dc.description.abstracteng | Spinal cord injuries severely change the life of affected people. Besides major brain reorganizations following the de-afferentiation of affected limbs, the resulting immobility is a major problem affecting life quality and self-efficacy. In tetraplegia, which is the paralysis of all four limbs, mobility is especially challenging. With traditional control interfaces, such as chin or head control and sip-and-puff controllers, 50% of individuals with tetraplegia find it hard or impossible to perform even simple wheelchair tasks. Drawbacks of these systems include complicated use in daily-life and severe interference with other important activities (e.g. communication or head orientation). Since these control systems depend on the integrity of dedicated functions (e.g. control over the diaphragm for sip-and-puff, over neck muscles for chin control) individuals with tetraplegia are often even left without options for control of assistive devices. Novel experimental control systems that are operated with the tongue or by sniffing are promising, but still interfere with other activities.
In order to address this problem, we have developed an EMG-based novel control option, which is operated with the posterior auricular muscles (PAM). These small muscles behind the ear are especially suited for myoelectric control, because they comply with three prerequisites: first, because they are innervated by the facial nerve, they are still available in a number of sever neurological impairments, including spinal cord injuries. Second, unlike other facial muscles, the auricular muscles in humans are not involved in any relevant motor function, which secures minimal interference with other activities (e.g. facial movements, talking). Third, the isolated position of the posterior auricular muscle (PAM) ensures good accessibility and minimal crosstalk, rendering it particularly suitable for a myoelectric control interface.
In this thesis, I will present first proof-of-principles of the suitability of the auricular muscles as a signal generator for rehabilitation devices. We then used the developed system to address neuronal plasticity from a new perspective.
In the first study, ten healthy subjects and two subjects with tetraplegia were trained for five consecutive days. For the first time, it was shown with several parameters that subjects are able to learn to activate their auricular muscles. With appropriate training, they were able to produce fine graded activations and achieve bilateral distinction, i.e. only activate one side. This ability is important for wheelchair driving, and accordingly, all subjects were able to control an electric wheelchair with their auricular muscles on the fifth training day.
In the second study, the control system was applied to another rehabilitation device, an arm prosthesis. The conventional myoelectric control scheme is robust and easy to use, but does not allow for simultaneous activation of e.g. opening and turning the prosthetic hand. The combination of the auricular system with a forearm EMG system solves this problem. It was tested with 10 healthy subjects and one subject with arm amputation. Without training, the hybrid auricular control system was significantly faster and more accurate than two other established myoelectric control modes in an established test.
The third study aimed to exploit the potentials of the control system in a longer testing time. Nine tetraplegic subjects were trained in auricular control of an electric wheelchair for two weeks. Additionally, we aimed to investigate motor training effects on the brain reorganization that has been observed in SCI subjects. Subjects underwent fMRI and TMS measuring before and after training. We found significant improvements of the driving abilities with the electric wheelchair in complex tasks. Importantly, we found excessive motor task related brain activation, that has been suggested as maladaptive in the literature. Data indicates that it is on the contrary a phenomenon that enhances motor performance. Nevertheless, we show for the first time that it can be reduced through training, indicating a normalization of brain activity. These novel findings have important implications for physical rehabilitation of individuals with SCI.
In conclusion, the results of this thesis provide several new insights about the auricular muscles, e.g. that voluntary activation can be learned and trained, and that they are appropriate for fine-tuned signal generation. The auricular control system has several advantages over other control systems, e.g. continuous and proportional signal generation. Results show that the application of this control system to different rehabilitation devices is easy and performance results are promising. Importantly, results provide new information about brain reorganization after SCI. Training of the auricular muscles can reduce over-activation of motor related cortex areas in tetraplegic subjects, which has been suggested to be maladaptive. | de |