Tuesday, Jun 25: 12:00 PM - 1:15 PM
Oral Sessions
COEX
Room: ASEM Ballroom 202
Presentations
The brain is a network of functionally interacting neural populations. Studying the functional architecture of the brain in awake humans is possible with multiple imaging technologies, although these technologies are often biased towards the cortex where signal quality is highest [1]. Perhaps the biggest missing piece of modern in vivo brain network reconstruction is the brainstem. This early evolutionary structure is crucial for survival and consciousness, and integrates signals from throughout the nervous system. Furthermore, multiple neurotransmitter systems originate in brainstem nuclei and project throughout the cortex, shaping cortical activity [2]. However, knowledge about brainstem function primarily comes from either lesion studies or studies in model organisms, and these studies are often limited to specific nuclei or pathways. Exciting recent imaging advances have improved the feasibility of measuring brainstem activity, making it now possible to augment the cortical functional connectome with an anatomically comprehensive representation of the brainstem [3-5].
Recent studies have suggested that naturalistic stimuli, such as movie clips, outperform rest and conventional tasks in phenotype prediction [1, 2]. Despite their promise, the impact of stimulus selection on phenotype prediction remains largely unclear. Most existing datasets of naturalistic conditions lack sufficient justifications for selecting specific stimuli, and many studies so far have used one single stimulus. Here, we investigate the impact of stimulus selection on phenotype prediction from two aspects, namely brain states (inter-subject synchrony) [3] and stimulus features. We focus on the paradigmatic case of sex classification due to the robust and well-established nature of the brain-sex relationship.
Presenter
Xuan Li, Research Centre Juelich
INM-7
Jülich, Select a State
Germany
Despite the well-known fact that environmental exposures play a critical role in influencing behaviors [1], we have limited understanding of how these exposures interact with the brain and in turn shape our behaviors, especially during adolescence with rapid development of brain and behaviors [2].
Presenter
Dongmei Zhi, Beijing Normal University Beijing, Select a State
China
Morphometricity is the proportion of phenotypic variation that can be explained by macroscopic brain morphology. It is estimated in a manner similar to heritability, with intersubject similarity of brain images replacing genetic relatedness [4]. It provides a simple approach to summarize the link between a phenotype and high-dimensional brain data with a single value. However, recent results have found unexpectedly large morphometricity values, e.g. brain structure explaining over 90% variation in BMI [2]. In this work we explore the role of smoothness in morphometricity in theory, simulation and real data evaluations, showing that image smoothness induces a positive bias that can help explain these unusual results.
Presenter
Nicolas Salvy, University of Oxford
Nuffield Department of Population Health
Oxford
United Kingdom
Explainable machine learning of complex multimodal data in neuroscience research is revolutionizing precision neuropsychiatry. Interpretable clustering of patients into distinct subtypes can enhance personalized prognosis, diagnosis, and treatment. However, training on biomedical data poses challenges due to high dimensionality, clustering, and limited sample size. To address this, we propose a scalable approach for cluster-aware embedding, incorporating a convex clustering penalty. This approach facilitates hierarchical clustering of principal component analysis (PCA), locally linear embedding (LLE), and canonical correlation analysis (CCA). Our method improves upon existing techniques and offers a modular framework for interpretable biomarker discovery in precision medicine. We apply this approach to identify neurocognitive subtypes in the Adolescent Brain Cognitive Development (ABCD) and Autism Brain Imaging Data Exchange (ABIDE) datasets.
Presenter
Amanda Buch, Weill Cornell Medicine, Cornell University New York, NY
United States
Understanding individual differences in brain-behavior relationships is a central goal of neuroscience. As such, machine learning approaches using neuroimaging data, such as functional connectivity, have grown increasingly popular in predicting numerous phenotypes. The reproducibility of such studies is hindered by data leakage, where information about the test data is introduced into the model during training (1). Although leakage is never a correct practice, quantifying the effects of leakage in neuroimaging data is important due to its pervasiveness. Here, we evaluate the effects of leakage on functional connectome-based machine learning in four large datasets for the prediction of three phenotypes.