Phenotyping superagers using machine learning algorithms on whole brain connectivity resting-state fMRI studies

Author:

de Godoy Laiz Laura1,de Paula Demetrius Ribeiro2,Min Wenqi3,Studart-Neto Adalberto4,Green Nathan5,Arantes Paula4,Chaim Khallil Taverna4,Moraes Natália Cristina4,Yassuda Mônica Sanches4,Nitrini Ricardo4,Leite Claudia da Costa4,Soddu Andrea6,Bisdas Sotirios5,Panovska-Griffiths Jasmina3

Affiliation:

1. University of Pennsylvania

2. Radboud University Nijmegen Medical Centre

3. University of Oxford

4. Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo

5. University College London

6. Western University

Abstract

Abstract

Superagers, older adults with memory performance similar to middle-aged individuals, were studied to identify key neural networks responsible for their brain function connectivity. Using a previously published resting-state fMRI (rs-fMRI) dataset from 31 participants (14 superagers and 17 controls) examined at 3 and 7 Tesla (T) scanners, we cross-validated the findings from an Elastic Net regression model using a Random Forest algorithm. Important nodes were identified based on Mean Decrease Gini and Mean Decrease Accuracy measures. Superagers were initially phenotyped in six key preselected networks and then across eleven whole-brain networks. The study confirmed the importance of the salience and default mode networks in classifying superagers, identifying significant nodes in the precuneus, posterior cingulate cortex, prefrontal cortex, temporo-occipital junction, and extrastriate superior cortex. Whole-brain analysis highlighted novel relevant networks, including auditory, visual-lateral, and visual-medial networks. Results showed that 7T rs-fMRI provided more discriminative nodes and better predictive performance than 3T. The findings underscore the role of particular brain regions and networks related to memory and cognition in superagers and suggest that additional nodes in auditory and visual networks contribute to their cognitive resilience. These insights enhance understanding of brain resilience and preserved cognitive abilities in older adults.

Publisher

Springer Science and Business Media LLC

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3