Article
Methodological considerations on predicting the sweet spot in deep brain stimulation based on probabilistic tractography
Methodische Überlegungen zur Vorhersage des „sweet spots“ für die tiefe Hirnstimulation basierend auf probabilistischer Traktographie
Search Medline for
Authors
Published: | June 4, 2021 |
---|
Outline
Text
Objective: Deep Brain Stimulation (DBS) has evolved to a standard treatment of various movement disorders. As DBS of specific neuronal structures also affects distant areas connected via fibres travelling through the volume of tissue activated, these fibre tracts themselves are also gaining specific interest for planning of DBS surgery and could serve as a base for patient-specific targeting. Probabilistic tractography is a powerful tool to reliably predict subcortical fibre tracts based on diffusion tensor imaging. We investigated possible methodologic influences and pit-falls in probabilistic tractography and their implications for clinical studies in DBS patients.
Methods: We retrospectively analysed 14 patients suffering from Essential Tremor and 19 patients suffering from Parkinson’s disease having received implantation of DBS systems at our centre. A workflow for probabilistic tractography was established using FSL 6.0.3 (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/). DWI-images were acquired at a 3T MRI scanner (Magnetom Skyra, SIEMENS) with 64 gradient directions. Patients were under general anaesthesia during image acquisition. We defined subcortical seed- and target regions to exemplarily track the dentato-rubro-thalamic tract (DRTT). DBS electrodes were reconstructed using LeadDBS (https://www.lead-dbs.org/). We evaluated possible influences in the workflow and their implication on measured distances to the specific electrode poles.
Results: Several pit-falls were identified in the workflow. Main influences on the measured distance of electrode poles to the DRTT were normalization into the MNI standard space and the method of definition of thresholds of the obtained probability maps. Additionally, the definition of criteria for evaluation of thresholds resulted in non-linear effects regarding distances from electrode poles. The way of measuring these distances either manually in a semi-objective manner or based on automated algorithms yielding to local distance maps resulted in slightly different findings.
Conclusion: Probabilistic tractography has been used multiply with respect to different questions in the field of DBS. As a variety of different parameters has to be defined for various steps in the workflow, special attention should be paid to possible influences on measurements used for these studies, which are often underestimated. Due to the lack of standardization, scientists should be aware of these influences regarding adequate interindividual comparability of results and objectively define individual standards.