Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Overview of Neuroimaging Phenotypes
Transcript of Overview of Neuroimaging Phenotypes
Anderson M. Winkler
consists of the observable biochemical, physiological or morphological characteristics of an individual determined by his genotype and environmental influences.
Establishing and using a phenotype
Complete and well planned.
Same results when repeated.
Measurements reflect what they are supposed to measure.
Clinical and research objectives differ
Use of clinical data for research requires careful judgement:
Is the clinical data
(e.g., all relevant MRI sequences collected)?
(last minute sequences or contrasts)?
(radiological diagnoses vs. quantification)?
Quantitative vs. Qualitative
traits are more
and should be preferred, except when less reliable.
Replace a diagnosis for a
tightly associated with the diagnosis whenever possible.
Schulze and McMahon 2004;
Bartlett and Frost, 2008.
multiple disorder subtypes.
Grouping or splitting syndromes:
stratification of phenotype is good if matches genotype.
Variable expressivity, penetrance;
Pleiotropy, locus heterogeneity;
GxG (epistasis) and GxE interactions.
Associated with illness;
Co-segregate with illness.
Gottesman and Gould, 2003.
Patterns of inheritance
single locus (autosomic/sex-linked, dominant/recessive).
multiple loci, any of them causing the observable phenotype.
single locus, multiple observable phenotypes.
multiple loci and environmental factors affecting the phenotype, interactions between genes and phenotypes, often with mixed pleiotropy and genetic heterogeneity.
anticipation, mosaicism, uniparental disomy, genomic imprinting, mitochondrial inheritance.
Levels of expression
Endophenotypes relate to
Bearden et al., in press.
Overview of main imaging phenotypes
thickness, area, volume, complexity, shape, tractography.
FA, MD, AD, RD, myelin content, GM/WM ratio, relaxometry, spectroscopy, susceptibility, magnetization transfer.
task-based, task-free, connectivity (BOLD-fMRI, EEG, MEG, NIRS).
The systematic standardization of measures hypothesized to represent the complete phenotypic space for a given biological system, and their assessment in all members of a study population.
Freimer and Sabatti, 2003.
Congdon et al., 2010.
Winkler et al. 2010.
Distance between pia-mater and the gray-white interface.
Captures the variability of all 6 cortical layers combined and myelination.
Heritable, and controlled regionally by distinct genetic factors.
Rimol et al., 2010.
Area of the gray-matter interface, or of the pia-mater.
Regionalization/compartmentalization of the cortex:
Absolute values (e.g., in mm²)
Relative values (use of a reference brain)
For regions, depends on the definition of the borders
For vertexwise/facewise, depends on the features that drive the registration (morphology, function, myelination).
Like thickness, heritable and regionally influenced by distinct genetic factors.
Winkler et al., 2012;
Chen et al., 2012.
Measurable in voxel-based (e.g. VBM) and surface-based representations of the brain.
Mixture of thickness and area, in variable proportions across the cortical mantle, but reflects mostly surface area.
Ambiguous: susceptible to misregistration, misclassification, folding/gyrification, thickness, amount of GM, deformation fields.
Should be avoided.
Eyler et al., 2012
Winkler et al., 2010.
Chen et al., 2012.
Ratio of the pial surface and the convex hull. Can be computed for each vertex (LGI).
Thompson et al., 1996.
Cortical organization, arealization.
Potential to drive the registration: good for areal analyses, and a completely
new way to do fMRI
Covariate/ancillary information for cortical thickness studies.
1/T1 imaging, T1/T2 ratio.
Ratio of the logs of the resolution scaling and the measurement scaling.
Indicator of a biological process different than cortical thickness.
Decreases with age and certain disorders, varies spatially, heritable.
May confound thickness measurements.
Westlye et al., 2009.
Panizzon et al., 2012.
Can be the total volume or volume of GM or WM.
Use a global measurement of brain volume to account for brain size (brain volume, ICV).
Kremen et al., 2010.
Eyler et al, 2011.
Mesh and/or parameterizable model the structures to analyse.
Multiple possibilities for analysis and interpretations depending on the method and the structure.
Styner et al., 2006.
Patenaude et al., 2011.
Pievani et al., 2011.
Fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), axial diffusivity (AD)
Relate to coherent direction of axons, the integrity of their walls, axonal transport systems, and myelin.
DTI, DKI, DSI, HARDI.
Kochunov et al., 2011.
Tries to reconstruct fiber tracts using diffusion and anisotropy.
Do not reflect true tract locations, but likelihoods for tracts.
Results per subject.
Can be used to analyze termination points, to project anisotropy measurements (FA, MD, RD, AD), to define regions and for shape analyses.
Jbabdi and Johansen-Berg, 2011;
Shi et al., 2013.
Brain structure: Other methods
Measurement of decay. Index of inflamation and perhaps other tissue changes.
Angiography. [Henkelman, 2001].
sensitive to venous blood, hemorrhage and iron storage [Haacke et al., 2009; Mittal et al., 2009].
Metabolites are heritable [Batouli et al, 2012].
Many of these may reflect mostly environmental influences, but potential use in phenomics/imaging genetics needs to be explored.
Brain activity and physiology
Smith et al., 2009.
Wager et al., 2013.
PET, ASL, EEG, MEG, NIRS.
Rubinov and Sporns, 2010
(but see Zalesky et al., 2012).
Trait split into many pieces:
Voxels, vertices, faces:
best localizing power, noisier, multiple testing problem more severe.
powerful if matches size of the effect, less noise, less localizing power.
Summarize data within region using average or first eigenvariate.
Use methods that preserve the variance.
Bilder et al, 2009.
Tom Nichols & J.-B. Poline
Jack Kent Jr.
D. Reese McKay, Emma Sprooten, Emma Knowles
Wojczynski and Tiwari, 2008.
when defining regions. Use data from previous studies, or from an atlas, or from a database, e.g., BrainMap, NeuroSynth, Allen Brain.
Endophenotypic Ranking Value
Glahn et al., 2011.
(see also Toro et al, 2008)