ohbm 2017: practical intensity based meta-analysis
TRANSCRIPT
Practical intensity-based meta-analysis
Camille Maumet OHBM Neuroimaging Meta-Analysis Educational course, 25 June 2017
Coordinate-based meta-analysis Image-based meta-analysis
Coordinate-Based & Image-Based Meta-Analyses
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 3
Neuroimaging meta-analysesAcquisition Analysis
Experiment Raw data Results
Acquisition Analysis
Experiment Raw data Results
…
Publication
Publication
Paper
Paper
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 4
Acquisition Analysis
Experiment Raw data Results
Acquisition Analysis
Experiment Raw data Results
…
Publication
Publication
Paper
Paper
Coordinate-based meta-analysis
Coordinate-based meta-analysis
Neuroimaging meta-analyses
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 5
Image-based meta-analysis
Shared resultsData sharing
Acquisition Analysis
Experiment Raw data Results
Acquisition Analysis
Experiment Raw data Results
…
Publication
Publication
Paper
Paper
Coordinate-based meta-analysis
Coordinate-based meta-analysis Image-based meta-analysis
Neuroimaging meta-analyses
How to perform an image-based meta-analysis?
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 7
InferenceDetections
(subject-level)Pre-processed
dataSub
ject
1
Model fitting and estimation Contrast and
std. err. maps
Image-based meta-analysis
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 8
InferenceDetections
(subject-level)
InferenceDetections
(subject-level)
Image-based meta-analysis
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
…
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 9
InferenceDetections
(study-level)
Image-based meta-analysis
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 10
InferenceDetections
(study-level)
InferenceDetections
(study-level)
Image-based meta-analysis
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 11
Image-based meta-analysis
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimation Contrast and
std. err. maps
Inference
Detections (meta-analysis)
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 12
InferenceDetections
(subject-level)
InferenceDetections
(subject-level)
InferenceDetections
(study-level)
InferenceDetections
(study-level)
Meta-analysis levelStudy levelSubject level
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Pre-processed dataS
ubje
ct 1
Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimationPre-processed
dataSub
ject
n
Contrast and std. err. maps
… Model fitting and estimation Contrast and
std. err. maps
Model fitting and estimation Contrast and
std. err. maps
Inference
Detections (meta-analysis)
Image-based meta-analysis
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 13
• Gold standard: Third-level Mixed-Effects GLM• Requirements
– study-level Contrast estimates and Standard error maps.
– Same units
Contrast and std. err. maps
Image-based meta-analysis
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 14
Units of contrast estimatesPre-processed
data
Model fitting and estimation Contrast and
std. err. maps
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 15
Pre-processed data
Model fitting and estimation Contrast and
std. err. mapsPre-processed
data
Data scalingScaled
pre-proc. data
Model parameter estimation Parameter
estimates
Contrast estimation Contrast and
std. err. maps
Units of contrast estimates
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 16
Units depend on mean estimation and scaling target.
Pre-processed data
Data scalingScaled
pre-proc. data
Model parameter estimation Parameter
estimates
Contrast estimation Contrast and
std. err. maps
Data scaling
Units of contrast estimates
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 17
Y = β +
Units depend on scaling of explanatory variables
Pre-processed data
Data scalingScaled
pre-proc. data
Model parameter estimation Parameter
estimates
Contrast estimation Contrast and
std. err. maps
Model parameter estimation
Units of contrast estimates
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 18
• Contrast Estimation– Linear combination of parameter estimates– Final statistics invariant to scale
• e.g. [ 1 1 1 1 ] gives same T’s & P’s as [ ¼ ¼ ¼ ¼ ]
Units depend on contrast vector– Rule for contrasts to preserve units
• Positive elements sum to 1• Negative elements sum to -1
Pre-processed data
Data scalingScaled
pre-proc. data
Model parameter estimation Parameter
estimates
Contrast estimation Contrast and
std. err. maps
Contrast estimation
Units of contrast estimates
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 19
• Gold standard:
• But…– Units will depend on:
• The scaling of the data (subject-level)• The scaling of the predictor(s) (subject- and study-level)• The scaling of the contrast (subject- and study-level).
– Contrast estimates and standard error maps are rarely shared…
Third-level Mixed-Effects GLM
Units of contrast estimates
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 20
3dMEMA_result+tlrc.BRIK[[0]][from contrast & stat maps]
Which images for IBMA?
Contrast & std. err. maps
Statistic mapE.g. Z-map
Contrast map
SPM FSL AFNI
con_0001.nii[SPM.mat]
cope1.niivarcope1.nii (squared)
3dMEMA_result+tlrc.BRIK[[1]]spmT_0001.nii tstat1.nii.gzzstat1.nii.gz
3dMEMA_result+tlrc.BRIK[[0]]con_0001.nii cope1.nii
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 21
• Fisher's
– Sum of −log P-values (from T/Z’s converted to P’s)
• Stouffer’s
– Average Z, rescaled to N(0,1)
• “Stouffer's Random Effects (RFX)”
– Submit Z’s to one-sample t-test
IBMA on Z maps
(Slide adapted from Thomas Nichols, OHBM 2015)
Statistic mapE.g. Z-map
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 22
• Weighted Stouffer’s
– Z’s from bigger studies get bigger weights
Statistic mapE.g. Z-map
IBMA on Z maps + N + N
(Slide adapted from Thomas Nichols, OHBM 2015)
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 23
• Random Effects (RFX) GLM
– Analyze per-study contrasts as “data”
Contrast + standard error maps• Fixed-Effects (FFX) GLM
– Don’t estimate variance, just take from first level
IBMA on Contrast mapsContrast map
Contrast & std. err. maps
(Slide adapted from Thomas Nichols, OHBM 2015)
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 24
Implementations
• Not all of these options are easily usedMeta-Analysis Method Inputs Neuroimaging
Implementation‘Gold Standard’ MFX Con’s + SE’s FSL’s FEAT
SPM spm_mfxAFNI 3dMEMA
RFX GLMStouffer’s RFX
Con’sZ’s
FSL, SPM, AFNI, etc…
FFX GLMFisher’sStouffer’sStouffer’s Weighted
Con’s +SE’sZ’sZ’sZ’s + N’s
n/a
(Slide from Thomas Nichols, OHBM 2015)
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 25
Self Promotion Alert: IBMA toolbox
• SPM Extension• Still in beta!
– But welcome all feedback
• Available on GitHub https://github.com/NeuroimagingMetaAnalysis/ibma
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 26
Meta-analysis of 21 pain studies
• Results– GLM methods similar– Z-based methods similar
• But FFX Z methods more sensitive (as expected)
RFX
Data: Tracey pain group, FMRIB, Oxford.
How to publish your statistic maps?
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 28
Share your statistic maps
http://neurovault.org
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 29
Share your statistic maps
http://neurovault.org
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 30
From SPM & FSL
NIDM-Results
http://nidm.nidash.org/getting-started/
• When data available, Image-Based preferred to Coordinate-Based meta-analysis
Conclusions
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 31
• When data available, Image-Based preferred to Coordinate-Based meta-analysis
• In practice, it is difficult to use the gold standard Mixed-Effects GLM
Conclusions
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 32
• When data available, Image-Based preferred to Coordinate-Based meta-analysis
• In practice, it is difficult to use the gold standard Mixed-Effects GLM
• When only contrast estimates are available, RFX GLM is a practical & valid approach
Conclusions
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 33
• When data available, Image-Based preferred to Coordinate-Based meta-analysis
• In practice, it is difficult to use the gold standard Mixed-Effects GLM
• When only contrast estimates are available, RFX GLM is a practical & valid approach
• Few tools for Z-based IBMA, but underway…
Conclusions
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 34
• When data available, Image-Based preferred to Coordinate-Based meta-analysis
• In practice, it is difficult to use the gold standard Mixed-Effects GLM
• When only contrast estimates are available, RFX GLM is a practical & valid approach
• Few tools for Z-based IBMA, but underway…
• Data sharing tools: NeuroVault, NIDM-Results
Conclusions
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 35
Camille Maumet - OHBM Neuroimaging Meta-Analysis Educational course 25 June 2017 36
Thank you!
This work is supported by