Biology subjects

Truong, N.

Publications and source records attributed to Truong, N..

6 recordsLinked to original sources

bioRxiv2026

Reassessing Number-Detector Units in Convolutional Neural Networks

Convolutional neural networks (CNNs) have become essential models for predicting neural activity and behavior in visual tasks. However, their ability to capture higher-level cognitive functions, such as numerosity discrimination, remains debated. Numerosity, the ability to perceive and estimate the number of items in a visual scene, is often proposed to rely on specialized number-detector units within CNNs, analogous to number-selective neurons observed in the brain. In this study, we use CORnet, a biologically inspired CNN architecture inspired by the organization of the primate visual system. To address a limitation of classical Representational Similarity Analysis (RSA)--its assumption that all units contribute equally--we apply pruning, a feature selection approach that identifies the units most relevant for explaining behavioral similarity structure. Our results show that number-detector units are not critical for population-level representations of numerosity, challenging their proposed role in previous studies.

neuroscience↗

bioRxiv2025

Inflammatory reprogramming of human brain endothelial cells compromises blood-brain barrier integrity in Alzheimer's disease

Blood-brain barrier (BBB) dysfunction is an early feature of Alzheimers disease (AD), yet the endothelial gene-regulatory programs involved remain incompletely understood. We integrate postmortem human single-nucleus transcriptomics with iPSC-based BBB models to define a conserved, inflammation-driven pathway that compromises barrier integrity. We identify an NF-{kappa}B-associated endothelial gene module endoM2 that is elevated in AD, inversely correlated with cognition, and enriched for inflammation and endothelial-to-mesenchymal transition signatures. Cytokine stimulation of iPSC-derived brain endothelial cells induces morphological remodeling, lipid accumulation, junctional disruption, and transcriptomic shifts that mirror endoM2. A targeted drug screen identifies the NF-{kappa}B inhibitor BAY11-7082 as protective against cytokine-induced changes. In our perfusable iPSC-derived BBB-Chip that recapitulates human BBB signatures, single-cell profiling reveals inflammatory endothelial state-specific programs reflecting those in AD brains and demonstrates that BAY11-7082 suppresses cytokine-triggered dysfunction and reverses inflammation-associated gene activation. Together, these findings position cerebrovascular inflammation as a therapeutic target to preserve BBB integrity in AD.

neuroscience↗

bioRxiv2025

Vascular-Perfusable Human 3D Brain-on-Chip

Development and delivery of treatments for neurological diseases are limited by the tight and selective human blood-brain barrier (BBB). Although animal models have been important research and preclinical tools, the rodent BBB exhibits species differences and fails to capture the complexity of human genetics. Microphysiological systems incorporating human-derived cells hold great potential for modeling disease and therapeutic development, with advantages in screening throughput, real-time monitoring, and tunable genetic backgrounds when combined with induced pluripotent stem cell (iPSC) technology. Existing 3D BBB-on-chip systems have incorporated iPSC-derived endothelial cells but not the other major brain cell types from iPSCs, each of which contributes to brain physiology and disease. Here we developed a 3D Brain-Chip system incorporating endothelial cells, pericytes, astrocytes, neurons, microglia, and oligodendroglia from iPSCs. To enable this multicellular 3D co-culture in-chip, we designed a GelChip microfluidic platform using a 3D printing-based approach and dextran-based engineered hydrogel. Leveraging this platform, we co-cultured and characterized iPSC-derived brain-on-chips and modeled the brain microvasculature of APOE4, the strongest known genetic risk factor for sporadic Alzheimers disease. These 3D brain-on-chips provide a versatile system to assess BBB vascular morphology and function, investigate downstream neurological effects in disease, and screen therapeutics to optimize delivery to the brain. Significance StatementThe blood-brain barrier (BBB) is both a contributing factor to neurological disease and a major obstacle to its treatment, yet human-relevant models remain limited. Most existing brain-on-chip systems incorporate only subsets of BBB cell types and cannot capture the full cellular complexity of the human neurovascular unit. Here, we establish a vascular-perfusable 3D Brain-Chip using human induced pluripotent stem cell-derived brain cells including endothelial cells, pericytes, astrocytes, neurons, microglia, and oligodendroglia. This system enables systematic analysis of human genetic risk factors, such as APOE4 in Alzheimers disease, and provides a powerful platform to investigate BBB function and dysfunction and accelerate the development of more effective neurological therapies.

neuroscience↗

bioRxiv2025

Investigating action topography in visual cortex and deep artificial neural networks

High-level visual cortex contains category-selective areas embedded within larger-scale topographic maps like animacy and real-world size. Here, we propose action as a key organizing factor shaping visual cortex topography and assess the ability of topographic deep artificial neural networks (DANNs) in capturing this organization. Using fMRI, we examined responses to images of body-parts and objects with different degrees of action properties. In left lateral occipitotemporal cortex, we identified a topographically-organized action gradient, with overlapping activations for bodies, hands, tools, and manipulable objects along a dorsal-posterior to ventral-anterior axis, culminating at the intersection of body parts and objects exhibiting higher action properties. Multivariate analyses confirmed action as a crucial organizing principle, while shape and animacy dominated ventral occipitotemporal cortex and DANNs, which exhibited no action-based organization. Our proposed action dimension serves as a further organizing principle of object categories, advancing understanding of visual cortex organization and its divergence from DANN-based models.

neuroscience↗

bioRxiv2023

Engineered 3D Immuno-Glial-Neurovascular Human Brain Model

Patient-specific, human-based cellular models integrating a biomimetic blood-brain barrier (BBB), immune, and myelinated neuron components are critically needed to enable accelerated, translationally relevant discovery of neurological disease mechanisms and interventions. By engineering a novel brain-mimicking 3D hydrogel and co-culturing all six major brain cell types derived from patient iPSCs, we have constructed, characterized, and utilized a multicellular integrated brain (miBrain) immuno-glial-neurovascular model with in vivo-like hallmarks inclusive of neuronal activity, functional connectivity, barrier function, myelin-producing oligodendrocyte engagement with neurons, multicellular interactions, and transcriptomic profiles. We implemented the model to study Alzheimers Disease pathologies associated with APOE4 genetic risk. APOE4 miBrains differentially exhibit amyloid aggregation, tau phosphorylation, and astrocytic GFAP. Unlike the co-emergent fate specification of glia and neurons in organoids, miBrains integrate independently differentiated cell types, a feature we harnessed to identify that APOE4 in astrocytes promotes neuronal tau pathogenesis and dysregulation through crosstalk with microglia.

neuroscience↗

bioRxiv2022

Lipid-polymer hybrid nanoparticles utilize B cells and dendritic cells to elicit distinct antigen-specific CD4+ and CD8+ T cell responses

Antigen presenting cells (APCs) have been extensively studied for treating cancers and autoimmune diseases. Dendritic cells (DCs) are potent APCs that uptake and present antigens (Ags) to activate immunity or tolerance. Despite their active use in cellular immunotherapies, DCs face several challenges that hinder clinical translation, such as inability to control Ag dosing for tuning immune responses and low abundance in peripheral blood. B cells are a potential alternative to DCs, but their poor non-specific Ag uptake capabilities compromise controllable priming of T cells. We developed phospholipid-conjugated Ags (L-Ags) and lipid-polymer hybrid nanoparticles (L/P-Ag NPs) as Ag delivery platforms to expand the range of accessible APCs for use in priming CD4+ and CD8+ T cells. These delivery platforms were evaluated using DCs, CD40-activated B cells, and resting B cells as a diverse set of APCs to understand the impact of various Ag delivery mechanisms for generation of Ag-specific T cell responses. L-Ag delivery (termed depoting) of MHC class I and II-restricted Ags successfully loaded all APC types in a tunable manner and primed both Ag-specific CD8+ and CD4+ T cells, respectively. Incorporating L-Ags and polymer-conjugated Ags (P-Ag) into NPs can direct Ags to different uptake pathways to engineer the dynamics of presentation and shape T cell responses. DCs were capable of processing and presenting Ag delivered from both L- and P-Ag NPs yet B cells could only utilize Ag delivered from L-Ag NPs. Multivariate analysis of cytokines secreted from APC:T cell co-cultures indicated that L-Ag NPs primed different T cell responses than P-Ag NPs. Altogether, we show that L-Ags and P-Ags can be rationally paired within a single NP to leverage distinct delivery mechanisms to access multiple Ag processing pathways in two APC types, offering a modular delivery platform for engineering immunotherapies.

bioengineering↗