Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT...
Transcript of Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT...
![Page 1: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/1.jpg)
Statistical Inference I GLM, Contrasts & RFT
Guillaume Flandin
Wellcome Trust Centre for Neuroimaging
University College London
SPM Course
Lausanne, November 2013
![Page 2: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/2.jpg)
Normalisation
Statistical Parametric Map
Image time-series
Parameter estimates
General Linear Model Realignment Smoothing
Design matrix
Anatomical
reference
Spatial filter
Statistical
Inference RFT
p <0.05
![Page 3: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/3.jpg)
Normalisation
Statistical Parametric Map
Image time-series
Parameter estimates
General Linear Model Realignment Smoothing
Design matrix
Anatomical
reference
Spatial filter
Statistical
Inference RFT
p <0.05
![Page 4: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/4.jpg)
Passive word
listening
versus rest
7 cycles of
rest and listening
Blocks of 6 scans
with 7 sec TR
Question: Is there a change in the BOLD
response between listening and rest?
Stimulus function
One session
fMRI experiment example
![Page 5: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/5.jpg)
BOLD signal
Tim
e
single voxel
time series
Voxel-wise time series analysis
Model
specification
Parameter
estimation
Hypothesis
Statistic
SPM
![Page 6: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/6.jpg)
BOLD signal
Tim
e
= 1 2 + + err
or
x1 x2 e
Single voxel regression model
exxy 2211
![Page 7: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/7.jpg)
Mass-univariate analysis: voxel-wise GLM
=
e+ y X
N
1
N N
1 1p
p
Model is specified by
1. Design matrix X
2. Assumptions about e
N: number of scans
p: number of
regressors
eXy
The design matrix embodies all available knowledge about
experimentally controlled factors and potential confounds.
),0(~ 2INe
![Page 8: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/8.jpg)
• one sample t-test
• two sample t-test
• paired t-test
• Analysis of Variance
(ANOVA)
• Analysis of Covariance
(ANCoVA)
• correlation
• linear regression
• multiple regression
GLM: a flexible framework for parametric analyses
![Page 9: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/9.jpg)
Parameter estimation
eXy
= +
e
2
1
Ordinary least
squares estimation
(OLS) (assuming i.i.d.
error):
yXXX TT 1)(ˆ
Objective:
estimate
parameters to
minimize
N
t
te1
2
y X
![Page 10: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/10.jpg)
Problems of this model with fMRI time series
1. The BOLD response has a delayed and dispersed shape.
2. The BOLD signal includes substantial amounts of low-frequency noise (eg due to scanner drift).
3. Due to breathing, heartbeat & unmodeled neuronal activity, the errors are serially correlated. This violates the assumptions of the noise model in the GLM.
![Page 11: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/11.jpg)
Boynton et al, NeuroImage, 2012.
Scaling
Additivity
Shift
invariance
Problem 1: BOLD response
Hemodynamic response function (HRF):
Linear time-invariant (LTI) system:
u(t) x(t) hrf(t)
Convolution operator:
![Page 12: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/12.jpg)
Problem 1: BOLD response Solution: Convolution model
![Page 13: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/13.jpg)
Convolution model of the BOLD response
Convolve stimulus function
with a canonical
hemodynamic response
function (HRF):
HRF
t
dtgftgf0
)()()(
![Page 14: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/14.jpg)
Brief
Stimulus
Undershoot
Peak
Informed Basis Set Canonical HRF
Hemodynamic Response Temporal Basis Set
Canonical
Temporal
Dispersion
![Page 15: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/15.jpg)
blue = data
black = mean + low-frequency drift
green = predicted response, taking into
account low-frequency drift
red = predicted response, NOT taking
into account low-frequency drift
Problem 2: Low-frequency noise Solution: High pass filtering
discrete cosine
transform (DCT)
set
![Page 16: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/16.jpg)
)(eCov
autocovariance
function
N
N
Problem 3: Serial correlations
i.i.d:
![Page 17: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/17.jpg)
Multiple covariance components
= 1 + 2
Q1 Q2
Estimation of hyperparameters with ReML (Restricted Maximum Likelihood).
V
enhanced noise model at voxel i
error covariance components Q
and hyperparameters
jj
ii
QV
VC
2
),0(~ ii CNe
![Page 18: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/18.jpg)
Summary: a mass-univariate approach
![Page 19: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/19.jpg)
Summary: Estimation of the parameters
i.i.d. assumptions:
OLS estimates:
![Page 20: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/20.jpg)
Normalisation
Statistical Parametric Map
Image time-series
Parameter estimates
General Linear Model Realignment Smoothing
Design matrix
Anatomical
reference
Spatial filter
Statistical
Inference RFT
p <0.05
![Page 21: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/21.jpg)
Contrasts
[1 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 1 -1 0 0 0 0 0 0 0 0 0 0 0]
![Page 22: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/22.jpg)
Hypothesis Testing
Null Hypothesis H0
Typically what we want to disprove (no effect).
The Alternative Hypothesis HA expresses outcome of interest.
To test an hypothesis, we construct “test statistics”.
Test Statistic T
The test statistic summarises evidence
about H0.
Typically, test statistic is small in
magnitude when the hypothesis H0 is true
and large when false.
We need to know the distribution of T
under the null hypothesis. Null Distribution of T
![Page 23: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/23.jpg)
cT = 1 0 0 0 0 0 0 0
T =
contrast of
estimated
parameters
variance
estimate
box-car amplitude > 0 ?
=
1 = cT> 0 ?
1 2 3 4 5 ...
T-test - one dimensional contrasts – SPM{t}
Question:
Null hypothesis: H0: cT=0
Test statistic:
pN
TT
T
T
T
t
cXXc
c
c
cT
~
ˆ
ˆ
)ˆvar(
ˆ
12
![Page 24: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/24.jpg)
T-test: one dimensional contrasts – SPM{t}
T-test is a signal-to-noise measure (ratio of estimate to
standard deviation of estimate).
T-contrasts are simple combinations of the betas; the T-statistic does not depend on the scaling of the regressors or the scaling of the contrast.
H0: 0Tc vs HA: 0Tc
Alternative hypothesis:
![Page 25: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/25.jpg)
F-test - the extra-sum-of-squares principle
Model comparison:
Null Hypothesis H0: True model is X0 (reduced model)
Full model ?
X1 X0
or Reduced model?
X0 Test statistic: ratio of
explained variability and
unexplained variability (error)
1 = rank(X) – rank(X0)
2 = N – rank(X)
RSS
2ˆfull
RSS0
2ˆreduced
![Page 26: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/26.jpg)
F-test - multidimensional contrasts – SPM{F}
Tests multiple linear hypotheses:
0 0 0 1 0 0 0 0 0
0 0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 1
cT =
H0: 4 = 5 = ... = 9 = 0
X1 (4-9) X0
Full model? Reduced model?
H0: True model is X0
X0
test H0 : cT = 0 ?
SPM{F6,322}
![Page 27: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/27.jpg)
F-test example: movement related effects
Design matrix
2 4 6 8
10
20
30
40
50
60
70
80
contrast(s)
Design matrix 2 4 6 8
10
20
30
40
50
60
70
80
contrast(s)
![Page 28: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/28.jpg)
F-test: summary
F-tests can be viewed as testing for the additional variance explained by a larger model wrt a simpler (nested) model model comparison.
0000
0100
0010
0001
In testing uni-dimensional contrast with an F-test, for example 1 – 2, the result will be the same as testing 2 – 1. It will be exactly the square of the t-test, testing for both positive and negative effects.
F tests a weighted sum of squares of one or several combinations of the regression coefficients .
Hypotheses:
0 : Hypothesis Null 3210 H
0 oneleast at : Hypothesis eAlternativ kAH
![Page 29: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/29.jpg)
Orthogonal regressors
Variability in Y
![Page 30: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/30.jpg)
Correlated regressors
Shared variance
Variability in Y
![Page 31: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/31.jpg)
Correlated regressors
Variability in Y
![Page 32: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/32.jpg)
Correlated regressors
Variability in Y
![Page 33: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/33.jpg)
Correlated regressors
Variability in Y
![Page 34: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/34.jpg)
Correlated regressors
Variability in Y
![Page 35: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/35.jpg)
Correlated regressors
Variability in Y
![Page 36: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/36.jpg)
Correlated regressors
Variability in Y
![Page 37: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/37.jpg)
Normalisation
Statistical Parametric Map
Image time-series
Parameter estimates
General Linear Model Realignment Smoothing
Design matrix
Anatomical
reference
Spatial filter
Statistical
Inference RFT
p <0.05
![Page 38: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/38.jpg)
Inference at a single voxel
Null distribution of test statistic T
u
Decision rule (threshold) u:
determines false positive
rate α
Null Hypothesis H0:
zero activation
Choose u to give acceptable
α under H0
![Page 39: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/39.jpg)
Multiple tests
t
u
t
u
t
u
t
u
t
u
t
u
Signal
If we have 100,000 voxels,
α=0.05 5,000 false positive voxels.
This is clearly undesirable; to correct
for this we can define a null hypothesis
for a collection of tests.
Noise
![Page 40: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/40.jpg)
Multiple tests
t
u
t
u
t
u
t
u
t
u
t
u
11.3% 11.3% 12.5% 10.8% 11.5% 10.0% 10.7% 11.2% 10.2% 9.5%
Use of ‘uncorrected’ p-value, α =0.1
Percentage of Null Pixels that are False Positives
If we have 100,000 voxels,
α=0.05 5,000 false positive voxels.
This is clearly undesirable; to correct
for this we can define a null hypothesis
for a collection of tests.
![Page 41: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/41.jpg)
Family-Wise Null Hypothesis
FWE
Use of ‘corrected’ p-value, α =0.1
Use of ‘uncorrected’ p-value, α =0.1
Family-Wise Null Hypothesis:
Activation is zero everywhere
If we reject a voxel null hypothesis at any voxel,
we reject the family-wise Null hypothesis
A FP anywhere in the image gives a Family Wise Error (FWE)
Family-Wise Error rate (FWER) = ‘corrected’ p-value
![Page 42: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/42.jpg)
Bonferroni correction
The Family-Wise Error rate (FWER), αFWE, for a family of N
tests follows the inequality:
where α is the test-wise error rate.
Therefore, to ensure a particular FWER choose:
This correction does not require the tests to be independent but
becomes very stringent if dependence.
![Page 43: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/43.jpg)
Spatial correlations
100 x 100 independent tests Spatially correlated tests (FWHM=10)
Bonferroni is too conservative for spatial correlated data.
Discrete data Spatially extended data
![Page 44: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/44.jpg)
Topological inference
Topological feature: Peak height
space
Peak level inference
![Page 45: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/45.jpg)
Topological inference
Topological feature: Cluster extent
space
uclus
uclus : cluster-forming threshold
Cluster level inference
![Page 46: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/46.jpg)
Topological inference
Topological feature: Number of clusters
space
uclus
uclus : cluster-forming threshold
c
Set level inference
![Page 47: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/47.jpg)
RFT and Euler Characteristic
Search volume Roughness
(1/smoothness) Threshold
![Page 48: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/48.jpg)
Expected Euler Characteristic
![Page 49: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/49.jpg)
Random Field Theory
The statistic image is assumed to be a good lattice
representation of an underlying continuous stationary
random field.
Typically, FWHM > 3 voxels
(combination of intrinsic and extrinsic smoothing)
A priori hypothesis about where an activation should be,
reduce search volume Small Volume Correction:
• mask defined by (probabilistic) anatomical atlases
• mask defined by separate "functional localisers"
• mask defined by orthogonal contrasts
• (spherical) search volume around previously reported coordinates
![Page 50: Statistical Inference I - UNIL · 2017-08-31 · Statistical Inference I GLM, Contrasts & RFT Guillaume Flandin Wellcome Trust Centre for Neuroimaging University College London SPM](https://reader035.fdocuments.us/reader035/viewer/2022070711/5ec98813cd21621731505a27/html5/thumbnails/50.jpg)
Pre- processings
General Linear Model
Statistical Inference
Contrast c
Random
Field Theory