Single-Subject Cortical Morphological Brain Networks: Phenotypic Associations and Neurobiological Substrates

Neurology Biochemistry Genetics Basic Medical Sciences Biophysics Medical Imaging Life Sciences Immunology Medicine Cell Biology
morphological brain network structural magnetic resonance imaging cortical thickness sex neurotransmitter

This paper is a study on the phenotypic associations and neurobiological bases of single-subject morphological brain networks. Combining multimodal and multiscale data, this study reveals the differences in morphological brain networks between genders, their potential as individual-specific markers, and their relationships with gene expression, layer-specific cellular structures, and chemical structures. These findings deepen our understanding of the role and origin of single-subject morphological brain networks and provide strong support for their application in future personalized brain connectome research.

Research Background and Problem Statement

Morphological brain networks refer to the morphological relationships between brain regions estimated based on structural magnetic resonance imaging (sMRI). The earliest studies estimated these relationships by calculating the inter-regional covariance of a morphological indicator (e.g., gray matter volume, cortical thickness, or surface area) within a group. However, this group-based approach neglects inter-individual differences, which leads to the unclear neurobiological significance of the morphological brain network. In recent years, there has been methodological progress in constructing individual-level morphological brain networks, making it possible to understand the neurobiological basis of morphological brain networks by studying their roles and origins. However, there are still many issues regarding the phenotypic associations and neurobiological bases of single-subject morphological brain networks. These issues include whether morphological brain networks differ between sexes, whether they can serve as fingerprints for predicting individual behavior and cognition, and whether they have heritability.

Research Source

This study was jointly completed by Zhen Li, Junle Li, Ningkai Wang, Yating Lü, Qihong Zou, and Jinhui Wang from the Institute for Brain Research and Rehabilitation at South China Normal University and was published in the journal NeuroImage. The article was published online on October 30, 2023.

Research Process

The study utilized several large publicly available datasets, including the Human Connectome Project (hcp) s1200 dataset, the Beijing Normal University Test-Retest Dataset, the longitudinal brain correlation dataset for children’s multisensory vocabulary processing (LBCMLPC Dataset), and the Allen Human Brain Atlas Dataset.

Process Overview:

  1. Data Preprocessing and Morphological Map Extraction: Extract cortical thickness, fractal dimension, gyrification index, and sulcal depth from each T1-weighted structural image.
  2. Brain Region Parcellation: The cortical surface was parcellated into 148 regions of interest (ROIs) using the Destrieux Atlas.
  3. Morphological Similarity Estimation: Evaluate the inter-regional similarity for each morphological indicator to construct four types of single-subject morphological brain networks (CTN, FDN, GIN, and SDN).
  4. Construction of Multi-Modal Morphological Brain Network: Combine networks of different morphological indicators to form a multi-modal network.
  5. Community Detection and Module Parcellation: Perform community detection at the group level to identify modules (communities) within the brain network.
  6. Behavioral and Cognitive Correlation: Use multivariate variance component models and BBS models to evaluate the degree to which morphological connections can explain and predict individual behavioral and cognitive variations.
  7. Individual Identification: Use network matching methods to investigate the potential of morphological brain networks as fingerprints for identifying individuals.
  8. Heritability Assessment: Use the genetic ACE model to quantify the degree of genetic control over morphological brain networks.
  9. Gene Association and Cellular Structure Analysis: Associate morphological brain networks with gene expression, cellular structure, and neurotransmitter networks to identify genes, cellular features, and neurotransmitter receptors significantly contributing to these associations.

Main Results

  1. Sex Differences: There are significant differences in morphological brain networks between males and females, particularly in CTN and SDN.
  2. Behavioral and Cognitive Correlation and Prediction: Single-subject morphological brain networks show significant effects in explaining and predicting individual performance in cognitive and motor domains. CTN and SDN show strong correlations and predictive power in these domains. Additionally, morphological connections between and within modules exhibit similar efficacy in behavioral and cognitive correlation and prediction.
  3. Individual Identification: Morphological connections show high accuracy in identifying individuals, including twin pairs.
  4. Comparison of Multi-Modal and Single-Layer Networks: Compared to networks of single morphological indicators, multi-modal morphological networks perform better in behavioral and cognitive correlation and prediction.
  5. Heritability: Morphological connections generally have moderate heritability, with networks of different morphological indicators showing varying levels of heritability. Inter-module connections are more heritable than intra-module connections.
  6. Gene and Cellular Structure Association: Significant positive correlations are observed between morphological brain networks and gene expression and cellular structure. CTNs show significant correlations with gene transcription and cellular characteristics, with genes mainly enriched in the upper and granular layers of the cortical surface.
  7. Chemical Structure Association: Morphological brain networks significantly correlate with chemical structures (e.g., neurotransmitter receptors), with different types of morphological networks driven by different neurotransmitter receptors.

Significance and Value of the Research

This study reveals the phenotypic associations and neurobiological bases of single-subject morphological brain networks, highlighting the importance of gender in morphological brain network research and demonstrating their potential as predictors of individual behavioral and cognitive performance, effectiveness in individual identification, and genetic and neurobiological bases. These results not only deepen our understanding of the structural connectome of the human brain but also provide new perspectives for the application of brain networks in personalized cognitive and clinical neuroscience research.

Openness of Data and Tools

All data and tools used in this study are open and available, and further research and validation are expected to consolidate and expand these important findings. These findings have significant implications not only for basic research in the field of neuroscience but also offer new ideas for clinical applications and personalized medicine.