Johannes Mehrer
I am a postdoctoral researcher at EPFL’s NeuroAI Lab headed by Martin Schrimpf. My research sits at the intersection of computational neuroscience and artificial intelligence: I build and evaluate deep neural network models of the brain, and use these models to design targeted causal interventions in biological neural systems.
A central theme of my work is the development of model-guided approaches to neuroscience — using computational models not only to predict, but to actively steer neural and behavioral outcomes. This includes work on model-guided microstimulation in the primate visual system, topographic organization in language models, and ecologically valid training datasets for vision models.
I received my PhD from the University of Cambridge (2015–2019), supported by a Cambridge Trust Vice-Chancellor’s Award, and subsequently led a data science team at weclapp SE (2020–2022) before joining EPFL in 2023.
Research Interests
- Brain-like computation in deep neural networks
- Model-guided causal interventions in neural systems
- Topographic organization of sensory cortex
- Ecologically valid models of vision and language
- Computational approaches to neurology and psychiatry
News
April 2026 — Two papers accepted at ICLR 2026: model-guided microstimulation and inducing dyslexia in vision-language models.
2025 — TopoLM selected as a top-2% oral paper at ICLR 2025. Featured in an interview with Swiss national radio RTS.
2025 — Co-organized workshop on “Modeling the Physical Brain: Spatial Organization and Biophysical Constraints” at CCN 2025.