Brain machine interface development has historically been hindered by slow prototyping due to the reliance on either expensive, labor-intensive animal experimentation or the static data obtained through such means. This work develops a Motor Cortex Simulator (MCS) that augments popular neural simulators by constructing complete motor cortex modules via fractal methods.

The MCS is unique in its ability to automatically generate models of three dimensional cortical cell populations with biologically plausible distributions of neurons, preferred directions, and synaptic connections. This is accomplished by entering a cortical region’s cellular and directional statistical data into fractal terrain mapping algorithms. The resulting distribution maps are layered to form a virtual scaffold in which cells are embedded along with their morphological definitions. By this method, a structure of hundreds to thousands of various classes of neurons may be generated pre-runtime from an anatomical definition.

This research integrates with two major, freely available resources: (1) a simulator, NEURON, that provides efficient system management and execution of the complex neural network, and (2) the online repositories ModelDB and NeuroMorpho that catalog previously published cellular models.

View the entire proposal in PDF format (MCS Thesis Proposal).