Related Experiment Video
Updated: Aug 5, 2026
09:06
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
Predicting future brain atrophy based on longitudinal MRI
Maryam Hadji1, Elaheh Moradi1, Jussi Tohka1
1A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland.
Journal of Alzheimer'S Disease : JAD
|July 28, 2026
Summary
Predicting future brain atrophy using longitudinal MRI scans and risk factors improves accuracy for neurodegenerative disease research, aiding in early detection and trial design for conditions like Alzheimer's disease.
Area of Science:
- Neuroimaging
- Neurodegeneration
- Biomarkers
Background:
- Neuron loss and brain atrophy are key indicators of neurodegenerative diseases, detectable via magnetic resonance imaging (MRI).
- Accurate prediction of future brain atrophy is crucial for Alzheimer's disease (AD) research and related dementias.
- Longitudinal MRI data can potentially offer improved insights into disease progression.
Purpose of the Study:
- To predict annualized changes in brain volumes (hippocampal, ventricular, total gray matter) in individuals with varying cognitive statuses.
- To assess if longitudinal MRI measures enhance prediction accuracy compared to single-time-point MRI data.
- To evaluate the impact of incorporating risk factors (age, sex, APOE4, diagnosis) into prediction models.
Main Methods:
- Elastic net regression was employed to compare prediction models.
- Models were based on single-timepoint MRI data versus longitudinal MRI-derived change measures.
- Models were evaluated with MRI data alone and combined with risk factors.
Main Results:
- The longitudinal MRI plus risk-factor model demonstrated the highest predictive performance (Pearson correlations: 0.62 for hippocampus, 0.51 for ventricles, 0.41 for TGM).
- Longitudinal models consistently outperformed single time-point models.
- Predicted atrophy rates were more effective than current volumes in identifying cognitive decline progression.
Conclusions:
- Longitudinal MRI features significantly enhance the prediction of brain atrophy.
- Predicted atrophy rates serve as sensitive markers for future cognitive decline.
- These findings support the use of predicted atrophy in clinical trial design and cohort enrichment for neurodegenerative diseases.
