van Meegen Lab

Statistical Physics of Information Processing in Neuronal Networks

alex_van_meegen.jpg
Prof. Alexander van Meegen

Modulbau Physik, Office 112

Otto-Blumenthal-Str. 12

52074 Aachen, Germany

Welcome!

We are a research group working at the intersection of Computational Neuroscience, Machine Learning, and Statistical Physics. The group is part of the Department of Physics at RWTH Aachen University.

Our goal is to develop a theoretical understanding of the principles underlying information processing in neural networks. We investigate both biological and artificial networks, with a current focus on the latter due to their analytical and computational tractability. Using tools from statistical physics, we study how these networks learn, represent, and transform information.

In the long term, we hope to identify generalizable principles that can inform translational research and, ultimately, clinical applications.

If our research, teaching, or outreach interest you, drop us an email or come by!

News

Mar 14, 2026 Louis Pezon presented a postern on our joint work “Hallmarks of computation via shared latent dynamical motifs” at COSYNE.
Mar 01, 2026 Max Wollgarten joined the lab as a PhD student. Welcome Max!
Feb 01, 2026 Official start of the lab at RWTH Aachen University.

Selected publications

  1. Unified field theoretical approach to deep and recurrent neuronal networks
    Kai Segadlo, Bastian Epping, Alexander van Meegen, David Dahmen, Michael Krämer, and Moritz Helias
    Journal of Statistical Mechanics: Theory and Experiment, 2022
  2. Fixed point geometry in chaotic neural networks
    Jakob Stubenrauch, Christian Keup, Anno C. Kurth, Moritz Helias, and Alexander van Meegen
    Physical Review Research, 2025
  3. Coding schemes in neural networks learning classification tasks
    Alexander van Meegen and Haim Sompolinsky
    Nature Communications, 2025
  4. Flat Channels to Infinity in Neural Loss Landscapes
    Flavio Martinelli, Alexander van Meegen, Berfin Simsek, Wulfram Gerstner, and Johanni Brea
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025