The multi-frequency decomposition entropy learning for nonlinear fMRI data analysis DOI Creative Commons
Di Han, Yuhu Shi, Lei Wang

et al.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Journal Year: 2024, Volume and Issue: 33, P. 68 - 80

Published: Dec. 11, 2024

Functional magnetic resonance imaging (fMRI) have been widely adopted to explore the underlying neural mechanisms between psychiatric disorders which share common neurobiology and clinical manifestations. However, existing studies mainly focus on linear relationships ignore nonlinear contributions. To address above issues, we propose a new method named multi-frequency decomposition entropy (MDE) learning for inferring functional connectivity brain regions. Firstly, variational mode was used divide fMRI data into five groups of frequency. Next, copula calculate relationship regions in each frequency group, then best important were screen out by using statistical t-test. Lastly, gyrus importance index proposed reflect distribution trend gyri different groups. The results applying MDE analysis schizophrenia, bipolar disorder, attention-deficit hyperactivity disorder showed that difference three patient healthy control is large at hub nodes, weak when they are same node. In addition, disease exhibits unique characteristics compared with other diseases control. word, differences commonalities reveal possible discriminating biomarkers among mental diseases.

Language: Английский

Modeling Alzheimer’s disease: Bayesian copula graphical model from demographic, cognitive, and neuroimaging data DOI

Lucas Vogels,

Reza Mohammadi, Marit Schoonhoven

et al.

Journal of Alzheimer s Disease, Journal Year: 2025, Volume and Issue: unknown

Published: May 4, 2025

Background The early detection of Alzheimer’s disease (AD) requires an understanding the relationships between a wide range features. Conditional independencies and partial correlations are suitable measures for these relationships, because they can identify effects confounding mediating variables. Objective To estimate conditional dependencies relevant features in AD using Bayesian approach to Gaussian copula graphical models (GCGMs). This has two key advantages. First, it includes binary, discrete, continuous Second, quantifies uncertainty estimates. Despite advantages, GCGMs have not been applied research yet. Methods We design GCGM find among brain-region specific gray matter volume glucose uptake, amyloid-beta levels, demographic information, cognitive test scores. our model 1 022 participants, including healthy cognitively impaired, across different stages AD. Results found that aging reduces cognition through three indirect pathways: hippocampal loss, posterior cingulate cortex (PCC) accumulation. positive correlation being woman cognition, but also discovered four pathways dampen this association women: lower volume, PCC more accumulation, less education. limited relations uptake hippocampus volumes related cognition. Conclusions study shows use offers valuable insights into pathogenesis.

Language: Английский

Citations

0
Bayesian Scalable Precision Factor Analysis for Gaussian Graphical Models DOI Open Access
Noirrit Kiran Chandra,

Peter Müller,

Abhra Sarkar

et al.

Bayesian Analysis, Journal Year: 2024, Volume and Issue: -1(-1)

Published: Jan. 1, 2024

Language: Английский

Citations

2
The multi-frequency decomposition entropy learning for nonlinear fMRI data analysis DOI Creative Commons
Di Han, Yuhu Shi, Lei Wang

et al.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Journal Year: 2024, Volume and Issue: 33, P. 68 - 80

Published: Dec. 11, 2024

Functional magnetic resonance imaging (fMRI) have been widely adopted to explore the underlying neural mechanisms between psychiatric disorders which share common neurobiology and clinical manifestations. However, existing studies mainly focus on linear relationships ignore nonlinear contributions. To address above issues, we propose a new method named multi-frequency decomposition entropy (MDE) learning for inferring functional connectivity brain regions. Firstly, variational mode was used divide fMRI data into five groups of frequency. Next, copula calculate relationship regions in each frequency group, then best important were screen out by using statistical t-test. Lastly, gyrus importance index proposed reflect distribution trend gyri different groups. The results applying MDE analysis schizophrenia, bipolar disorder, attention-deficit hyperactivity disorder showed that difference three patient healthy control is large at hub nodes, weak when they are same node. In addition, disease exhibits unique characteristics compared with other diseases control. word, differences commonalities reveal possible discriminating biomarkers among mental diseases.

Language: Английский

Citations

0