Discovering Mechanistic Models of Neural Activity: System Identification in an in Silico Zebrafish

TL;DR

This paper introduces an in silico zebrafish testbed for validating neural circuit models, demonstrating that LLM-based tree search can discover accurate, interpretable models that generalize well when guided by structural priors.

Contribution

The study shows that combining LLM-based search with structural priors enables robust discovery of mechanistic neural models in a simulated zebrafish system.

Findings

LLM-based tree search outperforms traditional forecasting baselines.

Structural priors are crucial for out-of-distribution generalization.

Models can recover interpretable mechanistic insights.

Abstract

Constructing mechanistic models of neural circuits is a fundamental goal of neuroscience, yet verifying such models is limited by the lack of ground truth. To rigorously test model discovery, we establish an in silico testbed using neuromechanical simulations of a larval zebrafish as a transparent ground truth. We find that LLM-based tree search autonomously discovers predictive models that significantly outperform established forecasting baselines. Conditioning on sensory drive is necessary but not sufficient for faithful system identification, as models exploit statistical shortcuts. Structural priors prove essential for enabling robust out-of-distribution generalization and recovery of interpretable mechanistic models. Our insights provide guidance for modeling real-world neural recordings and offer a broader template for AI-driven scientific discovery.