Background: The glymphatic system plays a critical role in brain waste clearance and health. Diffusion tensor imaging along the perivascular space (DTI-ALPS) is an emerging approach to assess glymphatic function, but manual analysis is limited by its subjectivity and laboriousness in clinical practice. To address these challenges, we developed a deep learning-enhanced DTI-ALPS (dALPS) method that automates and enhances measurement of DTI-ALPS in large-scale cohorts, enabling us to uncover its genetic and environmental determinants. Methods: We proposed an automated workflow combining convolutional neural network (CNN) and You Only Look Once (YOLO) for region-of-interest detection in DTI images. Using this method, we calculated dALPS index for over 65,000 participants from UK Biobank and multiple cohorts, and performed a genome-wide association study (GWAS). Additionally, we conducted transcriptome-wide association study (TWAS) and proteome-wide association study (PWAS) to explore the genetic and molecular underpinnings of glymphatic function. Associations between dALPS and demographic, lifestyle, and clinical traits were comprehensively evaluated. Mediation analysis was conducted to explore the potential mediating role of pharmacological treatments, including antidepressants and sleep medications, in the relationship between disease status and dALPS outcomes. Findings: Our automated dALPS index showed excellent reliability and reproducibility compared to conventional manual techniques (intraclass correlation coefficient = 0·95). We observed that the dALPS index was associated with a wide range of body composition measures and brain structures across different age groups and sex. GWAS identified five significant genetic loci associated with dALPS, two of which were replicated in an independent dataset. Subsequent TWAS and PWAS analyses highlighted potential causal genes and proteins linked to brain fluid dynamics. We found that higher healthy lifestyle index (HLI) was positively correlated with improved dALPS, and confirmed the associations between reduced dALPS and various central nervous system (CNS) disorders, including depression, anxiety and neurodegenerative diseases. Notably, mediation analysis indicated that antidepressants were a risk factor for lower brain glymphatic function (P = 0·004) by partly mediating the risk factor of depression. Interpretation: The dALPS analysis provides a reliable, precise, and automated biomarker for assessing brain glymphatic function. Our findings illuminate the genetic and environmental determinants of glymphatic activity, underscoring the potential of dALPS in clinical assessment, disease prediction and targeted therapeutic strategies. Funding: G.L.’s work is supported by National Natural Science Foundation of China (No. 32470708, No. 32270701), Shenzhen Fundamental Research Program (JCYJ20240813151132042), Young Talent Recruitment Project of Guangdong (2019QN01Y139), Shenzhen Key Laboratory for Systems Medicine in Inflammatory Diseases (ZDSYS20220606100803007) and The Science and Technology Planning Project of Guangdong Province (2023B1212060018). Z.P.’s work is supported by National Natural Science Foundation of China (No. 82271266) and The Shenzhen Medical Research Fund (C2501030).

Deep learning enhanced ALPS reveals genetic and environmental factors of brain glymphatic function

Toja A.;
2026-01-01

Abstract

Background: The glymphatic system plays a critical role in brain waste clearance and health. Diffusion tensor imaging along the perivascular space (DTI-ALPS) is an emerging approach to assess glymphatic function, but manual analysis is limited by its subjectivity and laboriousness in clinical practice. To address these challenges, we developed a deep learning-enhanced DTI-ALPS (dALPS) method that automates and enhances measurement of DTI-ALPS in large-scale cohorts, enabling us to uncover its genetic and environmental determinants. Methods: We proposed an automated workflow combining convolutional neural network (CNN) and You Only Look Once (YOLO) for region-of-interest detection in DTI images. Using this method, we calculated dALPS index for over 65,000 participants from UK Biobank and multiple cohorts, and performed a genome-wide association study (GWAS). Additionally, we conducted transcriptome-wide association study (TWAS) and proteome-wide association study (PWAS) to explore the genetic and molecular underpinnings of glymphatic function. Associations between dALPS and demographic, lifestyle, and clinical traits were comprehensively evaluated. Mediation analysis was conducted to explore the potential mediating role of pharmacological treatments, including antidepressants and sleep medications, in the relationship between disease status and dALPS outcomes. Findings: Our automated dALPS index showed excellent reliability and reproducibility compared to conventional manual techniques (intraclass correlation coefficient = 0·95). We observed that the dALPS index was associated with a wide range of body composition measures and brain structures across different age groups and sex. GWAS identified five significant genetic loci associated with dALPS, two of which were replicated in an independent dataset. Subsequent TWAS and PWAS analyses highlighted potential causal genes and proteins linked to brain fluid dynamics. We found that higher healthy lifestyle index (HLI) was positively correlated with improved dALPS, and confirmed the associations between reduced dALPS and various central nervous system (CNS) disorders, including depression, anxiety and neurodegenerative diseases. Notably, mediation analysis indicated that antidepressants were a risk factor for lower brain glymphatic function (P = 0·004) by partly mediating the risk factor of depression. Interpretation: The dALPS analysis provides a reliable, precise, and automated biomarker for assessing brain glymphatic function. Our findings illuminate the genetic and environmental determinants of glymphatic activity, underscoring the potential of dALPS in clinical assessment, disease prediction and targeted therapeutic strategies. Funding: G.L.’s work is supported by National Natural Science Foundation of China (No. 32470708, No. 32270701), Shenzhen Fundamental Research Program (JCYJ20240813151132042), Young Talent Recruitment Project of Guangdong (2019QN01Y139), Shenzhen Key Laboratory for Systems Medicine in Inflammatory Diseases (ZDSYS20220606100803007) and The Science and Technology Planning Project of Guangdong Province (2023B1212060018). Z.P.’s work is supported by National Natural Science Foundation of China (No. 82271266) and The Shenzhen Medical Research Fund (C2501030).
File in questo prodotto:
File Dimensione Formato  
main.pdf

accesso aperto

Tipologia: Full Text
Licenza: DRM non definito
Dimensione 2.1 MB
Formato Adobe PDF
2.1 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11379/647445
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact