Addressing missing nodes as missing data in dynamic causal modeling

Christopher L. Wyatt, Shaza Zaghlool

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Dynamic Causal Modeling (DCM) uses dynamical systems to represent the high-level neural processing strategy for a given cognitive task. The logical network topology of the model is specified by a combination of prior knowledge and statistical analysis of the neuro-imaging signals. Parameters of this a-priori model are then estimated and competing models are compared to determine the most likely model given experimental data. Inter-subject analysis using DCM requires considerable judgement on the part of the experimenter to decide on the validity of assumptions used in the modeling and statistical analysis; in particular, the dropping of subjects with insufficient activity in a region of the model and ignoring activations not included in the model. This manual data filtering is required so that the model's network size is consistent across subjects. A solution to this problem would ease using DCM in population studies and reduce potential sources of experimental bias. The paper describes and compares three different approaches to allow inter-subject comparisons by treating variation in the network size as a missing data problem. These approaches are compared with respect to accuracy in classifying and predicting subject DCMs using simulated data.

Original languageEnglish
Title of host publicationProceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012
Pages81-84
Number of pages4
DOIs
Publication statusPublished - 29 Oct 2012
Externally publishedYes
Event2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012 - London, United Kingdom
Duration: 2 Jul 20124 Jul 2012

Other

Other2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012
CountryUnited Kingdom
CityLondon
Period2/7/124/7/12

Fingerprint

Statistical methods
Neuroimaging
Dynamic analysis
Dynamical systems
Chemical activation
Topology
Processing

Keywords

  • Dymanic Causal Modeling
  • Graph Topology
  • Missing Data

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Cite this

Wyatt, C. L., & Zaghlool, S. (2012). Addressing missing nodes as missing data in dynamic causal modeling. In Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012 (pp. 81-84). [6295933] https://doi.org/10.1109/PRNI.2012.29

Addressing missing nodes as missing data in dynamic causal modeling. / Wyatt, Christopher L.; Zaghlool, Shaza.

Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012. 2012. p. 81-84 6295933.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Wyatt, CL & Zaghlool, S 2012, Addressing missing nodes as missing data in dynamic causal modeling. in Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012., 6295933, pp. 81-84, 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012, London, United Kingdom, 2/7/12. https://doi.org/10.1109/PRNI.2012.29
Wyatt CL, Zaghlool S. Addressing missing nodes as missing data in dynamic causal modeling. In Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012. 2012. p. 81-84. 6295933 https://doi.org/10.1109/PRNI.2012.29
Wyatt, Christopher L. ; Zaghlool, Shaza. / Addressing missing nodes as missing data in dynamic causal modeling. Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012. 2012. pp. 81-84
@inproceedings{356d9b4ca52c4e46ad3836ca575e4b5f,
title = "Addressing missing nodes as missing data in dynamic causal modeling",
abstract = "Dynamic Causal Modeling (DCM) uses dynamical systems to represent the high-level neural processing strategy for a given cognitive task. The logical network topology of the model is specified by a combination of prior knowledge and statistical analysis of the neuro-imaging signals. Parameters of this a-priori model are then estimated and competing models are compared to determine the most likely model given experimental data. Inter-subject analysis using DCM requires considerable judgement on the part of the experimenter to decide on the validity of assumptions used in the modeling and statistical analysis; in particular, the dropping of subjects with insufficient activity in a region of the model and ignoring activations not included in the model. This manual data filtering is required so that the model's network size is consistent across subjects. A solution to this problem would ease using DCM in population studies and reduce potential sources of experimental bias. The paper describes and compares three different approaches to allow inter-subject comparisons by treating variation in the network size as a missing data problem. These approaches are compared with respect to accuracy in classifying and predicting subject DCMs using simulated data.",
keywords = "Dymanic Causal Modeling, Graph Topology, Missing Data",
author = "Wyatt, {Christopher L.} and Shaza Zaghlool",
year = "2012",
month = "10",
day = "29",
doi = "10.1109/PRNI.2012.29",
language = "English",
isbn = "9780769547657",
pages = "81--84",
booktitle = "Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012",

}

TY - GEN

T1 - Addressing missing nodes as missing data in dynamic causal modeling

AU - Wyatt, Christopher L.

AU - Zaghlool, Shaza

PY - 2012/10/29

Y1 - 2012/10/29

N2 - Dynamic Causal Modeling (DCM) uses dynamical systems to represent the high-level neural processing strategy for a given cognitive task. The logical network topology of the model is specified by a combination of prior knowledge and statistical analysis of the neuro-imaging signals. Parameters of this a-priori model are then estimated and competing models are compared to determine the most likely model given experimental data. Inter-subject analysis using DCM requires considerable judgement on the part of the experimenter to decide on the validity of assumptions used in the modeling and statistical analysis; in particular, the dropping of subjects with insufficient activity in a region of the model and ignoring activations not included in the model. This manual data filtering is required so that the model's network size is consistent across subjects. A solution to this problem would ease using DCM in population studies and reduce potential sources of experimental bias. The paper describes and compares three different approaches to allow inter-subject comparisons by treating variation in the network size as a missing data problem. These approaches are compared with respect to accuracy in classifying and predicting subject DCMs using simulated data.

AB - Dynamic Causal Modeling (DCM) uses dynamical systems to represent the high-level neural processing strategy for a given cognitive task. The logical network topology of the model is specified by a combination of prior knowledge and statistical analysis of the neuro-imaging signals. Parameters of this a-priori model are then estimated and competing models are compared to determine the most likely model given experimental data. Inter-subject analysis using DCM requires considerable judgement on the part of the experimenter to decide on the validity of assumptions used in the modeling and statistical analysis; in particular, the dropping of subjects with insufficient activity in a region of the model and ignoring activations not included in the model. This manual data filtering is required so that the model's network size is consistent across subjects. A solution to this problem would ease using DCM in population studies and reduce potential sources of experimental bias. The paper describes and compares three different approaches to allow inter-subject comparisons by treating variation in the network size as a missing data problem. These approaches are compared with respect to accuracy in classifying and predicting subject DCMs using simulated data.

KW - Dymanic Causal Modeling

KW - Graph Topology

KW - Missing Data

UR - http://www.scopus.com/inward/record.url?scp=84867830108&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867830108&partnerID=8YFLogxK

U2 - 10.1109/PRNI.2012.29

DO - 10.1109/PRNI.2012.29

M3 - Conference contribution

SN - 9780769547657

SP - 81

EP - 84

BT - Proceedings - 2012 2nd International Workshop on Pattern Recognition in NeuroImaging, PRNI 2012

ER -