Can Tractography Predict Electrophysiological Connectivity?

Motivation: To infer electrophysiological functional connectivity from tractography-based structural connectivity. Goal(s): Determine the accuracy of electrophysiological connectivity prediction using tractography and vice versa. Approach: Five patients with refractory focal epilepsy were recruited. After 3T diffusion tensor imaging, the causal electrophysiological connectivity was measured by applying single-pulse electrical stimulation to ~200 implanted depth electrodes and recording the pulse-evoked potentials. Structural connectivity was computed using probabilistic streamline counts and white matter tracts defined by a tractography atlas. We study the correspondence between these structural and electrophysiological connectivities. Results: Electrode pairs that are structurally connected are electrophysiologically connected with probabilities 0.75-0.94 (mean:0.86, std:0.07). Impact: If we can predict the probability of electrophysiological signals via tractography then we have a non-invasive tool for predicting electrophysiological signaling in the brain with broad scientific and clinical applications including guidance for placement of electrodes for functional neurosurgeries.