Guest lecture 10.3. Jussi Tohka, UEF

10.3. at 9:45-10:30
Guest lecture
Agora C234.1

Professor Jussi Tohka, University of Eastern Finland:
Machine learning for predictive brain imaging

Brain imaging can reveal minute changes in brain structure and function that precede clinical symptoms of brain diseases. As a result identifying biomarkers of brain disorders from neuroimaging data has become a rapidly growing research area at the intersection of biomedical engineering, machine learning, and neuroscience. Specifically, a fruitful research strategy for developing predictive models is to combine earlier brain imaging data with more recent diagnostic information using machine learning.

Tohka's group develops ML methods to overcome methodological challenges in this area, including:

1) combining imaging data across different centers and cohorts,
2) poor interpretability of the derived surrogate markers, and
3) combining data from multiple imaging techniques and behavioral, lifestyle, genetic and demographic information.

Tohka's research program aims to develop an interpretable ML framework capable of integrating multiple types of imaging and other data into prediction models while effectively controlling for multiple nuisance factors.