Prediction of Verbal Abilities From Brain Connectivity Data Across the Lifespan Using a Machine Learning Approach
Deborah Früh1,2, Camilla Mendl-Heinisch1,2, Nora Bittner1,2
1Institute of Neuroscience and Medicine (INM-1), Research Centre Jülich, Jülich, Germany.
Brain connectivity data cannot reliably predict verbal abilities like fluency or vocabulary across the lifespan. However, nonverbal cognitive functions showed moderate predictability, particularly in younger adults, suggesting differences in how brain networks relate to distinct cognitive skills.
Area of Science:
- Neuroscience
- Cognitive Science
- Brain Imaging
- Machine Learning
Background:
- Language abilities generally remain stable longer than nonverbal cognition, but specific verbal processes may differ across the lifespan.
- Alterations in brain network architecture, particularly within the frontoparietal network (FPN) and default mode network (DMN), are potential explanations for differing verbal function trajectories.
- Previous research on predicting language abilities from brain connectivity (functional connectivity - FC, and structural connectivity - SC) has yielded mixed results.
Purpose of the Study:
- To investigate the predictability of verbal fluency and vocabulary knowledge using brain connectivity data from the DMN, FPN, and whole brain in a lifespan sample.
- To compare prediction performance across different cognitive abilities (verbal vs. nonverbal), data modalities (FC vs. SC), feature sets (network-specific vs. whole brain), and age groups (total, younger, older).
Main Methods:
- Utilized machine learning (ML) approaches on a lifespan sample (N=717, ages 18-85) from the 1000BRAINS study.
- Analyzed resting-state functional connectivity (FC) and structural connectivity (SC) data.
- Systematically compared prediction performance for verbal fluency, vocabulary knowledge, processing speed, and visual working memory across various parameters.
Main Results:
- Verbal abilities (verbal fluency and vocabulary knowledge) could not be reliably predicted from FC and SC data across all tested feature sets and age groups.
- Nonverbal abilities (processing speed and visual working memory) were moderately predictable from connectivity data, especially SC, in the total and younger adult groups.
- Prediction performance for nonverbal functions was not satisfactory in the older adult group.
Conclusions:
- Verbal functions appear more challenging to predict from domain-general brain connectivity networks compared to nonverbal abilities across the lifespan.
- The findings highlight potential differences in the relationship between brain network architecture and the predictability of distinct cognitive functions.
- Further research is warranted to explore these differences in predictability across cognitive domains and age groups.
More Related Videos
11:14A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants
Published on: October 4, 2015
08:05Measuring Statistical Learning Across Modalities and Domains in School-Aged Children Via an Online Platform and Neuroimaging Techniques
Published on: June 30, 2020
