Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models

Abstract Amyotrophic lateral sclerosis (ALS) is a fatal neurological disease marked by motor deterioration and cognitive decline. Early diagnosis is challenging due to the complexity of sporadic ALS and the lack of a defined risk population. In this study, we developed Miniset-DenseSENet, a convolut...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuqing Xia, Jenna M. Gregory, Fergal M. Waldron, Holly Spence, Marta Vallejo
Format: Article
Language:English
Published: Nature Portfolio 2025-02-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-90881-9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850185240822153216
author Yuqing Xia
Jenna M. Gregory
Fergal M. Waldron
Holly Spence
Marta Vallejo
author_facet Yuqing Xia
Jenna M. Gregory
Fergal M. Waldron
Holly Spence
Marta Vallejo
author_sort Yuqing Xia
collection DOAJ
description Abstract Amyotrophic lateral sclerosis (ALS) is a fatal neurological disease marked by motor deterioration and cognitive decline. Early diagnosis is challenging due to the complexity of sporadic ALS and the lack of a defined risk population. In this study, we developed Miniset-DenseSENet, a convolutional neural network combining DenseNet121 with a Squeeze-and-Excitation attention mechanism, using 190 autopsy brain images from the Gregory Laboratory at the University of Aberdeen. The model distinguishes controls, ALS patients with no cognitive impairment, and ALS patients with cognitive impairment (ALS-frontotemporal dementia) with 97.37% accuracy, addressing a significant challenge in overlapping neurodegenerative disorders involving TDP-43 proteinopathy. Miniset-DenseSENet outperformed other transfer learning models, achieving a sensitivity of 1 and specificity of 0.95. These findings suggest that integrating transfer learning and attention mechanisms into neuroimaging can enhance diagnostic accuracy, enabling earlier ALS detection and improving patient stratification. This model has the potential to guide clinical decisions and support personalied therapeutic strategies.
format Article
id doaj-art-18d247b5d3494e1eaa87334af221e3fc
institution OA Journals
issn 2045-2322
language English
publishDate 2025-02-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-18d247b5d3494e1eaa87334af221e3fc2025-08-20T02:16:48ZengNature PortfolioScientific Reports2045-23222025-02-0115111310.1038/s41598-025-90881-9Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning modelsYuqing Xia0Jenna M. Gregory1Fergal M. Waldron2Holly Spence3Marta Vallejo4School of Engineering and Physical Sciences, Heriot-Watt UniversityInstitute of Medical Sciences, University of AberdeenInstitute of Medical Sciences, University of AberdeenInstitute of Medical Sciences, University of AberdeenSchool of Mathematical and Computer Sciences, Heriot-Watt UniversityAbstract Amyotrophic lateral sclerosis (ALS) is a fatal neurological disease marked by motor deterioration and cognitive decline. Early diagnosis is challenging due to the complexity of sporadic ALS and the lack of a defined risk population. In this study, we developed Miniset-DenseSENet, a convolutional neural network combining DenseNet121 with a Squeeze-and-Excitation attention mechanism, using 190 autopsy brain images from the Gregory Laboratory at the University of Aberdeen. The model distinguishes controls, ALS patients with no cognitive impairment, and ALS patients with cognitive impairment (ALS-frontotemporal dementia) with 97.37% accuracy, addressing a significant challenge in overlapping neurodegenerative disorders involving TDP-43 proteinopathy. Miniset-DenseSENet outperformed other transfer learning models, achieving a sensitivity of 1 and specificity of 0.95. These findings suggest that integrating transfer learning and attention mechanisms into neuroimaging can enhance diagnostic accuracy, enabling earlier ALS detection and improving patient stratification. This model has the potential to guide clinical decisions and support personalied therapeutic strategies.https://doi.org/10.1038/s41598-025-90881-9Amyotrophic lateral sclerosisTDP-43 proteinCognitive impairmentTransfer learningAttention mechanisms
spellingShingle Yuqing Xia
Jenna M. Gregory
Fergal M. Waldron
Holly Spence
Marta Vallejo
Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
Scientific Reports
Amyotrophic lateral sclerosis
TDP-43 protein
Cognitive impairment
Transfer learning
Attention mechanisms
title Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
title_full Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
title_fullStr Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
title_full_unstemmed Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
title_short Improving ALS detection and cognitive impairment stratification with attention-enhanced deep learning models
title_sort improving als detection and cognitive impairment stratification with attention enhanced deep learning models
topic Amyotrophic lateral sclerosis
TDP-43 protein
Cognitive impairment
Transfer learning
Attention mechanisms
url https://doi.org/10.1038/s41598-025-90881-9
work_keys_str_mv AT yuqingxia improvingalsdetectionandcognitiveimpairmentstratificationwithattentionenhanceddeeplearningmodels
AT jennamgregory improvingalsdetectionandcognitiveimpairmentstratificationwithattentionenhanceddeeplearningmodels
AT fergalmwaldron improvingalsdetectionandcognitiveimpairmentstratificationwithattentionenhanceddeeplearningmodels
AT hollyspence improvingalsdetectionandcognitiveimpairmentstratificationwithattentionenhanceddeeplearningmodels
AT martavallejo improvingalsdetectionandcognitiveimpairmentstratificationwithattentionenhanceddeeplearningmodels