ZAPBench: A Benchmark for Whole-Brain Activity Prediction in Zebrafish
🤖 AI Summary
The absence of standardized benchmarks hinders rigorous evaluation of whole-brain neural activity prediction in living vertebrates. Method: We introduce ZAPBench—the first cell-resolution benchmark for whole-brain neural activity prediction in zebrafish—built upon >70,000-neuron 4D light-sheet microscopy data. It integrates motion correction, voxel-level cell segmentation annotations, and high-fidelity spatiotemporal alignment to enable fair assessment of both temporal and 3D video modeling approaches. Contribution/Results: ZAPBench enables the first quantitative, cell-level assessment of predictability for dynamic whole-brain activity in a living vertebrate and provides an extensible interface to synaptic-resolution structural atlases, facilitating structure–function joint modeling. Empirical evaluation shows that state-of-the-art temporal and video models significantly outperform naive baselines, yet substantial room for improvement remains. ZAPBench establishes a reproducible, comparable, and scalable evaluation standard for brain activity prediction, advancing systematic benchmarking in computational neuroscience.
Application Category
📝 Abstract
Data-driven benchmarks have led to significant progress in key scientific modeling domains including weather and structural biology. Here, we introduce the Zebrafish Activity Prediction Benchmark (ZAPBench) to measure progress on the problem of predicting cellular-resolution neural activity throughout an entire vertebrate brain. The benchmark is based on a novel dataset containing 4d light-sheet microscopy recordings of over 70,000 neurons in a larval zebrafish brain, along with motion stabilized and voxel-level cell segmentations of these data that facilitate development of a variety of forecasting methods. Initial results from a selection of time series and volumetric video modeling approaches achieve better performance than naive baseline methods, but also show room for further improvement. The specific brain used in the activity recording is also undergoing synaptic-level anatomical mapping, which will enable future integration of detailed structural information into forecasting methods.
Problem
Research questions and friction points this paper is trying to address.
Predicting whole-brain neural activity in zebrafish.
Developing forecasting methods using 4d microscopy data.
Integrating synaptic-level anatomical mapping into activity prediction.
Innovation
Methods, ideas, or system contributions that make the work stand out.
ZAPBench predicts whole-brain neural activity
Uses 4d light-sheet microscopy recordings
Integrates synaptic-level anatomical mapping
🔎 Similar Papers
No similar papers found.
