Principal component analysis of event-related potentials: A note on misallocation of variance

Joachim Möcks; Rolf Verleger

doi:10.1016/0168-5597(86)90018-3

Principal component analysis of event-related potentials: A note on misallocation of variance

Joachim Möcks^*, Rolf Verleger

^*Corresponding author for this work

Department of Neurology

University of Mannheim

84 Citations (Scopus)

Abstract

Wood and McCarthy (1984) found a 'misallocation of variance' when applying PCA, including Varimax rotation, to simulated data. Here it is demonstrated that this effect can be produced by Varimax rotation, without PCA as an intervening step. PCA does not distort or lose information when extracting components, since it is shown that the prototypes may be perfectly reconstructed from the unrotated solution. However, it is stressed that infinitely many sets of prototypes may render the same final solution, a fact which cannot be overcome by any method. The role of the rotation step within this framework is discussed.

Original language	English
Journal	Electroencephalography and Clinical Neurophysiology/ Evoked Potentials
Volume	65
Issue number	5
Pages (from-to)	393-398
Number of pages	6
ISSN	0168-5597
DOIs	https://doi.org/10.1016/0168-5597(86)90018-3
Publication status	Published - 01.01.1986

Funding

This work was made possible by financial support from the DeutscheF orschungsgemeinschaft (SFB 116, SFB 307). We thank Theo Gasser, Christel Jennen-Steinmetz, and Dieter Schellberg for their helpful comments.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being
SDG 10 Reduced Inequalities

Research Areas and Centers

Academic Focus: Center for Brain, Behavior and Metabolism (CBBM)

Access to Document

10.1016/0168-5597(86)90018-3

Cite this

@article{973caaf97f3341e28832bbcca77a38ab,

title = "Principal component analysis of event-related potentials: A note on misallocation of variance",

abstract = "Wood and McCarthy (1984) found a 'misallocation of variance' when applying PCA, including Varimax rotation, to simulated data. Here it is demonstrated that this effect can be produced by Varimax rotation, without PCA as an intervening step. PCA does not distort or lose information when extracting components, since it is shown that the prototypes may be perfectly reconstructed from the unrotated solution. However, it is stressed that infinitely many sets of prototypes may render the same final solution, a fact which cannot be overcome by any method. The role of the rotation step within this framework is discussed.",

author = "Joachim M{\"o}cks and Rolf Verleger",

year = "1986",

month = jan,

day = "1",

doi = "10.1016/0168-5597(86)90018-3",

language = "English",

volume = "65",

pages = "393--398",

journal = "Electroencephalography and Clinical Neurophysiology/ Evoked Potentials",

issn = "0168-5597",

publisher = "Elsevier B.V.",

number = "5",

}

TY - JOUR

T1 - Principal component analysis of event-related potentials: A note on misallocation of variance

AU - Möcks, Joachim

AU - Verleger, Rolf

PY - 1986/1/1

Y1 - 1986/1/1

N2 - Wood and McCarthy (1984) found a 'misallocation of variance' when applying PCA, including Varimax rotation, to simulated data. Here it is demonstrated that this effect can be produced by Varimax rotation, without PCA as an intervening step. PCA does not distort or lose information when extracting components, since it is shown that the prototypes may be perfectly reconstructed from the unrotated solution. However, it is stressed that infinitely many sets of prototypes may render the same final solution, a fact which cannot be overcome by any method. The role of the rotation step within this framework is discussed.

AB - Wood and McCarthy (1984) found a 'misallocation of variance' when applying PCA, including Varimax rotation, to simulated data. Here it is demonstrated that this effect can be produced by Varimax rotation, without PCA as an intervening step. PCA does not distort or lose information when extracting components, since it is shown that the prototypes may be perfectly reconstructed from the unrotated solution. However, it is stressed that infinitely many sets of prototypes may render the same final solution, a fact which cannot be overcome by any method. The role of the rotation step within this framework is discussed.

UR - https://www.scopus.com/pages/publications/0022549636

U2 - 10.1016/0168-5597(86)90018-3

DO - 10.1016/0168-5597(86)90018-3

M3 - Journal articles

C2 - 2427330

AN - SCOPUS:0022549636

SN - 0168-5597

VL - 65

SP - 393

EP - 398

JO - Electroencephalography and Clinical Neurophysiology/ Evoked Potentials

JF - Electroencephalography and Clinical Neurophysiology/ Evoked Potentials

IS - 5

ER -

Principal component analysis of event-related potentials: A note on misallocation of variance

Abstract

Funding

UN SDGs

Research Areas and Centers

Access to Document

Other files and links

Fingerprint

Cite this