Toward whole-brain maps of neural connections: Logical framework and fast implementation

MRI tractography is the only method that noninvasively maps neural connections in the brain. Interest in its use for diagnosis and treatment of neurological disease is growing rapidly. Probabilistic tractography provides quantitative measures that can be interpreted as the strength or reliability of connections, but Monte Carlo implementations can require impractical computation times and have difficulty identifying connections between distal regions. Here, we develop a generic logical framework for probabilistic tractography with minimal assumptions that lends itself to solution by standard finite-difference methods. We demonstrate an implementation that outperforms Monte Carlo approaches in terms of computation time and identifying distal connections. The generality of the logic and the speed of the implementation indicate the potential of this approach for real-time mapping of whole-brain neural connections.