Fuzzy clustering of fMRI data: toward a theoretical basis for choosing the fuzziness index

Martin Buerki; Helmut Oswald; Gerhard Schroth

doi:10.1117/12.467054

9 May 2002 Fuzzy clustering of fMRI data: toward a theoretical basis for choosing the fuzziness index

Martin Buerki, Helmut Oswald, Gerhard Schroth

Proceedings Volume 4684, Medical Imaging 2002: Image Processing; (2002) https://doi.org/10.1117/12.467054
Event: Medical Imaging 2002, 2002, San Diego, California, United States

Abstract

The fuzzy clustering algorithm (FCA) is a promising approach for the unsupervised analysis of complex fMRI studies with unknown input functions. Among the few parameters required by the FCA, the fuzziness index m plays an important role and the outcome of the clustering depends strongly on it. Unfortunately, there is no theoretical basis currently known for choosing the value of m and so far, empirical approaches have been carried out to find a reasonable value. The theoretical approach presented here calculates the probability distribution of the membership values u_ij during one iteration of the FCA and judges the regularity of this distribution, therefore indicating the degree of fuzziness of the resulting partition. This allows us to estimate the compactness of the clusters. It turns our that this probability does not only depend on the fuzziness index m, but also on the length of the time courses, a fact that was until now not noticed. Consequently, a reasonable choice of the fuzziness index depends on the signal to noise ratio and the temporal dimension of the data.

Citation Download Citation

Martin Buerki, Helmut Oswald, and Gerhard Schroth "Fuzzy clustering of fMRI data: toward a theoretical basis for choosing the fuzziness index", Proc. SPIE 4684, Medical Imaging 2002: Image Processing, (9 May 2002); https://doi.org/10.1117/12.467054

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available