/MCI

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer's disease

Primary LanguagePythonMIT LicenseMIT

MCI

A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to Alzheimer's disease: Spasov et al. https://www.biorxiv.org/content/early/2018/11/15/383687

Overview

We provide our experimentation code for predicting Mild Cognitive Impairment to Alzheimer's Disease conversion.

  • mci_train.py runs our evaluation experiments
  • utils.py contains:
    • Out data loading and preprocessing procedures for the structural MRI, Jacobian Determinant images and clinical variables (preprocess.py)
    • The data iterator (augmentation.py)
    • A custom implementation of 3D separable convolutions (sepconv3D.py)
    • The network model architecture (models.py)

Dependencies

The code has been tested under Python 2.7.13, with the following packages installed (along with their dependencies):

  • numpy==1.14.5
  • scipy==1.0.0
  • sklearn==0.19.1
  • nibabel==2.1.0
  • tensorflow-gpu==1.10.1 In addition, CUDA 9.0 and cuDNN 7 have been used.

Network Architecture

ROC curve

ROC curve for pMCI/sMCI classification using our custom MRI template (and demographic, genetic and cognitive measure)

License

MIT