Computer simulations are essential for gaining insight into the relative contributions of anatomy and biophysics to the function of neurons and circuits. NEURON, a program which we have developed and provide freely for MacOS, MSWindows, and UNIX, has simplified this task for non-specialists in numerical methods and programming. It can simulate individual neurons and networks of neurons which include, but are not limited to, complex branching morphology, multiple channel types, in homogenous channel distribution, ionic diffusion, and restricted extracellular space. NEURON is being applied by neuroscientists around the world to examine cellular and network mechanisms that are affected by diseases such as epilepsy, multiple sclerosis, and disorders of learning and memory. To meet the evolving requirements of neuroscience research, NEURON must grow in several critical areas. New, exiting developments in large parameter space optimization could revolutionize computational neuroscience, but these techniques have to be organized and made available in a convenient manner. Optimization places new demands on numerical efficiency, we will continue to augment the class of feasible stimulations through the use of local variable time step methods. Maximizing the beneficial impact of these enhancements on neuroscience research requires dissemination of NEURON and the models that are constructed with it. Translation from physical system to conceptual model to simulation instance often requires collaboration between experimental and computational neuroscientists. To this end, we will provide user support through activities such as personal collaborations, short courses, and writing articles and manuals. Users also need in organizing models for effective use by others, which is indispensable for communication and reproducibility of results. This aspect of the project will employ database infrastructures developed through other initiatives such as the Human Brain Project.
Showing the most recent 10 out of 50 publications
