Yanniklas Kravutske

Yanniklas Kravutske

Home: Medizinische Universität Wien
Host: Stanford University School of Medicine
Topic: Enhanced Detection of Pediatric Epileptogenic Cortical Malformations Using MindGlide: A Deep Learning Approach

Professionally, the integration has been smooth. The workplace is exceptionally well structured, and he receives support whenever needed. He was introduced to the radiology department early on, which helped him feel well integrated. Much of his work is independent, but colleagues and staff are always available for guidance, creating a positive balance between autonomy and support.

His research project is progressing steadily. He is currently evaluating the performance of MindGlide, a deep-learning segmentation algorithm originally developed for multiple sclerosis, in the context of focal cortical dysplasia type IIb (FCD IIb). The project investigates whether the model can accurately detect and segment the transmantle sign, a hallmark imaging feature in FCD IIb, without epilepsy-specific retraining. At this stage, they are applying the model to a dataset of manually annotated MR exams and analyzing detection rates and segmentation accuracy