University of Sussex
Browse

File(s) not publicly available

Three-dimensional textural analysis of brain images reveals distributed grey-matter abnormalities in schizophrenia

journal contribution
posted on 2023-06-07, 16:03 authored by Balaji Ganeshan, Kenneth A Miles, Rupert YoungRupert Young, Chris ChatwinChris Chatwin, Hugh M D Gurling, Hugo CritchleyHugo Critchley
Objectives: Three-dimensional (3-D) selective- and relative-scale texture analysis (TA) was applied to structural magnetic resonance (MR) brain images to quantify the presence of grey-matter (GM) and white-matter (WM) textural abnormalities associated with schizophrenia. Materials and methods: Brain TA comprised volume filtration using the Laplacian of Gaussian filter to highlight fine, medium and coarse textures within GM and WM, followed by texture quantification. Relative TA (e.g. ratio of fine to medium) was also computed. T1-weighted MR whole-brain images from 32 participants with diagnosis of schizophrenia (n?=?10) and healthy controls (n?=?22) were examined. Five patients possessed marker alleles (SZ8) associated with schizophrenia on chromosome 8 in the pericentriolar material 1 gene while the remaining five had not inherited any of the alleles (SZ0). Results: Filtered fine GM texture (mean grey-level intensity; MGI) most significantly differentiated schizophrenic patients from controls (P?=?0.0058; area under the receiver-operating characteristic curve?=?0.809, sensitivity?=?90%, specificity?=?70%). WM measurements did not distinguish the two groups. Filtered GM and WM textures (MGI) correlated with total GM and WM volume respectively. Medium-to-coarse GM entropy distinguished SZ0 from controls (P?=?0.0069) while measures from SZ8 were intermediate between the two. Conclusions: 3-D TA of brain MR enables detection of subtle distributed morphological features associated with schizophrenia, determined partly by susceptibility genes.

History

Publication status

  • Published

Journal

European Radiology

ISSN

0938-7994

Publisher

Springer Verlag

Issue

4

Volume

20

Page range

941-948

Department affiliated with

  • Clinical and Experimental Medicine Publications

Full text available

  • No

Peer reviewed?

  • Yes

Legacy Posted Date

2011-08-10

Usage metrics

    University of Sussex (Publications)

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC