An optimization approach to signal extraction from noisy multivariate data

T. Yokoo, B. W. Knight, L. Sirovich

Research output: Contribution to journalArticle

55 Citations (Scopus)

Abstract

We consider a problem of blind signal extraction from noisy multivariate data, in which each datum represents a system's response, observed under a particular experimental condition. Our prototype example is multipixel functional images of brain activity in response to a set of prescribed experimental stimuli. We present a novel multivariate analysis technique, which identifies the different activity patterns (signals) that are attributable to specific experimental conditions, without a priori knowledge about the signal or the noise characteristics. The extracted signals, which we terra the generalized indicator functions, are optimal in the sense that they maximize a weighted difference between the signal variance and the noise variance. With an appropriate choice of the weighting parameter, the method returns a set of images whose signal-to-noise ratios satisfy some user-defined level of significance. We demonstrate the performance of our method in optical intrinsic signal imaging of cat cortical area 17. We find that the method performs effectively and robustly in all tested data, which include both real experimental data and numerically simulated data. The method of generalized indicator functions is related to canonical variate analysis, a multivariate analysis technique that directly solves for the maxima of the signal-to-noise ratio, but important theoretical and practical differences exist, which can make our method more appropriate in certain situations.

Original languageEnglish (US)
Pages (from-to)1309-1326
Number of pages18
JournalNeuroImage
Volume14
Issue number6
DOIs
StatePublished - 2001

Fingerprint

Signal-To-Noise Ratio
Noise
Multivariate Analysis
Cats
Brain

Keywords

  • Functional imaging
  • Image analysis
  • Multivariate analysis
  • Optical imaging
  • Signal analysis

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

An optimization approach to signal extraction from noisy multivariate data. / Yokoo, T.; Knight, B. W.; Sirovich, L.

In: NeuroImage, Vol. 14, No. 6, 2001, p. 1309-1326.

Research output: Contribution to journalArticle

Yokoo, T. ; Knight, B. W. ; Sirovich, L. / An optimization approach to signal extraction from noisy multivariate data. In: NeuroImage. 2001 ; Vol. 14, No. 6. pp. 1309-1326.
@article{775b6647b50245c697f799b0f817ac3b,
title = "An optimization approach to signal extraction from noisy multivariate data",
abstract = "We consider a problem of blind signal extraction from noisy multivariate data, in which each datum represents a system's response, observed under a particular experimental condition. Our prototype example is multipixel functional images of brain activity in response to a set of prescribed experimental stimuli. We present a novel multivariate analysis technique, which identifies the different activity patterns (signals) that are attributable to specific experimental conditions, without a priori knowledge about the signal or the noise characteristics. The extracted signals, which we terra the generalized indicator functions, are optimal in the sense that they maximize a weighted difference between the signal variance and the noise variance. With an appropriate choice of the weighting parameter, the method returns a set of images whose signal-to-noise ratios satisfy some user-defined level of significance. We demonstrate the performance of our method in optical intrinsic signal imaging of cat cortical area 17. We find that the method performs effectively and robustly in all tested data, which include both real experimental data and numerically simulated data. The method of generalized indicator functions is related to canonical variate analysis, a multivariate analysis technique that directly solves for the maxima of the signal-to-noise ratio, but important theoretical and practical differences exist, which can make our method more appropriate in certain situations.",
keywords = "Functional imaging, Image analysis, Multivariate analysis, Optical imaging, Signal analysis",
author = "T. Yokoo and Knight, {B. W.} and L. Sirovich",
year = "2001",
doi = "10.1006/nimg.2001.0950",
language = "English (US)",
volume = "14",
pages = "1309--1326",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",
number = "6",

}

TY - JOUR

T1 - An optimization approach to signal extraction from noisy multivariate data

AU - Yokoo, T.

AU - Knight, B. W.

AU - Sirovich, L.

PY - 2001

Y1 - 2001

N2 - We consider a problem of blind signal extraction from noisy multivariate data, in which each datum represents a system's response, observed under a particular experimental condition. Our prototype example is multipixel functional images of brain activity in response to a set of prescribed experimental stimuli. We present a novel multivariate analysis technique, which identifies the different activity patterns (signals) that are attributable to specific experimental conditions, without a priori knowledge about the signal or the noise characteristics. The extracted signals, which we terra the generalized indicator functions, are optimal in the sense that they maximize a weighted difference between the signal variance and the noise variance. With an appropriate choice of the weighting parameter, the method returns a set of images whose signal-to-noise ratios satisfy some user-defined level of significance. We demonstrate the performance of our method in optical intrinsic signal imaging of cat cortical area 17. We find that the method performs effectively and robustly in all tested data, which include both real experimental data and numerically simulated data. The method of generalized indicator functions is related to canonical variate analysis, a multivariate analysis technique that directly solves for the maxima of the signal-to-noise ratio, but important theoretical and practical differences exist, which can make our method more appropriate in certain situations.

AB - We consider a problem of blind signal extraction from noisy multivariate data, in which each datum represents a system's response, observed under a particular experimental condition. Our prototype example is multipixel functional images of brain activity in response to a set of prescribed experimental stimuli. We present a novel multivariate analysis technique, which identifies the different activity patterns (signals) that are attributable to specific experimental conditions, without a priori knowledge about the signal or the noise characteristics. The extracted signals, which we terra the generalized indicator functions, are optimal in the sense that they maximize a weighted difference between the signal variance and the noise variance. With an appropriate choice of the weighting parameter, the method returns a set of images whose signal-to-noise ratios satisfy some user-defined level of significance. We demonstrate the performance of our method in optical intrinsic signal imaging of cat cortical area 17. We find that the method performs effectively and robustly in all tested data, which include both real experimental data and numerically simulated data. The method of generalized indicator functions is related to canonical variate analysis, a multivariate analysis technique that directly solves for the maxima of the signal-to-noise ratio, but important theoretical and practical differences exist, which can make our method more appropriate in certain situations.

KW - Functional imaging

KW - Image analysis

KW - Multivariate analysis

KW - Optical imaging

KW - Signal analysis

UR - http://www.scopus.com/inward/record.url?scp=0035205339&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0035205339&partnerID=8YFLogxK

U2 - 10.1006/nimg.2001.0950

DO - 10.1006/nimg.2001.0950

M3 - Article

C2 - 11707087

AN - SCOPUS:0035205339

VL - 14

SP - 1309

EP - 1326

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

IS - 6

ER -