Skip to content
PaperFren

Can a 3D MRI network spot and forecast Parkinson’s?

Open paper intelligence

A 3D CNN using T1 MRI separated moderate-to-severe Parkinson’s disease from controls at 74% accuracy; transfer learning reached 64% for mild PD, and MRI plus clinical data predicted 2-year progression at >70%.

Source

Multi-Center 3D CNN for Parkinson's disease diagnosis and prognosis using clinical and T1-weighted MRI data

Basaia S, Sarasso E, Sciancalepore F, et al. · NeuroImage. Clinical · 2025

doi.org/10.1016/j.nicl.2025.103859Read the full paper ↗4 citationscc by

Study at a glance

Design
Computational / modelling — Multi-cohort 3D CNN on T1-weighted MRI (± clinical data) with transfer learning and k-means progression clusters
N
N=312 · Main: 86 mild PD, 62 moderate-to-severe PD, 60 controls; PPMI 14+14; de novo 38+38 (312 MRI/clinical participants)
Population
People with Parkinson’s disease (mild, moderate-to-severe, de novo) and controls from three cohorts including PPMI
Outcome
CNN accuracy for PD vs controls and for faster vs slower motor progression over ~2 years

Structured fields used in claim comparison tables when every cited study has a complete layer.

What they did

Trained a 3D CNN on three cohorts (86 mild, 62 moderate-to-severe PD, 60 controls; PPMI 14+14; de novo 38+38), clustered PD motor progression with k-means on baseline/follow-up UPDRS-III, and tested transfer learning plus MRI+clinical fusion.

What they found

MRI alone: 74% accuracy for moderate-to-severe PD vs controls. Transfer learning improved mild-PD vs control classification to 64%. Progression prediction exceeded 70% with MRI plus clinical features. Activation maps highlighted influential brain regions.

The limits

What it doesn't show

Mid-70s accuracy is not a standalone diagnostic test; UPDRS-based clusters and multi-centre MRI do not replace clinical exam or prove who will need a specific treatment.

Key terms

3D CNN
Three-dimensional convolutional neural network trained on T1-weighted MRI volumes.
Transfer learning
Reusing weights from the easier moderate-PD task to improve mild-PD vs control classification.
UPDRS-III
Motor exam score used (labeled UDPRS-III in the paper) to k-means cluster progression.
PPMI
Parkinson’s Progression Markers Initiative, one of the external mild-PD cohorts.
Activation maps
Visualizations of MRI regions that most influenced CNN decisions.
De novo PD
Recently diagnosed, largely untreated mild PD (n=38) with matched controls.

Flashcards

1 / 10

Research intelligence for this paper

See its role on concept claims, tensions it is part of, placement history, and related discoveries.

Open paper intelligence

Quiz yourself

1 / 5

Moderate-to-severe PD vs control MRI accuracy:

Common questions

Main cohort sizes?

86 mild PD, 62 moderate-to-severe PD, 60 controls.

Moderate PD vs control accuracy?

74% with MRI alone.

Mild PD vs control?

64% after transfer learning.

Progression prediction?

>70% accuracy using MRI plus clinical data.

More on Neurodegeneration