|
|
|
|
|
|
|
Research Labs Neurosensor Control & Eye Movement The long term objective of this research is the continued development of a neural network model for an integrated head and eye movement neurosensory controller driven by visual, auditory and vestibular stimuli. This is a multiple input, multiple output system that will be addressed using theoretical and experimental techniques. The long term goal of this research is to produce a three dimensional theoretical model based on physiological and anatomical data that will accurately describe the head and eye movement responses to any combination of visual, auditory, and vestibular inputs. The proposed study attempts to quantify and model the central nervous system's response to multisensory information. The objective of this proposed research is the rigorous development of a homeomorphic neural network (system) whose basic elements will rely on single neuron models and muscle models from the sarcomere level. The neural network developed will quantitatively describe the role and activity of individual neurons in the neural network, and its topology (connectivity) will reflect known CNS anatomy. The neuron model will be a Genesis type molecular-chemical model, simulated in MATLAB's SIMULINK, and will include the dendrites, cell body and axon. The neuron and network model will be viewable from the molecular to the large system level. Each neuron population in the network will be quantified with experimental data, and represent specific nuclei properties. The quantitative muscle model will be developed at the sarcomere level, and have global characteristics based on the muscle model created by Enderle et. al. The cross-bridge level muscle model will be simulated in SIMULINK, quantified with experimental data. The complete head and eye movement system will be constructed in Pro/Engineer 2000 and Pro/Mechanica2000 formed from SIMULINK simulations. The anatomical insertion points of the muscles, as well as their effective pulley locations, will be carefully located and accounted for. The output of Pro/Engineer 2000 will provide an image of the system in which muscles are contracting and changing shape moving the eyes and head, all driven by the neural network. Since the system is driven by subsystems, modeled mathematically and parameterized by experimental data, one can drill down to the molecular level to observe performance. Such a system will be an invaluable tool in designing new physiological experiments, but also an excellent educational tool. The insight gained in this study is directly applicable to many facets of clinical applications or sensory tests: e.g. guide future "brain repair" via targeted gene therapy or altered circuitry, assess effects of fatigue, aging, alcohol, drugs, or pathology on motoneuron activity, design control of artificial eye implants, or applications in robotics and virtual reality. Publications(PDF)
John Enderle |
|