Extension of NEURON simulator for simulation of reaction-diffusion in neurons

点击翻译按钮获取中文摘要

Development of simulation bioengineering techniques in medical neuroscience has been limited due to computational neuroscience's focus on the higher capacities of humans, such as the ability to play chess or go. That has led to the notion that the brain can best be understood by recourse to the concepts of computer science: artificial neural networks, information theory, Bayesian inference, pattern recognition, etc. Whether or not these tools are sufficient for understanding human neocortical function, they are clearly not sufficient for understanding the nervous system dysfunction seen in neurological and psychiatric pathology. Brain disease, like diseases of other organ systems, is disease of biological tissue, and must eventually consider energetics and oxygenation, blood and brain pressures, as well as chemical reactions and diffusion of toxins and pharmaceuticals. Such full-organ simulation will require linkages to other simulators. Through such linkages, as well as through internal extensions, we have been extending the widely-used NEURON simulator to handle reaction-diffusion in neural tissue in order to better understand brain signaling, neurodegenerative toxic cascades, and drug effects. For the current funding period we will further augment NEURON's NRxD module through 4 Aims: 1. Improve synaptic modeling techniques by considering presynaptic volume, cleft and postsynaptic volume together to handle ionotropic synapses, metabotropic synapses, gap junctions and complex combination synapses; along with reuptake, diffusion, electrodiffusion, neurotransmitters, second messengers and retrograde neurotransmitters in a coordinated way. 2. Allow more rapid and more extensive exploration of parameter space by improving simulation speed and by allowing full state variable saving for improved simulation initialization and recovery from high-performance computing failures. 3. Develop both Python-based and socket-based application programming interfaces (APIs) for easier linkage with other simulators, including for electrodiffusion and for various types of stochastic simulation. 4. Continue package dissemination and education in order to get more computational and experimental NRxD users. We will continue to offer twice-yearly tutorials, and to sponsor additional workshops at Computational Neuroscience and at other meetings. We will integrate online tutorials with the documentation to provide both Programmer's Reference and Biological Reference online manuals. We will develop a set of video tutorials associated with these that will eventually be linked together to make an online course. Overall, we expect that our novel simulation neurotechnology will be of increasing use and utilization both for research use, and for future clinical adoption in personalized medicine.