Xian Qiqi, Chen Zhe Sage
Dept. Psychiatry, Neuroscience Institute, NYU Grossman School of Medicine, New York, NY, USA.
Dept. Biomedical Engineering, NYU Tandon School of Engineering, Brooklyn, NY, USA.
Proc IEEE Int Conf Acoust Speech Signal Process. 2024 Apr;2024:2156-2159. doi: 10.1109/ICASSP48485.2024.10446801. Epub 2024 Mar 18.
Identifying directed information flow or Granger causality between multivariate time series is important for a wide range of applications in science and engineering. However, traditional data-driven approaches are insufficient to assess Granger causality between multimodal data with distinct temporal resolution. Here we propose a new analysis approach to address this challenge and present quantitative characterizations and statistical assessment on frequency-dependent directed information flow ("generalized spectral causality"). We validate our approach with intensive computer simulations on bivariate and trivariate systems for various conditions.
识别多元时间序列之间的定向信息流或格兰杰因果关系对于科学和工程中的广泛应用非常重要。然而，传统的数据驱动方法不足以评估具有不同时间分辨率的多模态数据之间的格兰杰因果关系。在此，我们提出一种新的分析方法来应对这一挑战，并对频率相关的定向信息流（“广义谱因果关系”）进行定量表征和统计评估。我们通过对双变量和三变量系统在各种条件下进行密集的计算机模拟来验证我们的方法。
