SciNet Education Site

Get familiar with how one can manipulate and transform neuroimaging data using Python s neuroimaging packages (nibabel, nilearn). Develop an understanding how MRI data is represented in Python and perform some hands-on tasks such as basic manipulation on both structural MR and functional MR. Then we will discuss the steps required to take minimally pre-processed MR data (fmriprep), to clean and workable data through the process of motion cleaning and dimensionality reduction. Finally, we will cover how to perform functional connectivity (FC) analysis to build a resting state connectivity matrix. All analyses will be performed using Jupyter notebooks in the spirit of reproducible and open science.

Practical Introduction to machine learning for neuroimaging: classifiers, dimensionality reduction, cross-validation and neuropredict. How to apply machine learning to your data, even if you do not know how to program. Learn what is machine learning and get a high-level overview of few popular types of classification and dimensionality reduction methods. Learn (without any math) how support vector machines work. Learn how to plan a predictive analysis study on your own data? What are the key steps of the workflow? What are the best practices, and which cross-validation scheme to choose? How to evaluate and report classification accuracy? Learn which toolboxes to use when, with a practical categorization of few toolboxes. This is followed by detailed demo of neuropredict, for automatic estimation of predictive power of different features or classifiers without needing to code at all.

The course will introduce participants to the neuroimaging data and best practices for the analysis of neuroimaging data using High Performance Clusters (HPC). We will introduce types of neuroimaging scanning modalities with instructions for how to organize these data using the Brain Imaging Data Structure (BIDS). We will then introduce Singularity container software (BIDS-apps) for the preprocessing of neuroimaging data (including mriqc and fmriprep) and demonstrate how to run them on the HPC. We will discuss general information about running Singularity containerized software on the HPC and how to construct custom containers for your own analysis using NeuroDocker. -- Prerequisites: basic Linux command line skills

How to find and use public datasets for neuroscience research with a focus on transcriptomics and neuroimaging. Introduction and guides to some of the largest datasets will be provided (Allen human brain atlases, BrainEAC, ADNI, and the Human Connectome Project).

Approaches to parallelization (utilizing SciNet systems) for image analysis.

Part of the 2017 Ontario Summer School.