1. Preprocessing of FMRI Data

What is preprocessing?

Signal, noise, and the General Linear Model

Signal-Noise-Ratio (SNR)

Preprocessing Steps

Tools for Preprocessing

Slice Timing Correction

Why do we correct for slice timing?

Effects of uncorrected slice timing

Base HDR: 2s TR

Base HDR + Noise

Base HDR + Slice Timing Errors

HDR + Noise + Slice Timing

Interpolation Strategies

Motion Correction

Head Motion: Good, Bad,…

… and  catastrophically bad

Why does head motion introduce problems?

Simulated Head Motion

Severe Head Motion: Simulation

Severe Head Motion: Real Data

Correcting Head Motion

Effects of Head Motion Correction

Limitations of Motion Correction

Prevention is the best medicine

Coregistration

Should you Coregister?

Normalization

Slide 29

Standardized Spaces

Normalization to Template

Anterior and Posterior Commissures

Should you normalize?

Slice-Based Normalization

Spatial Smoothing

Techniques for Smoothing

Effects of Smoothing on Activity

Slide 38

Should you spatially smooth?

Segmentation

Histogram of Voxel Intensities

Region of Interest Drawing

Why use an ROI-based approach?

Drawing ROIs

ROI Examples

Slide 46

Additional Resources

2. Issues in Experimental Design

What is Experimental Design?

Goals of Experimental Design

What are hypotheses?

Independent Variables

Dependent Variable: BOLD signal

Causal and non-causal relations between variables

Detection vs. Estimation

Detection

Estimation

Optimal Experimental Design

fMRI Design Types

1. Blocked Designs

What are Blocked Designs?

PET Designs

Choosing Length of Blocks

Effects of Block Interval upon HDR

What baseline should you choose?

Interpreting Baseline Activity

Non-Task Processing

Slide 68

Slide 69

Power in Blocked Designs

HDR Estimation: Blocked Designs

Power in Blocked Designs

Power in Blocked Designs

Power in Blocked Designs

Limitations of Blocked Designs

2. Event-Related Designs

What are Event-Related Designs?

Why use event-related designs?

Event-Related and Blocked Designs give Similar Results

2a. Periodic Single Trial Designs

Trial Spacing Effects: Periodic Designs

Slide 82

2b. Jittered Single Trial Designs

Randomization = Jittering

Extracting different task components

Effects of Jittering on Stimulus Variance

Effects of ISI on Power

2c. Staggered Single Trial

Slide 89

Slide 90

Post-Hoc Sorting of Trials

Limitations of Event-Related Designs

3. Mixed Designs

3a. Combination Blocked/Event

Slide 95

Mixed designs

3b. Variable Stimulus Probability

Random and Semi-Random Designs

Summary of Experiment Design