Visual Analytics of Cohort Study Data
Speaker: Pd Dr.-Ing.habil. Steffen Oetze Jafra (né Oetze) leads the research group Medicine and Digitalization (MedDigt) at the University Clinic of Neurology, Otto von Guericke University, Magdeburg, Germany.
Traditionally, a lot of medical visualization research has been focused on the visualization of data obtained from a single individual, i.e. a single, uni-modal patient dataset, being usually defined on a regular drid in 3D and capturing a selected part of the human anatomy.
In recent years, however, the most pressing challenges in medical visualization have broadened including the investigation of data obtained from populations.
Large pools of image and non-image data are acquired for hundreds to thousands of individuals and their analysis poses tremendous new challenges.
Data Science methods show a high potential in tackling these challenges.
They address the blending of analysis and visualization techniques to make sense out of the big data, the integrated analysis of image and non-image data, the integrated visualization of very heterogeneous data and the effective and efficient interactive exploration of the data.
Exemplary fields that acquire such data pools are epidemiology, where longitudinal studies of defined populations (cohorts) are conducted, and healthcare, where large patient databases are compiled from Electronic Healt Records (EHR) of hospitals.
In this talk, I will present results of collaborations with (1) epidemiologists from the Uniersity of Greifswald, Germany running the populations Study of Health in Pomerania (SHiP) and 2) neuroscientists as well as computer scientists from the University of Bergen, jointly analysing data from a cognitive aging study. Moreover, I will outline (3) and ongoing effort of the German federal state os Saxony-Anhalt aming at a continious registration and quantitative analysis of brain structure and function of all patients with neurological and neuropsychiatrict disorders.