Biology subjects

Byeon, K.

Publications and source records attributed to Byeon, K..

6 recordsLinked to original sources

bioRxiv2025

Statistically valid explainable black-box machine learning: applications in sex classification across species using brain imaging

Sex classification using neuroimaging data has the potential to revolutionize personalized diagnostics by revealing subtle structural brain differences that underlie sex-specific disease risks. Despite the promise of machine learning, traditional methods often fall short in providing both high classification accuracy and interpretable, statistically validated feature importance scores for high-dimensional imaging data. This gap is particularly evident when conventional techniques such as random forests, LIME, and SHAP are applied, as they struggle with complex feature interactions and managing noise in large datasets. We address this challenge by developing an integrated framework that combines Oblique Random Forests (ORFs) with a novel, permutation-based feature importance testing algorithm. ORFs extend traditional random forests by employing oblique decision boundaries through linear combinations of features, thereby capturing intricate interactions inherent in neuroimaging data. Our feature importance testing method, NEOFIT, rigorously quantifies the significance of each feature by generating null distributions and corrected p-values. We first validate our approach using simulated datasets, establishing its robustness and scalability under controlled conditions. We then apply our method to classify sex from both voxel-wise structural MRI and cortical thickness data in humans and macaques, facilitating direct cross-species comparisons. ORFs achieves AUC > 0.80 on human data, and > 0.70 on macaque data, while NEOFIT identifies statistically significant features aligned with sex-dimorphic neuroanatomy. Our results demonstrate that the proposed framework not only enhances classification performance but also provides clear, interpretable insights into the neuroanatomical features that distinguish sexes. These methodological advancements pave the way for improved diagnostic tools and contribute to a deeper understanding of the evolutionary basis of sex differences in brain structure.

neuroscience↗

bioRxiv2025

Developmental Variations in Recurrent Spatiotemporal Brain Propagations from Childhood to Adulthood

The brain undergoes profound structural and functional transformations from childhood to adolescence. Convergent evidence suggests that neurodevelopment proceeds in a hierarchical manner, characterized by heterogeneous maturation patterns across brain regions and networks. However, the maturation of the intrinsic spatiotemporal propagations of brain activity remains largely unexplored. This study aims to bridge this gap by delineating spatiotemporal propagations from childhood to early adulthood. By leveraging a recently developed approach that captures time-lag dynamic propagations, we characterized intrinsic dynamic propagations along three axes: sensory-association (S-A), task-positive to default networks (TP-D), and somatomotor-visual (SM-V) networks, which progress towards adult-like brain dynamics from childhood to early adulthood. Importantly, we demonstrated that as participants mature, there is a prolonged occurrence of the S-A and TP-D propagation states, indicating that they spend more time in these states. Conversely, the prevalence of SM-V propagation states declines during development. Notably, top-down propagations along the S-A axis exhibited an age-dependent increase in occurrence, serving as a superior predictor of cognitive scores compared to bottom-up S-A propagation. These findings were replicated across two independent cohorts (N = 677 in total), emphasizing the robustness and generalizability of these findings. Our results provide new insights into the emergence of adult-like functional dynamics during youth and their role in supporting cognition.

neuroscience↗

bioRxiv2024

Brain Charts for the Rhesus Macaque Lifespan

Recent efforts to chart human brain growth across the lifespan using large-scale MRI data have provided reference standards for human brain development. However, similar models for nonhuman primate (NHP) growth are lacking. The rhesus macaque, a widely used NHP in translational neuroscience due to its similarities in brain anatomy, phylogenetics, cognitive, and social behaviors to humans, serves as an ideal NHP model. This study aimed to create normative growth charts for brain structure across the macaque lifespan, enhancing our understanding of neurodevelopment and aging, and facilitating cross-species translational research. Leveraging data from the PRIMatE Data Exchange (PRIME-DE) and other sources, we aggregated 1,522 MRI scans from 1,024 rhesus macaques. We mapped non-linear developmental trajectories for global and regional brain structural changes in volume, cortical thickness, and surface area over the lifespan. Our findings provided normative charts with centile scores for macaque brain structures and revealed key developmental milestones from prenatal stages to aging, highlighting both species-specific and comparable brain maturation patterns between macaques and humans. The charts offer a valuable resource for future NHP studies, particularly those with small sample sizes. Furthermore, the interactive open resource (https://interspeciesmap.childmind.org) supports cross-species comparisons to advance translational neuroscience research.

neuroscience↗

bioRxiv2024

A shifting role of thalamocortical connectivity in the emergence of large-scale functional brain organization across early lifespan development

While cortical patterning has been a perennial research topic in neuroscience, the mechanism for its consequence, namely functional specialization at the macro scale, remains an open question in the human brain. Here, we focused on age-dependent changes of resting-state thalamocortical connectivity to investigate its role in the emergence of large-scale functional networks across infancy, childhood and young adulthood. We found that the thalamocortical connectivity during infancy reflects an early differentiation of sensorimotor networks and genetically-influenced axonal projection. This initial role of the thalamus, however, seems to change during childhood, by establishing connectivity with the salience network and decoupling externally- and internally-oriented functional processes. Developmental simulation and perturbation analyses corroborated these findings, demonstrating the highest contribution of thalamic connectivity, especially in the later age of youth, in the formation of key characteristics of the mature brain, such as functional gradient and cortical hierarchy. Our study highlights a developmentally shifting role of the thalamus in orchestrating complex brain organization and its potential implications for developmental conditions characterized by compromised internal and external processing.

neuroscience↗

bioRxiv2022

Subgroups of eating behavior traits independent of obesity defined using functional connectivity and feature representation learning

AO_SCPLOWBSTRACTC_SCPLOWEating behavior is highly heterogeneous across individuals, and thus, it cannot be fully explained using only the degree of obesity. We utilized unsupervised machine learning and functional connectivity measures to explore the heterogeneity of eating behaviors. This study was conducted on 424 healthy adults. We generated low-dimensional representations of functional connectivity defined using the resting-state functional magnetic resonance imaging, and calculated latent features using the feature representation capabilities of an autoencoder by nonlinearly compressing the functional connectivity information. The clustering approaches applied to latent features identified three distinct subgroups. The subgroups exhibited different disinhibition and hunger traits; however, their body mass indices were comparable. The model interpretation technique of integrated gradients revealed that these distinctions were associated with the functional reorganization in higher-order associations and limbic networks and reward-related subcortical structures. The cognitive decoding analysis revealed that these systems are associated with reward- and emotion-related systems. We replicated our findings using an independent dataset, thereby suggesting generalizability. Our findings provide insights into the macroscopic brain organization of eating behavior-related subgroups independent of obesity.

neuroscience↗

bioRxiv2020

Body mass variations relate to fractionated functional brain hierarchies

AO_SCPLOWBSTRACTC_SCPLOWVariations in body mass index (BMI) have been suggested to relate to atypical brain organization, yet connectome-level substrates of BMI and their neurobiological underpinnings remain unclear. Studying 325 healthy young adults, we examined association between functional connectome organization and BMI variations. We capitalized on connectome manifold learning techniques, which represent macroscale functional connectivity patterns along continuous hierarchical axes that dissociate low level and higher order brain systems. We observed an increased differentiation between unimodal and heteromodal association networks in individuals with higher BMI, indicative of an increasingly segregated modular architecture and a disruption in the hierarchical integration of different brain system. Transcriptomic decoding and subsequent gene enrichment analyses identified genes previously implicated in genome-wide associations to BMI and specific cortical, striatal, and cerebellar cell types. These findings provide novel insights for functional connectome substrates of BMI variations in healthy young adults and point to potential molecular associations.

neuroscience↗