Analysis of functional magnetic resonance imaging (fMRI) data

Instructor: Petr Janata

Class meetings: Wednesday, 6:30-9:30PM, Center for Mind and Brain conference room (Annex, 202 Cousteau Pl.)

CRN: 69518

Instructor info:

• e-mail: pjanata@ucdavis.edu
• phone: 297-4471
• office hours: by appointment

Course description/objectives:  

This course is structured as a laboratory experience.  By the end of the course, students should exhibit competence in structuring, implementing, and interpreting analyses of fMRI data.  Students will work on data analysis projects throughout the quarter in small groups.  Datasets for analysis will be selected from the fMRI Data Center based on group interests. Two common fMRI data analysis packages will be used, SPM and FSL.  One goal will be to assess the relative merits of each package in terms of their implementations of data preprocessing and statistical modeling. Various topics will be explored including general issues in analysis of time-series data, and more specific issues of region of interest, multivariate, and meta analyses. Batch processing techniques will be emphasized throughout.  While knowledge of Matlab programming and/or shell-scripting is highly recommended and will greatly facilitate the batch processing objectives, it is not absolutely required for the course.  Basic programming concepts will be introduced as needed, but they will not be explored in depth.

Requirements:

1. Data analysis assignments.
2. Group data analysis project with presentation to class and term paper

Grading:

• 50% for assignments
• 50% for project/paper

---

Schedule

(subject to change)

Week 1, January 4, 2006

• Organizational. Introduction to Unix/Linux. fMRI Data Center. Computing at a distance.
• Data analysis packages: FSL, SPM
• Data formats
• Data viewers
• Matlab: variables, datatypes (numeric, strings, structures, cells), functions, scripts
• Batch processing
• Assignment 1 (due Jan. 18): Identify datasets to analyze

Week 2, January 11, 2006

• Matlab programming
  • scripts & functions
  • flow control
  • setting up the infrastructure for batch processing
  • display images
• SPM graphical user interfaces
• Readying the data for analysis
  • Understanding the data's organization
  • Rudimentary quality checks - Is the data worthy of analysis?
  • Example - checking for scrambling of slices

Week 3, January 18, 2006

• Readying the data for analysis
• Logging into the fMRIDC servers
• AFS
• Running VNC remotely
• Data quality checking (check_fmridc_data.m)
• SPM5 viewer
• Loading 3D and 4D image volumes into matlab
• Assignment 2 (due Jan. 25): Examine and compare organization of the datasets.

Week 4, January 25, 2006

• Reports on dataset organization
• Batch script for dataset integrity checking
  • Are all the data present for all subjects
  • Handling organization heterogeneity within and between experiments
• Assignment 3 (due Feb. 1): Batch script to verify dataset integrity, identify missing data.

Week 5, February 1, 2006

• Reports on dataset integrity and anomalies.
• Continued work on batch scripts for data integrity checking.
• Handling variability across experiments.
• Assignment 4 (due Feb. 8): Batch script to check for aberrent volumes.

Week 6, February 8, 2006

• Reports on image anomalies.
• Image pre-processing
• Preprocessing with FSL
  • GUI
  • preprocessing commands
• Preprocessing with SPM5
  • GUI
  • Structure
  • Job manager
• Spatial
  • Motion estimation and correction
  • Spatial coregistration
  • Spatial normalization
• Temporal
• Slice-timing correction
• Pre-whitening

Assignment 5 (due February 15): Preprocess data (using at least 2 different strategies)

Week 7, February 15, 2006

• Design matrix construction and model evaluation
• Design matrix specification
  • Block designs
  • Event-related designs
  • GLM
  • Regressors
  • Contrasts
• Design matrix specification in SPM5
  • GUI
  • Batch
• Design matrix specification in FSL
• GUI
• Batch (.fsf)
• FEAT

Assignment 5 (due February 22): Construct design matrices

Week 8, February 22, 2006

• Region of interest (ROI) analyses
• Masking
• Anatomy toolbox
• Time-courses
• HRF estimation
• ROI presentations
• Multivariate methods - ICA, partial least-squares, coherence, effective connectivity
• Assignment (due February 22): ROI analyses

Week 9, March 1, 2006

• Flat-mapping & Meta-analysis

Week 10, March 8, 2006

Final Presentations