Project Page

Back to search results

ANDI-II-B Taking research tools into clinical Practice: Symmetric model generation

Dr Andrew Janke, Dr Nicolas Cherbuin, Dr Moyra Mortby
Research Centre DCRC Early Diagnosis and Prevention
Partner Institution Australian National University
Project Description

After basic filtering of input data, a model of the entire data set can be generated. The advantage of a nonlinear model over a more traditional linear model is that the resulting model will show structure where there is consistent structure in the population or sub-group. This means that wherever there is structure evident on the resulting model a level of inference can be drawn regarding the shape/size/etc of that structure with respect to other variables that describe the group of subjects that were used to generate the model (e.g.: age, cognitive score).

Work in this area would focus on developing robust modelling techniques that can handle multiple input data types/permutations.

There is also work to be done on methods that can decompose the variance automatically into brain subregions for further analysis. The sub-regions will be automatically defined at each fitting level based upon the local similarity of the structure across the group.

Project Portfolio

Back to search results