Overview

This graduate course provides a thorough grounding in the principles and practice of functional MRI. Students gain hands-on experience with analysis pipelines (FSL, SPM, nilearn) and understand the statistical foundations of each step. The lab component uses real datasets from the ACL@NCU.

Prerequisites: Graduate standing; basic statistics and programming background helpful.

Part 1: MRI Physics and the BOLD Signal

  • Proton magnetic resonance basics
  • Pulse sequences: EPI, gradient echo, spin echo
  • The BOLD signal: neural coupling, hemodynamic response function
  • Signal-to-noise and field strength considerations
Lab resources

Part 2: Experimental Design

  • Block, event-related, and mixed designs
  • Design efficiency and counterbalancing
  • Sample size and power in neuroimaging
  • Pre-registration and registered reports

Part 3: Preprocessing Pipeline

  • BIDS formatting and data organization
  • Slice-timing and motion correction
  • Spatial normalization to MNI space
  • Spatial smoothing and temporal filtering
  • fMRIPrep: a standardized preprocessing workflow

Part 4: The General Linear Model

  • Design matrix construction and HRF convolution
  • Contrast specification and statistical maps
  • Multiple comparison correction (FWE, FDR, cluster-based)
  • Random effects and second-level analysis

Part 5: Multivariate Pattern Analysis

  • The rationale for MVPA
  • Cross-validation schemes and bias avoidance
  • Searchlight analysis
  • Representational Similarity Analysis (RSA)

Part 6: Connectivity and Advanced Methods

  • Seed-based functional connectivity
  • Psychophysiological interaction (PPI)
  • Dynamic causal modelling (DCM)
  • Resting-state ICA and dual regression

Part 7: Reproducibility and Open Science

  • Data sharing: OpenNeuro, OSF
  • Code sharing and containerization (Docker, Singularity)
  • Effect size estimation in neuroimaging
  • Replication crisis and solutions