Abstract:

   Traditional statistical analysis of bioimaging data, e.g. based on GLM and t-tests, has provided several insights into working of the human brain. However, it cannot provide generalizable results, i.e. findings that apply to unseen parts of the population. Machine learning methods provide a way towards discovering such generalizable insights. In this talk, he will discuss the basics of machine learning methods and then showcase recent findings on personality and working memory prediction based on meta-analytically defined functional brain networks.

Bio sketch:
     He is mainly interested in understanding biological systems in an algorithmic fashion and develop methods for this purpose. In his current work at the Research Center Juelich, he is focusing on predictive modeling of neurobiological diseases. Especially, the Parkinson’s disease and the Alzheimer’s disease. The aim is to develop machine learning based methods for early diagnosis and disease progression prediction. For this purpose, He rely on functional connectivity and structural brain imaging data and a plethora of machine learning methods, especially classification and clustering. In addition to the research work, He have been serving as referee for several reputed journals and conferences as well as mentoring younger researchers. For example, He have been serving as a reviewer for several scientific journals, including Nat. Methods, Nat. Communications, PloS Comp Biol., PLoS One and IEEE TCBB.  He have been a member of the scientific program committee for several conferences, including the International Conference on Practical Applications of Computational Biology & Bioinformatics and the Intelligent Data Analysis conference.