The effect of latent confounding processes on the estimation of the strength of causal influences in chain-type networks

Helen Christine Shiells, Marco Thiel, Claude Wischik, Bjoern Schelter

Research output: Contribution to journalArticle

10 Downloads (Pure)

Abstract

Reliable recognition of casual interactions between processes is an issue particularly prevalent in the Neurosciences. When the structure of a network is not a priori known it is almost impossible to observe and measure all components of a system, and missing certain components could potentially lead to the inference of spurious interactions. The aim of this study is to demonstrate the effect of missing components of a network on the inferred strength of a spurious interaction. Our novel method uses vector autoregressive modelling and renormalised partial directed coherence to show how and why the inferred strength of causal interactions between processes changes when components in a network are missed. In cases where a latent confounder is influencing a network and consequently a spurious interaction appears, it is not possible to rely on estimates of the strength of this link as strength estimation methods are influenced by the noise of the latent confounder. Our novel approach demonstrates precisely how a latent confounder can affect the strength measure using analysis of vector autoregressive models. While it is possible to measure the strength of directed causal influences between processes the estimation of strength can be confounded if not all components of a system have been observed during measurement.
Original languageEnglish
Article number1298
JournalMedical Research Archives
Volume5
Issue number9
Publication statusPublished - 18 Sep 2017

Keywords

  • Granger causality
  • VAR modelling
  • rPDC
  • latent confounders

Cite this

@article{3b228d014a92494f9610287a486fb0a5,
title = "The effect of latent confounding processes on the estimation of the strength of causal influences in chain-type networks",
abstract = "Reliable recognition of casual interactions between processes is an issue particularly prevalent in the Neurosciences. When the structure of a network is not a priori known it is almost impossible to observe and measure all components of a system, and missing certain components could potentially lead to the inference of spurious interactions. The aim of this study is to demonstrate the effect of missing components of a network on the inferred strength of a spurious interaction. Our novel method uses vector autoregressive modelling and renormalised partial directed coherence to show how and why the inferred strength of causal interactions between processes changes when components in a network are missed. In cases where a latent confounder is influencing a network and consequently a spurious interaction appears, it is not possible to rely on estimates of the strength of this link as strength estimation methods are influenced by the noise of the latent confounder. Our novel approach demonstrates precisely how a latent confounder can affect the strength measure using analysis of vector autoregressive models. While it is possible to measure the strength of directed causal influences between processes the estimation of strength can be confounded if not all components of a system have been observed during measurement.",
keywords = "Granger causality, VAR modelling, rPDC, latent confounders",
author = "Shiells, {Helen Christine} and Marco Thiel and Claude Wischik and Bjoern Schelter",
note = "The authors acknowledge GTD TauRx Therapeutics centres for generous funding of this research.",
year = "2017",
month = "9",
day = "18",
language = "English",
volume = "5",
journal = "Medical Research Archives",
issn = "2375-1924",
number = "9",

}

TY - JOUR

T1 - The effect of latent confounding processes on the estimation of the strength of causal influences in chain-type networks

AU - Shiells, Helen Christine

AU - Thiel, Marco

AU - Wischik, Claude

AU - Schelter, Bjoern

N1 - The authors acknowledge GTD TauRx Therapeutics centres for generous funding of this research.

PY - 2017/9/18

Y1 - 2017/9/18

N2 - Reliable recognition of casual interactions between processes is an issue particularly prevalent in the Neurosciences. When the structure of a network is not a priori known it is almost impossible to observe and measure all components of a system, and missing certain components could potentially lead to the inference of spurious interactions. The aim of this study is to demonstrate the effect of missing components of a network on the inferred strength of a spurious interaction. Our novel method uses vector autoregressive modelling and renormalised partial directed coherence to show how and why the inferred strength of causal interactions between processes changes when components in a network are missed. In cases where a latent confounder is influencing a network and consequently a spurious interaction appears, it is not possible to rely on estimates of the strength of this link as strength estimation methods are influenced by the noise of the latent confounder. Our novel approach demonstrates precisely how a latent confounder can affect the strength measure using analysis of vector autoregressive models. While it is possible to measure the strength of directed causal influences between processes the estimation of strength can be confounded if not all components of a system have been observed during measurement.

AB - Reliable recognition of casual interactions between processes is an issue particularly prevalent in the Neurosciences. When the structure of a network is not a priori known it is almost impossible to observe and measure all components of a system, and missing certain components could potentially lead to the inference of spurious interactions. The aim of this study is to demonstrate the effect of missing components of a network on the inferred strength of a spurious interaction. Our novel method uses vector autoregressive modelling and renormalised partial directed coherence to show how and why the inferred strength of causal interactions between processes changes when components in a network are missed. In cases where a latent confounder is influencing a network and consequently a spurious interaction appears, it is not possible to rely on estimates of the strength of this link as strength estimation methods are influenced by the noise of the latent confounder. Our novel approach demonstrates precisely how a latent confounder can affect the strength measure using analysis of vector autoregressive models. While it is possible to measure the strength of directed causal influences between processes the estimation of strength can be confounded if not all components of a system have been observed during measurement.

KW - Granger causality

KW - VAR modelling

KW - rPDC

KW - latent confounders

M3 - Article

VL - 5

JO - Medical Research Archives

JF - Medical Research Archives

SN - 2375-1924

IS - 9

M1 - 1298

ER -