Poster No:
1421
Submission Type:
Abstract Submission
Authors:
Eunku Bae1, Kwangsun Yoo2, Song E Kim3, Chang-hyun Park3, Hyang Woon Lee3
Institutions:
1Ewha Womans University, Seoul, select a state, 2Sungkyunkwan University, Seoul, Korea, Republic of, 3Ewha Womans University, Seoul, Korea, Republic of
First Author:
Eunku Bae
Ewha Womans University
Seoul, select a state
Co-Author(s):
Kwangsun Yoo
Sungkyunkwan University
Seoul, Korea, Republic of
Song E Kim
Ewha Womans University
Seoul, Korea, Republic of
Introduction:
Investigating the delicate interplay between structural and functional brain connectivity (structure-function coupling) is crucial to understand individual differences in cognitive and behavioral capacities. Building on this premise, our research aimed to develop MRI-based predictive models, with a special emphasis on the integration of structural and functional brain connectivity patterns derived from MRI data. This model is designed to predict individual cognitive functions, considering the variances in cognitive traits not only among individuals but also across different socio-demographic factors, including sex, education, and racial differences. The resulting model is tailored for enhanced prediction accuracy within specific populations, providing a novel approach to cognitive function prediction in the realm of structural and functional brain connectivity studies.
Methods:
The study involved 191 healthy subjects aged 17 to 80 years old, who underwent both structural and functional MRI, alongside a battery of neuropsychological tests including executive functions, verbal and spatial memory. For structural MRI analysis, high resolution 3-dimensional T1 MR images with the FreeSurfer 6.0 was utilized to quantify cortical thickness and volumes. Functional connectivity was analyzed to extract salient features for developing predictive models using the Connectome-based Predictive Modeling (CPM) framework1. This process included creating individual connectivity matrices2, selecting features for modeling, and building a predictive model for validation on new subjects.
Results:
We measured multivariate brain connectivity models based on both structural and functional MRI and evaluated them for various cognitive variables among subjects to predict the individual cognitive capacities. The predictive model based on structural and functional MRI dataset showed distinctive patterns as substantial efficacy in forecasting individual cognitive functions. Notably, feature selection varied across different cognitive tests, suggesting a specific link between brain networks and specific cognitive functions.
Conclusions:
Our study demonstrates the potential of employing integrated brain structural and functional connectivity models to predict individual cognitive levels. Furthermore, it underscores the prospect of developing normative models capable of distinguishing between healthy individuals and patients as we recruit more subjects with certain neuropsychiatric diseases, thereby providing a comprehensive tool for assessing cognitive functions. This approach highlights the utility of population-specific models in the realm of cognitive function prediction and brain connectivity research.
Acknowledgements: This study is supported by the National Research Foundation of Korea (NRF) (No.2020R1A2C2013216, 2019M3C1B8090803, 2019M3C1B8090802, and RS-2023-00265524), Institute of Information & Communication Technology Planning & Evaluation (IITP) grant (No. RS-2022-00155966) by the Korea government (MSIT), and BK21-plus FOUR and Artificial Intelligence Convergence Innovation Human Resources Development programs of Ewha Womans University.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Connectivity (eg. functional, effective, structural) 2
Multivariate Approaches
Keywords:
Data analysis
FUNCTIONAL MRI
Machine Learning
MRI
Multivariate
Open-Source Software
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
1. Shen X et al. Using connectome-based predictive modeling to predict individual behavior from brain connectivity. Nat. Protoc 12, 506–518 (2017). [PubMed: 28182017]
2. Yoo, K., Rosenberg, M.D., Kwon, Y.H. et al. A brain-based general measure of attention. Nat Hum Behav 6, 782–795 (2022)