MR volume segmentation of gray matter and white matter using manual thresholding: Dependence on image brightness

G. J. Harris, P. E. Barta, L. W. Peng, S. Lee, P. D. Brettschneider, A. Shah, J. D. Henderer, T. E. Schlaepfer, G. D. Pearlson

Research output: Contribution to journalArticlepeer-review

59 Scopus citations

Abstract

PURPOSE: To describe a quantitative MR imaging segmentation method for determination of the volume of cerebrospinal fluid, gray matter, and white matter in living human brain, and to determine the method's reliability. METHODS: We developed a computer method that allows rapid, user-friendly determination of cerebrospinal fluid, gray matter, and white matter volumes in a reliable manner, both globally and regionally. This method was applied to a large control population (N = 57). RESULTS: Initially, image brightness had a strong correlation with the gray-white ratio (r = .78). Bright images tended to overestimate, dim images to underestimate gray matter volumes. This artifact was corrected for by offsetting each image to an approximately equal brightness. After brightness correction, gray-white ratio was correlated with age (r = -.35). The age-dependent gray-white ratio was similar to that for the same age range in a prior neuropathology report. Interrater reliability was high (.93 intraclass correlation coefficient). CONCLUSIONS: The method described here for gray matter, white matter, and cerebrospinal fluid volume calculation is reliable and valid. A correction method for an artifact related to image brightness was developed.

Original languageEnglish (US)
Pages (from-to)225-230
Number of pages6
JournalAmerican Journal of Neuroradiology
Volume15
Issue number2
StatePublished - 1994
Externally publishedYes

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Clinical Neurology

Fingerprint

Dive into the research topics of 'MR volume segmentation of gray matter and white matter using manual thresholding: Dependence on image brightness'. Together they form a unique fingerprint.

Cite this