Estimating brain functional connectivity with sparse multivariate autoregression

Pedro A. Valdés-Sosa, Jose Miguel Bornot, Agustín Lage-Castellanos, Mayrim Vega-Hernández, Jorge Bosch-Bayard, Lester Melie-García, Erick Canales-Rodríguez

    Research output: Contribution to journalArticle

    187 Citations (Scopus)

    Abstract

    There is much current interest in identifying the anatomical and functional circuits that are the basis of the brain's computations, with hope that functional neuroimaging techniques will allow the in vivo study of these neural processes through the statistical analysis of the time-series they produce. Ideally, the use of techniques such as multivariate autoregressive (MAR) modelling should allow the identification of effective connectivity by combining graphical modelling methods with the concept of Granger causality. Unfortunately, current time-series methods perform well only for the case that the length of the time-series Nt is much larger than p, the number of brain sites studied, which is exactly the reverse of the situation in neuroimaging for which relatively short time-series are measured over thousands of voxels. Methods are introduced for dealing with this situation by using sparse MAR models. These can be estimated in a two-stage process involving (i) penalized regression and (ii) pruning of unlikely connections by means of the local false discovery rate developed by Efron. Extensive simulations were performed with idealized cortical networks having small world topologies and stable dynamics. These show that the detection efficiency of connections of the proposed procedure is quite high. Application of the method to real data was illustrated by the identification of neural circuitry related to emotional processing as measured by BOLD.

    LanguageEnglish
    Pages969-981
    Number of pages13
    JournalPhilosophical Transactions of the Royal Society B: Biological Sciences
    Volume360
    Issue number1457
    DOIs
    Publication statusPublished - 1 Jan 2005

    Fingerprint

    Time series
    Brain
    time series analysis
    brain
    Functional neuroimaging
    Small-world networks
    Neuroimaging
    methodology
    Functional Neuroimaging
    Statistical methods
    in vivo studies
    application methods
    pruning
    Causality
    topology
    Topology
    statistical analysis
    Networks (circuits)
    Processing

    Keywords

    • fMRI
    • Functional connectivity
    • Graphical model
    • Sparse multivariate autoregressive model
    • Variable selection

    Cite this

    Valdés-Sosa, P. A., Bornot, J. M., Lage-Castellanos, A., Vega-Hernández, M., Bosch-Bayard, J., Melie-García, L., & Canales-Rodríguez, E. (2005). Estimating brain functional connectivity with sparse multivariate autoregression. 360(1457), 969-981. https://doi.org/10.1098/rstb.2005.1654
    Valdés-Sosa, Pedro A. ; Bornot, Jose Miguel ; Lage-Castellanos, Agustín ; Vega-Hernández, Mayrim ; Bosch-Bayard, Jorge ; Melie-García, Lester ; Canales-Rodríguez, Erick. / Estimating brain functional connectivity with sparse multivariate autoregression. 2005 ; Vol. 360, No. 1457. pp. 969-981.
    @article{5ca4a6c13ddf44aebb2b3ad3026a3d37,
    title = "Estimating brain functional connectivity with sparse multivariate autoregression",
    abstract = "There is much current interest in identifying the anatomical and functional circuits that are the basis of the brain's computations, with hope that functional neuroimaging techniques will allow the in vivo study of these neural processes through the statistical analysis of the time-series they produce. Ideally, the use of techniques such as multivariate autoregressive (MAR) modelling should allow the identification of effective connectivity by combining graphical modelling methods with the concept of Granger causality. Unfortunately, current time-series methods perform well only for the case that the length of the time-series Nt is much larger than p, the number of brain sites studied, which is exactly the reverse of the situation in neuroimaging for which relatively short time-series are measured over thousands of voxels. Methods are introduced for dealing with this situation by using sparse MAR models. These can be estimated in a two-stage process involving (i) penalized regression and (ii) pruning of unlikely connections by means of the local false discovery rate developed by Efron. Extensive simulations were performed with idealized cortical networks having small world topologies and stable dynamics. These show that the detection efficiency of connections of the proposed procedure is quite high. Application of the method to real data was illustrated by the identification of neural circuitry related to emotional processing as measured by BOLD.",
    keywords = "fMRI, Functional connectivity, Graphical model, Sparse multivariate autoregressive model, Variable selection",
    author = "Vald{\'e}s-Sosa, {Pedro A.} and Bornot, {Jose Miguel} and Agust{\'i}n Lage-Castellanos and Mayrim Vega-Hern{\'a}ndez and Jorge Bosch-Bayard and Lester Melie-Garc{\'i}a and Erick Canales-Rodr{\'i}guez",
    year = "2005",
    month = "1",
    day = "1",
    doi = "10.1098/rstb.2005.1654",
    language = "English",
    volume = "360",
    pages = "969--981",
    number = "1457",

    }

    Valdés-Sosa, PA, Bornot, JM, Lage-Castellanos, A, Vega-Hernández, M, Bosch-Bayard, J, Melie-García, L & Canales-Rodríguez, E 2005, 'Estimating brain functional connectivity with sparse multivariate autoregression', vol. 360, no. 1457, pp. 969-981. https://doi.org/10.1098/rstb.2005.1654

    Estimating brain functional connectivity with sparse multivariate autoregression. / Valdés-Sosa, Pedro A.; Bornot, Jose Miguel; Lage-Castellanos, Agustín; Vega-Hernández, Mayrim; Bosch-Bayard, Jorge; Melie-García, Lester; Canales-Rodríguez, Erick.

    Vol. 360, No. 1457, 01.01.2005, p. 969-981.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Estimating brain functional connectivity with sparse multivariate autoregression

    AU - Valdés-Sosa, Pedro A.

    AU - Bornot, Jose Miguel

    AU - Lage-Castellanos, Agustín

    AU - Vega-Hernández, Mayrim

    AU - Bosch-Bayard, Jorge

    AU - Melie-García, Lester

    AU - Canales-Rodríguez, Erick

    PY - 2005/1/1

    Y1 - 2005/1/1

    N2 - There is much current interest in identifying the anatomical and functional circuits that are the basis of the brain's computations, with hope that functional neuroimaging techniques will allow the in vivo study of these neural processes through the statistical analysis of the time-series they produce. Ideally, the use of techniques such as multivariate autoregressive (MAR) modelling should allow the identification of effective connectivity by combining graphical modelling methods with the concept of Granger causality. Unfortunately, current time-series methods perform well only for the case that the length of the time-series Nt is much larger than p, the number of brain sites studied, which is exactly the reverse of the situation in neuroimaging for which relatively short time-series are measured over thousands of voxels. Methods are introduced for dealing with this situation by using sparse MAR models. These can be estimated in a two-stage process involving (i) penalized regression and (ii) pruning of unlikely connections by means of the local false discovery rate developed by Efron. Extensive simulations were performed with idealized cortical networks having small world topologies and stable dynamics. These show that the detection efficiency of connections of the proposed procedure is quite high. Application of the method to real data was illustrated by the identification of neural circuitry related to emotional processing as measured by BOLD.

    AB - There is much current interest in identifying the anatomical and functional circuits that are the basis of the brain's computations, with hope that functional neuroimaging techniques will allow the in vivo study of these neural processes through the statistical analysis of the time-series they produce. Ideally, the use of techniques such as multivariate autoregressive (MAR) modelling should allow the identification of effective connectivity by combining graphical modelling methods with the concept of Granger causality. Unfortunately, current time-series methods perform well only for the case that the length of the time-series Nt is much larger than p, the number of brain sites studied, which is exactly the reverse of the situation in neuroimaging for which relatively short time-series are measured over thousands of voxels. Methods are introduced for dealing with this situation by using sparse MAR models. These can be estimated in a two-stage process involving (i) penalized regression and (ii) pruning of unlikely connections by means of the local false discovery rate developed by Efron. Extensive simulations were performed with idealized cortical networks having small world topologies and stable dynamics. These show that the detection efficiency of connections of the proposed procedure is quite high. Application of the method to real data was illustrated by the identification of neural circuitry related to emotional processing as measured by BOLD.

    KW - fMRI

    KW - Functional connectivity

    KW - Graphical model

    KW - Sparse multivariate autoregressive model

    KW - Variable selection

    UR - http://www.scopus.com/inward/record.url?scp=25444464693&partnerID=8YFLogxK

    U2 - 10.1098/rstb.2005.1654

    DO - 10.1098/rstb.2005.1654

    M3 - Article

    VL - 360

    SP - 969

    EP - 981

    IS - 1457

    ER -

    Valdés-Sosa PA, Bornot JM, Lage-Castellanos A, Vega-Hernández M, Bosch-Bayard J, Melie-García L et al. Estimating brain functional connectivity with sparse multivariate autoregression. 2005 Jan 1;360(1457):969-981. https://doi.org/10.1098/rstb.2005.1654