Ning Lan

Phantom finger perception evoked with transcutaneous electrical stimulation for sensory feedback of prosthetic hand

This paper reports a pilot study to explore the plausibility of using the phenomenon of phantom finger perception (PFP) in amputees to develop sensory feedback for prosthetic hand. PFP can be aroused by touching a specific part of the skin in the stump area of amputees. We hypothesized that transcutaneous electrical stimulation (TES) can evoke the similar PFP in amputees with electrodes placed at the same skin area. We tested this hypothesis in subjects with distal amputation at the forearm above the wrist. The areas of PFP on the stump skin of amputee subjects were first identified and labeled by manually touching the skin. Electrical stimulation was then applied to the same area with surface electrode. The subjects reported that the corresponding finger was touched with electrical stimulation, and gradation of sensation with increased strength of stimulation current was perceived in a similar way as reported in normal subjects. Peripheral nerve sprouting into the receptors in stump skin after amputation is speculated as the neural mechanism of PFP. Preliminary results support the hypothesis that PFP evoked by TES can be utilized to establish a direct sensory feedback from the fingers of prosthetic hand to the amputee.

Evaluation of Feedforward and Feedback Contributions to Hand Stiffness and Variability in Multijoint Arm Control

The purpose of this study is to validate a neuromechanical model of the virtual arm (VA) by comparing emerging behaviors of the model to those of experimental observations. Hand stiffness of the VA model was obtained by either theoretical computation or simulated perturbations. Variability in hand position of the VA was generated by adding signal dependent noise (SDN) to the motoneuron pools of muscles. Reflex circuits of Ia, Ib and Renshaw cells were included to regulate the motoneuron pool outputs. Evaluation of hand stiffness and variability was conducted in simulations with and without afferent feedback under different patterns of muscle activations during postural maintenance. The simulated hand stiffness and variability ellipses captured the experimentally observed features in shape, magnitude and orientation. Steady state afferent feedback contributed significantly to the increase in hand stiffness by 35.75 ± 16.99% in area, 18.37 ± 7.80% and 16.15 ± 7.15% in major and minor axes; and to the reduction of hand variability by 49.41 ± 21.19% in area, 36.89 ± 12.78% and 18.87 ± 23.32% in major and minor axes. The VA model reproduced the neuromechanical behaviors that were consistent with experimental data, and it could be a useful tool for study of neural control of posture and movement, as well as for application to rehabilitation.

Design of a high voltage stimulator chip for a stroke rehabilitation system

This paper describes the design of an 8-channel high voltage stimulator chip for rehabilitation of stroke patients through surface stimulation, which requires high stimulation currents and high compliance voltage. The chip gets stimulation control data through its Serial Peripheral Interface (SPI), and can accordingly generate biphasic stimulation currents with different amplitudes, duration, frequencies and polarities independently for each channel. The current driver is implemented with thick oxide devices with a supply voltage up to 90V. The chip is designed in a 0.35μm X-FAB high voltage process.

Development of a network FES system for stroke rehabilitation

This paper describes a Functional Electrical Stimulation (FES) system based on the distributed network structure for rehabilitation of stroke patients. This FES system performs surface stimulation to activate the nerve of paretic muscles for training stroke patients to relearn motor functions. The main components of the networked FES system include a master unit (MU), a distributed stimulation-sensor unit (DSSU), and a clinical computer. In this system, the MU can drive a set of DSSUs, which is located at the node on the distributed network structure. The MU also stores the stimulation plan of rehabilitation training prescribed by clinicians. The DSSU serves as a single channel stimulator whose current amplitude, duration and frequency can be modulated by the MU. This system has two distinctive characters. First, since a stimulator is designed as a node on the network, the number of stimulation channels could be expanded according to specific needs. Second, a sensor component can be incorporated in the DSSU to allow monitoring physiological variables. The two features of system design make the networked FES system practical and flexible in clinical applications. We have completed a prototype of system including hardware and software. The evaluation test indicates that the system performance meets design specifications.

Effects of electrical stimulation of cutaneous afferents on corticospinal transmission of tremor signals in patients with Parkinson's disease

It has been hypothesized that propriospinal neurons (PNs) in the C3-C4 spinal cord mediates cortical motor commands to the peripheral muscles during tremor in patients with Parkinsons disease (PD). However, there has been no direct evidence so far to support the role of PN in transmitting tremor commands. In this paper, we report the positive correlation of cutaneous afferents with reduction in tremor amplitude and frequency in PD patients. Resting tremor and EMGs of biceps, triceps, flexor digitorum superficialis (FDS), extensor digitorum (ED), flexor carpi ulnaris (FCU) and extensor carpi radialis (ECR) muscles were recorded while transcutaneous electrical stimulation (TES) was applied to the dorsal hand skin of PD subjects. We observed instant suppression of tremor amplitude and EMGs occurring immediately at the start of electrical stimulation. However at the end of stimulation, the tremor amplitude and EMGs showed a quick recovery to the level prior to stimulation. Spectral analysis indicated that cutaneous afferents also have a long-lasting memory effect on the tremor frequency, i.e. the tremor frequency post stimulation did not recover to the value prior to stimulation. Preliminary results implied that the reduction in tremor amplitude and EMGs could be due to the inhibitory effects of cutaneous afferents to PNs, supporting the hypothesis that the PN network is involved in transmission of cortical tremor commands to peripheral muscles.

A neuromuscular electrical stimulation strategy based on muscle synergy for stroke rehabilitation

Recent experiments have suggested that the central nervous system (CNS) makes use of muscle synergies as a neural strategy to simplify the control of a variety of movements by using a single pattern of neural command signal. This nature of muscle coordination could have great significance in the treatment and rehabilitation of upper limb impairments for hemiparetic patients post stroke.The use of neuromuscular electrical stimulation (NMES) for neural prosthetics or therapeutic applications has been demonstrated as a promising clinical intervention for stroke patients to recover motor function of the upper extremity. However, the existing NMES systems do not provide control methods for the patient to achieve an individualized and functional rehabilitation training. In this research work, muscle synergies from the flexionextension elbow antagonistic muscles were studied. Using motion information and EMG signals, muscle synergies were extracted using non-negative matrix factorization (NMF) method. Reconstructed signals obtained from the muscle synergies were then applied to the virtual arm (VA) model to test a synergy based NMES strategy. Results show close resemblance to the original elbow trajectory of normal movements and thus the feasibility to control movements in stroke patients for rehabilitation.

Validation of a Virtual Arm Model for Movement Control and Rehabilitation

The purpose of this study is to validate a computational model that can be applied to studies of movement control and rehabilitation. A two joint, six muscle, virtual arm (VA) model has been developed in previous work [Song et al. 2008a]. The VA model driven by internal noise of neural control of muscles, i.e. the signal dependent noise (SDN), displays a behavior of kinematic variability that is often observed in human motor performance. In the present study, simulations were carried out to generate variability behaviors of the VA model under open-loop conditions. The hand stiffness was evaluated through a theoretical calculation method. Simulation results were compared to the corresponding behaviors observed in human subjects. It was shown that the shape, magnitude, and orientation of the simulated hand stiffness and variability were consistent with those of experimental measurements cross a range of posture conditions. This general agreement proves that the computational model could be a viable approach to replicating the realistic human motor behaviors. The model could be a useful tool for simulation of motor deficits caused by centrally impaired functions, like Parkinsons disease and stroke, as well as for the development of rehabilitation strategies for these neurological disorders.

Dynamic simulation of perturbation responses in a closed-loop virtual arm model

A closed-loop virtual arm (VA) model has been developed in SIMULINK environment by adding spinal reflex circuits and propriospinal neural networks to the open-loop VA model developed in early study [1]. An improved virtual muscle model (VM4.0) is used to speed up simulation and to generate more precise recruitment of muscle force at low levels of muscle activation. Time delays in the reflex loops are determined by their synaptic connections and afferent transmission back to the spinal cord. Reflex gains are properly selected so that closed-loop responses are stable. With the closed-loop VA model, we are developing an approach to evaluate system behaviors by dynamic simulation of perturbation responses. Joint stiffness is calculated based on simulated perturbation responses by a least-squares algorithm in MATLAB. This method of dynamic simulation will be essential for further evaluation of feedforward and reflex control of arm movement and position.

A novel experimental method to evaluate motor task control in Parkinson's patients

In this paper, a novel experimental method was developed to study planar arm movement control in tremor dominant Parkinsons (PD) patients. The method utilized a ball-bearing supported fiberglass brace apparatus against gravity to maintain the upper extremity in the horizontal plane. Subjects can perform postural and movement tasks with minimum damping effects. Arm movements were recorded using the MotionMonitor II system concurrently with EMGs of multiple muscles. Testing results in normal subjects with and without the brace support showed that the inertia and damping effects were negligible for oscillatory arm movement at maximum voluntary frequency (MVF). The tremor behaviors in horizontal posture maintenance and reaching movement in three PD subjects were also obtained with this method. The average frequency of postural tremor was 4.34±0.15 Hz in all arm positions tested. However, the tremor magnitudes changed significantly with posture locations. In performing reaching movements, the tremor was inhibited prior to reaching, but resumed after reaching. These results may provide interesting insights into the pathological mechanisms of Parkinsonian tremor, as well as the modular nature of neural control of movements.

Analysis of muscle synergy for evaluation of task-specific performance in stroke patients

Muscle synergy represents a central neural module that organizes and activates a group of muscles when performing a certain task. However, whether muscle synergy is a good physiological indicator of motor ability in task performance for patients suffering stroke is not clear. The purpose of this study is to understand how information of task-specific muscle synergy in healthy subjects and patients post stroke can be used to evaluate their motor ability, and further to assist motor rehabilitation for stroke patients. Electromyography (EMG) signals and movement kinematics in reaching tasks were recorded in 5 healthy subjects and 4 stroke patients. Muscle synergies were extracted from EMGs and compared cross healthy and stroke subjects. Normal synergies displayed a characteristic pattern common in healthy subjects. But pathological synergies in stroke subjects lacked the characteristics of normal synergy without a common component, implicating varying extent of damage to the motor module due to lesion in cerebral circuits. Further analysis in stroke subjects showed that pathological patterns of synergy in stroke subjects corresponded to the abnormality in their movement control compared with healthy subjects. Data showed that task-specific muscle synergy did reveal a positive correlation to the ability of neural control of tasks. It was further observed that task-specific synergy was changed towards the normal pattern after intervention with functional electrical stimulation in patients post stroke.