Traumatic brain injury (TBI) is the signature injury of the ongoing military conflicts in the Middle East and Afghanistan, largely due to the use of improvised explosive devices (IEDs), which have affected soldiers and civilians alike. Blast-induced TBI (bTBI) biomechanics are complex and multiphasic. While research has clearly demonstrated the negative effects of penetrative (secondary blast) and inertia-driven (tertiary blast) injury, the effect of shock wave loading (primary blast) on the brain remains unclear. Combined primary-tertiary blast exposure in vivo has been reported previously to alter brain function, specifically hippocampal function; however, it is extremely difficult to deliver primary blast exposure in isolation with an in vivo injury model. The research presented in this thesis utilized a custom-designed in vitro blast injury model to deliver military-relevant shock wave exposures, in isolation, to organotypic hippocampal slice cultures (OHSCs). To contextualize blast-induced pathobiology with previous TBI studies, the first goal of this thesis was to experimentally characterize the deformation profile induced in OHSCs with our blast injury model. Using stereoscopic, high-speed cameras and digital image correlation to calculate strain, we found that our blast model induced low strain magnitudes (<9%) but at high strain rates (25-86s-1), which aligned closely with associated computational simulations of our model. The second aim was to determine if primary blast was capable of altering hippocampal electrophysiological function. We exposed OHSCs to a range of shock intensities and found, using a micro-electrode array system, that long-term potentiation (LTP), a measure of synaptic plasticity, was very sensitive to primary blast exposure; a threshold for disruption of LTP was found between 9 and 39 kPa•ms impulse. Alternative measures of basal electrophysiology were less sensitive than LTP. Blast exposure significantly reduced LTP between 1 and 24 hours post-injury, and this deficit persisted through 6 days post-injury. Depending on shock intensity, LTP spontaneously recovered 10 days post-injury. The third aim was to explore the cellular mechanisms for blast-induced LTP deficits. Using a chemical LTP induction protocol, blast exposure altered key proteins necessary for the induction of LTP by 24 hours post-injury including, postsynaptic density protein-95 (PSD-95), a major scaffolding protein that organizes the postsynaptic density (PSD), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid glutamate receptor 1 (AMPA-GluR1), and stargazin, an auxiliary GluR1 protein that binds AMPA-GluR1 to PSD-95. Modulation of the cyclic adenosine monophosphate (cAMP) pathway reversed the observed effects of blast on LTP. We theorized that blast-induced disruption of PSD-95 prevented translocation, and subsequent phosphorylation, of GluR1-containing AMPARs to the postsynaptic membrane, which, in turn, prevented potentiation. The final aim was to investigate the efficacy of phosphodiesterase-4 (PDE4) inhibitors, which block degradation of cAMP, as a therapeutic strategy. When delivered immediately following primary blast injury, multiple PDE4 inhibitors proved efficacious in restoring LTP measured 24 hours post-injury. Roflumilast, a Food and Drug Administration-approved PDE4 inhibitor, was effective when delivered at a clinically relevant concentration (1nM) and at a delayed time point (up to 6 hours). Roflumilast reversed blast-induced changes in expression/phosphorylation of the key LTP protein targets. We hypothesized that maintenance of PSD-95 drove the observed therapeutic effect. Greater work is necessary to determine how blast exposure degrades PSD-95 and how roflumilast prevented these detrimental effects. This thesis has shown that primary blast exposure can negatively alter neurological function, as well as protein expression and phosphorylation. These studies expand the understanding of primary blast injury mechanisms, provide computational models with important tissue-level tolerance criteria, inform protective equipment design, inform clinical care guidelines for bTBI, and present a promising therapeutic candidate for further clinical investigation.

Predicting and understanding the dynamic of human motion has many applications such as motion synthesis, augmented reality, security, education, reinforcement learning, autonomous vehicles, and many others. In this thesis, we create a novel end-to-end pipeline that can predict multiple future poses from the same input, and, in addition, can classify the entire sequence. Our focus is on the following two aspects of human motion understanding: Probabilistic human action prediction: Given a sequence of human poses as input, we sample multiple possible future poses from the same input sequence using a new GAN-based network. Human motion understanding: Given a sequence of human poses as input, we classify the actual action performed in the sequence and improve the classification performance using the presentation learned from the prediction network. We also demonstrate how to improve model training from noisy labels, using facial expression recognition as an example. More specifically, we have 10 taggers to label each input image, and compare four different approaches: majority voting, multi-label learning, probabilistic label drawing, and cross-entropy loss. We show that the traditional majority voting scheme does not perform as well as the last two approaches that fully leverage the label distribution. We shared the enhanced FER+ data set with multiple labels for each face image with the research community (https://github.com/Microsoft/FERPlus). For predicting and understanding of human motion, we propose a novel sequence-to-sequence model trained with an improved version of generative adversarial networks (GAN). Our model, which we call HP-GAN2, learns a probability density function of future human poses conditioned on previous poses. It predicts multiple sequences of possible future human poses, each from the same input sequence but seeded with a different vector z drawn from a random distribution. Moreover, to quantify the quality of the non-deterministic predictions, we simultaneously train a motion-quality-assessment model that learns the probability that a given skeleton pose sequence is a real or fake human motion. In order to classify the action performed in a video clip, we took two approaches. In the first approach, we train on a sequence of skeleton poses from scratch using random parameters initialization with the same network architecture used in the discriminator of the HP-GAN2 model. For the second approach, we use the discriminator of the HP-GAN2 network, extend it with an action classification branch, and fine tune the end-to-end model on the classification tasks, since the discriminator in HP-GAN2 learned to differentiate between fake and real human motion. So, our hypothesis is that if the discriminator network can differentiate between synthetic and real skeleton poses, then it also has learned some of the dynamics of a real human motion, and that those dynamics are useful in classification as well. We will show through multiple experiments that that is indeed the case. Therefore, our model learns to predict multiple future sequences of human poses from the same input sequence. We also show that the discriminator learns a general representation of human motion by using the learned features in an action recognition task. And we train a motion-quality-assessment network that measure the probability of a given sequence of poses are valid human poses or not. We test our model on two of the largest human pose datasets: NTURGB-D, and Human3.6M. We train on both single and multiple action types. The predictive power of our model for motion estimation is demonstrated by generating multiple plausible futures from the same input and showing the effect of each of the several loss functions in the ablation study. We also show the advantage of switching to GAN from WGAN-GP, which we used in our previous work. Furthermore, we show that it takes less than half the number of epochs to train an activity recognition network by using the features learned from the discriminator.

In modern life, human have become dependent on portable electronics, such as cell phones, MP3 and handheld computers, most of which are powered by batteries. Although the performance of batteries is being continuously improved, the limited energy storage and service life constrain the lasting use of these mobile electronics. Therefore it is desirable to find alternative or supplementary methods to solve this problem from its root cause. It is known that human body contains rich chemical energy, part of which is converted to mechanical energy up to 200W when in motion, so it is ideal to harvest a small fraction of the human kinetic energy to power mobile electronic devices. / In this thesis, first, the previous work done by other researchers on energy harvesting from human motion, especially from unintentional human motion, such as arm swing and leg moving, is reviewed. Then the fundamental principles to mechanically harvest motion energy are discussed, including the mechanical oscillating mechanisms and electromagnetic transduction. Derived from the general harvesting model, four different devices are designed and analyzed. / Shoe is important for human, one of which functions is to serve as shock-absorber to protect foot from the large impact force. As the foot strikes the ground, the shoe is subject to not only large force but also large displacement in the heel. The third new device is designed to insert in the shoe heel to harvest the kinetic energy from foot strike, and at the same time to function as a shock absorber for foot. Considering the stability and efficiency, a spring-slider-crank mechanism is used in this harvester to covert the up-down foot strike motion into unidirectional rotation to drive an AC generator. The spring and slider compose an oscillating system to absorb the foot strike motion, and crank and slider make up the conversion mechanism to transfer the bi-directional translation into unidirectional rotation. A set of gear is used to speed up the rotation. The kinematical performance of the harvester is also analyzed. / The first one is the automatic winding mechanism of mechanical movement. It consists of an oscillating weight, a ratchet mechanism, a gear set and a mainspring. The mechanism can be modeled as a double pendulum when worn on a user's wrist. Its kinematical performance is analyzed with experimental validation. / This thesis discusses the feasibility of mechanical power generators driven by human motion, with the focus on their architecture design and performance analysis. The main objective is to develop effective power generators for harvesting the energy from human motion, and use it to power portable electronic devices. / To directly convert the human arm motion to electricity, the second novel energy harvester is designed, analyzed and simulated. It mainly consists of an eccentric rotor made of permanent magnet, and a set of coils as a stator. The eccentric rotor, as a simple pendulum, acts as the kinetic energy harvester which absorbs the motion from human body in motion. With the permanent magnets on the rotor, the moving rotor produces a changing magnetic field, from which the stator induces electricity. In this design, a torsion spring is also added onto the rotor so that the harvester works even when the motion is on horizontal plane. / When foot strikes the ground, a large acceleration is produced. The fourth new energy harvester uses dual-oscillating mode. It contains two oscillating mechanisms: one is spring-mass oscillator to absorb the vibration from footstep motion, and the other is cantilever beam using the tip mass to amplify the vibration. Analysis shows that the dual-oscillating mechanism can be more effectively harvest the foot step motion. The energy conversion sub-mechanism is based on the electromagnetic induction, where the coils fixed at the tip end of the cantilever beam serves as the slider, and the fixed permanent magnets and yoke produce the changing magnetic field. Mathematical analysis and simulation are included. / Xie, Longhan. / Adviser: Ruxu Du. / Source: Dissertation Abstracts International, Volume: 72-04, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2010. / Includes bibliographical references (leaves 124-128). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

Part 1. Characteristics of foot movement in Tai Chi exercise. The performance of 16 experienced Tai Chi practitioners demonstrating a whole set of 42-form Tai Chi movements were recorded with two cameras. The APAS motion analysis system was used to identify the foot supporting and stepping characteristics during the practice. Seven foot support patterns and six step directions were identified. The results revealed that compared with normal walking, Tai Chi movement had more double support and less single support total duration. The duration of each support pattern was longer and movement from one pattern to the next was slow. The duration of each step direction was short, and changes of direction were frequent. It was expected that support patterns changed slowly, and combined with various step directions, they were found to be better than those of walking in simulating the gait challenges that may be encountered in daily activities. / Part 2: The plantar pressure distribution during Tai Chi exercise. The purpose of this study is to describe and quantify the plantar pressure distribution characteristics during Tai Chi exercise and to explain the beneficial effects of Tai Chi on balance control and muscle strength as compared with normal walking. Sixteen experienced Tai Chi practitioners participated in this study. Five typical Tai Chi movements represented by stepping forward, backward, sideways, up-down, and fixing could be isolated from the whole set of 42-form Tai Chi. The pressure-time integral, ground reaction force, displacement of center of pressure during the performance of the five typical movements were recorded and analyzed by the Pedar-X insole system (Germany). Results showed that during Tai Chi movements, the loading of the first metatarsal head and the great toe were significantly greater than in other regions (p&lt;0.05). The ground reaction forces varied between the Tai Chi movements and normal walking. Compared with normal walking, the locations of the center of pressure in the Tai Chi movements were significantly more medial and posterior at initial contact (p&lt;0.05), and were significantly more medial and anterior at the end of contact with the ground (p&lt;0.05). The displacements of the center of pressure were significantly wider (p&lt;0.05) in the mediolateral direction in the forward, backward and sideways Tai Chi movements. The displacement was significantly larger (p&lt;0.05) in the anterposterior direction in the forward movement. The plantar pressure characteristics of Tai Chi movements found in this study may be one of the important factors for Tai Chi's improvement of balance control and muscle strength. / Part 3. The duration and plantar pressure distribution during one-leg stance in Tai Chi exercise. The aim of this study is to quantify the one-leg stance duration and plantar pressure distribution during the one-leg stance in Tai Chi and to try to elaborate on its probable effects on the ability to balance on one leg. Sixteen experienced Tai Chi practitioners participated in this study. The Novel Pedar-X insole system (Germany) was used to record the plantar forces during the execution of a set of 42-form Tai Chi movements and during normal walking. The one-leg stance duration and plantar pressure distribution during the one-leg stance were analyzed. Results showed that in Tai Chi, the total duration spent in the one-leg stance was less (p&lt;0.05), the duration of each one-leg stance was longer (p&lt;0.01) and the medial-lateral displacement of the center of pressure was greater (p&lt;0.05) than during normal walking. The peak pressure and pressure-time integral of the second and third metatarsal heads and the fourth and fifth metatarsal heads were significantly greater (p&lt;0.05) than those of other plantar regions during the one-leg stance in normal walking; whereas the peak pressure and pressure-time integral of the first metatarsal head and the great toe were significantly greater (p&lt;0.05) than those of other plantar regions during the one-leg stance in Tai Chi. The longer duration of each one-leg stance and the plantar pressure distribution characteristics during the one-leg stance in Tai Chi may be associated with an improved ability to balance on one leg. / Part 4. The muscle contraction characteristics of the lower extremities during Tai Chi exercise. The objective of this study is to examine the muscle contraction characteristics of the lower extremities during Tai Chi exercise and to explain the beneficial effect of Tai Chi on the improvement of muscle strength. Sixteen experienced Tai Chi practitioners participated in this study. Five typical Tai Chi movements, represented by stepping in forward, backward, sideways, up-down and fixing were selected. The electromyographic activity of the rectus femoris, semitendinosus, gastrocnemius, and anterior tibialis muscles were recorded by Delsys electromyography measurement system (USA) during the performance of five typical Tai Chi movements. (Abstract shortened by UMI.) / Mao Dewei. / "August 2006." / Adviser: Youlian Hong. / Source: Dissertation Abstracts International, Volume: 68-03, Section: B, page: 1598. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2006. / Includes bibliographical references (p. 102-112). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

The need to characterize normal and pathological human movement has consistently driven researchers to develop new tracking devices and to improve movement analysis systems. Movement has traditionally been captured by either optical, magnetic, mechanical, structured light, or acoustic systems. All of these systems have inherent limitations. Optical systems are costly, require fixed cameras in a controlled environment, and suffer from problems of occlusion. Similarly, acoustic and structured light systems suffer from the occlusion problem. Magnetic and radio frequency systems suffer from electromagnetic disturbances, noise and multipath problems. Mechanical systems have physical constraints that limit the natural body movement. Recently, the availability of low-cost wearable inertial sensors containing accelerometers, gyroscopes, and magnetometers has provided an alternative means to overcome the limitations of other motion capture systems. Inertial sensors can be used to track human movement in and outside of a laboratory, cannot be occluded, and are low cost. To calculate changes in orientation, researchers often integrate the angular velocity. However, a relatively small error or drift in the measured angular velocity leads to large integration errors. This restricts the time of accurate measurement and tracking to a few seconds. To compensate that drift, complementary data from accelerometers and magnetometers are normally integrated in tracking systems that utilize the Kalman filter (KF) or the extended Kalman filter (EKF) to fuse the nonlinear inertial data. Orientation estimates are only accurate for brief moments when the body is not moving and acceleration is only due to gravity. Moreover, success of using magnetometers to compensate drift about the vertical axis is limited by magnetic field disturbance. We combine kinematic models designed for control of robotic arms with state space methods to estimate angles of the human shoulder and elbow using two wireless wearable inertial measurement units. The same method can be used to track movement of other joints using a minimal sensor configuration with one sensor on each segment. Each limb is modeled as one kinematic chain. Velocity and acceleration are recursively tracked and propagated from one limb segment to another using Newton-Euler equations implemented in state space form. To mitigate the effect of sensor drift on the tracking accuracy, our system incorporates natural physical constraints on the range of motion for each joint, models gyroscope and accelerometer random drift, and uses zero-velocity updates. The combined effect of imposing physical constraints on state estimates and modeling the sensor random drift results in superior joint angles estimates. The tracker utilizes the unscented Kalman filter (UKF) which is an improvement to the EKF. This removes the need for linearization of the system equations which introduces tracking errors. We validate the performance of the inertial tracking system over long durations of slow, normal, and fast movements. Joint angles obtained from our inertial tracker are compared to those obtained from an optical tracking system and a high-precision industrial robot arm. Results show an excellent agreement between joint angles estimated by the inertial tracker and those obtained from the two reference systems.

Upper body control is required to complete many daily tasks. One needs to stabilize the head and trunk over the pelvis, as one shifts the center of mass to interact with the world. While healthy individuals can perform activities that require leaning, reaching, and grasping readily, those with neurological and musculoskeletal disorders present with control deficits. These deficits can lead to difficulty in shifting the body center of mass away from the stable midline, leading to functional limitations and a decline in the quality of activity. Often these patient groups use canes, walkers, and wheelchairs for support, leading to occasional strapping or joint locking of the body for trunk stabilization. Current rehabilitation strategies focus on isolated components of stability. This includes strengthening, isometric exercises, hand-eye coordination tasks, isolated movement, and proprioceptive training. Although all these components are evidence based and directly correlate to better stability, motor learning theories such as those by Nikolai Bernstein, suggest that task and context specific training can lead to better outcomes. In specific, based on our experimentation, we believe functional postural exploration, while encompassing aspects of strengthening, hand-eye coordination, and proprioceptive feedback can provide better results. In this work, we present two novel cable robotic platforms for seated and standing posture training. The Trunk Support Trainer (TruST) is a platform for seated posture rehabilitation that provides controlled external wrench on the human trunk in any direction in real-time. The Stand Trainer is a platform for standing posture rehabilitation that can control the trunk, pelvis, and knees, simultaneously. The system works through the use of novel force-field algorithms that are modular and user-specific. The control uses an assist-as-needed strategy to apply forces on the user during regions of postural instability. The device also allows perturbations for postural reactive training. We have conducted several studies using healthy adult populations and pilot studies on patient groups including cerebral palsy, cerebellar ataxia, and spinal cord injury. We propose new training methods that incorporate motor learning theory and objective interventions for improving posture control. We identify novel methods to characterize posture in form of the “8-point star test”. This is to assess the postural workspace. We also demonstrate novel methods for functional training of posture and balance. Our results show that training with our robotic platforms can change the trunk kinematics. Specifically, healthy adults are able to translate the trunk further and rotate the trunk more anteriorly in the seated position. In the standing position, they can alter their reach strategy to maintain the upper trunk more vertically while reaching. Similarly, Cerebral Palsy patients improve their trunk translations, reaching workspace, and maintain a more vertical posture after training, in the seated position. Our results also showed that an Ataxia patient was able to improve their reaching workspace and trunk translations in the standing position. Finally, our results show that the robotic platforms can successfully reduce trunk and pelvis sway in spinal cord injury patients. The results of the pilot studies suggest that training with our robotic platforms and methods is beneficial in improving trunk control.

TThis thesis looks at the development of a Hand and Wrist Strength Gauge. The gauge was developed by the author at the Flinders Medical Centre Occupational Therapy Department in 1997. / thesis (MApSc(OccupationalTherapy))--University of South Australia, 2003.

The hypothesis underlying this research is that hand activity produces changes in the tissues of the hand which are reflected in the various functions of those tissues. Understanding the effect of hand activity upon hand function would allow occupational therapists to assess the efficacy of therapeutic interventions upon a clients ability to perform hand activity without damage to the tissues of the hand. Such information could assist in the design of safe and sustainable work tasks. The first step towards understanding how activity affects the hand is to measure its effects. The aim of this research is to determine which instruments can measure the effects of activity upon the hand.

Thesis (Ph. D.)--Ohio State University, 2006. / Title from first page of PDF file. Includes bibliographical references (p. 178-191).

Have you ever attempted to rub your abdomen with one hand while tapping your head with the other? Separately these movements are easy to perform but doing them together (bilateral task) requires motor adaptation. Motor adaptation is the process through which the Central Nervous System improves upon performance. Transfer of learning is the process through which learning a motor task in one condition improves performance in another condition. The purpose of this study was to determine whether transfer of learning occurs during bilateral goal-directed reaching tasks. It was hypothesized that transfer of learning would occur from the non-dominant to the dominant arm during bilateral tasks and that position and load feedback from the arms would affect the rate of adaptation and transfer of learning. During the experiments, subjects reached with one or both their index finger(s) to eight targets while a velocity dependent force perturbation (force environment) was applied to the arm(s). Three groups of bilateral tasks were examined: (1) unilateral reaching, where one arm learned to reach in a force environment, while the other arm remained stationary and therefore did not provide movement related position or load feedback; (2) bilateral reaching single load, where both arms performed reaching movements but only one arm learned a force environment and therefore the other arm provided movement related position feedback but not load feedback; (3) bilateral reaching two loads, where both arms performed reaching movements and both learned a force environment, while providing movement related position and load feedback. The rate of adaptation of the force environment was quantified as the speed at which the perturbed index finger trajectory became straight over the course of repeated task performance. The rate of adaptation was significantly slower for the dominant arm during the unilateral reaching tasks than during the bilateral reaching single load tasks (p<0.05). This indicates that the movement related position feedback from the non-dominant arm improved significantly the motor adaptation of the dominant arm; therefore transfer of learning occurred from the non-dominant to the dominant arm. The rate of adaptation for the non-dominant arm did not differ significantly (p>0.05) between the unilateral reaching and bilateral reaching single load tasks. Results also indicated that the rate of adaptation was significantly (p<0.05) faster for both the non-dominant and the dominant arms during the bilateral reaching two loads tasks than during the bilateral reaching single load tasks. The latter results indicate that transfer of learning occurred in both directions - from the dominant to the non-dominant arm and from the non-dominant to the dominant arm - when position and load feedback was available from both arms, but only when the force environment acted in the same joint direction. This study demonstrated that transfer of learning does occur during bilateral reaching tasks and that the direction and degree of transfer of learning may be modulated by the position and load feedback that is available to the central nervous system. This information may be used by physical therapists in order to improve rehabilitation strategies for the upper extremity.