Statistical Signal Processing for fMRI Douglas N. Greve Mark Vangel Anastasia Yendiki Overview First-Level Univariate Analysis Signal Modeling Nuisance Modeling Noise Modeling Hypothesis Testing Correction for Multiple Comparisons Cross-Subject/Higher Level Analysis
Lab HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Analysis Goals Quantify Neural Correlates in fMRI Amplitude of Hemodynamic Response Delay/Shape of Hemodynamic Response Extent/Size of Activation Localization of function Quantify Uncertainty Cross-subject (within group) Cross-group eg, Normals, Clinical Populations Within-subject EEG/MEG/Optical/Surgical Planning
HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Challenges Large Noise thermal, physiological, motion Small Signal delay, dispersion Structural/Functional Alignment within subject Intersubject Alignment Copious amounts of data eg, 20 subjects, 5 runs per
subject, 100 time points per run, 64x64x30 volume = 1.2G data points More spatial voxels than time points (multiple comparisons problem). Model Validation HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Method Correlational synchronized stimulus and acquistion Linear/Gaussian Assumptions GLM General Linear Model MSE Minimum Square Error LMS Least Mean Squares
Massively Univariate HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Hemodynamic Response (BOLD) Time-to-Peak (~6sec) Dispersion TR (~2sec) Undershoot Delay (~1-2sec) HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve
Equilibrium (~16-32sec) fMRI Noise Synthetic data. HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Averaging Synthetic data.
HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Typical Analysis Stream Preprocessing Univariate First-Level GLM Analysis Univariate Higher-Level GLM Analysis Multivariate Analysis Packages: SPM Statistical Parametric Mapping AFNI Analysis of Functional NeuroImages FSL fMRI Software Library FS-FAST FreeSurfer Functional Analysis STream HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve
Preprocessing k-Space reconstruction Slice-Timing Correction (?) Motion Correction Spatial Filtering (Smoothing - FWHM) Intensity Normalization Temporal Filtering (or in analysis) Per-run, within subject HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Univariate First-Level Analysis
Per-voxel, per-subject Postulate model of the observable (ie raw time course) Signal model (eg, hemodynamic response) Noise model (eg, autocorrelation function) Drift (eg, mean offset, linear, quadratic) General Linear Model (GLM) Parameterized Linear (superposition) Least-mean-square estimation of parameters Hypothesis Test = Contrast of Parameters Assemble into a map HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Univariate High-level Analysis
Per-voxel, Cross-subject Requires intersubject registration Dave Kennedy Uses information from First/Lower Levels GLM to describe relationship Random Effects Fixed Effects HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Multivariate Statistics Cross-voxel (within map) Thresholding and multiple comparisons problem Gaussian Random Fields (GRF)
Principal Component Analysis (PCA/SVD) Independent Component Analysis Region-of-Interest HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Hemodynamic Response Model Time-to-Peak (~6sec) Dispersion TR (~2sec)
Undershoot Delay (~1-2sec) HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Equilibrium (~16-32sec) Visual Activation Paradigm Flickering Checkerboard Visual, Auditory, Motor, Tactile, Pain, Perceptual, Recognition, Memory, Emotion, Reward/Punishment,
Olfactory, Taste, Gastral, Gambling, Economic, Acupuncture, Meditation, The Pepsi Challenge, HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Scientific Clinical Pharmaceutical Blood Oxygen Level Dependence (BOLD) Oxygenated Hemoglobin (DiaMagnetic) Neurons
Lungs Oxygen HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve CO2 Deoxygenated Hemoglobin (ParaMagnetic) Functional MRI (fMRI) Stimulus
Localized Neural Firing Localized Increased Blood Flow Localized BOLD Changes Sample BOLD response in 4D
Space (3D) voxels (64x64x35, 3x3x5mm^3) Time (1D) time points (100, 2 sec) Time 1 Time 2 Time 3 HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Analysis Goals Given: raw fMRI time course and
stimulus presentation times Compute: Hemodynamic Response (HRF) Amplitude HRF Confidence Interval Quantify Uncertainty P-Value Noise Amplitude } HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Final Results: Maps Assign values to each voxel
Display as pseudo-color images Threshold? HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve Final Results: Tables List of active regions Cluster TalX Number (mm) TalY (mm) TalZ
(mm) Volume Sig (mm^3) (log10) 1 -30.5 13.2 0.2 125.6
5.7 2 4.5 9.7 -20.2 878.1 4.1
3 2.9 -18.0 17.7 400.3 3.2 HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve
Final Results: Waveforms Average raw data over time and space HST583: Statistical Signal Processing for fMRI -- Douglas N. Greve
