Information criteria determine the number of active sources

T.R. Knosche; E.M. Berends; M.J. Peters; H.R.A. Jagers

Information criteria determine the number of active sources

T.R. Knosche, E.M. Berends, M.J. Peters, H.R.A. Jagers

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

With the neuroelectromagnetic inverse problem, the optimal choice of the number of sources is a difficult problem, especially in the presence of correlated noise. In this paper we present a number of information criteria that help to solve this problem. They are based on the probability density function of the measurements or their eigenvalues. Make use of the Akaike or MDL (minimum description length) correction term and all employ some sort of noise information. By extensive simulations we investigated the conditions under which these criteria yield reliable estimations. We were able to quantify two major factors of influence: (1) the precision of the noise information and (2) the signal-to-noise ratio (SNR). Here defined as the ratio of the smallest signal eigenvalues and the average of the noise eigenvalues. Furthermore, we found that the Akaike correction term tends to overestimate, due to its greater sensibility to the precision of the noise information.

Original language	English
Title of host publication	Quantitative and topologival EEG and MEG analysis
Subtitle of host publication	Proceedings of the 3rd Hans Berger Congress
Editors	Herbert Witte
Place of Publication	Jena, Germany
Publisher	Karl Mayer GmbH
Pages	321-323
Number of pages	3
ISBN (Print)	3-925978-67-4
Publication status	Published - 14 Dec 1997
Event	Third International Hans Berger Congress - Jena, Duitsland Duration: 4 Oct 1996 → …

Conference

Conference	Third International Hans Berger Congress
City	Jena, Duitsland
Period	4/10/96 → …

1 Conference contribution

Information criteria determine the number of active sources
Knosche, T. R., Berends, E. M., Peters, M. J. & Jagers, H. R. A., 30 Sept 1996, Proceedings of 18th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Los Alamitos, CA: IEEE, Vol. 2. p. 83-87 5 p.
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Open Access
File
142 Downloads (Pure)

Cite this

@inproceedings{7239a46708ca4b23a9cd6e976b7958e9,

title = "Information criteria determine the number of active sources",

abstract = "With the neuroelectromagnetic inverse problem, the optimal choice of the number of sources is a difficult problem, especially in the presence of correlated noise. In this paper we present a number of information criteria that help to solve this problem. They are based on the probability density function of the measurements or their eigenvalues. Make use of the Akaike or MDL (minimum description length) correction term and all employ some sort of noise information. By extensive simulations we investigated the conditions under which these criteria yield reliable estimations. We were able to quantify two major factors of influence: (1) the precision of the noise information and (2) the signal-to-noise ratio (SNR). Here defined as the ratio of the smallest signal eigenvalues and the average of the noise eigenvalues. Furthermore, we found that the Akaike correction term tends to overestimate, due to its greater sensibility to the precision of the noise information.",

author = "T.R. Knosche and E.M. Berends and M.J. Peters and H.R.A. Jagers",

year = "1997",

month = dec,

day = "14",

language = "English",

isbn = "3-925978-67-4",

pages = "321--323",

editor = "Herbert Witte",

booktitle = "Quantitative and topologival EEG and MEG analysis",

publisher = "Karl Mayer GmbH",

address = "Germany",

note = "Third International Hans Berger Congress ; Conference date: 04-10-1996",

}

Information criteria determine the number of active sources. / Knosche, T.R.; Berends, E.M.; Peters, M.J. et al.
Quantitative and topologival EEG and MEG analysis: Proceedings of the 3rd Hans Berger Congress. ed. / Herbert Witte. Jena, Germany: Karl Mayer GmbH, 1997. p. 321-323.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Information criteria determine the number of active sources

AU - Knosche, T.R.

AU - Berends, E.M.

AU - Peters, M.J.

AU - Jagers, H.R.A.

PY - 1997/12/14

Y1 - 1997/12/14

N2 - With the neuroelectromagnetic inverse problem, the optimal choice of the number of sources is a difficult problem, especially in the presence of correlated noise. In this paper we present a number of information criteria that help to solve this problem. They are based on the probability density function of the measurements or their eigenvalues. Make use of the Akaike or MDL (minimum description length) correction term and all employ some sort of noise information. By extensive simulations we investigated the conditions under which these criteria yield reliable estimations. We were able to quantify two major factors of influence: (1) the precision of the noise information and (2) the signal-to-noise ratio (SNR). Here defined as the ratio of the smallest signal eigenvalues and the average of the noise eigenvalues. Furthermore, we found that the Akaike correction term tends to overestimate, due to its greater sensibility to the precision of the noise information.

AB - With the neuroelectromagnetic inverse problem, the optimal choice of the number of sources is a difficult problem, especially in the presence of correlated noise. In this paper we present a number of information criteria that help to solve this problem. They are based on the probability density function of the measurements or their eigenvalues. Make use of the Akaike or MDL (minimum description length) correction term and all employ some sort of noise information. By extensive simulations we investigated the conditions under which these criteria yield reliable estimations. We were able to quantify two major factors of influence: (1) the precision of the noise information and (2) the signal-to-noise ratio (SNR). Here defined as the ratio of the smallest signal eigenvalues and the average of the noise eigenvalues. Furthermore, we found that the Akaike correction term tends to overestimate, due to its greater sensibility to the precision of the noise information.

M3 - Conference contribution

SN - 3-925978-67-4

SP - 321

EP - 323

BT - Quantitative and topologival EEG and MEG analysis

A2 - Witte, Herbert

PB - Karl Mayer GmbH

CY - Jena, Germany

T2 - Third International Hans Berger Congress

Y2 - 4 October 1996

ER -

Information criteria determine the number of active sources

Abstract

Conference

Fingerprint

Research output

Information criteria determine the number of active sources

Cite this