Related Experiment Video
Updated: Jul 30, 2026
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.6K
Deciphering white matter microstructural alterations in catatonia according to ICD-11: replication and machine
Robin Peretzke1,2, Peter F Neher1,3,4, Geva A Brandt5,6
1Division of Medical Image Computing, German Cancer Research Center, Heidelberg, Germany.
Molecular Psychiatry
|December 2, 2024
Summary
This study reveals that microstructural white matter alterations in the corpus callosum are key to understanding catatonia. Machine learning models using tractomics show promise for improved catatonia diagnosis.
Area of Science:
- Neuroimaging
- Psychiatry
- Computational Neuroscience
Background:
- Catatonia is a severe psychomotor disorder with poorly understood underlying mechanisms.
- Previous research suggests white matter (WM) dysconnectivity in catatonia, but its role in classification remains unclear.
Purpose of the Study:
- To investigate microstructural WM alterations in catatonia using diffusion-weighted MRI.
- To develop and evaluate machine learning models for classifying catatonia patients based on WM features.
Main Methods:
- Diffusion-weighted MRI data from two independent cohorts (whiteCAT and replication) of catatonia patients and controls.
- Tract-based spatial statistics (TBSS), tractometry (TractSeg), and machine learning (ML) with a novel tool (RadTract) for WM feature extraction.
- Classification performance assessed using Area Under the Receiver Operating Characteristics (AUROC) curves.
Main Results:
- Catatonia patients exhibited fractional anisotropy (FA) alterations in multiple segments of the corpus callosum (CC).
- ML models trained on tractomics features achieved higher classification performance (AUROC ~0.79/0.76) than those trained on FA tractometry (AUROC ~0.66/0.51).
- Tractomics features from CC_6 demonstrated robust classification across both cohorts.
Conclusions:
- Microstructural alterations in the corpus callosum WM are significantly implicated in catatonia pathophysiology.
- ML-based classification using tractomics offers a promising approach to enhance diagnostic precision for catatonia.
