The effect of conflicting sensorimotor memories on optimal force strategies was explored. Subjects operated a virtual object controlled by a physical handle to complete a simple straight-line task. Perturbations applied to the handle induced a period of increased error in subject accuracy. After two blocks of 33 trials, perturbations switched direction, inducing increased error from the previous trials. Subjects returned after a 24-hour period to complete a similar protocol, but beginning with the second context and ending with the first. Interference from the first context on each day caused an increase in initial error for the second (P < 0.05). Following the rest period, subjects showed retention of the sensorimotor memory from the previous day through significantly decreased initial error (P = 3x10-6). However, subjects showed an increase in forces for each new context resulting from a sub-optimal motor strategy. Higher levels of total effort (P < 0.05) and a lack of separation between force values for opposing and non-opposing digits (P > 0.05) indicated a strategy that used more energy to complete the task, even when rates of learning appeared identical or improved. Two possible mechanisms for this lack of energy conservation have been proposed.

Research on human grasp typically involves the grasp of objects designed for the study of fingertip forces. Instrumented objects for such studies have often been designed for the simulation of functional tasks, such as feeding oneself, or for rigidity such that the objects do not deform when grasped. The goal of this thesis was to design a collapsible, instrumented object to study grasp of breakable objects. Such an object would enable experiments on human grip responses to unexpected finger-object events as well as anticipatory mechanisms once object fragility has been observed. The collapsible object was designed to be modular to allow for properties such as friction and breaking force to be altered. The instrumented object could be used to study both human and artificial grasp.

Telestroke networks reduce disparities in acute stroke care between metropolitan primary stroke centers and remote hospitals. Current technologies used to conduct remote patient assessments have very high start-up costs, yet they cannot consistently establish quality connection in a timely manner. Smartphgones can be used for high quality video teleconferencing (HQ-VTC). They are relatively inexpensive and widley used among healthcare providers. We aimed to study the reliability of HQ-VTC using smartphones for conducting the NIHSS. Two vascular neurologists (VNs) assessed 83 stroke patients with the NIHSS. The remote VN assessed patients using videoconferencing on a smartphone with the assistance of a bedside medical aide. The bedside VN rated patients ontemporaneously. Each VN was blinded to the other's NIHSS scores. We tested the inter-method agreement and physician satisfaction with the device. We demonstrated high total NIHSS score correlation between the methods (r=0.941, p<0.001). The mean total NIHSS scores for bedside and remote assessments were 7.3 plus or minus 7.9 and 6.7 plus or minus 7.6 with ranges of 0-30 and 0-37, respectively. Seven NIHSS categories had significantly high agreement beyond chance: LOC-questions, LOC-commands, visual fields, motor left arm, motor right arm, motor left leg, motor right leg; seven categories had moderate agreement: LOC-consciousness, best gaze, facial palsy, sensory, best language, dysarthria, extinction/inattention; one category had poor agreement: ataxia. There was high physician satisfaction with the device. The VNs rated 96% of the assessments as good or very good for "image quality," "sound quality," "ease of use," and "ability to assess subject using NIHSS," and 84% of the assesssments as good or very good for "reception in hospital." The smartphones with HQ-VTC is reliable, easy to use, and affordable for telestroke NIHSS administration. This device has high physician satisfaction. With the variety of smartphones and professional medical applications available today, the telestroke practitioner has all the tools necessary for fast clinical decision-makingby accessing electronic medial records, viewing images, and tracking patient vitals.

Startle-evoked-movement (SEM), the involuntary release of a planned movement via a startling stimulus, has gained significant attention recently for its ability to probe motor planning as well as enhance movement of the upper extremity following stroke. We recently showed that hand movements are susceptible to SEM. Interestingly, only coordinated movements of the hand (grasp) but not individuated movements of the finger (finger abduction) were susceptible. It was suggested that this resulted from different neural mechanisms involved in each task; however it is possible this was the result of task familiarity. The objective of this study was to evaluate a more familiar individuated finger movement, typing, to determine if this task was susceptible to SEM. We hypothesized that typing movements will be susceptible to SEM in all fingers. These results indicate that individuated movements of the fingers are susceptible to SEM when the task involves a more familiar task, since the electromyogram (EMG) latency is faster in SCM+ trials compared to SCM- trials. However, the middle finger does not show a difference in terms of the keystroke voltage signal, suggesting the middle finger is less susceptible to SEM. Given that SEM is thought to be mediated by the brainstem, specifically the reticulospinal tract, this suggest that the brainstem may play a role in movements of the distal limb when those movements are very familiar, and the independence of each finger might also have a significant on the effect of SEM. Further research includes understanding SEM in fingers in the stroke population. The implications of this research can impact the way upper extremity rehabilitation is delivered.

The purpose of this paper was to systematically review current literature regarding the effect of hand splints on aesthetic outcomes for individuals with acquired hand deformities. Hand splints vary in form and function, and are used to maintain or ameliorate hand function and aesthetics. A literature search was performed on peer-reviewed publications that used splinting as an intervention for conservative hand improvement. Evidence from ten randomized clinical trials (published from 2003 to 2015) was evaluated for aesthetic improvement among a total of 659 subjects. Cosmetic outcomes were analyzed by a change in angle measurements, such as extensor lag, ulnar deviation, and passive and active range of motion. Of these ten studies, five focused on hand deformities caused by neurological impairment, while the other five measured those with musculoskeletal complications. Only two of the ten studies concluded that splinting could aesthetically improve the hands, and only one of these reporting statistical significance in its data. The data was not only limited in quantity, but was presented in heterogeneous formats. There was an extensive variation in measured outcomes, intervention protocols, follow-up times, and many other aspects of the studies; this dissimilarity led to difficulty in performing a systematic assessment. The majority of evidence concludes that splinting does not improve the appearance of deformities, however none directly investigated this measure. Therefore, further RCTs that include measurements of cosmetic traits are necessary to better quantify the effect of splinting for management of hand deformities. This review was the first of its kind to evaluate the correction of hand deformities using splints as an intervention.

As robots become more prevalent, the need is growing for efficient yet stable control systems for applications with humans in the loop. As such, it is a challenge for scientists and engineers to develop robust and agile systems that are capable of detecting instability in teleoperated systems. Despite how much research has been done to characterize the spatiotemporal parameters of human arm motions for reaching and gasping, not much has been done to characterize the behavior of human arm motion in response to control errors in a system. The scope of this investigation is to investigate human corrective actions in response to error in an anthropomorphic teleoperated robot limb. Characterizing human corrective actions contributes to the development of control strategies that are capable of mitigating potential instabilities inherent in human-machine control interfaces. Characterization of human corrective actions requires the simulation of a teleoperated anthropomorphic armature and the comparison of a human subject's arm kinematics, in response to error, against the human arm kinematics without error. This was achieved using OpenGL software to simulate a teleoperated robot arm and an NDI motion tracking system to acquire the subject's arm position and orientation. Error was intermittently and programmatically introduced to the virtual robot's joints as the subject attempted to reach for several targets located around the arm. The comparison of error free human arm kinematics to error prone human arm kinematics revealed an addition of a bell shaped velocity peak into the human subject's tangential velocity profile. The size, extent, and location of the additional velocity peak depended on target location and join angle error. Some joint angle and target location combinations do not produce an additional peak but simply maintain the end effector velocity at a low value until the target is reached. Additional joint angle error parameters and degrees of freedom are needed to continue this investigation.

Electrochemical sensors function by detecting electroactive species at the electrode surface of a screen printed sensor. As more force is applied, the concentration of electroactive species at the surface of the sensor increases and a larger current is measured. Thus, when all conditions including voltage are made constant, as in Amp i-t, a quantifiable current can be read and the force applied can be calculated. Two common electrochemical techniques in which current is measured, cyclic voltammetry(CV) and amperometric i-t(Amp i-t), were used. A compressible sensor capable of transducing a force and acquiring feedback was created.

In 1937 Canadian neurosurgeon Wilder Penfield made the first to attempt to map the sensorimotor cortex of the human brain in his paper entitled Somatic Motor and Sensory Representation in the Cerebral Cortex of Man as Studied by Electrical Stimulation. While analogous experimentation had been carried out previously using animal subjects, Penfield sought to understand the delicate and complex neuronal pathways that served as the hidden control mechanisms for human activity. The motor homunculus that followed from his findings has been widely accepted as the standard model for the relative spatial representation of the functionality of the motor cortex, and has been virtually unaltered since its inception. While Penfield took measures to collect cortical data in a manner as accurately as scientifically possible for the time period, his original model is deserving of further analysis using modern techniques. This study uses functional magnetic resonance imaging (fMRI) to quantitatively determine motor function volumes and spatial relationships for four motor tasks: toe, finger, eyebrow, and tongue. Although Penfield's general representation of the superior-to-inferior spatial distribution of the motor cortex was replicated with reasonable accuracy, relative mean task volumes seem to differ from Penfield's original model. The data was first analyzed in each individual patient's native anatomical space for task comparison within a single subject. The volumes of the motor cortex devoted to the eyebrow and toe tasks, which comprise only small portions of the Penfield homunculus, are shown to be relatively large in their fMRI representation compared to finger and tongue. However, these tasks have large deviation values, indicating a lack of consistency in task volume size among patients. Behaviorally, toe movement may include whole foot movement in some individuals, and eyebrows may include face movement, causing distributions that are more widespread. The data was then analyzed in the Montreal Neurological Institute (MNI) space, which is mathematically normalized for task comparison between different subjects. Tongue and finger tasks were the largest in volume, much like Penfield's model. However, they also had substantial deviation, again indicating task volume size inconsistencies. Since the Penfield model is only a qualitative spatial evaluation of motor function along the precentral gyrus, numerical deviation from the model cannot necessarily be quantified. Hence, the results of this study can be interpreted standalone without a current comparison. While future research will serve to further validate these distances and volumes, this quantitative model of the functionality of the motor cortex will be of great utility for future neurological research and during preoperative evaluations of neurosurgical patients.

I worked on the human-machine interface to improve human physical capability. This work was done in the Human Oriented Robotics and Control Lab (HORC) towards the creation of an advanced, EMG-controlled exoskeleton. The project was new, and any work on the human- machine interface needs the physical interface itself. So I designed and fabricated a human-robot coupling device with a novel safety feature. The validation testing of this coupling proved very successful, and the device was granted a provisional patent as well as published to facilitate its spread to other human-machine interface applications, where it could be of major benefit. I then employed this coupling in experimentation towards understanding impedance, with the end goal being the creation of an EMG-based impedance exoskeleton control system. I modified a previously established robot-to-human perturbation method for use in my novel, three- dimensional (3D) impedance measurement experiment. Upon execution of this experiment, I was able to successfully characterize passive, static human arm stiffness in 3D, and in doing so validated the aforementioned method. This establishes an important foundation for promising future work on understanding impedance and the creation of the proposed control scheme, thereby furthering the field of human-robot interaction.

The role of retention and forgetting of context dependent sensorimotor memory of dexterous manipulation was explored. Human subjects manipulated a U-shaped object by switching the handle to be grasped (context) three times, and then came back two weeks later to lift the same object in the opposite context relative to that experience on the last block. On each context switch, an interference of the previous block of trials was found resulting in manipulation errors (object tilt). However, no significant re-learning was found two weeks later for the first block of trials (p = 0.826), indicating that the previously observed interference among contexts lasted a very short time. Interestingly, upon switching to the other context, sensorimotor memories again interfered with visually-based planning. This means that the memory of lifting in the first context somehow blocked the memory of lifting in the second context. In addition, the performance in the first trial two weeks later and the previous trial of the same context were not significantly different (p = 0.159). This means that subjects are able to retain long-term sensorimotor memories. Lastly, the last four trials in which subjects switched contexts were not significantly different from each other (p = 0.334). This means that the interference from sensorimotor memories of lifting in opposite contexts was weaker, thus eventually leading to the attainment of steady performance.