Postdoctoral researcher

I am a postdoctoral researcher currently working on Theoretical and Computational neuroscience with a background on stochastic dynamics, stochastic analysis, and nonlinear dynamics.

I am broadly interested in how learning and synaptic plasticity shape collective population dynamics and representational geometries, and how these changes manifest in behaviour. For more specific current directions, feel free to reach out.

I completed my PhD with summa cum laude¹ under the mentorship of Prof. Manfred Opper in March 2023 at the Technical University of Berlin (degree)² and under close colaboration with the Institute of Mathematics at the University of Potsdam. There by drawing insights from machine learning, statistical learning theory and statistical physics, we developed three frameworks for efficient sampling, control, and inference of overdamped Langevin dynamics ( Maoutsa, Reich, Opper; Entropy; 2020, Maoutsa, Opper; MLPS-NeurIPS; 2021, Maoutsa, Opper; PRRev; 2022, Maoutsa; MLPS-NeurIPS; 2022, Maoutsa; Physics4ML-ICLR; 2023 ).
The main contributions of my thesis (thesis reviews) were an interacting particle system with deterministic dynamics for samling solutions of Fokker-Planck equations, a feedback control algorithm for stochastic dynamics, and a feedback control-inspired framework for introducing geometric inductive biases into the dynamical inference of stochastic systems.

Before my PhD, I worked in the Network Dynamics (Timme) group at the Max Planck Institute for Dynamics and Self-Organisation in Göttingen, after and during my master’s on Computational Neuroscience at the University of Göttingen. During my stay there, I developed a method for identifying synaptic interactions from spike trains by proposing a mapping of the spiking activity to high-dimensional event spaces that effectively reveal the underlying neuronal interactions (Casadiego*, Maoutsa*, Timme; PRL; 2018), and further studied phase transitions of autonomous intersections.

I made a short stint of ~1 year (July 2023 - Sept. 2024) at the Computation in Neural Circuits lab within the School of Life Sciences at Technical University of Munich working on topics relating structure and function of neural circuits. During that time I worked on identifying potential circuit mechanisms and design constraints that explain the differential representation of familiar and novel stimuli in layer 2/3 mouse V1 [see more], and (co-)supervised three Master’s students working on:

• identifying biologically plausible plasticity rules for learning tasks with low-dimensional representations in the neuronal activity space [see more],

• investigating functional properties of different spike-timing dependent plasticity rules on nonlinear Hawkes networks [see more],

• revealing structural and functional properties of brain circuits from their responses to optogenetic stimulation [see more].

( For more details see mentoring and projects )