data modeling general linear model & statistical inference
DESCRIPTION
Data Modeling General Linear Model & Statistical Inference. Thomas Nichols, Ph.D. Assistant Professor Department of Biostatistics http://www.sph.umich.edu/~nichols Brain Function and fMRI ISMRM Educational Course July 11, 2002. Motivations. Data Modeling Characterize Signal - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/1.jpg)
1
Data ModelingGeneral Linear Model &
Statistical InferenceThomas Nichols, Ph.D.
Assistant ProfessorDepartment of Biostatistics
http://www.sph.umich.edu/~nichols
Brain Function and fMRIISMRM Educational Course
July 11, 2002
![Page 2: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/2.jpg)
2
Motivations
• Data Modeling– Characterize Signal– Characterize Noise
• Statistical Inference– Detect signal– Localization (Where’s the blob?)
![Page 3: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/3.jpg)
3
Outline
• Data Modeling– General Linear Model – Linear Model Predictors– Temporal Autocorrelation – Random Effects Models
• Statistical Inference– Statistic Images & Hypothesis Testing– Multiple Testing Problem
![Page 4: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/4.jpg)
4
Basic fMRI Example
• Data at one voxel– Rest vs.
passive word listening
• Is there an effect?
![Page 5: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/5.jpg)
5
A Linear Model
IntensityT
ime = 1 2+ + er
ror
x1 x2
• “Linear” in parameters 1 & 2
![Page 6: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/6.jpg)
6
Linear model, in image form…
= + +1 2
Y 11x 22x
![Page 7: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/7.jpg)
7
Linear model, in image form…
= + +1̂ 2̂
Y ̂ 11̂x 22ˆ x
Estimated
![Page 8: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/8.jpg)
8
… in image matrix form…
= +
2
1
ˆ
ˆ
Y ̂ X ̂
![Page 9: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/9.jpg)
9
… in matrix form.
XY
=
+YY X
N
1
N N
1 1p
p
N: Number of scans, p: Number of regressors
![Page 10: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/10.jpg)
10
Linear Model Predictors
• Signal Predictors– Block designs– Event-related responses
• Nuisance Predictors– Drift– Regression parameters
![Page 11: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/11.jpg)
11
Signal Predictors
• Linear Time-Invariant system
• LTI specified solely by– Stimulus function of
experiment
– Hemodynamic ResponseFunction (HRF)
• Response to instantaneousimpulse
Blocks
Events
![Page 12: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/12.jpg)
12
Convolution Examples
Event-Related
Hemodynamic Response Function
Predicted Response
Block Design
Experimental Stimulus Function
![Page 13: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/13.jpg)
13
HRF Models
• Canonical HRF– Most sensitive
if it is correct– If wrong, leads to
bias and/or poor fit• E.g. True response
may be faster/slower
• E.g. True response may have smaller/bigger undershoot
SPM’s HRF
![Page 14: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/14.jpg)
14
HRF Models
• Smooth Basis HRFs– More flexible– Less interpretable
• No one parameter explains the response
– Less sensitive relativeto canonical (only if canonical is correct)
Gamma Basis
Fourier Basis
![Page 15: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/15.jpg)
15
HRF Models
• Deconvolution– Most flexible
• Allows any shape
• Even bizarre, non-sensical ones
– Least sensitive relativeto canonical (again, if canonical is correct) Deconvolution Basis
![Page 16: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/16.jpg)
16
Drift Models
• Drift– Slowly varying– Nuisance variability
• Models– Linear, quadratic– Discrete Cosine Transform
Discrete Cosine Transform Basis
![Page 17: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/17.jpg)
17
General Linear ModelRecap
• Fits data Y as linear combination of predictor columns of X
• Very “General”– Correlation, ANOVA, ANCOVA, …
• Only as good as your X matrix
XY
![Page 18: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/18.jpg)
18
Temporal Autocorrelation
• Standard statistical methods assume independent errors– Error i tells you nothing about j i j
• fMRI errors not independent– Autocorrelation due to– Physiological effects– Scanner instability
![Page 19: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/19.jpg)
19
Temporal AutocorrelationIn Brief
• Independence
• Precoloring
• Prewhitening
![Page 20: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/20.jpg)
20
Autocorrelation: Independence Model
• Ignore autocorrelation
• Leads to – Under-estimation of variance– Over-estimation of significance– Too many false positives
![Page 21: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/21.jpg)
21
Autocorrelation:Precoloring
• Temporally blur, smooth your data– This induces more dependence!– But we exactly know the form of the
dependence induced– Assume that intrinsic autocorrelation is
negligible relative to smoothing
• Then we know autocorrelation exactly• Correct GLM inferences based on “known”
autocorrelation
[Friston, et al., “To smooth or not to smooth…” NI 12:196-208 2000]
![Page 22: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/22.jpg)
22
Autocorrelation:Prewhitening
• Statistically optimal solution
• If know true autocorrelation exactly, canundo the dependence– De-correlate your data, your model– Then proceed as with independent data
• Problem is obtaining accurate estimates of autocorrelation– Some sort of regularization is required
• Spatial smoothing of some sort
![Page 23: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/23.jpg)
23
Autocorrelation Redux
Advantage Disadvantage Software
Indep. Simple Inflated significance
All
Precoloring Avoids autocorr. est.
Statistically inefficient
SPM99
Whitening Statistically optimal
Requires precise autocorr. est.
FSL, SPM2
![Page 24: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/24.jpg)
24
Autocorrelation: Models
• Autoregressive– Error is fraction of previous error plus
“new” error
– AR(1): i = i-1 + I
• Software: fmristat, SPM99
• AR + White Noise or ARMA(1,1)– AR plus an independent WN series
• Software: SPM2
• Arbitrary autocorrelation function k = corr( i, i-k )
• Software: FSL’s FEAT
![Page 25: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/25.jpg)
25
Statistic Images &Hypothesis Testing
• For each voxel– Fit GLM, estimate betas
• Write b for estimate of – But usually not interested in all betas
• Recall is a length-p vector
XY
![Page 26: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/26.jpg)
26
Building Statistic Images
=
+
= +Y X
Predictor of interest
![Page 27: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/27.jpg)
27
Building Statistic Images
• Contrast– A linear combination
of parameters– c’
T =
contrast ofestimated
parameters
varianceestimate
T =
ss22c’(X’X)c’(X’X)++cc
c’bc’b
c’ = 1 0 0 0 0 0 0 0
b b b b b ....
![Page 28: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/28.jpg)
28
Hypothesis Test
• So now have a value T for our statistic
• How big is big– Is T=2 big? T=20?
![Page 29: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/29.jpg)
29
Hypothesis Testing
• Assume Null Hypothesis of no signal
• Given that there is nosignal, how likely is our measured T?
• P-value measures this– Probability of obtaining T
as large or larger
level– Acceptable false positive rate
P-val
T
![Page 30: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/30.jpg)
30
Random Effects Models
• GLM has only one source of randomness
– Residual error
• But people are another source of error– Everyone activates somewhat differently…
XY
![Page 31: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/31.jpg)
31
Subj. 1
Subj. 2
Subj. 3
Subj. 4
Subj. 5
Subj. 6
0
Fixed vs.RandomEffects
• Fixed Effects– Intra-subject
variation suggests all these subjects different from zero
• Random Effects– Intersubject
variation suggests population not very different from zero
Distribution of each subject’s effect
![Page 32: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/32.jpg)
32
Random Effects for fMRI• Summary Statistic Approach
– Easy• Create contrast images for each subject• Analyze contrast images with one-sample t
– Limited• Only allows one scan per subject• Assumes balanced designs and homogeneous meas. error.
• Full Mixed Effects Analysis– Hard
• Requires iterative fitting• REML to estimate inter- and intra subject variance
– SPM2 & FSL implement this, very differently
– Very flexible
![Page 33: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/33.jpg)
33
Random Effects for fMRIRandom vs. Fixed
• Fixed isn’t “wrong”, just usually isn’t of interest• If it is sufficient to say
“I can see this effect in this cohort”then fixed effects are OK
• If need to say“If I were to sample a new cohort from the population I would get the same result”
then random effects are needed
![Page 34: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/34.jpg)
34
Multiple Testing Problem
• Inference on statistic images– Fit GLM at each voxel– Create statistic images of effect
• Which of 100,000 voxels are significant? =0.05 5,000 false positives!
t > 0.5 t > 1.5 t > 2.5 t > 3.5 t > 4.5 t > 5.5 t > 6.5
![Page 35: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/35.jpg)
35
MCP Solutions:Measuring False Positives
• Familywise Error Rate (FWER)– Familywise Error
• Existence of one or more false positives
– FWER is probability of familywise error
• False Discovery Rate (FDR)– R voxels declared active, V falsely so
• Observed false discovery rate: V/R
– FDR = E(V/R)
![Page 36: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/36.jpg)
36
FWER MCP Solutions
• Bonferroni
• Maximum Distribution Methods– Random Field Theory– Permutation
![Page 37: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/37.jpg)
37
FWER MCP Solutions
• Bonferroni
• Maximum Distribution Methods– Random Field Theory– Permutation
![Page 38: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/38.jpg)
38
FWER MCP Solutions: Controlling FWER w/ Max
• FWER & distribution of maximum
FWER= P(FWE)= P(One or more voxels u |
Ho)= P(Max voxel u | Ho)
• 100(1-)%ile of max distn controls FWERFWER = P(Max voxel u | Ho)
u
![Page 39: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/39.jpg)
39
FWER MCP Solutions:Random Field Theory
• Euler Characteristic u
– Topological Measure• #blobs - #holes
– At high thresholds,just counts blobs
– FWER = P(Max voxel u | Ho)= P(One or more blobs | Ho) P(u 1 | Ho) E(u | Ho)
Random Field
Suprathreshold Sets
Threshold
![Page 40: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/40.jpg)
40
Controlling FWER: Permutation Test
• Parametric methods– Assume distribution of
max statistic under nullhypothesis
• Nonparametric methods– Use data to find
distribution of max statisticunder null hypothesis
– Any max statistic!
5%
Parametric Null Max Distribution
5%
Nonparametric Null Max Distribution
![Page 41: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/41.jpg)
41
Measuring False Positives
• Familywise Error Rate (FWER)– Familywise Error
• Existence of one or more false positives
– FWER is probability of familywise error
• False Discovery Rate (FDR)– R voxels declared active, V falsely so
• Observed false discovery rate: V/R
– FDR = E(V/R)
![Page 42: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/42.jpg)
42
Measuring False PositivesFWER vs FDR
Signal
Signal+Noise
Noise
![Page 43: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/43.jpg)
43
FWE
6.7% 10.4% 14.9% 9.3% 16.2% 13.8% 14.0% 10.5% 12.2% 8.7%
Control of Familywise Error Rate at 10%
11.3% 11.3% 12.5% 10.8% 11.5% 10.0% 10.7% 11.2% 10.2% 9.5%
Control of Per Comparison Rate at 10%
Percentage of Null Pixels that are False Positives
Control of False Discovery Rate at 10%
Occurrence of Familywise Error
Percentage of Activated Pixels that are False Positives
![Page 44: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/44.jpg)
44
Controlling FDR:Benjamini & Hochberg
• Select desired limit q on E(FDR)• Order p-values, p(1) p(2) ... p(V)
• Let r be largest i such that
• Reject all hypotheses corresponding to p(1), ... , p(r).p(i) i/V q p(i)
i/V
i/V qp-
valu
e
0 1
01
![Page 45: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/45.jpg)
45
Conclusions
• Analyzing fMRI Data– Need linear regression basics– Lots of disk space, and time– Watch for MTP (no fishing!)
![Page 46: Data Modeling General Linear Model & Statistical Inference](https://reader035.vdocuments.net/reader035/viewer/2022062314/56812eaa550346895d944a10/html5/thumbnails/46.jpg)
46
Thanks
• Slide help– Stefan Keibel, Rik Henson, JB Poline, Andrew
Holmes