Clustered cNMF for fMRI data analysis

Xiaoxiang Wang; Jie Tian; Lei Yang; Jin Hu

doi:10.1117/12.596023

Clustered cNMF for fMRI data analysis

Xiaoxiang Wang
, Jie Tian
, Lei Yang
, Jin Hu

Chinese Academy of Sciences

Research output: Contribution to journal › Conference article › peer-review

4 Scopus citations

Abstract

This paper introduces a framework for the application of constrained non-negative matrix factorization (cNMF) to estimate the statistically distinct neural responses in a sequence of functional magnetic resonance images (fMRI). While an improved objective function has been defined to make the representation suitable for task-related brain activation detection, in this paper we explore various methods for better detection and efficient computation, placing particular emphasis on the initialization of the constrained NMF algorithm. The K-means algorithm performs this structured initialization and the information theoretic criterion of minimum description length (MDL) is used to estimate the number of clusters. We illustrate the method by a set of functional neuroimages from a motor activation study.

Original language	English
Article number	73
Pages (from-to)	631-638
Number of pages	8
Journal	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	5746
Issue number	II
DOIs	https://doi.org/10.1117/12.596023
State	Published - 2005
Externally published	Yes
Event	Medical Imaging 2005 - Physiology, Function, and Structure from Medical Images - San Diego, CA, United States Duration: 13 Feb 2005 → 15 Feb 2005

Keywords

BOLD
K-means clustering
Minimum Description Length (MDL)
Non-negative Matrix Factorization (NMF)
fMRI

Access to Document

10.1117/12.596023

Cite this

@article{9932f29504db4dd8884b60f3a72e4724,

title = "Clustered cNMF for fMRI data analysis",

abstract = "This paper introduces a framework for the application of constrained non-negative matrix factorization (cNMF) to estimate the statistically distinct neural responses in a sequence of functional magnetic resonance images (fMRI). While an improved objective function has been defined to make the representation suitable for task-related brain activation detection, in this paper we explore various methods for better detection and efficient computation, placing particular emphasis on the initialization of the constrained NMF algorithm. The K-means algorithm performs this structured initialization and the information theoretic criterion of minimum description length (MDL) is used to estimate the number of clusters. We illustrate the method by a set of functional neuroimages from a motor activation study.",

keywords = "BOLD, K-means clustering, Minimum Description Length (MDL), Non-negative Matrix Factorization (NMF), fMRI",

author = "Xiaoxiang Wang and Jie Tian and Lei Yang and Jin Hu",

year = "2005",

doi = "10.1117/12.596023",

language = "英语",

volume = "5746",

pages = "631--638",

journal = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

issn = "1605-7422",

publisher = "SPIE",

number = "II",

note = "Medical Imaging 2005 - Physiology, Function, and Structure from Medical Images ; Conference date: 13-02-2005 Through 15-02-2005",

}

TY - JOUR

T1 - Clustered cNMF for fMRI data analysis

AU - Wang, Xiaoxiang

AU - Tian, Jie

AU - Yang, Lei

AU - Hu, Jin

PY - 2005

Y1 - 2005

N2 - This paper introduces a framework for the application of constrained non-negative matrix factorization (cNMF) to estimate the statistically distinct neural responses in a sequence of functional magnetic resonance images (fMRI). While an improved objective function has been defined to make the representation suitable for task-related brain activation detection, in this paper we explore various methods for better detection and efficient computation, placing particular emphasis on the initialization of the constrained NMF algorithm. The K-means algorithm performs this structured initialization and the information theoretic criterion of minimum description length (MDL) is used to estimate the number of clusters. We illustrate the method by a set of functional neuroimages from a motor activation study.

AB - This paper introduces a framework for the application of constrained non-negative matrix factorization (cNMF) to estimate the statistically distinct neural responses in a sequence of functional magnetic resonance images (fMRI). While an improved objective function has been defined to make the representation suitable for task-related brain activation detection, in this paper we explore various methods for better detection and efficient computation, placing particular emphasis on the initialization of the constrained NMF algorithm. The K-means algorithm performs this structured initialization and the information theoretic criterion of minimum description length (MDL) is used to estimate the number of clusters. We illustrate the method by a set of functional neuroimages from a motor activation study.

KW - BOLD

KW - K-means clustering

KW - Minimum Description Length (MDL)

KW - Non-negative Matrix Factorization (NMF)

KW - fMRI

UR - https://www.scopus.com/pages/publications/23844487491

U2 - 10.1117/12.596023

DO - 10.1117/12.596023

M3 - 会议文章

AN - SCOPUS:23844487491

SN - 1605-7422

VL - 5746

SP - 631

EP - 638

JO - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

JF - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

IS - II

M1 - 73

T2 - Medical Imaging 2005 - Physiology, Function, and Structure from Medical Images

Y2 - 13 February 2005 through 15 February 2005

ER -

Clustered cNMF for fMRI data analysis

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this