ALS diagnosis and characterization based on T1w MRI

The goal of the research is to investigate the structural 3D brain T1-weighted MRI as tool to infer information about Amyotrophic Lateral Sclerosis (ALS). The project poses a challenge because T1-weighted MRI has traditionally been overlooked by clinicians, as it does not provide visual information that can be easily perceived by the human eye.

This work investigates two MRI representations to extract meaningful features for ALS diagnosis and characterization.

• Radiomic features
  This includes features of First Order Statistics, 3D Shape-based and 2D Shape-based features, Grey Level Co-occurrence Matrix features, 14 Gray Level Dependence Matrix features, Grey Level Run Length Matrix features, Gray Level Size Zone Matrix features, and Neighbouring Gray Tone Difference Matrix features.
• Deep Learning (DL) feature
  The DL features are obtained by leveraging Transfer Learning. More specifically, a pre-trained 3D Convolutional Neural Network is used as feature extractor. The CNN model has been pre-trained on T1-weighted MRI images to discriminate between healthy subjects and patients with Alzheimer's Disease. 

Partner: University of Turin - Department of Neuroscience, Istituto Neurologico "Carlo Besta" (Milan, Italy). 

PRIN 2017.