortion. In the second aim, we showed intra- subject decoding works in all the combinations (e.g., inter-subject decoding of different terrains, level ground walking only, treadmill walking, etc.), but inter- subject decoding only works for electromyography (EMG) to kinematics decoding. To deal with this problem, several methods were used to improve inter- subject decoding. Of these methods, transfer learning achieved the most promising results. The work in this dissertation contributes to a greater understanding of the decoding models and their performance/generalizability on non-invasive gait decoding.Item A Noninvasive Neural Interface for Control of a Powered Lower Limb Prosthesis(2019-12) Brantley, Justin Alexander; Contreras-Vidal, Jose L.; Prasad, Saurabh; Faghih, Rose T.; Zhang, Yingchun; Yau, Jeffery Min-In; Howell, JaredLimb amputation results in a physical disability that causes activities of daily living to become difficult or impossible for the amputee. Current lower-limb prostheses provide limited control and result in only modest improvements in mobility for the amputee. Recent advancements in powered lower limb prostheses allow for more intelligent control and better walking functions. The incorporation of neural signals, specifically muscle and brain, may offer a viable method for improved volitional control by directly interpreting signals from muscle and cortical brain activations. In this study, we employ a multimodal neuroimaging approach to determine if noninvasively recorded brain signals can be used within a lower limb prosthesis control scheme. First, we use a mobile brain/body imaging (MoBI) approach to identify the neural correlates of walking during terrain transitions between level ground and stair ascent. These data are then used to demonstrate the feasibility of predicting the two terrains directly from EEG signals, with cross-validation accuracies achieving greater than 80% in offline decoding. Second, we utilize functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) to study the brain of amputees during isolated movements of the intact and phantom limb. These data are used to: (1) identify neural correlates of movement during isolated limb movements in the amputee population, and (2) demonstrate the feasibility of control of a powered lower limb prosthesis using neural signals from the brain and muscles. We observed that the representation of the phantom limb is preserved in the deprived cortex. Additionally, using a Kalman Filtering approach, we achieved moderate reconstruction accuracy for predicting movements of the phantom and intact limb directly from EEG. The work in this dissertation contributes to a greater understanding of the neural signals associated with phantom limb movements in lower limb amputees and presents a strategy for neural control of powered prostheses.Item Automation Process of 3D Scan Based Brace Design(2020-09-29) Desabhotla, Krishna Sarvani; Steele, Alexander G.; Eguren, DavidApproximately 83,700 children across the United States and 1.7 million children worldwide are affected by cerebral palsy and have limited or no walking ability. [1] Customizable exoskeletons could greatly improve rehabilitation outcomes and quality of life for children suffering from diseases such as cerebral palsy and spina bifida which limit mobility. Currently, it takes an expert 5-6 hours to create a single set of exoskeleton braces from one 3D scan imported into SolidWorks. In this project, we developed a process that enables braces created from a template 3D scan to readjust to another with minimal steps. This is critical as the braces must be modified as children grow. The process of solid modelling typically involves orienting objects, such as the braces, in reference to 3D space. Instead, by defining the braces using global variables and joint references like the hip, knee, and ankle, braces can transfer from the template scan to another scan with automatic size and orientation adjustment. The methodology of brace transfer involves changing the reference surface and joint of the template brace from one scan to another and can be done within minutes. This technique standardizes the design and drastically shortens the time it takes to create custom-fitted braces while reducing human error. Future work would expand the proof of concept to create a complete set of braces that transfers between scans. Applications for this research are not limited to exoskeletons but can be applied to any customizable 3D scan based orthotic/prosthetic. [1]https://www.cdc.gov/ncbddd/cp/data.htmlItem Autonomic Markers of Visual Awareness(2013-12) Li, Ziyang 1987-; Sheth, Bhavin R.; Ogmen, Haluk; Contreras-Vidal, Jose L.; Francis, David J.; Stevenson, Scott B.The mind–body problem in philosophy examines the relationship between mind and matter, and in particular, the relationship between consciousness and the brain. In order to provide a scientific footing to this centuries old philosophical problem, an investigation into the interaction between consciousness and the autonomic nervous system, which controls the internal viscera, is required. However, this issue has received scant attention to date. Here, I investigate the response of the autonomic system and its sympathetic and parasympathetic components, to visual awareness using classical paradigms of binocular rivalry and visual detection, using a combination of electrocardiography (ECG), impedance cardiography (ICG) and pupillometry to examine cardiac autonomic functions, namely heart rate, the high-frequency component of heart rate variability (HRV), pre-ejection period (PEP) and change in pupil area. My studies reveal that the parasympathetic component dominates the autonomic response to visual awareness; physical alternation of stimuli has effects on the autonomic activation that go above and beyond alternations in percept; and uncertainty of subjective judgment drives the dynamics of autonomic response. The present studies, from the autonomic pathway, demonstrate that “mind affects body in action”, which leads to a more integrative view of sensory awareness and suggests the involvement of structures in the nervous system above and beyond the cortex.Item Bayesian Decoder Design for Investigation of Cognitive Arousal and Performance Using Physiological and Behavioral Data(2021-12) Khazaei, Saman; Faghih, Rose T.; Contreras-Vidal, Jose L.; Grigoriadis, Karolos M.Human arousal is one of the indicators of internal stress that has effects on productivity and performance. Poor stress management may lead to reduced performance. Hence, decoding the unobserved arousal and performance, and identifying the arousal-performance relationship are challenging topics of study. On the other hand, cognitive performance can be affected by several external factors such as working environment and surroundings. Decoding human cognitive performance during a cognitive task, based on the environmental variations, is an important topic of interest in the cognitive neuroengineering area. Our study has a great potential to transform workplaces and educational systems. In this study, using the state-space approach within an expectation-maximization (EM) framework, we first obtain the arousal and performance states separately. We investigate the feasibility of using the Yerkes–Dodson law from psychology to link arousal to cognitive performance. Thereafter, we develop a novel arousal decoder based on the corresponding performance. Next, in order to capture the external effects on performance, we design the performance state-space model such that it becomes adaptive to the environmental changes. We develop a time-varying state-space model by applying the autoregressive conditional heteroskedasticity (ARCH) framework to our problem. Subsequently, we present a particle filtering approach and track the human performance through a cognitive task under the EM scheme. Our findings manifest the background music effects on the obtained arousal and performance states. The arousal and performance relationship reveals the existence of Yerkes–Dodson law. The novel arousal decoder offers us a better agreement between the arousal-performance relationship and the Yerkes–Dodson law. Furthermore, our investigations indicate that applying the ARCH framework within the state-space model results in an improved performance state estimation and it outperforms the previous model in terms of capturing the environmental impacts on human performance. Our study can be implemented directly in designing non-invasive closed-loop systems, future smart workplaces, and online educational systems to regulate stress for maximizing performance and productivity.Item Brain Machine Interface with Closed-Loop Neuromuscular Stimulation for Grasping in Stroke and Spinal Cord Injury Survivors(2017-12) Bhagat, Nikunj Arunkumar; Contreras-Vidal, Jose L.; Becker, Aaron T.; Thrasher, Timothy Adam; Francisco, Gerard E.; Ogmen, HalukSixty percent of elderly hand movements involve grasping, which is unarguably why grasp restoration is a major component of upper-limb rehabilitation therapy. Neuromuscular, or functional electrical stimulation (FES), can help retrain grasping by using short bursts of electrical pulses to artificially contract paralyzed muscles. However, current home-use FES requires users to operate a keypad or coordinate body movements for initiating the stimulation, which is often challenging and inefficient for paralyzed patients, as well as unfeasible for severely impaired patients. Conversely, therapeutic FES devices that are controlled by a therapist or pre-programmed, fail to engage patients and ultimately undermine the therapy outcomes. Besides, commercially available FES devices are open-loop systems that require frequent parameter adjustments, which disrupts their continuous use. To increase engagement and ensure accessibility to severely impaired patients, several researchers have suggested non-invasive electroencephalography (EEG)-based brain-machine interfaces (BMI) that allow patients to operate FES devices using their brain activity. However, EEG’s weak signal-to-noise ratio and inherent trial-to-trial variability, have deteriorated the performance of EEG-based BMIs and compromised their long-term reliability. Likewise, closed-loop FES for grasping is promising, but its long-term efficacy is debatable due to lack of effective muscle models, subject variability, and implementation challenges. To address EEG’s challenges, we developed a novel BMI design using optimal adaptive windows to extract movement related cortical potentials, which is a widely studied neural correlate of movement intention. A pilot study with four chronic stroke survivors demonstrated consistent above chance-level BMI performance (65% true and 28% false positives) across two days. In a subsequent study involving two stroke, one spinal injury, and two control subjects, we evaluated the efficacy of integrating BMI with closed-loop FES in order to restore grasping. A custom-built FES prototype using feedback control was developed to automatically adjust the stimulation intensities during grasping and was further validated in an isometric force tracking task. After three sessions, it was concluded that the normalized tracking errors were significantly smaller during closed-loop stimulation (25 ± 15%) versus open-loop stimulation (31 ± 24%), (F (748.03, 1) = 23.22, p < 0.001). These findings will benefit future designs of BMI with closed-loop FES and help determine the clinical efficacy of BMI-FES therapy in motor rehabilitation, following stroke or spinal cord injury.Item Classification and Transfer Learning of EEG during a Kinesthetic Motor Imagery Task using Deep Convolutional Neural Networks(2018-12) Craik, Alexander R.; Contreras-Vidal, Jose L.; Prasad, Saurabh; Pollonini, LucaThe reliable classification of Electroencephalogram (EEG) signals is a crucial step towards making EEG-controlled non-invasive Brain-Machine exoskeleton rehabilitation a practical reality. EEG signals collected during motor imagery tasks have been proposed to act as a control signal for exoskeleton applications. Here, a Deep Convolutional Neural Network (DCNN) was optimized to classify a two-class kinesthetic motor imagery EEG dataset, leading to an architecture consisting of four convolutional layers and three fully connected layers. Transfer learning, or the leveraging of data from past subjects to classify the intentions of a new subject, is important for rehabilitation as it helps to minimize the number of training sessions required from disabled subjects, who lack full motor functionality. The transfer learning training paradigm investigated through this thesis utilized region criticality trends to reduce the number of new subject training sessions and increase the classification performance when compared against a single-subject non-transfer-learning classifier.Item Closed-Loop Control of Brain States using Physiological Signals from Wearable Devices(2021-08) Rezende Fiuza Branco, Luciano; Faghih, Rose T.; Contreras-Vidal, Jose L.; Franchek, Matthew A.Emotions exert powerful influences on all aspects of cognition. In terms of emotional valence, negative emotions can hinder our ability to concentrate and recall, while long periods of overly positive emotions can take a toll in rational decision-making. In addition, lower levels of emotional arousal impair our motivation and productivity, while high levels of this cognitive stress greatly impacts quality of life and life expectancy. Our goal here is to investigate the closed-loop control of emotional levels, as this could improve current and future medical and psychiatric diagnoses and treatments. Open-loop methods are employed in traditional neurostimulation approaches, in which there is no feedback of the internal brain state, impeding the controller of automatically adjusting to neurophysiological changes. To address this issue, we use physiological measurements attained with wearable devices to infer hidden brain states. Specifically, we use electrodermal activity and facial electromyogram as feedback biomarkers for arousal and valence levels, respectively. Using a systematic approach to estimate and track hidden emotional levels, we develop and employ a simulation environment in which we recreate states of low or high valence as well as low or high arousal within the brain model. From the simulated physiological responses, we extract binary and continuous features before using a Bayesian filter to estimate brain states in real-time. We close the loop with a fuzzy logic controller (FLC) optimized with a genetic algorithm. FLCs have been widely used because of how well they mitigate the inaccuracies and uncertainties of the real world. However, manually tuning the parameters can be time-consuming. To this end, we employ a genetic algorithm to optimize the parameters and adapt the controller to different subjects. Moreover, we discuss the investigation of non-invasive stimuli in cognitive processes. Specifically, the use of deep-learning music generation as an interesting type of brain stimulation for closed-loop control therapies. Final results illustrate the feasibility of our approach in recovering, tracking and regulating the hidden emotional states using physiological measurements obtained with wearable devices. Ultimately, this study suggests initial prospects of regulating mental states with data collected non-invasively, offering potential towards novel treatments to various mental illnesses.Item Closed-Loop Regulation of Internal Brain States using Wearable Brain Machine Interface Architectures with Real-World Experimental Implementation(2021-12) Fekri Azgomi, Hamid; Faghih, Rose T.; Francis, Joseph T.; Mayerich, David; Cescon, Marzia; Contreras-Vidal, Jose L.The brain is a control system with a strong impact on all human functions. Inspired by the recent advances in wearable technologies, we design wearable-machine interface (WMI) architectures for controlling brain responses. The WMI architecture encompasses collecting physiological data using wearable devices, inferring neural stimuli underlying pulsatile signals, estimating an unobserved state based on the underlying stimuli, designing the control, and closing the loop. In this thesis, we design WMI architectures for regulating human’s cognitive stress state and controlling energy levels in patients with hypercortisolism. Hypercortisolism, which corresponds to the excessive levels of cortisol hormone, is associated with tiredness and fatigue during the day and disturbed sleep at night. Automating the use of medications that are effective by either elevating or lowering the energy levels might help patients with hypercortisolism to experience more balanced energy cycles required for their daily activities and better sleep patterns at night. Keeping cognitive stress at a healthy range can improve the overall quality of life by helping the subjects to decrease their high levels of arousal to relax them and elevate their low levels of arousal to increase the engagement. Skin conductance data provides us with valuable information regarding one's cognitive stress-related state. We propose to use this physiological data collected via wearable devices to infer individuals' arousal state. In the first part of this research, we simulate multi-day cortisol profile data for multiple subjects both in healthy conditions and with Cushing's disease. Then, we present a state-space model to relate an internal hidden cognitive energy state to subject's cortisol secretion patterns. Particularly, we consider circadian upper and lower bound envelopes on cortisol levels, and timings of hypothalamic pulsatile activity underlying cortisol secretions as continuous and binary observations, respectively. By estimating the hidden energy state and incorporating the simulated hypothetical medication dynamics, we design a knowledge-based control system and close the loop. In the second part of this research, we design a simulation environment to control a cognitive stress-related state in a closed-loop manner. Hence, using the state-space approach, we relate internal cognitive stress state to the changes in skin conductance. Then, we estimate the hidden stress state and close the loop by designing a fuzzy controller. Next, we propose supervised control architectures to enhance the closed-loop performance in cognitive stress regulation. To further enhance the closed-loop design, we consider adaptive and robust control systems to handle model uncertainty and additional disturbance input. Finally, we design and perform multiple human-subject experiments to further explore safe actuation to regulate internal hidden brain states in real-world. In these novel experiments, we employ wearable technologies and publish data sets that could be further investigated to model the dynamics of proposed safe actuation. These studies are the first steps toward the goal of treating similar mental and hormone-related disorders in real-world situations. Analyzing the human subjects’ responses to the effective safe actuation would further enhance the efficiency of proposed approaches and lead us to a practical automated personalized closed-loop architecture.Item CONTINUOUS AND DISCRETE DECODING OF OVERT SPEECH WITH ELECTROENCEPHALOGRAPHY(2023-08) Craik, Alexander Robert; Contreras-Vidal, Jose L.; Dial, Heather R.; Prasad, Saurabh; Buckner, Cameron J.; Nguyen, Hien VanNeurological disorders affecting speech production impair the quality of life for over 7 million individuals in the US. Traditional speech interfaces like eye-tracking devices and P300 spellers are slow and unnatural for these patients. An alternative solution, speech Brain-Computer Interfaces (BCIs), directly decodes speech characteristics, offering a more natural communication mechanism. This research explores the feasibility of decoding speech features using non-invasive EEG. Nine neurologically intact participants were equipped with a 63-channel EEG system with additional sensors to eliminate eye artifacts. Participants read aloud sentences displayed on a screen selected for phonetic similarity to the English language. Pre-processing techniques, including filtering, line noise removal, and eye artifact removal, were applied prior to assessment of methods for the removal of facial electromyography (EMG) contamination. Four Blind Source Separation cleaning methods, including Canonical Correlation Analysis, Independent Component Analysis, and two-stage EMG removal methods employing Ensemble Empirical Mode Decomposition, were evaluated. Three implementations of these methods were selected for further decoding analysis based on signal characteristic metrics and their correlation with decoding performance. Deep learning models, including Convolutional Neural Networks and Recurrent Neural Networks with/without attention modules, were optimized with a focus on minimizing trainable parameters and utilizing small input window sizes. These models were employed for discrete and continuous speech decoding tasks, achieving above-chance participant-independent decoding performance for discrete classes and continuous characteristics of the produced audio signal. A frequency sub-band analysis highlighted the significance of certain frequency bands (delta, theta, and gamma) for decoding performance, and a perturbation analysis identified crucial channels. Assessed channel selection methods did not significantly improve performance, but they still outperformed chance levels, suggesting high-density EEG systems might not be warranted for speech BCIs. Transfer learning demonstrated the possibility of utilizing common speech neural correlates, reducing data collection requirements from individual participants. The successful classification of continuously-produced phonemes and regression of acoustic characteristics signify progress in non-invasive speech BCI development. This research presents promising steps towards developing a universal non-invasive speech BCI control signal, offering opportunities for commercial applications as a replacement for task-specific protocols. Improved speech BCIs hold the potential to improve the overall quality of life of individuals living with neurological speech disorders.Item Contribution of the Fronto-parietal Cortical Dynamics to Grip Force Control(2022-05-05) Rao, Nishant; Parikh, Pranav J.; Gorniak, Stacey L.; Layne, Charles S.; Contreras-Vidal, Jose L.; Li, ShengWhile holding a coffee mug filled to the brim, we strive to avoid spilling the coffee. This ability relies on the interaction between the control of finger forces on a moment-to-moment basis and the visual information about the object. Such sensorimotor interaction is affected in patients with stroke, Parkinson’s disease, and cerebral palsy. Studies investigating force control have shown that fluctuations in the exerted force are not mere noise but arise from systematic physiological processes. Most recent evidence points toward a link between neural activity within the fronto-parietal brain regions including primary motor cortex (M1) and the fluctuations in grip force. However, specific contribution of the cortical activity to regulation of grip force remains unclear. This is a significant research gap because it limits our understanding about how the brain enables efficient control of grip forces during grasping. The current dissertation focused on bridging this research gap using noninvasive neuromodulation and neuroimaging approaches via two specific aims. In Aim-1, we determined the causal involvement of M1 in regulating grip force variability using transcranial magnetic stimulation (TMS) among healthy young individuals. Consistent with our hypothesis, temporary disruption of M1 resulted in upregulation of the grip force variability when compared to that post sham (placebo) stimulation. Interestingly, this upregulation was observed when visual feedback of the exerted grip force was available, but not when the visual feedback of the exerted grip force was removed, indicating the critical role of M1 in integrating visuomotor information for regulating grip force variability. In Aim-2, we examined the dependence of lateralized fronto-parietal neural activity on grip force magnitude during a grip force control task using noninvasive electroencephalography (EEG). Accumulating evidence suggests mechanistic role of neural variability in cognitive processes that scale with task demands. Consequently, we hypothesized laterally specific modulation in EEG variability with increasing magnitude of the grip force exerted during an isometric grip force control task in healthy young individuals. Consistent with our hypothesis, the neural variability was found to be lateralized, topographically constrained, and functionally dependent on the grip force magnitude thereby, showcasing the influence of force-dependent behavioral processes on neural variability. Taken together, this dissertation underscores the integral role of M1 and associated fronto-parietal cortical activity during grip force control. We highlight the relevance of these findings to the rehabilitation of upper extremity motor functions among patients with sensorimotor deficits and propose directions for future studies investigating neural correlates of digit force control.Item Cortical Control of Human Upright Stance(2016-05) Ozdemir, Recep Ali; Paloski, William H.; Layne, Charles S.; Thrasher, Timothy Adam; Contreras-Vidal, Jose L.This dissertation examined, for the first time, differences between young and elderly volunteers in cortical representations of human posture control during (1) quiet stance with normal and altered sensory stimulation, (2) biomechanical perturbations, and (3) dual tasking. The primary focus of the first part was to monitor changes in cortical activity when unexpectedly altering the sensory conditions of upright stance, such as switching from stable (eyes open, fixed support surface) to less-stable (eyes closed, sway-referenced support surface) conditions (experiment 1). Our results demonstrate increased cortical activations in delta (0.2-4Hz) and gamma (30-50 Hz) oscillations, primarily over central-frontal, central and central parietal cortices during challenging postural conditions. While increased delta rhythms were observed in both groups during challenging sensory conditions, elderly individuals also showed increased gamma band activity over sensorimotor and parietal cortices, when compared to the younger group. Correlation analyses also suggest that increased cerebral activity became more relevant to the control of Center of Mass dynamics when upright stance was threatened, especially in the elderly group. The second part studied compensatory postural responses to unexpected perturbations while simultaneously recording Electroencephalography, Electromyography, and Center of Mass dynamics (experiment 2). Our results also suggest that, rather than motor system malfunctioning, impairments in perceptual processing of sensory afferences forms the basis of prolonged postural responses to perturbed stance conditions in non-faller older adults. In general, our results are not only consistent with previous reports suggesting involvement of cerebral cortices in human upright stance control, but also extend them by showing ageing related cortical activity modulations during challenging postural tasks. The third part focused on performance changes in posture control-cognition dual tasking as well as cortical representations of these performance changes both in cognitive and posture control tasks (experiment 3). Postural and cognitive data analyses showed that elderly people had no performance deficits during single postural task conditions, but decreased cognitive performance even during challenging single cognitive tasks. Dual tasking analyses indicated that working memory impairments in the elderly group can be observed when a challenging cognitive task is performed in any postural condition, while postural control performance differences only became significant during dual tasking with challenging postural and cognitive task conditions. EEG analyses showed increased delta, theta and gamma oscillations, primarily over frontal, central-frontal, central and central-parietal cortices during challenging dual tasking conditions. While delta oscillations are more responsive to challenging postural conditions, theta rhythms are found to be changing as a function of cognitive task difficulty in both groups, with more pronounced increases in the young subjects. These results, in general, indicate that elderly subjects may adopt a non-automated conscious control strategy and prioritize postural performance over cognitive performance to maintain upright stance only when the cognitive load is low. High cognitive loads, on the other hand, dramatically increase postural sway, thus the risk of falling, in the elderly people. Regarding the cortical basis of age related performance differences during dual tasking conditions, EEG analyses suggest that while increased theta over frontal and central-frontal cortices may underlie the cortical correlates of high level cognitive computations including encoding and retrieval, delta oscillations, in general, maybe underlie cortical monitoring of changes in postural state of the body when sensory conditions of upright stance is compromised.Item Data Collection for a Longitudinal Mobile Brain-Body Imaging (MoBI) Study of the Creative Process Over the Span of 18 Months in Real-World Settings(2019) Bellman, Devon E.; Alarcon, Christian BernardUnderstanding human creativity remains one of the fundamental questions linking art, science, and engineering. Contemporary neuroscience studies investigating the brain in relation to creativity have been limited to single-session laboratory settings that fail to capture the progressive nature of the creative process in complex settings. To overcome these limitations, we deployed a combination of context-aware documentation (video and personal journal) and mobile brain-body imaging (MoBI) technology to monitor and record a Houston-based multimedia installation artist in real-world-settings. We make available the first longitudinal MoBI dataset using dry-electrode electroencephalography (EEG) as she performs label-specific tasks.Item Data-Driven, Label Consistent, Dictionary Learning Methods for Analysis of Biological Datasets(2016-08) Megjhani, Murad; Roysam, Badrinath; Contreras-Vidal, Jose L.; Shih, Wei-Chuan; Mayerich, David; Leasure, J. Leigh; Burks, JaredThe goal of this thesis is to develop a data-driven, label consistent, and dictionary learning based framework that can be applied on a variety of signal analysis problems. Current methods based on analytical models do not adequately take the variability within and across datasets into consideration when designing signal analysis algorithms. This variability can be added as a morphological constraint to improve the signal analysis algorithms. In particular, this work focuses on three different applications: 1) we present a method for large-scale automated three-dimensional (3-D) reconstruction and profiling of microglia populations in extended regions of brain tissue for quantifying arbor morphology, sensing activation states, and analyzing the spatial distributions of cell activation patterns in tissue; this work provided an opportunity to profile the distribution of microglia in the controlled and device implanted brain. 2) we present a novel morphological constrained spectral unmixing (MCSU) algorithm that combines the spectral and morphological cues in the multispectral image data cube to improve the unmixing quality, this work provided an opportunity to identify new therapeutic opportunities for pancreatic ductal adenocarcinoma (PDAC) from the images collected from humans; and finally, 3) we developed a framework to analyze neuronal response from electroencephalography (EEG) datasets acquired from the infants ranging from 6-24 months. We demonstrated that combining different frequency bands from different spatial locations, yields better classification results, instead of the traditional approach where either one or two frequency bands are used. Using an adaptation of Tibshirani’s Sparse Group LASSO algorithm, we uncovered different spatial and bio markers for understanding a human infant’s brain. These bio-markers can be used for developmental stages of infants and further analysis is required to study the clinical aspects of infant’s social and cognitive development. This work establishes the fundamental mathematical basis for the next generation of algorithms that can leverage the morphological cues from the biological datasets. The algorithm has been embedded into the open source FARSIGHT toolkit with an intuitive graphical user interface.Item DESIGN AND EVALUATION OF A POWERED LOWER-LIMB EXOSKELETON FOR PEDIATRIC GAIT REHABILITATION AND MOBILITY(2022-12-14) Eguren, David; Contreras-Vidal, Jose L.; Francisco, Gerard E.; Han, Zhu; Prasad, Saurabh; Hall, StaceyPowered exoskeletons for gait rehabilitation and mobility assistance are currently available for the adult population and hold great promise for children with mobility limiting conditions. Described here is the development and bench testing evaluation of a modular, lightweight and customizable powered exoskeleton for over-ground walking and gait rehabilitation. The pediatric lower-extremity gait system (P-LEGS) exoskeleton contains bilaterally active hip, knee and ankle joints to provide movement support for walking standing and sitting to young children with lower-limb disabilities such as those present in Cerebral Palsy, Spina Bifida and Spinal Cord Injured populations. The system consists of six joint control modules, one at each hip, knee and ankle joint. The joint control module, features an actuator, gear, motor driver, microcontroller and system monitoring sensors. Bench-testing results of the joint actuator assembly as well as full device evaluation for a 10-meter walk test and repeated sit-stand transition evaluation under unloaded conditions as well as conditions loaded with a mannequin to simulate a young child are presented and discussed. The device is being prepared to enter clinical testing with able-bodied and children from the target populations and customized cuffs have been developed through 3D scanning and 3D printing for the able-bodied subjects to begin testing.Item Developing Explainable Deep Learning Models Using EEG for Brain Machine Interface Systems(2021-12) Sujatha Ravindran, Akshay; Contreras-Vidal, Jose L.; Parikh, Pranav J.; Buckner, Cameron J.; Mayerich, David; Nguyen, Hien VanDeep learning (DL) based decoders for Brain-Computer-Interfaces (BCI) using Electroencephalography (EEG) have gained immense popularity recently. However, the interpretability of DL models remains an under-explored area. This thesis aims to develop and validate computational neuroscience approaches to make DL models more robust and explainable. First, a simulation framework was developed to evaluate the robustness and sensitivity of twelve back-propagation-based visualization methods. Comparing to ground truth features, after randomizing model weights and labels, multiple methods had reliability issues: e.g., the gradient approach, which is the most used visualization technique in EEG, was not class or model-specific. Overall, DeepLift was the most reliable and robust method. Second, we demonstrated how model explanations combined with a clustering approach can be used to complement the analysis of DL models applied to measured EEG in three tasks. In the first task, DeepLift identified the EEG spatial patterns associated with hand motor imagery in a data-driven manner from a database of 54 individuals. Explanations identified different strategies used by individuals and exposed the issues in limiting decoding to the sensorimotor channels. The clustering approach improved the decoding in high-performing subjects. In the second task, we used GradCAM to explain the Convolutional Neural Network’s (CNN) decision associated with detecting balance perturbations while wearing an exoskeleton, deployable for fall prevention. Perturbation evoked potentials (PEP) in EEG (∼75 ms) preceded both the peak in electromyography (∼180 ms) and the center of pressure (∼350 ms). Explanation showed that the model utilized electro-cortical components in the PEP and was not driven by artifacts. Explanations aligned with dynamic functional connectivity measures and prior studies supporting the feasibility of using BCI-exoskeleton systems for fall prevention. In the third task, the susceptibility of DL models to eyeblink artifacts was evaluated. The frequent presence of blinks (in 50% trials or more), whether they bias a particular class or not, leads to a significant difference in decoding when using CNN. In conclusion, the thesis contributes towards improving the BCI decoders using DL models by using model explanation approaches. Specific recommendations and best practices for the use of back-propagation-based visualization methods for BCI decoder design are discussed.Item Development of a Wearable Device That Provides Haptic Feedback Based on Force Distribution of Gait Stance for Pediatric Applications(2018-12) Arunkumar, Anirudh; Contreras-Vidal, Jose L.; Pollonini, Luca; Zhang, YingchunLower limb disorders reduce the quality of life by making walking extremely difficult, especially for children. In this study, a wearable device that provides vibrotactile feedback on the forearms in a pattern that emulates the sensation of pressures from the force distribution on the feet was developed to determine whether children can intuitively associate the vibration pattern with their gait, with the ultimate aiding in walking rehabilitation. This device was tested on multiple children by having them walk multiple trials at different speeds and directions, and then replaying the vibrations of each trial in a different order while the subject attempted to verify the speed and direction of the gait based on the vibrations. Results showed low accuracies for children’s ability to verify the gait based on the vibrations, be-cause of inconsistent sensor readings and the need for more training for the subjects for intuitive association.Item Examining the Improvisational Creative Process in the Visual Arts: A Mobile Brain Body Imaging Approach(2017) Cruz-Garza, Jesus G.; Chatufale, Girija; Contreras-Vidal, Jose L.Mobile brain/body imaging (MoBI) for the study of the human improvisational creative process. In the spirit of the "Exquisite Corpse," an improvisational creative game, three artists created three art-pieces. The artists were equipped with 64 channel wireless EEG (2 channel EOG) and 3 inertial measurement units on forearms and head. We report the most relevant features for offline classification both in motion and EEG data pertaining to baseline, planning, and execution phases, of the improvisational creative process. The angular velocity in the left-right direction and the magnitude ratio (right / left hand) of the movement jerk were features that consistently shared the most mutual information with the class labels. The most relevant features for classification in the EEG data varied for each artist, and relate to their approach to the artwork. These features were mostly found in the parietal, central and frontal electrodes across frequency bands and time-domain features.
- «
- 1 (current)
- 2
- 3
- »