Statistical methods for neuroimaging data
Presenter: Russell Shinohara
When: Monday, July 11, 2016 Time: 11:00 AM - 12:30 PM
Room: Oak Bay 1-2 (Level 1)
Session Synopsis:
Statistical methods for neuroimaging analyses of multi-center studies
Magnetic resonance imaging (MRI) intensities are acquired in arbitrary units, making scans non-comparable across sites and between subjects and sites. Intensity normalization is a first step for the improvement of comparability of the images across subjects. However, we show that unwanted inter-scan variability associated with imaging site, scanner effect and other technical artifacts are still present after standard intensity normalization in large multi-site neuroimaging studies. We propose RAVEL (Removal of Artificial Voxel Effect by Linear regression), a tool to remove residual technical variability after intensity normalization. We compare RAVEL to intensity-normalization-only methods including histogram matching, and White Stripe. We show that RAVEL performs best at improving the replicability of the brain regions that are empirically found to be most associated with Alzheimer's disease (AD), and that these regions are significantly more present in structures impacted by AD (hippocampus, amygdala, parahippocampal gyrus, enthorinal area and fornix stria terminals). In addition, we show that the RAVEL-corrected intensities have the best performance in distinguishing between MCI subjects and healthy subjects by using the mean hippocampal intensity, a marked improvement compared to results from intensity normalization alone. RAVEL is generalizable to many imaging modalities, and shows promise for longitudinal studies. Additionally, because the choice of the control region is left to the user, RAVEL can be applied in studies of many brain disorders.
Statistical methods for neuroimaging data
Presenter: Todd Ogden
When: Monday, July 11, 2016 Time: 11:00 AM - 12:30 PM
Room: Oak Bay 1-2 (Level 1)
Session Synopsis:
Functional data modeling of dynamic PET data
One major goal of dynamic positron emission tomography (PET) imaging, with particular relevance to the study of mental and neurological disorders, is the estimation of the special density of specific proteins throughout the brain. This estimation has been done almost exlusively using parametric models that require fairly strong assumptions and scalar-valued summaries. We will describe extensions of the analysis in two different directions: a nonparametric approach to the modeling of the observed PET data, and a functional data analytic (FDA) approach to modeling the results across subjects. We demonstrate the application of this approach and compare the results with those derived from standard parametric approaches.
Statistical methods for neuroimaging data
Presenter: Ana-Maria Staicu
When: Monday, July 11, 2016 Time: 11:00 AM - 12:30 PM
Room: Oak Bay 1-2 (Level 1)
Session Synopsis:
Scalar-on-Image Regression via Soft-Thresholded Gaussian Processes
We propose a novel Bayesian framework to do scalar-on-image regression by modeling the regression coefficient through a thresholded latent Gaussian processes. The thresholding of the process will ensure that the result is sparse while its smoothness property will guarantee the smoothness of the effect. Our methods are computationally tractable, accommodate additional covariates, and can be extended to generalized responses. We illustrate the approach through simulations and apply it to the data from an electroencephalography (EEG) study of alcoholism, where we study the relation between the alcoholism status and the electrical brain activity over time.
Statistical methods for neuroimaging data
Presenter: Timothy Johnson
When: Monday, July 11, 2016 Time: 11:00 AM - 12:30 PM
Room: Oak Bay 1-2 (Level 1)
Session Synopsis:
Modeling fMRI data for pre-surgical planning via a conditionally-weighted adaptive-smoothing model
Spatial smoothing is an essential step in the analysis of functional magnetic resonance imaging (fMRI) data. The standard method is to convolve the image data (at each point in the time-series) with a three-dimensional Gaussian kernel that applies a fixed amount of smoothing to the entire image. In pre-surgical planning, using fMRI to de- termine functionally eloquent, peritumoral regions of the brain image, spatial accuracy is paramount. Thus methods that rely on global smoothing may not be reasonable as global smoothing can blur the boundaries between activated and non-activated regions of the brain. Moreover, in a standard fMRI analysis strict false positive control is desired. For pre-surgical planning false negatives may be of greater concern. Our model is a spatially adaptive, conditionally autoregressive model. This model reduces smoothing at boundaries between regions of no activation and activation in the Z-statistic image. After we fit the model to the data, we take a Bayesian decision theoretic approach that allows false negatives and false positives to be penalized dierently.
