NeuroDyn
A Python toolbox for fitting interpretable latent dynamics to large-scale neural recordings.
PythonJAXstate-space
Dr. Elena Rivers
Associate Professor of Computational Neuroscience, Institute for Brain & Cognitive Sciences
I build models of learning in biological and artificial neural systems.
About
I study how networks of neurons—biological and artificial—reorganize themselves to learn from experience. My lab combines large-scale electrophysiology, theory, and machine learning to ask a single question: what are the algorithms the brain uses to update its own wiring?
Before joining the Institute for Brain & Cognitive Sciences I completed my PhD at Stanford and a postdoc at the Gatsby Computational Neuroscience Unit. I care deeply about open, reproducible science and about training the next generation of quantitative neuroscientists.
I’m always happy to hear from prospective students, collaborators, and anyone working at the intersection of neuroscience and machine learning.
Research
Theory
Credit assignment in the brain
Biologically plausible alternatives to backpropagation, and what cortical microcircuits can tell us about how synaptic credit is actually assigned.
Experiment
Representational drift
Why do neural codes change over days even when behavior is stable? We track thousands of neurons over weeks to find out.
Methods
Latent dynamics
State-space and sequence models that recover interpretable low-dimensional dynamics from large neural recordings.
Selected publications
All publications →News
More →- New preprint on representational drift is out — see Publications.
- Elena gave a keynote at the Cosyne 2026 main meeting.
- Welcome to two new PhD students joining the lab this spring!
- Our paper on local credit assignment was accepted at NeurIPS 2025 (spotlight).
- Awarded an NIH R01 to study learning across timescales. Thank you, reviewers!