[CaBSSem] Lectures on neuroimaging data analysis by Timothy Johnson of U Michigan

Stephen Lindsay slindsay at uvic.ca
Tue Jun 3 10:31:19 PDT 2014

Two lectures by Timothy Johnson (Department of Biostatistics, University of Michigan) will be presented on June 19 and 20. These lectures are co-sponsored by the Department of Mathematics and Statistics and by the Cognition and Brain Sciences program in Psychology.  The titles and abstracts for the lectures are given below.  Anyone interested in these topics is welcome to attend.

TITLE : A Spatial Generalized Linear Mixed Model and the Estimation of Spatially-Varying Coefficients with Application to Multiple Sclerosis MRI Data
DATE/TIME: Thursday June 19th at 3PM
LOCATION: Hickman Building (HHB) 120


Multiple Sclerosis (MS) is an autoimmune disease affecting the central nervous system (CNS) by disrupting nerve transmission. This disruption is caused by damage to the myelin sheath surrounding nerves that acts as an insulator. Patients with MS have a multitude of symptoms that depend on where lesions occur in the brain and/or spinal cord. MS has no cure and in order to help manage the disease, physicians subtype MS patients into 5 categories that depend on the pattern of MS episodes. Patient symptoms are rated by the Kurtzke Functional Systems (FS) scores and the paced auditory serial addition test (PASAT) score. The eight functional systems (CNS areas or circuits that regulate body functions) are: 1) pyramidal; 2) cerebellar; 3) brainstem; 4) sensory; 5) bowel and bladder; 6) visual; 7) cerebral; and 8) other. Of interest to Neurologists is whether lesion locations can be predicted using these FS and PASAT scores and whether the data can help predict MS subtype in a newly diagnosed patient.
To help answer the above questions, we propose an autoprobit regression model with both spatially varying random coefficients. The data of interest are digitized binary images of the brain derived from high-resolution T2-weight MRI images en- coded such that, at each voxel, 1 indicates the presence of a lesion and 0 denotes absence of a lesion. These binary lesion maps are taken as the dependent variables. The model incorporates both spatially varying covariates as well as patient specific, non-spatially varying covariates. In contrast to most spatial applications, in which only one realization of a process is observed, we have multiple, independent realiza- tions, one from each patient. This allows the modeling and estimation of spatially varying parameters of patient level covariates such as age, gender, FS and PASAT scores. Maps of these spatially varying parameters over the brain allow us to spa- tially predict lesion probabilities over the brain given covariates. We show that via Bayes Theorem, the model can accurately predict the MS subtype of a new subject.

TITLE : Bringing New Techniques from Statistics into Neuroimaging Meta-Analysis
DATE/TIME: Friday June 20th at 3PM
LOCATION: Hickman Building (HHB) 120

Functional neuroimaging has provided a wealth of information on the cerebral localization of mental functions. In spite of its success, several limitations restrict the amount of knowledge that may be obtained from each individual experiment. These include a small sample size, limited reliability of an indirect signal, and the need to infer knowledge from specific contrasts. Such limitations have raised some concerns whether neuroimaging can provide fundamental insights into problems from cognitive psychology or clinical neurosciences. In turn, however, these limitations have also encouraged the development of quantitative meta-analysis approaches that allow summaries of a vast amount of neuroimaging findings across a large number of participants and diverse experimental settings. Such integration enables statistically defensible generalizations on the neural basis of psychological processes in health and disease. Quantitative meta-analysis therefore represents a powerful tool to gain a synoptic view of distributed neuroimaging findings in an objective and impartial fashion and addresses the above concerns.
In this talk I will briefly summarize the historical developments of neuroimag- ing meta-analyses. I will then present an overview of the currently most prevalent approach of coordinate-based meta-analyses, explaining concepts and current implementations. Lastly, I will provide an overview on innovative new approaches that are currently being developed and future directions.


D. Stephen Lindsay, Ph.D.


Department of Psychology

University of Victoria

P.O. Box 1700 STN CSC

Victoria, B.C. V8W 2Y2

v: (250) 721-8593

f: (250) 721-8929

w:  http://web.uvic.ca/~dslind/
Check out the newly published Sage Handbook of Applied Memory<http://www.uk.sagepub.com/books/Book237290?subject=K00&sortBy=defaultPubDate%20desc&fs=1#tabview=title> co-edited with Tim Perfect.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.uvic.ca/pipermail/cabssem/attachments/20140603/394485ce/attachment.html>

More information about the CaBSSem mailing list