Subcortical electrophysiology deformed into a patient-specific brain atlas.

摘要

Stereotactic neurosurgery for movement disorders involves elements of radiological, anatomical, and electrophysiological identification of anatomy deep within the human brain. Most of the anatomical structures targeted stereotactically are not visualized well by standard medical imaging techniques, as regions of different functionality do not necessarily appear as differences in gray/white matter contrast. These indiscernible structures must be approximated on medical images with the aid of antiquated anatomical atlases of the brain ill-suited for the task. Exhaustive electrophysiologic exploration of the target anatomy refines the atlas-approximated target into a conclusive target for placement of a therapeutic lesion or chronic deep brain stimulator. While minimal, each passage of the exploratory electrode increases the risk of intracerebral hemorrhage and inducing morbidity.; Hundreds of thousands, if not millions, of electrophysiologic data are generated at surgical centres around the world as a byproduct of functional stereotactic neurosurgery. Post-operatively, these data could provide rare insight into the workings of the conscious human brain. However, in their raw form, they cannot be analyzed collectively with any accuracy nor can they be directly applied to the planning of subsequent surgeries.; This thesis reviews the development, contents, and validation of a novel probabilistic functional brain atlas of normalized intra-operative electrophysiologic observations. This research contributes to the field of image-guided stereotactic functional neurosurgery in three significant ways: (1) it presents a software framework that enables the display and analysis of single patient and pooled population physiologic data. The software suite is extensible, tailored for stereotactic surgical planning and guidance, and addresses several deficiencies of current surgical planning systems; (2) it describes the development of the first 3D probabilistic functional atlas of human electrophysiologic observations that now holds over 13 000 individually coded observations that may be directly applied to stereotactic surgical planning, guidance, and analysis. The probabilistic atlas is expandable, searchable, and capable of intersubject nonrigid registration to patient medical images; and (3) presents two validation studies of the probabilistic functional atlas. The first demonstrates its prognostic ability for predicting clinical outcome using a retrospective lesion-analysis technique. The second analyzes the somatotopic organization of thalamic nuclei represented by the atlas and presents unambiguous clustering of thalamic functional data relevant to surgical planning.; In addition, an extensible software platform, ASP, is presented that is sufficiently generic to allow rapid development and prototyping of surgical guidance software applications for intra-operative data acquisition and guidance. This platform is open-source, hardware independent, and has spawned the generation of a diverse collection of surgical guidance applications.