Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review

被引：937

作者：

Krishnan, Anjali ^{[2
]}

Williams, Lynne J. ^{[3
]}

McIntosh, Anthony Randal ^{[1
]}

Abdi, Herve ^{[2
]}

机构：

[1] Univ Toronto, Dept Psychol, Toronto, ON M5S 3G3, Canada

[2] Univ Texas Dallas, Sch Behav & Brain Sci, Richardson, TX 75080 USA

[3] Rotrnan Res Inst, Kunen Luenfeld Appl Res Unit, Toronto, ON M6A 2E1, Canada

来源：

NEUROIMAGE | 2011年 / 56卷 / 02期

关键词：

Partial least squares correlation; Partial least squares regression; Partial least squares path modeling; PLS; Symmetric PLS; Asymmetric PLS; Task PLS; Behavior PLS; Seed PLS; Multi-block PLS; Multi-table PLS; Canonical variate analysis; Co-inertia analysis; Multiple factor analysis; STATIS; Barycentric discriminant analysis; Common factor analysis; FUNCTIONAL BRAIN IMAGES; EPISODIC MEMORY; PRINCIPAL COMPONENT; REGRESSION; NETWORKS; RETRIEVAL; SYSTEMS; FMRI; CLASSIFICATION; PERFORMANCE;

D O I：

10.1016/j.neuroimage.2010.07.034

中图分类号：

Q189 [神经科学];

学科分类号：

071006 ;

摘要：

Partial Least Squares (PLS) methods are particularly suited to the analysis of relationships between measures of brain activity and of behavior or experimental design. In neuroimaging, PLS refers to two related methods: (1) symmetric PLS or Partial Least Squares Correlation (PLSC), and (2) asymmetric PLS or Partial Least Squares Regression (PLSR). The most popular (by far) version of PLS for neuroimaging is PLSC. It exists in several varieties based on the type of data that are related to brain activity: behavior PLSC analyzes the relationship between brain activity and behavioral data, task PLSC analyzes how brain activity relates to predefined categories or experimental design, seed PLSC analyzes the pattern of connectivity between brain regions, and multi-block or multi-table PLSC integrates one or more of these varieties in a common analysis. PLSR, in contrast to PLSC, is a predictive technique which, typically, predicts behavior (or design) from brain activity. For both PLS methods, statistical inferences are implemented using cross-validation techniques to identify significant patterns of voxel activation. This paper presents both PLS methods and illustrates them with small numerical examples and typical applications in neuroimaging. (C) 2010 Elsevier Inc. All rights reserved.

引用

页码：455 / 475

页数：21

共 83 条

[1]

Abdi H., 2009, Experimental design and analysis for psychology

[2]

Abdi H., 2007, ENCY MEASUREMENT STA, P657

[3]

Abdi H., 2005, Proceedings of the ieee computer society: International conference on computer vision and pattern recognition, P43

[4]

Abdi H., 2010, Encyclopedia of Research Design, P64

[5]

Abdi H, 2007, Encyclopedia of Measurement and Statistics, P907, DOI DOI 10.4135/9781412952644.N413

[6] Principal component analysis [J].

Abdi, Herve ;

Williams, Lynne J. .

WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL STATISTICS, 2010, 2 (04) :433-459

[7] Partial least squares regression and projection on latent structure regression (PLS Regression) [J].

Abdi, Herve .

WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL STATISTICS, 2010, 2 (01) :97-106

[8] How to compute reliability estimates and display confidence and tolerance intervals for pattern classifiers using the Bootstrap and 3-way multidimensional scaling (DISTATIS) [J].

Abdi, Herve ;

Dunlop, Joseph P. ;

Williams, Lynne J. .

NEUROIMAGE, 2009, 45 (01) :89-95

[9] Characterizing spatial and temporal features of autobiographical memory retrieval networks: a partial least squares approach [J].

Addis, DR ;

McIntosh, AR ;

Moscovitch, M ;

Crawley, AP ;

McAndrews, MP .

NEUROIMAGE, 2004, 23 (04) :1460-1471

[10]

[Anonymous], 1996, Rev. Stat. Appl.

← 1 2 3 4 5 6 7 8 9 →