Back

Machine Learning for Longitudinal Brain-Age Prediction

Wegmann, M.; Ganz, M.; Svensson, J. E.; Plaven-Sigray, P.; Dörfel, R. P.

2025-12-11 health informatics
10.64898/2025.12.10.25341964 medRxiv
Show abstract

Cross-sectional brain age models have demonstrated high accuracy and reliability for predicting chronological age based on structural brain features derived from single MRI scans. However, these models cannot separate baseline variation from true aging-related changes or noise. Longitudinal models address this limitation by predicting inter-scan intervals from paired MRI scans, controlling for baseline factors through repeated measurements. Using OASIS-3 data, we compare a cross-sectional 3D CNN against three longitudinal architectures for predicting inter-scan intervals: LILAC (Siamese neural network), LILAC+ (enhanced Siamese network with multi-layer perceptron), and AM (variational autoencoder). Longitudinal models substantially outperformed the cross-sectional approach, with LILAC+ achieving best performance (MSE = 1.97 years2, MAE = 0.99 years, r = 0.86, R2 = 0.71). Our results suggest that direct modeling of longitudinal change is more effective at capturing individual aging trajectories than deriving intervals from cross-sectional predictions.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
Human Brain Mapping
○ 329 papers in training set
Top 0.1%
18.4%
Similar papers in this journal
2
Brain Informatics
○ 10 papers in training set
Top 0.1%
11.9%
Similar papers in this journal
3
Biology Methods and Protocols
◐ 61 papers in training set
Top 0.1%
7.2%
Similar papers in this journal
4
Neurobiology of Aging
○ 107 papers in training set
Top 0.3%
5.5%
Similar papers in this journal
5
Scientific Reports
○ 3612 papers in training set
Top 22%
4.4%
Similar papers in this journal
6
Medical Image Analysis
○ 35 papers in training set
Top 0.2%
3.2%
50% of probability mass above
7
Frontiers in Neuroscience
○ 256 papers in training set
Top 1%
3.2%
8
BMC Medical Research Methodology
○ 47 papers in training set
Top 0.4%
3.2%
9
PLOS ONE
● 5266 papers in training set
Top 37%
3.1%
10
Frontiers in Aging Neuroscience
○ 74 papers in training set
Top 0.6%
2.4%
11
Scientific Data
○ 209 papers in training set
Top 1%
1.7%
12
IEEE Transactions on Biomedical Engineering
● 40 papers in training set
Top 0.6%
1.7%
13
NeuroImage: Clinical
○ 144 papers in training set
Top 1%
1.7%
14
GeroScience
○ 109 papers in training set
Top 1%
1.7%
15
PLOS Digital Health
● 106 papers in training set
Top 3%
1.5%
16
IEEE Journal of Biomedical and Health Informatics
● 37 papers in training set
Top 0.8%
1.3%
17
Computers in Biology and Medicine
○ 128 papers in training set
Top 3%
1.3%
18
Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring
○ 42 papers in training set
Top 0.8%
1.1%
19
NMR in Biomedicine
○ 28 papers in training set
Top 0.3%
1.1%
20
Imaging Neuroscience
● 282 papers in training set
Top 3%
1.1%
21
Frontiers in Psychiatry
○ 87 papers in training set
Top 2%
1.1%
22
Informatics in Medicine Unlocked
○ 22 papers in training set
Top 1.0%
1.0%
23
PLOS Computational Biology
● 1863 papers in training set
Top 20%
0.8%
24
Bioinformatics
◐ 1204 papers in training set
Top 9%
0.8%
25
Neurophotonics
● 42 papers in training set
Top 0.6%
0.8%
26
IEEE Transactions on Neural Systems and Rehabilitation Engineering
● 49 papers in training set
Top 0.9%
0.8%
27
Alzheimer's Research & Therapy
○ 57 papers in training set
Top 1%
0.8%
28
BMJ Open
● 601 papers in training set
Top 12%
0.8%
29
JAMIA Open
◐ 42 papers in training set
Top 2%
0.6%
30
Artificial Intelligence in Medicine
○ 17 papers in training set
Top 0.9%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.