Perspectives of high-resolution 3D reconstruction for automated mapping of cytoarchitectonic human brain areas

The human brain can be subdivided into structural regions based on the organization of neuronal cells. To identify these regions, postmortem human brains are cut into thin histological slices, scanned using light-microscopic scanners, and analyzed by neuroscientists. However, established analysis methods are time and labour intensive and do not scale to the large amounts of acquired image data. This motivates the ongoing development of Deep Learning methods to automatically identify brain regions from microscopic images.
Analysis of individual 2D images is not always sufficient to identify brain regions, as some relevant tissue properties can only be recovered from 3D context. Recent work on 3D brain reconstruction at cellular resolution now enables Deep Learning methods to exploit 3D context information for improved brain region classification. This project will investigate how 3D reconstructed brain tissue can benefit Deep Learning-based classification of brain regions and thus lay the foundation for next-generation brain mapping techniques.

Faculty Supervisor:

Alan Evans

Student:

Partner:

Forschungszentrum Jülich

Discipline:

Life Sciences

Sector:

Artificial Intelligence; Health and Related Sciences & Technology

University:

McGill University

Program: