| Spinal cord injuries (SCI) disrupt both axonal pathways and segmental spinal cordcircuitry, producing severe impairments of motor, sensory, and autonomic function at andbelow the level of the injury. The recovery of walking function is one of the main goals ofpatients after spinal cord injury. However, there is no treatment available that restores theinjury-induced loss of function. Most people who experience SCI are destined to spend theremainder of their life in a wheelchair. The consequences of SCI are devastatingphysically and socially.Recent progress in clinic has shown epidural spinal cord stimulation (ESCS) combinedwith partial weight bearing training can facilitate standing and locomotion in patients withalmost intact neural circuits in the lumbar-sacral segment after spinal cord injuries.However, only a few spinal cord injured subjects were participated in these clinical trials.It is desirable to conduct further research in animals to decipher the potential mechanismsfor the exploration of optimal protocols and stimulating parameters to guide furtherclinical application of this promising treatment for motor function recovery after severeneural injury. There are no previous reports of the development of implantable ESCSdevice for animal research yet.The final option for restoring function to those with motor impairments is to provide thebrain with a new, non-muscular communication and control channel, a directbrain-computer interface (BCI) for conveying messages and commands to the externalword. Electroencephalography (EEG) based BCI can function in most environments,require relatively simple and inexpensive equipment, and non-invasive, offer thepossibility of a practical BCI. BCI technique makes it possible for paralyzed SCI patientsto control the ESCS device by their motor intention. The feasibility of this novel motorneural system reconstruction method largely depends on the techniques of personal motorintention acquisition and the signal processing.The research work has been done in this dissertation includes the following aspects:(1) The dissertation has investigated the feature extraction and the classificationmethods of EEG based BCI. Discrete wavelet transformation has been used for theextraction of features from sub-bands D2(16-32Hz) and D3(8-16Hz). The proposedfuzzy vector machine (FSVM) classifier and support vector machine (SVM) classifier onthe presented feature extraction methods achieve the high classification accuracy and largemutual information rate. (2) The dissertation presents the development of a fully implantable voltage-regulatedstimulator with bi-directional wireless communication for investigating underlying neuralmechanisms of ESCS facilitating motor function improvement. The stimulation systemconsists of a computer, an external controller, an implantable pulse generator (IPG), amagnet, the extension leads and a stimulation electrode. The IPG is coated with conformalcoating and silicone elastomer to keep tightness and biological compatibility. Thestimulation electrode is manufactured with flexible circuit board technique, and has threeround electrode contacts to target rat spinal cord structure. The contacts are separated with8mm center-to-center distance. The telemetry transmission between the IPG and theexternal controller is achieved by a commercially available transceiver chip with2.4GHzcarrier band. The magnet is used to activate the IPG only when necessary to minimize thepower consumption. The encapsulated IPG measures33mm×24mm×8mm, with a totalmass of~12.6g. The IPG supplied by a primary3V button battery can enable the chronicstimulation of spinal cord for about two weeks in the pilot study.(3) We develop a computer based integrated field-neuron model to study the influenceof pulse duration on the threshold of dorsal column fibers, dorsal root fibers and ventralroot fibers. Furthermore, the recruited neural structures under spinal cord stimulationcondition are discussed.(4) Animal experiments are conducted in Sprague-Dawley rats to validate the functionof the stimulator, and to investigate the relationship between ESCS parameters andhindlimb electromyography (EMG) responses. In our experiments, ESCS at lumbar-sacrallevel were effective for inducing hindlimb extension movements in normal anesthetizedrats. Stimulation under motor threshold did not induce observable changes in EMGresponse. However, when the stimulation amplitude was above motor threshold, the EMGresponses of the tibialis anterior (TA) muscle showed a progressive increase in thepeak-to-peak amplitude with increased stimulation amplitude. There was a progressivedecrease in the mean peak-to-peak amplitude with increasing frequency of stimulation. Incontrast, there was a progressive increase in the EMG amplitude with an increase instimulation pulse duration. |
PDF Full Text Request