Neurophysiology-Driven Parameter Selection in nTMS-Based DTI Tractography: A Multidimensional Mathematical Model

Object: There is an increasing interest in preoperative diffusion tensor imaging-based fiber tracking (DTI-FT) to preserve function during surgeries in motor eloquent brain regions. However, DTI tractography is challenged by inherent presumptions during particular tracking steps [e.g., deterministic vs. probabilistic DTI, fractional anisotropy (FA) and fiber length (FL) thresholding] and the missing “ground truth” information. In the present study, we intended to establish an objective, neurophysiology-driven approach for parameter selection during DTI-FT of the corticospinal tract integrating both imaging and neurophysiological information. Methods: In ten patients with lesions in eloquent motor areas, preoperative navigated transcranial magnetic stimulation (nTMS) was performed, followed by individual deterministic DTI-FT from a grid of cortical seed points. We investigated over 300 combinations of FA and FL thresholds and applied subsequently a multidimensional mathematical modeling of this empirical data. Optimal DTI parameters were determined by the relationship between DTI-FT (i.e., number of fibers, NoF) and nTMS (i.e., amplitudes of motor-evoked potentials) results. Finally, neurophysiological DTI parameters and the resulting tractography were compared to the current standard approaches of deterministic DTI fiber tracking with a 75% and 50% FA and a FL threshold of 110 mm as well as with intraoperative direct cortical and subcortical stimulation. Results: There was a good goodness-of-fit for the mathematical model (r(2) = 0.68 ± 0 13; range: 0.59–0.97; n = 8) except of two cases. Neurophysiology-driven parameter selection showed a high correlation between DTI-FT and nTMS results (r = 0.73 ± 0.16; range: 0.38–0.93). In comparison to the standard approach, the mathematically calculated thresholds resulted in a higher NoF in 75% of patients. In 50% of patients this approach helped to clarify the exact tract location or to detect additional functional tracts, which were not identified by the standard approach. This was confirmed by direct cortical or subcortical stimulation. Conclusion: The present study evaluates a novel user-independent method to extract objective DTI-FT parameters that were completely based on neurophysiological data. The findings suggest that this method may improve the specificity and sensitivity of DTI-FT and thereby overcome the disadvantages of current approaches.