Ruben P. Dörfel
PhD Student at the Centre for Psychiatry Research, Karolinska Institutet and Neurobiology Research Unit, Rigshospitalet.
Welcome to my personal website! I am a data scientist passionate about developing robust and explainable models for safe and fair applications in medicine and neuroscience. Generally, I’m interested in applied problems in the combined fields of computer science, mathematics and neuroscience. My expertise lies in machine learning, neuro-image analysis (M/EEG, MRI, PET), software development, and applied statistics. My current research focuses on developing and validating biomarkers for biological brain aging.
Short Bio
I have a Bachelor of Science in Biomedical Engineering from the Technical University of Ilmenau, Germany, and a Master of Science from the Technical University Denmark and Copenhagen University. During my Bachelor’s I did internships at the MEG Lab at the MGH/HST Athinoula A. Martinos Center for Biomedical Imaging where I worked on real-time movement estimation and correction for MEG measurements. A significant part of this work has been contributed to the open-source software MNE-CPP. Additionally, I’ve been granted the BND Summer of Code scholarship to contribute to this project.
During my Master’s, I worked as a student software engineer at Demant A/S, a globally leading hearing aid provider. Additionally, I did research projects at the Neurobiology Research Unit, Rigshospitalet with a research stay at the Centre for Psychiatry Research, Karolinska Institutet. I continued this appointment as a research assistant after graduating and am now enrolled as a PhD student at those institutions. My primary focus is developing and evaluating new and existing models for brain-age prediction. Brain age is a summary measure that is thought to reflect aging-related biology in the brain. Specifically, I’m assessing the accuracy, reliability, and validity of pre-trained publicly available methods for structural MRI-based brain-age estimation. Such evaluation is crucial to establish brain age as a clinically useful biomarker. Furthermore, I am using machine learning to model age-related changes in serotonin 2A receptor binding to improve existing methods for MRI-based brain-age estimation.