Data Science for Vision Research - from retinal computations to clinical diagnostics

01/01/2018 to 31/12/2020

New neurotechnologies allow an unprecedented look at the cellular components of the visual system and the sophisticated computations they perform.At the same time, clinical diagnosis and epidemiological studies in ophthalmology are becoming increasingly multimodal, combining traditional methods such as fundus imaging with full-field or 3D ocular coherence tomography and genetic sequencing, allowing to assess the clinical state and time course of patients with eye diseases to be monitored with unprecedented accuracy. Large epidemiological studies have resulted in datasets including tens of thousands of patients. These rich datasets in both basic research of retinal neuroscience as well as clinical research in ophthalmic diseases pose a challenge for data analysis and demand algorithmic solutions for integration and interpretation. My goal is to contribute towards this “data science for vision research” and develop algorithms based on advances in machine learning, computational statistics and data management that allow combining datasets of multiple modalities in a principled way.