arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.12992q-bio.NCcs.LG
收藏

SpikeProphecy:大规模用于自回归神经群体预测的基准测试

SpikeProphecy: A Large-Scale Benchmark for Autoregressive Neural Population Forecasting

Department of Electrical and Computer Engineering, University of California, Santa Cruz, CA, USA(加州大学圣克ruz分校电气与计算机工程系);UC Santa Cruz Genomics Institute, University of California, Santa Cruz, CA, USA(加州大学圣克ruz分校基因组研究所);Department of Computer Science and Engineering, University of California, Santa Cruz, CA, USA(加州大学圣克ruz分校计算机科学与工程系);Department of Applied Mathematics, University of California, Santa Cruz, CA, USA(加州大学圣克ruz分校应用数学系);Department of Biomolecular Engineering, University of California, Santa Cruz, CA, USA(加州大学圣克ruz分校生物分子工程系)

John R. Minnick, Jinghui Geng, Kamran Hussain, Jesus Gonzalez-Ferrer, Ash Robbins, Mohammed A. Mostajo-Radji, David Haussler, Jason K. Eshraghian, Mircea Teodorescu

AI总结:

本文提出SpikeProphecy基准测试,用于评估自回归神经群体预测,通过分解指标揭示数据结构,验证了不同架构在预测脑区可预测性上的表现。

摘要:

神经群体模型通常通过预测和实际尖峰计数之间的单个皮尔逊相关系数来评估,该系数掩盖了关键结构。我们主张评估尖峰预测与构建同样重要,并引入SpikeProphecy,首个大规模基准测试,用于在真实电生理记录上进行因果、自回归尖峰计数预测。我们的核心贡献是群体指标分解,将总体性能分解为时间保真度、空间模式准确性和幅度不变对齐。分解揭示了底层数据的方面,这些方面被单个标量折叠在一起。我们应用该协议到105个Neuropixels会话(Steinmetz 2019 + IBL Repeated Site;~89,800个神经元)上,使用七个架构基线,涵盖四种结构家族:四个SSM(三个对角线和一个非对角线)、Transformer、LSTM和脉冲网络。分解揭示了一个脑区可预测性排名,该排名在所有七个基线中重现,并在考虑放电统计学约束的ANCOVA校正后仍然有效(区域ΔR² = 0.018以上)。它还暴露了一个亚泊松评估地板,其中严格指标与真实生物物理约束相结合,揭示了在泊松计数域中ANN到SNN转换的KL-输出率蒸馏的负结果。

英文摘要:

Neural population models, which predict the joint firing of many simultaneously recorded neurons forward in time, are typically evaluated by a single aggregate Pearson correlation $r$ between predicted and actual spike counts, a number that masks critical structure. We argue that how we evaluate spike forecasting matters as much as what we build, and introduce SpikeProphecy, the first large-scale benchmark for causal, autoregressive spike-count forecasting on real electrophysiology recordings. Our core contribution is a population metric decomposition that separates aggregate performance into temporal fidelity, spatial pattern accuracy, and magnitude-invariant alignment. The decomposition surfaces aspects of the underlying data that an aggregate scalar collapses together. We apply the protocol to 105 Neuropixels sessions (Steinmetz 2019 + IBL Repeated Site; ~89,800 neurons) with seven architecture baselines spanning four structural families: four SSMs (three diagonal and one non-diagonal), a Transformer, an LSTM, and a spiking network. The decomposition surfaces a brain-region predictability ranking that reproduces across all seven baselines and survives ANCOVA correction for firing-statistics constraints (region $ΔR^2 = 0.018$ above the firing-statistics covariates). It also exposes a sub-Poisson evaluation floor where rigorous metrics combine with genuine biophysical constraints on regular spike trains, and yields a negative result on KL-on-output-rates distillation for ANN-to-SNN transfer in this Poisson count domain.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑