Skip to main content

Detecting Subtle Changes in White Matter Volume Using Deep Learning Approaches

Primary supervisor

Yasmeen George

Co-supervisors

  • David Gonsalvez
  • Yasith Mathangasinghe

Background:

Imagine the human brain as a complex electrical grid, with over 80 billion neurons (nerve cells) acting as power stations. These power stations need to send electrical signals to each other efficiently. Myelin, a special lipid sheath, wraps around the neuron processes (axons) like insulation around electrical wires. This insulation ensures that the signals travel quickly and without losing strength, giving the brain’s “white matter” its name (Figure 1A).

When diseases like multiple sclerosis strike, it is as if the insulation around the wires gets damaged, causing the electrical signals to slow down. Over time, this damage can also affect the wires themselves, leading to even more significant problems, much like an electrical grid experiencing power outages.

To understand how white matter develops and is affected by disease, we use advanced imaging techniques like Magnetic Resonance Imaging (MRI) (Figure 1A). By utilising AI and deep learning, we can detect subtle changes in white matter volume. This is akin to having a sophisticated system that can monitor the health of the electrical grid, detecting even the smallest issues in the insulation, and providing valuable insights into the development and progression of white matter diseases.

Volumetric quantification of white matter in brain. A. Magnetic Resonance Imaging (MRI) machine (left) and an MRI of the brain, demonstrating white matter (right). B. A schematic of AI-based segmentation process of a major white matter tract.
Figure 1: Volumetric quantification of white matter in brain. A. Magnetic Resonance Imaging (MRI) machine (left) and an MRI of the brain, demonstrating white matter (right). B. A schematic of AI-based segmentation process of a major white matter tract.

 

 

Student cohort

Double Semester

Aim/outline

Project Aim:

This project aims to create an automated pipeline to analyse volumetric changes in major white matter tracts throughout the lifespan and understand how these changes are impacted by multiple sclerosis (Figure 1B). 

Required knowledge

Python, PyTorch, Foundations of Image Analysis