|
Activity Number:
|
628
- Complex Data Analysis with Mental Health Applications
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Mental Health Statistics Section
|
|
Abstract #328934
|
Presentation
|
|
Title:
|
Repeated Within-Subject Distributions with Covariates and Censoring: a Neuroscience Application
|
|
Author(s):
|
Ryan Kelly* and Allan Sampson and Rob Sweet and Ken Fish and David Lewis
|
|
Companies:
|
University of Pittsburgh and University of Pittsburgh and University of Pittsburgh and University of Pittsburgh and University of Pittsburgh
|
|
Keywords:
|
Repeated Measures;
Within-subject Distribution;
Censoring/Truncation;
Neuroscience Application
|
|
Abstract:
|
Neuroscientists use post-mortem brain tissue to study schizophrenia's relationship to dendritic spine development in several brain regions. Focus is on the possible factors that influence the distribution of dendritic spine sizes across subjects with schizophrenia as compared to control subjects. One analytic approach to this problem is to discretize each spine size into size "bins" and analyze numbers of spines within each of these bins. This approach not only has multiple testing concerns, but also interpretability issues. To circumvent these issues, we develop a repeated measures model for the distribution of spine sizes within a subject that takes into account subject level covariates that arise in post-mortem tissue studies. We take care to account for truncation and censoring in the data which arises from the chosen method of imaging. Our models are shown to fit the observed data and alleviate many problems from the discretized analysis. However, the methodology introduces analytic difficulties and interpretation concerns of its own.
|
Authors who are presenting talks have a * after their name.
