Spatiotemporal deep learning for early detection of isolated REM sleep behavior disorder and Parkinson’s disease using functional MRI data
10 affiliations
- 1Neuroimaging Research Unit, Division of Neuroscience, IRCCS San Raffaele Scientific Institute, Milan, Italy
- 2Neurotech Hub, Vita-Salute San Raffaele University, Milan, Italy
- 3Neurology Unit, IRCCS San Raffaele Scientific Institute, Milan, Italy
- 4Department of Neuroscience, Rehabilitation, Ophthalmology, Genetics and Maternal Child Health, University of Genoa, Genoa, Italy
- 5Biolab, PoliTo(BIO)Med Lab, Department of Electronics and Telecommunications, Politecnico di Torino, Torino, Italy
- 6Vita-Salute San Raffaele University, Milan, Italy
- 7Sleep Disorders Center, Division of Neuroscience, IRCCS San Raffaele Scientific Institute, Milan, Italy
- 8Neurorehabilitation Unit, IRCCS San Raffaele Scientific Institute, Milan, Italy
- 9Department of Advanced Medical and Surgical Sciences, University of Campania “Luigi Vanvitelli”, Napoli, Italy
- 10Neurophysiology Service, IRCCS San Raffaele Scientific Institute, Milan, Italy
Abstract
This study aimed to develop and evaluate a spatiotemporal deep-neural-network (stDNN) using resting-state fMRI (rs-fMRI) data to identify brain biomarkers associated with isolated REM sleep behavior disorder (iRBD) and Parkinson’s disease (PD) and to differentiate these conditions from controls. The final sample included 771 subjects, comprising 423 patients with PD, 144 with iRBD, and 204 healthy controls. stDNN model was applied to mean timeseries extracted for each subject from rs-fMRI data. By integrating spatio-temporal features, the network classified subjects based on distinct neural patterns. Model generalizability was assessed using subject-wise k-fold cross-validation. Explainable artificial intelligence (XAI) methods were applied. stDNN achieved balanced accuracy rates of 71.0% in distinguishing controls from PD and up to 71.9% in middle-stage PD cases. It also demonstrated over 80% accuracy in differentiating healthy controls from iRBD. XAI analysis highlighted the involvement of fronto-parietal and temporal regions, including the dorsolateral prefrontal cortex, and anterior temporal gyri, in distinguishing controls from PD. In the comparison with iRBD, key contributing areas included the bilateral superior and medial frontal gyri, dorsolateral prefrontal cortex, parietal and occipital regions (lingual gyri and cuneus). This study demonstrates the potential of stDNN to differentiate between iRBD, PD, and controls using rs-fMRI data.
Related
- Improved detection and spatiotemporal spectral analysis of neural traveling waves
- Patient-specific functional networks track early cognitive alterations and minor hallucinations in Parkinson’s disease
- Whole genome sequencing, assembly, and functional annotation of Hyphopichia burtonii Y951 isolated from Vitis vinifera L. cv Cabernet Sauvignon grape must
- Window leakage in clinical EEG machine learning: matched subject-independent evaluation across ADHD and Parkinson cohorts
- Genomic and functional characterization of Bacillus siamensis AFSC23 isolated from the gut of European conger (Conger conger) as a probiotic candidate for aquaculture.
BibTeX
@article{basaia2026spatiotemporal,
title = {{Spatiotemporal deep learning for early detection of isolated REM sleep behavior disorder and Parkinson’s disease using functional MRI data}},
author = {Silvia Basaia and Stefano Pisano and Elisabetta Sarasso and Tania Filaferro and Andrea Gardoni and Sara Marelli and Roberta Balestrino and Tommaso Cusolito and Alessandra Castelnuovo and Andrea Grassi and Luigi Ferini-Strambi and Rosita De Micco and Alessandro Tessitore and Massimo Salvi and Filippo Molinari and Federica Agosta and Massimo Filippi},
journal = {NPJ Parkinson's Disease},
year = {2026},
month = jul,
doi = {10.1038/s41531-026-01477-7},
url = {https://doi.org/10.1038/s41531-026-01477-7},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Spatiotemporal deep learning for early detection of isolated REM sleep behavior disorder and Parkinson’s disease using functional MRI data
AU - Silvia Basaia
AU - Stefano Pisano
AU - Elisabetta Sarasso
AU - Tania Filaferro
AU - Andrea Gardoni
AU - Sara Marelli
AU - Roberta Balestrino
AU - Tommaso Cusolito
AU - Alessandra Castelnuovo
AU - Andrea Grassi
AU - Luigi Ferini-Strambi
AU - Rosita De Micco
AU - Alessandro Tessitore
AU - Massimo Salvi
AU - Filippo Molinari
AU - Federica Agosta
AU - Massimo Filippi
PY - 2026
DA - 2026/07/14
JO - NPJ Parkinson's Disease
DO - 10.1038/s41531-026-01477-7
UR - https://doi.org/10.1038/s41531-026-01477-7
AB - This study aimed to develop and evaluate a spatiotemporal deep-neural-network (stDNN) using resting-state fMRI (rs-fMRI) data to identify brain biomarkers associated with isolated REM sleep behavior disorder (iRBD) and Parkinson’s disease (PD) and to differentiate these conditions from controls. The final sample included 771 subjects, comprising 423 patients with PD, 144 with iRBD, and 204 healthy controls. stDNN model was applied to mean timeseries extracted for each subject from rs-fMRI data. By integrating spatio-temporal features, the network classified subjects based on distinct neural patterns. Model generalizability was assessed using subject-wise k-fold cross-validation. Explainable artificial intelligence (XAI) methods were applied. stDNN achieved balanced accuracy rates of 71.0% in distinguishing controls from PD and up to 71.9% in middle-stage PD cases. It also demonstrated over 80% accuracy in differentiating healthy controls from iRBD. XAI analysis highlighted the involvement of fronto-parietal and temporal regions, including the dorsolateral prefrontal cortex, and anterior temporal gyri, in distinguishing controls from PD. In the comparison with iRBD, key contributing areas included the bilateral superior and medial frontal gyri, dorsolateral prefrontal cortex, parietal and occipital regions (lingual gyri and cuneus). This study demonstrates the potential of stDNN to differentiate between iRBD, PD, and controls using rs-fMRI data.
SN - 2373-8057
ER -