Meeting Banner
Abstract #3235

Assessment of Atlas Warping of Small Basal Ganglia on Colin27

Lara Bailey1, Yiming Xiao1, Mallar M. Chakravarty2, Abbas F. Sadikot3, D. Louis Collins1

1McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; 2Kimel Family Translational Imaging-Genetics Laboratory, Research Imaging Centre, Centre for Addiction and Mental Health, Toronto, Canada; 3Division of Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, Canada

Localization of basal ganglia (BG) is necessary for deep brain stimulation, but is difficult to achieve with clinical MRI. A method for warping a histological atlas to a T1w target contrast was previously validated for larger BG structures, but not smaller structures only visible on T2w contrasts. Here, we manually labelled T2w Colin12 and compared it with the atlas segmentation. Overall, there was agreement between the labels, yet the accuracy is not sufficient to delineate borders of the subthalamic nucleus, substantia nigra, and red nucleus. In the future, we will improve registration of these structures by including a T2w contrast.

Abstract PDF

Keywords

abstract accuracy accurate achieve addiction additional agree agreement allows although anterior approximately assessment asymmetric asymmetry atlas available axis basal beneficial beyond blue borders brain calculating chin clinical coefficient contrast coordinate course covered deep demonstrably despite developed dice difficult disagreement disease distance distinct division drive easily ensure even examine exception exist fact family finally future ganglia generating genetics geometric gold ground hand health hemisphere histological identified important individual inferior inflation inspection institute issue label labels laboratory lack lateral limited localization location longer lowest manual manually matches medial medical metric metrics native neurological neurosurgery note nuclei nucleus overall overlap painted painting patient placement poor position posterior proportional proportionally proportions pseudo quantitatively rater reflect reflects registered registering registration reliability respect rigidly segmentations sessions shape shrunken significantly smaller snap space statistics structural structure structures subjects sufficient superior table task thalamus though translational truth twice upon validated variable visibility visible visual visualization volume volumes warping worse