Benchmarking Probabilistic Time Series Forecasting Models on Neural Activity

TL;DR

This paper systematically evaluates eight probabilistic deep learning models for neural activity forecasting, demonstrating their superior performance over classical models and highlighting their potential for neural control and understanding.

Contribution

It provides a comprehensive benchmarking of modern deep learning models on neural activity data, revealing their advantages and limitations compared to traditional approaches.

Findings

Deep learning models outperform classical methods in neural forecasting.

Best models forecast up to 1.5 seconds ahead with informative accuracy.

Results suggest potential for neural control applications.

Abstract

Neural activity forecasting is central to understanding neural systems and enabling closed-loop control. While deep learning has recently advanced the state-of-the-art in the time series forecasting literature, its application to neural activity forecasting remains limited. To bridge this gap, we systematically evaluated eight probabilistic deep learning models, including two foundation models, that have demonstrated strong performance on general forecasting benchmarks. We compared them against four classical statistical models and two baseline methods on spontaneous neural activity recorded from mouse cortex via widefield imaging. Across prediction horizons, several deep learning models consistently outperformed classical approaches, with the best model producing informative forecasts up to 1.5 seconds into the future. Our findings point toward future control applications and open new avenues for probing the intrinsic temporal structure of neural activity.