An Interpretable Deep Learning Optimized Wearable Daily Detection System for Parkinson’s Disease DOI Creative Commons
Min Chen,

Zhanfang Sun,

Xin Tao

et al.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Journal Year: 2023, Volume and Issue: 31, P. 3937 - 3946

Published: Jan. 1, 2023

Walking detection in the daily life of patients with Parkinson's disease (PD) is great significance for tracking progress disease. This study aims to implement an accurate, objective, and passive algorithm optimized based on interpretable deep learning architecture walking PD explore most representative spatiotemporal motor features. Five inertial measurement units attached wrist, ankle, waist are used collect motion data from 100 subjects during a 10-meter test. The raw each sensor subjected continuous wavelet transform train classification model constructed 6-channel convolutional neural network (CNN). results show that located at has best performance accuracy 98.01%±0.85% area under receiver operating characteristic curve (AUC) 0.9981±0.0017 ten-fold cross-validation. gradient-weighted class activation mapping shows feature points greater contribution were concentrated lower frequency band (0.5~3Hz) compared healthy controls. visual maps 3D CNN only three out six time series have contribution, which as basis further optimize input, greatly reducing processing costs (50%) while ensuring its (AUC=0.9929±0.0019). To our knowledge, this first consider interpretation-based optimization intelligent diagnosis PD.

Language: Английский

The risk of prenatal bisphenol A exposure in early life neurodevelopment: Insights from epigenetic regulation DOI
Norazirah Mat Nayan, Andrean Husin, Rosfaiizah Siran

et al.

Early Human Development, Journal Year: 2024, Volume and Issue: 198, P. 106120 - 106120

Published: Sept. 11, 2024

Language: Английский

Citations

3
Epigenome‐wide association study, meta‐analysis, and multiscore profiling of whole blood in Parkinson's disease DOI Creative Commons

Ingeborg Haugesag Lie,

Manuela Tan,

Maren Stolp Andersen

et al.

Annals of Clinical and Translational Neurology, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 5, 2025

An increasing body of evidence indicates altered DNA methylation in Parkinson's disease, yet the reproducibility and utility such changes are largely unexplored. We aimed to further elucidate role dysregulated disease evaluate biomarker potential methylation-based profiling. conducted an epigenome-wide association study (EWAS) whole blood, including 280 279 control participants from Oslo, Norway. Next, we took advantage data Progression Markers Initiative (PPMI) a previously published EWAS conduct blood meta-analysis incorporating results total 3068 participants. Finally, generated multiple scores for each Oslo PPMI participant tested their with status, individually joint multiscore model. In meta-analysis, confirm SLC7A11 hypermethylation nominate novel differentially methylated CpG near LPIN1. A model polygenic risk estimates epigenetic risk, smoking, leukocyte proportions differentiated patients area under receiver-operator curve 0.82 cohort 0.65 PPMI. Our highlight power profiling capture aspects indicating precision medicine neurodegenerative disorders. The specific CpGs across sets was limited but may improve if future studies designed account stage incorporate environmental exposure data.

Language: Английский

Citations

0
Nuclear DNA and Mitochondrial Damage of the Cooked Meat Carcinogen 2-Amino-1-methyl-6-phenylimidazo[4,5-b]pyridine in Human Neuroblastoma Cells DOI Open Access
Medjda Bellamri, Kyle Brandt,

Kari Cammerrer

et al.

Chemical Research in Toxicology, Journal Year: 2023, Volume and Issue: 36(8), P. 1361 - 1373

Published: July 8, 2023

Animal fat and iron-rich diets are risk factors for Parkinson's disease (PD). The heterocyclic aromatic amines (HAAs) harman norharman neurotoxicants formed in many foods beverages, including cooked meats, suggesting a role red meat PD. structurally related carcinogenic HAAs 2-amino-1-methyl-6-phenylimidazo[4,5-

Language: Английский

Citations

8
Genetic variation and pesticide exposure influence blood DNA methylation signatures in females with early-stage Parkinson’s disease DOI Creative Commons
Samantha L. Schaffner, William Casazza, Fanny Artaud

et al.

npj Parkinson s Disease, Journal Year: 2024, Volume and Issue: 10(1)

Published: May 7, 2024

Although sex, genetics, and exposures can individually influence risk for sporadic Parkinson's disease (PD), the joint contributions of these factors to epigenetic etiology PD have not been comprehensively assessed. Here, we profiled sex-stratified genome-wide blood DNAm patterns, SNP genotype, pesticide exposure in agricultural workers (71 early-stage cases, 147 controls) explored replication three independent samples varying demographics (n = 218, 222, 872). Using a region-based approach, found more associations with females (69 regions) than males (2 regions, Δβadj| ≥0.03, padj ≤ 0.05). For 48 regions females, models including genotype or substantially improved explaining interindividual variation (padj 0.05), accounting variables decreased estimated effect on DNAm. The results suggested that lesser degree, genotype-exposure interactions contributed PD-associated Our findings should be further larger study populations experimental systems, preferably precise measures exposure.

Language: Английский

Citations

3
An Interpretable Deep Learning Optimized Wearable Daily Detection System for Parkinson’s Disease DOI Creative Commons
Min Chen,

Zhanfang Sun,

Xin Tao

et al.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Journal Year: 2023, Volume and Issue: 31, P. 3937 - 3946

Published: Jan. 1, 2023

Walking detection in the daily life of patients with Parkinson's disease (PD) is great significance for tracking progress disease. This study aims to implement an accurate, objective, and passive algorithm optimized based on interpretable deep learning architecture walking PD explore most representative spatiotemporal motor features. Five inertial measurement units attached wrist, ankle, waist are used collect motion data from 100 subjects during a 10-meter test. The raw each sensor subjected continuous wavelet transform train classification model constructed 6-channel convolutional neural network (CNN). results show that located at has best performance accuracy 98.01%±0.85% area under receiver operating characteristic curve (AUC) 0.9981±0.0017 ten-fold cross-validation. gradient-weighted class activation mapping shows feature points greater contribution were concentrated lower frequency band (0.5~3Hz) compared healthy controls. visual maps 3D CNN only three out six time series have contribution, which as basis further optimize input, greatly reducing processing costs (50%) while ensuring its (AUC=0.9929±0.0019). To our knowledge, this first consider interpretation-based optimization intelligent diagnosis PD.

Language: Английский

Citations

7