AI Insight
Researchers developed CAPT (Continuous Autoregressive Population Transformer), an AI model that predicts brain cell activity from calcium imaging data across multiple species including mice, zebrafish, and C. elegans. The model uses a continuous autoregressive approach that can be pre-trained on one dataset and successfully transferred to predict neural activity in different animals and experimental settings from other laboratories. CAPT consistently outperformed both specialized and general-purpose baseline models in forecasting neural population activity and decoding behavior, demonstrating that a single model architecture can learn reusable representations of brain activity patterns across diverse datasets and species.
Why it matters
This approach could significantly accelerate neuroscience research by enabling scientists to leverage existing calcium imaging datasets to analyze new experiments, even across different species. The ability to transfer learned representations between laboratories and experimental paradigms could reduce the computational resources and data required for studying brain activity patterns in various organisms.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
-cross
Abstract: Calcium imaging has enabled large-scale recordings of neural population activity across diverse animals, laboratories, and experimental paradigms. However, these datasets remain highly fragmented: Existing approaches are often designed for individual datasets or tasks, making it unclear whether their learned representations are reusable for new calcium imaging trace datasets. To tackle this gap, we present textbf{CAPT}, a textbf{C}ontinuous textbf{A}utoregressive textbf{P}opulation textbf{T}ransformer for calcium population dynamics. CAPT models continuous calcium traces directly through a continuous patch tokenization strategy and is trained autoregressively with a mean squared error objective, enabling end-to-end training and adaptation to diverse downstream tasks. We first pretrain CAPT on a large-scale mouse calcium imaging dataset and evaluate its transferability across independent mouse, larval zebrafish, and textit{C. elegans} datasets collected by different laboratories. In these transfer settings, the pretrained backbone is frozen and only adaptation modules, i.e., neuron and session embeddings or task-specific heads, are updated. Across neural population forecasting and behavior decoding tasks, CAPT consistently outperforms specialized and general-purpose baselines, demonstrating that a single continuous autoregressive backbone can learn representations that transfer beyond the original pretraining distribution. Beyond predictive performance, multimodal analyses using NeuroPAL annotations in textit{C. elegans} datasets show that CAPT embeddings form a shared functional space across datasets and capture cell-identity-related structure. Together, these results suggest that continuous autoregressive modeling provides an effective framework for learning reusable calcium population representations that transfer across datasets, experimental paradigms, and species.