Commentary - Journal of Neuroinformatics and Neuroimaging (2021) Volume 6, Issue 6

Neuroimaging data association with behavioral variables: A class of multivariate methods and their comparison using multi-task fmri data.

Luke Morgan^*

Editorial Office, Journal of Neuroinformatics and Neuroimaging, London, United Kingdom

*Corresponding Author:

Luke Morgan
Editorial Office
Journal of Neuroinformatics and Neuroimaging
London
United Kingdom
E-mail: info@alliedacademies.org

Accepted date: December 19, 2021

Citation: Morgan L. Neuroimaging data association with behavioral variables: A class of multivariate methods and their comparison using multi-task fmri data. Neuroinform Neuroimaging 2021;6(6):1.

Introduction

In cognitive neuroimaging, the availability of various datasets that give complementary information is becoming more prevalent. The creation of fusion methods in which the datasets may completely interact with and inform each other when selecting interesting characteristics for further investigation is a key difficulty when studying such datasets. Latent variable models are used in matrix and tensor decomposition methods, allowing for such interactions across datasets. Because these approaches ensure uniqueness with few assumptions, the estimated latent variables (components) are immediately interpretable, that is, they can be used to explain the link between datasets and populations. Because these approaches ensure uniqueness with few assumptions, the estimated latent variables (components) are immediately interpretable, that is, they can be used to explain the link between datasets and populations.

Independent Component Analysis (ICA), a prominent matrix decomposition-based approach on which we focus in this study, offers an appealing basis for completely multivariate data fusion. The use of ICA and its extensions developed for the fusion of multiple datasets can help explain the underlying relationship across datasets by starting from the assumption of latent variable independence (component independence) and are successful when there is a good model match, that is, the model's assumptions are satisfied. Independent Vector Analysis (IVA), for example, generalises ICA across many datasets by using existing statistical information inside and across datasets, which is crucial for multivariate data fusion. IVA enables datasets to completely interact with each other in a symmetric manner by allowing them to play a comparable function in the decomposition and build association only when it is accessible. As a result, IVA offers a completely multivariate strategy for analysing various neuroimaging datasets. Independent Component Analysis (ICA), a prominent matrix decomposition-based approach on which we focus in this study, offers an appealing basis for completely multivariate data fusion. The use of ICA and its extensions developed for the fusion of multiple datasets can help explain the underlying relationship across datasets by starting from the assumption of latent variable independence (component independence) and are successful when there is a good model match, that is, the model's assumptions are satisfied. Independent Vector Analysis (IVA), for example, generalises ICA across many datasets by using existing statistical information inside and across datasets, which is crucial for multivariate data fusion. IVA enables datasets to completely interact with each other in a symmetric manner by allowing them to play a comparable function in the decomposition and build association only when it is accessible. As a result, IVA offers a completely multivariate strategy for analysing various neuroimaging datasets