Professional Documents
Culture Documents
J. Daunizeau
• normalization:
a=2
• marginalization:
• conditioning :
b=5 (Bayes rule)
a=2
Bayesian paradigm
deriving the likelihood function
y f e.g., GLM: f X
P 4 0.05
f
1
p exp 2 2
2
→ Distribution of data, given fixed parameters:
1 2
p y exp 2 y f
2
Bayesian paradigm
likelihood, priors and the model evidence
Likelihood:
Prior:
Bayes rule:
generative model m
Bayesian paradigm
forward and inverse problems
forward problem
p y , m
likelihood
posterior distribution
p y , m
inverse problem
Bayesian paradigm
model comparison
Principle of parsimony :
« plurality should not be assumed without necessity »
Model evidence:
y = f(x)
“Occam’s razor” :
•••
hierarchy
causality
Hierarchical models
directed acyclic graphs (DAGs)
Hierarchical models
univariate linear hierarchical model
p t H 0 p y
P H0 y
P t t * H0
t t Y
t*
• estimate parameters (obtain test stat.) • define the null, e.g.: H 0 : 0
• apply decision rule, i.e.: • apply decision rule, i.e.:
if P t t * H 0 then reject H0
if P H 0 y then accept H0
• define the null and the alternative hypothesis in terms of priors, e.g.:
p Y H 0
1 if 0
H 0 : p H 0 p Y H1
0 otherwise
H1 : p H1 N 0,
Y
y space of all datasets
P H0 y
• apply decision rule, i.e.: if 1 then reject H0
P H1 y
p 0 y, H1
p y H 0 p y H1
p 0 H1
Overview of the talk
1 Probabilistic modelling and representation of uncertainty
1.1 Bayesian paradigm
1.2 Hierarchical models
1.3 Frequentist versus Bayesian inference
2 Numerical Bayesian inference methods
2.1 Sampling methods
2.2 Variational methods (ReML, EM, VB)
3 SPM applications
3.1 aMRI segmentation
3.2 Decoding of brain images
3.3 Model-based fMRI analysis (with spatial priors)
3.4 Dynamic causal modelling
Sampling methods
MCMC example: Gibbs sampling
Variational methods
VB / EM / ReML
→ VB : maximize the free energy F(q) w.r.t. the “variational” posterior q(θ)
under some (e.g., mean field, Laplace) approximation
p 1 , 2 y, m
p 1 or 2 y, m
q 1 or 2
2
1
Overview of the talk
1 Probabilistic modelling and representation of uncertainty
1.1 Bayesian paradigm
1.2 Hierarchical models
1.3 Frequentist versus Bayesian inference
2 Numerical Bayesian inference methods
2.1 Sampling methods
2.2 Variational methods (ReML, EM, VB)
3 SPM applications
3.1 aMRI segmentation
3.2 Decoding of brain images
3.3 Model-based fMRI analysis (with spatial priors)
3.4 Dynamic causal modelling
segmentation posterior probability dynamic causal multivariate
and normalisation maps (PPMs) modelling decoding
statistical Gaussian
inference field theory
normalisation
p <0.05
template
aMRI segmentation
mixture of Gaussians (MoG) model
class variances
1 2 … k
1 ith voxel
label
2 yi ci
ith voxel class
…
value frequencies
k
class
means
+ >>
prior variance
of GLM coeff
PPM: regions best explained
prior variance AR coeff by long-term memory model
of data noise (correlated noise) long-term memory
design matrix (X)
GLM coeff
m1 m2 m3 m4
attention attention attention attention
attention
models marginal likelihood estimated
ln p y m
0.10 effective synaptic strengths
15
for best model (m4)
PPC 0.39
0.26
10 1.25
0.26
stim V1 0.13 V5
5
0.46
0
m1 m2 m 3 m4
DCMs and DAGs
a note on causality
1 2
21
1 2 3
1 2
32
1 2 3 time 13
t t 3
3 t 0 13u 3u
t
x f ( x, u , ) u
t t
Dynamic Causal Modelling
model comparison for group studies
ln p y m1 ln p y m2
differences in log- model evidences
m1
m2
subjects
p y H1
u
p y H0
is the most powerful test of size p u H 0 to test the null.
• what is the threshold u, above which the Bayes factor test yields a error I rate of 5%?
ROC analysis
1 - error II rate
CCA (F-statistics)
F=2.20, power=20%
error I rate