You are here : Home > The Institute > News > Linear mixed models for association and prediction from vertex-wise brain MRI measurements

Conference | Brain


NeuroSpin Conferences

Linear mixed models for association and prediction from vertex-wise brain MRI measurements

From 5/31/2021 to 5/31/2021
webinar

Baptiste COUVY-DUCHESNE (University of Queensland, Brisbane, Australia / Brain and Spine Institute, Paris, France,) will give a talk on Zoom on May 31st.

https://prairie-institute.fr/chairs/couvy-duchesne-baptiste/ 


Invited by: 

Short abstract:

The recent availability of large brain MRI samples, calls for new methods to perform association analyses and build predictors from thousands of participants but hundred of thousands of brain measurements. We introduce a set of analyses that fits all brain measurements into random variable(s) in order to overcome the dimensionality problem. We demonstrated how such model allows for association testing, as well as to derive prediction from MRI images. Finally, we also showed how LMM can be used instead of the current state of the art GLMs (generalised linear models) to provide a more parsimonious representation of the localised brain regions associated with the trait/disease.. 

Practical information


replay




Top page