Automatic extraction of sulcal lines on cortical surfaces based on anisotropic geodesic distance

Jun Kyung Seong, Kiho Im, Sang Wook Yoo, Sang Won Seo, Duk L. Na, Jong M. Lee

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Analyzing cortical sulci is important for studying cortical morphology and brain functions. Although sulcal lines on cortical surfaces can be defined in various ways, it is critical in a neuroimaging study to define a sulcal line along the valley of a cortical surface with a high curvature within a sulcus. To extract the sulcal lines automatically, we present a new geometric algorithm based on the computation of anisotropic skeletons of sulcal regions. Because anisotropic skeletons are highly adaptive to the anisotropic nature of the surface shape, the resulting sulcal lines lie accurately on the valleys of the sulcal areas. Our sulcal lines remain unchanged under local shape variabilities in different human brains. Through experiments, we show that the errors of the sulcal lines for both synthetic data and real cortical surfaces were nearly as constant as the function of random noise. By measuring the changes in sulcal shape in Alzheimer's disease (AD) patients, we further investigated the effectiveness of the accuracy of our sulcal lines using a large sample of MRI data. This study involved 70 normal controls (n [men/women]: 29/41, age [mean ± SD]: 71.7 ± 4.9 years), and 100 AD subjects (37/63, 72.3 ± 5.5). We observe significantly lower absolute average mean curvature and shallower sulcal depth in AD subjects, where the group difference becomes more significant if we measure the quantities along the sulcal lines rather than over the entire sulcal area. The most remarkable difference in the AD patients was the average sulcal depth (control: 11.70 and AD: 11.34).

Original languageEnglish
Pages (from-to)293-302
Number of pages10
JournalNeuroImage
Volume49
Issue number1
DOIs
Publication statusPublished - 2010 Jan 1
Externally publishedYes

Fingerprint

Alzheimer Disease
Skeleton
Brain
Neuroimaging

Keywords

  • Anisotropy
  • Cortical surface
  • Curvature tensor
  • Geodesic
  • Hamilton-Jacobi
  • Skeleton
  • Sulcal lines

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Automatic extraction of sulcal lines on cortical surfaces based on anisotropic geodesic distance. / Seong, Jun Kyung; Im, Kiho; Yoo, Sang Wook; Seo, Sang Won; Na, Duk L.; Lee, Jong M.

In: NeuroImage, Vol. 49, No. 1, 01.01.2010, p. 293-302.

Research output: Contribution to journalArticle

Seong, Jun Kyung ; Im, Kiho ; Yoo, Sang Wook ; Seo, Sang Won ; Na, Duk L. ; Lee, Jong M. / Automatic extraction of sulcal lines on cortical surfaces based on anisotropic geodesic distance. In: NeuroImage. 2010 ; Vol. 49, No. 1. pp. 293-302.
@article{60576c0113094c0f8c50011b3065cd27,
title = "Automatic extraction of sulcal lines on cortical surfaces based on anisotropic geodesic distance",
abstract = "Analyzing cortical sulci is important for studying cortical morphology and brain functions. Although sulcal lines on cortical surfaces can be defined in various ways, it is critical in a neuroimaging study to define a sulcal line along the valley of a cortical surface with a high curvature within a sulcus. To extract the sulcal lines automatically, we present a new geometric algorithm based on the computation of anisotropic skeletons of sulcal regions. Because anisotropic skeletons are highly adaptive to the anisotropic nature of the surface shape, the resulting sulcal lines lie accurately on the valleys of the sulcal areas. Our sulcal lines remain unchanged under local shape variabilities in different human brains. Through experiments, we show that the errors of the sulcal lines for both synthetic data and real cortical surfaces were nearly as constant as the function of random noise. By measuring the changes in sulcal shape in Alzheimer's disease (AD) patients, we further investigated the effectiveness of the accuracy of our sulcal lines using a large sample of MRI data. This study involved 70 normal controls (n [men/women]: 29/41, age [mean ± SD]: 71.7 ± 4.9 years), and 100 AD subjects (37/63, 72.3 ± 5.5). We observe significantly lower absolute average mean curvature and shallower sulcal depth in AD subjects, where the group difference becomes more significant if we measure the quantities along the sulcal lines rather than over the entire sulcal area. The most remarkable difference in the AD patients was the average sulcal depth (control: 11.70 and AD: 11.34).",
keywords = "Anisotropy, Cortical surface, Curvature tensor, Geodesic, Hamilton-Jacobi, Skeleton, Sulcal lines",
author = "Seong, {Jun Kyung} and Kiho Im and Yoo, {Sang Wook} and Seo, {Sang Won} and Na, {Duk L.} and Lee, {Jong M.}",
year = "2010",
month = "1",
day = "1",
doi = "10.1016/j.neuroimage.2009.08.013",
language = "English",
volume = "49",
pages = "293--302",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",
number = "1",

}

TY - JOUR

T1 - Automatic extraction of sulcal lines on cortical surfaces based on anisotropic geodesic distance

AU - Seong, Jun Kyung

AU - Im, Kiho

AU - Yoo, Sang Wook

AU - Seo, Sang Won

AU - Na, Duk L.

AU - Lee, Jong M.

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Analyzing cortical sulci is important for studying cortical morphology and brain functions. Although sulcal lines on cortical surfaces can be defined in various ways, it is critical in a neuroimaging study to define a sulcal line along the valley of a cortical surface with a high curvature within a sulcus. To extract the sulcal lines automatically, we present a new geometric algorithm based on the computation of anisotropic skeletons of sulcal regions. Because anisotropic skeletons are highly adaptive to the anisotropic nature of the surface shape, the resulting sulcal lines lie accurately on the valleys of the sulcal areas. Our sulcal lines remain unchanged under local shape variabilities in different human brains. Through experiments, we show that the errors of the sulcal lines for both synthetic data and real cortical surfaces were nearly as constant as the function of random noise. By measuring the changes in sulcal shape in Alzheimer's disease (AD) patients, we further investigated the effectiveness of the accuracy of our sulcal lines using a large sample of MRI data. This study involved 70 normal controls (n [men/women]: 29/41, age [mean ± SD]: 71.7 ± 4.9 years), and 100 AD subjects (37/63, 72.3 ± 5.5). We observe significantly lower absolute average mean curvature and shallower sulcal depth in AD subjects, where the group difference becomes more significant if we measure the quantities along the sulcal lines rather than over the entire sulcal area. The most remarkable difference in the AD patients was the average sulcal depth (control: 11.70 and AD: 11.34).

AB - Analyzing cortical sulci is important for studying cortical morphology and brain functions. Although sulcal lines on cortical surfaces can be defined in various ways, it is critical in a neuroimaging study to define a sulcal line along the valley of a cortical surface with a high curvature within a sulcus. To extract the sulcal lines automatically, we present a new geometric algorithm based on the computation of anisotropic skeletons of sulcal regions. Because anisotropic skeletons are highly adaptive to the anisotropic nature of the surface shape, the resulting sulcal lines lie accurately on the valleys of the sulcal areas. Our sulcal lines remain unchanged under local shape variabilities in different human brains. Through experiments, we show that the errors of the sulcal lines for both synthetic data and real cortical surfaces were nearly as constant as the function of random noise. By measuring the changes in sulcal shape in Alzheimer's disease (AD) patients, we further investigated the effectiveness of the accuracy of our sulcal lines using a large sample of MRI data. This study involved 70 normal controls (n [men/women]: 29/41, age [mean ± SD]: 71.7 ± 4.9 years), and 100 AD subjects (37/63, 72.3 ± 5.5). We observe significantly lower absolute average mean curvature and shallower sulcal depth in AD subjects, where the group difference becomes more significant if we measure the quantities along the sulcal lines rather than over the entire sulcal area. The most remarkable difference in the AD patients was the average sulcal depth (control: 11.70 and AD: 11.34).

KW - Anisotropy

KW - Cortical surface

KW - Curvature tensor

KW - Geodesic

KW - Hamilton-Jacobi

KW - Skeleton

KW - Sulcal lines

UR - http://www.scopus.com/inward/record.url?scp=70349973770&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70349973770&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2009.08.013

DO - 10.1016/j.neuroimage.2009.08.013

M3 - Article

C2 - 19683580

AN - SCOPUS:70349973770

VL - 49

SP - 293

EP - 302

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

IS - 1

ER -