Gon Khang

A Gait Rehabilitation and Training System based on Task Specific Repetitive Approach

Practice, effort and specificity have been proved as the major principles for gait rehabilitation. The best way to improve performance of a motor task is to execute that specific motor task again and again. To assure the consistency of a task specific repetitive gait rehabilitation training with accuracy, from both the patient and therapist ends for a long training session, automated gait trainer can be a solution for providing a patient having paretic or artificial leg, with a natural gait-like movement. This paper describes the design and implementation of a gait trajectory guiding device based on two automated foot-boards which move following a computer simulated patient-specific walking trajectory. Replication of a true walking pattern, robust and compact mechanical design and acting in harmony with the patients successive improvement were the main concerns in developing this device. An elaborate gait-database collected from various sources has been used to determine patient specific gait trajectory. Along with the guided foot panels, a body weight suspension system has been implemented for supporting the partial weight of the patient depending on feedbacks of the load sensors mounted on each foot-board. The main consideration of the software interface was to make sure a good graphical representation of continuous feedback-data from the device and patient history of consecutive training sessions.

Influence of the anisotropic mechanical properties of the skull in low-intensity focused ultrasound towards neuromodulation of the brain

Lately, neuromodulation of the brain is considered one of the promising applications of ultrasound technology in which low-intensity focused ultrasound (LIFU) is used noninvasively to excite or inhibit neuronal activity. In LIFU, one of critical barriers in the propagation of ultrasound wave is the skull, which is known to be highly anisotropic mechanically: this affects the ultrasound focusing, thereby neuromodulation effects. This study aims to investigate the influence of the anisotropic properties of the skull on the LIFU via finite element head models incorporating the anisotropic properties of the skull. We have examined the pressure and stress distributions within the head in LIFU. Our results show that though most of the pressure that reaches to the brain is due to the longitudinal wave propagation through the skull, the normal stress in the transverse direction of the wave propagation has the main role to control the pressure profile inside the brain more than the shear stress. The results also show that the anisotropic properties of skull contribute in broadening the focal zone in comparison to that of the isotropic skull.

A spanning tree-based human activity prediction system using life logs from depth silhouette-based human activity recognition

In this work, we propose a Human Activity Prediction (HAP) system using activity sequence spanning trees constructed from a life-log created by a video sensor-based daily Human Activity Recognition (HAR) system using time-sequential Independent Component (IC)-based depth silhouette features with Hidden Markov Models (HMMs). In the daily HAR system, the IC features are extracted from the collection of the depth silhouettes containing various daily human activities such as walking, sitting, lying, cooking, eating etc. Using these features, HMMs are used to model the time sequential features and recognize various human activities. The depth silhouette-based human activity recognition system is used to recognize daily human activities automatically in real time, which creates a life-log of daily activity events. In this work, we propose a method for human activity prediction using fixed-length activity sequence spanning trees based on the life-log. Utilizing the consecutive activities recorded in an activity sequence database (i.e. life-log) for a specific period of time of each day over a period such as a month, the fixed-length spanning trees can be constructed for the sequences starting with each activity where the leaf nodes contain the frequency of the fixed-length consecutive activity sequences. Once the trees are constructed, to predict an activity after a sequence of activities, we traverse the spanning trees until a path up to the previous node of the leaf nodes is matched with the testing pattern. Finally, we can predict the next activity based on the highest frequency of the leaf nodes along the matched path. The prediction experiments over the computer simulated data which is based on the daily logs show satisfactory results. Our video sensor-based human activity recognition and prediction systems can be utilized for practical applications such as smart and proactive healthcare.

Perception Caused by Current Amplitude Variation in Electro-Tactile Stimulation

Prosthetic hand devices are used to overcome the difficulties accompanying loss of the tactile abilities. We target the recreation of tactile sensations using electric stimulation as part of a feedback system for prosthetic devices. The experi- ment here described employed electric current with fixed parameters and variable amplitude for the excitation of the nerve fibers attached to the Merkel Disks Receptors of the skin. The amplitude thresholds defining the current interval to work with were determined for 10 subjects. These values ranged between 1.00-3.50 mA and 6.00-15.00 mA, respec- tively, different among the subjects, which indicated the indi- vidual differences in perceptual functioning. The discrimina- tion ability of the subjects, specifying the required amplitude variation between stimuli to be differentiated, was character- ized using a linear regression model. The resulting discrimina- tion ability value varied between 1.00-2.50 mA/sensation. A set of tactile descriptors related to the electrically-elicited sensations were quantified in accordance to the number of times they were reported. The result suggested that the elic- ited-sensations were comparable to the mechanically- generated sensations, being also easily describable and localiz- able.

A method for size estimation of amorphous pupil in 3-dimensional geometry

Measuring pupil size is a noninvasive method for evaluation of autonomic nervous system (ANS) activity because the pupil is exclusively controlled by the autonomic nervous system. To evaluate the activity of ANS, an accurate pupil size calculation method different from those of pupillometer and eye tracker should be suggested. This paper presents a three-dimensional model of eye and pupil and also suggests an accurate estimation method of pupil size from images projected on two-dimensional image plane with distortions.

Patient-specific walking pattern simulation in a gait trajectory guiding device

Repetitive training is of much importance for restoring full-fledged gait ability. At present, task-specific repetitive approach has been proved to be the most effective motor learning concept. In this regard, a gait trajectory guiding device with partial body weight support system can be a solution for gait rehabilitation. This paper presents a complete gait study with an objective to implement the motion of a natural walking pattern in the automated foot-boards of a gait trajectory guiding device. In our developed motion algorithm of foot-boards we have concentrated on adaptation of patient-specific true walking trajectory, determination of variable velocity pattern along different degrees of freedom and timedivision for simulating different phases of a complete gait cycle. Gait database, collected from disparate sources and previous gait-studies have been used for kinetic and kinematic analysis of human walking. We have modeled those data based on the previous researches done in this area and adopt them for our motion algorithm. A precise velocity pattern and time-division have been described along different axes so that patient’s biofeedback and postural stability in different walking phases can be recorded accordingly and motion-correction of the footboards can be done in consecutive cycles through iterative learning control algorithm with the help of motion sensors.

Measurment of Strain Generated Potential Near Bone and Implant Interface for Assessment of Osseointegration

In this study, a minimally invasive assessment using bone strain generated potential (SGP) was developed to examine the amount of osseointegration (OI) at bone-implant interface. SGP is generated by interstitial fluid flow in porous bone structure. Four experimental white New Zealand rabbits underwent pure titanium implant insertion surgery to tibia after amputation. After surgery, two animals were kept in small cages with minimal movement (Group 1). In contrast, the other rabbits were kept in a large cage that was large enough for jumping and walking (Group 2). At the end of the 5 weeks, all experimental animals were euthanized and the amputated tibia-implants were harvested. Then, a quasi-static force was applied to a bone site near the bone-implant interface for each tibia-implant specimen. Also, SGPs were measured near the interface using needle or probe electrodes. After the measurements, digital radiographs were taken to check the amount of OI for the interfaces. Full OI was observed for animals in Group 1. However, incomplete OI was found for animals in Group 2. Also, significant difference was found for mean SGP values between Group 1 and 2. The results could imply that SGP could be used as a minimally invasive assessment method to check the OI at the bone-implant interface.

The Elastic Moduli and Fatigue Properties of Canine Trabecular Bone Tissue

The elastic modulus and fatigue properties of canine and human trabecular bone tissues (single trabeculae) were experimentally determined on a microstructural level using four-point bending cyclic test, and they were compared based on microstructural characteristics and mineral density. The results showed that canine trabecular bone tissue had significantly lower modulus and lower fatigue strength than human tissue. The observed microstructural differences between the two tissues may be more responsible for the differences, although the lower mineral density in canine tissue might also have contributed to the lower modulus and fatigue strength.

Model of activation dynamics for an FES-induced muscle fatigue

A computer model of activation dynamics for an FES-induced muscle fatigue, which is one of the major problems in functional electrical stimulation (FES) has developed. Model consists of activation dynamics and Hill-type musculotendon dynamics. The authors modified the conventional activation dynamics to represent reduction and recovery of muscle force. By considering that FES can cause the shortening velocity of muscle to decline the authors modified the activation dynamics to affect the muscle shortening velocity. To model muscle fatigue the authors introduced the fatigue variable and the fatigue admittance assuming that the activation changes with the electrical admittance of muscle fibers. The fatigue variable relates the stimulation time to the applying pulses. The fatigue admittance changes the Ca/sup 2+/ uptake rate and thus the muscle shortening velocity as a function of the fatigue variable. The computer simulation results showed that the proposed admittance-type model match well with the experimental observations reported in literatures in terms of force reduction and recovery as well.

A Feasibility Study to Elicit Tactile Sensations by Electrical Stimulation

In this paper we investigated the feasibility of electrically-elicited tactile sensations. Eighteen subjects participated in the experiment and the constant current stimulation was applied to the subject’s index finger pad. The cathodic monophasic rectangular pulse train was used throughout the experiment. The pulse amplitude was increased in every 2.5 seconds in such a way that the charge increment per pulse was kept constant at 70 nanoC. Four different pulse widths (200, 500, 700 and 1,000 us) and three different frequencies (20, 50 and 200 Hz) were selected to investigate the effect of the pulse parameters on the elicited tactile sensations.