Biology subjects

Teodorescu, M.

Publications and source records attributed to Teodorescu, M..

At least 19 recordsLinked to original sources

bioRxiv2026

A foundation model for brain organoid network dynamics uncovers intrinsic computation and convergent disease phenotypes

Brain organoids develop self-organizing neural networks that can be used to model neurodevelopment and disease. Here we introduce PULSE, a foundation model developed using over 900 hours of high-density microelectrode array recordings of in vitro and ex vivo neural tissue. We show that this single model can address two open questions: how disease-associated variants alter network development, and how naive circuits encode information. PULSE revealed accelerated network maturation in held-out human 16p11.2-deletion organoids. It also revealed a change in network coordination shared with SHANK3 deficiency. Because organoids develop without sensory input, they let us ask how circuits encode information before any representation is learned. Closed-loop control of a virtual inverted pendulum engaged existing spontaneous activity patterns rather than creating new ones. PULSE provides a common reference frame for neural circuits across genotypes, protocols and tasks.

neuroscience↗

bioRxiv2026

Decoding brain anatomy from neuronal neighborhoods with MYCEL

Extracellular recordings carry information about neuronal location, and pooling per-unit predictions across neighboring electrodes improves its readout, but whether this local pooling should be learned is untested. We present MYCEL (Message-passing Yields Cellular Embeddings of Location), a graph neural network whose nodes are spike-sorted units carrying waveform and spike-timing features selected on cell-type ground truth, and whose edges encode recording geometry. On the Allen Brain Observatory Visual Coding Neuropixels dataset, learned message passing outperforms fixed aggregation, with the largest gains where single units are least informative. On the International Brain Laboratory Brain Wide Map, MYCEL matches the strongest population model at one-tenth the pairwise interactions. On held-out animals, the embeddings recover laminar depth within visual cortex. Retrained with only the edge geometry changed, the same framework decodes anatomy in macaque visual cortex and human medial temporal lobe. Thus, how neighboring neurons are combined matters as much as the features that describe them.

neuroscience↗

bioRxiv2026

Servo-Actuated 3D-Printed Disposable Microvalves for Automated, Scalable Organoid Culture in Standard Incubators

Automation of organoid and cell culture processes is essential for achieving scalable and standardized experimentation in regenerative medicine and stem cell research. However, existing microfluidic platforms often rely on complex setups, limiting their integration within standard incubator environments. To address these challenges, we developed a compact, scalable multi-well platform featuring 3D-printed, servo-actuated disposable microvalves for fully automated media and drug exchange. This design eliminates the need for external pressure sources and control channels, providing a simplified and cost-effective solution for organoid culture. The platform integrates an internet-connected microscopy module with a motorized XYZ stage, allowing continuous, real-time imaging of individual wells directly within the incubator. It supports precise and reliable fluid handling under physiological conditions, improving throughput, reproducibility, and accessibility. We validate the platform through bench-top testing and in both mouse and human organoid models. Morphological analysis, immunohistochemistry (IHC), and qPCR demonstrate comparable viability, growth, and gene expression profiles between automated and manual culture conditions. These results establish a robust and scalable framework for fully automated organoid culture, offering a simplified and accessible alternative to conventional microfluidic systems with broad applications in regenerative medicine, drug discovery, and scalable biological screening. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=101 SRC="FIGDIR/small/732526v1_ufig1.gif" ALT="Figure 1"> View larger version (42K): org.highwire.dtl.DTLVardef@10297f1org.highwire.dtl.DTLVardef@1458d48org.highwire.dtl.DTLVardef@13e7eeeorg.highwire.dtl.DTLVardef@1ff4e5c_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗

bioRxiv2026

Spatially defined axonal guidance in neural organoids with micropatterned microfluidic channels

Three-dimensional stem cell-derived neural organoids provide a promising platform for investigating early brain development and interregional circuit formation. Although co-culture of region-specific organoids into assembloids has enabled the study of cortical and subcortical interactions, these models lack directional specificity and spatial control, limiting their ability to recapitulate canonical circuit architecture. Here, we present a microfluidic platform for constructing directional and tunable interregional circuits while preserving anatomical distinction. This system, which we term "directoids" incorporates micropatterned polydimethylsiloxane (PDMS) microstructures to control uni- and bidirectional axonal growth between cortical and thalamic organoids. We observed a 70.4% success rate of axons traversing the full channel length in the permissive direction and reaching the opposing organoid, whereas no neurites successfully crossed the probative direction. These results demonstrate robust directionally bias in axon outgrowth and establish a scalable, reproducible strategy for controlling asymmetric connectivity between anatomically distinct neural organoids. Using high-density CMOS microelectrode arrays, we further validated directional tuning of extracellular action potential propagation within directoid microchannels, a feature not observed in straight-channel connectoid controls. Directoids also exhibited significant asymmetry in firing rates between channel entry and exit sites, consistent with engineered bias in signal flow. This provides an experimental paradigm for dissecting how anatomical connectivity and functional activity converge to shape neuronal networks. Together, these findings establish a microfluidic platform for investigating the mechanisms underlying hierarchical circuit formation, regional specification, and functional integration in developing human neural organoid models at cellular resolution not possible in vivo.

bioengineering↗

bioRxiv2026

SpikeLab: Agentic tools for spike data analysis

Large language models have the potential to transform scientific research and analysis, but without domain-specific structure they produce silent methodological errors, unreported decisions, and irreproducible results. Here we present SpikeLab, a text-to-analysis framework for neural spike data that combines composable data structures with a skill-based agentic system enforcing bounded autonomy: mandatory use of expert-vetted methods, correctness over efficiency, and clarification-seeking on ambiguous requests. In a controlled benchmark on electrophysiology data, Sonnet 4.6 with SpikeLab produced correct and reproducible results across all tasks, outperforming both the unassisted Sonnet and the more capable Opus 4.6, which exhibited deterministic failures including ad hoc method invention, silent data reduction, and inconsistent experimental designs. We demonstrate versatility across in vivo mouse, human, and in vitro brain organoid recordings, and apply the framework to a pharmacological dose-response study spanning single-unit dynamics, pairwise network structure, burst-level temporal sequences, and latent population states, all through natural language prompts without writing analysis code.

neuroscience↗

bioRxiv2026

Microfluidic Control of Dorsal-Ventral Patterning Within a Single Forebrain Organoid

How distinct regional identities emerge within a single developing brain remains poorly understood. Current in vitro models address this by fusing independently generated organoids, but this introduces variability in size, maturation state, and connectivity, confounding the study of regionalization itself. Here, we present a microfluidic platform that supports the co-development of different tissue identities within a single, continuous 3D culture domain. The device integrates controlled microfluidic flow with real-time fluorescence imaging, providing stable perfusion and high-resolution tracking of molecular transport without the need for embedded sensors or disruptive sampling. By delivering SAG, a Sonic hedgehog pathway agonist, to one surface of mouse forebrain organoids, we induced spatially segregated ventral (Nkx2.1+) and dorsal (Pax6+) domains within a unified tissue architecture. Controlled morphogen delivery is sufficient to drive region-specific fate specification without organoid fusion, offering a practical, scalable alternative for studying tissue regionalization in vitro.

bioengineering↗

bioRxiv2026

A Modular In-Incubator Microscope for Longitudinal Live Cell Microscopy

Longitudinal live cell imaging is valuable for characterizing dynamic morphological and phenotypic changes in biological systems. However, conventional approaches rely on manual microscope operation, which is labor-intensive, limits imaging frequency, and disrupts the cellular environment. These constraints reduce scalability, increase experimental variability, and restrict both the duration and temporal resolution of continuous imaging. Although automated imaging platforms partially address these limitations, existing solutions are often constrained by the cost, footprint, and inflexibility of in-incubator microscopes or stage-top incubators. Here, we present an automated in-incubator epifluorescence microscope designed for long-term operation. The system features a modular architecture with optional multi-fluorescence imaging, automated plate scanning, configurable light sources, and compatibility with multiple plate formats, including integration with fluidic automation devices. By positioning the light sources and control electronics outside the incubator, the platform improves thermal stability and long-term operational reliability. This approach enables continuous, high-frequency imaging over extended durations, providing a source of rich data for quantifying time-dependent tissue phenotypes, morphological remodeling, and transient biological processes.

bioengineering↗

bioRxiv2026

Cloud-Connected Pluripotent Stem Cell Platform Enhances Scientific Identity in Underrepresented Students

Stem cell research offers unique opportunities for authentic scientific engagement, yet infrastructure requirements have confined participation to elite institutions, perpetuating workforce disparities. We developed an integrated framework combining engineered biology, cloud-connected microscopy, and validated psychometric assessment to make pluripotent stem cell (PSC) experimentation widely accessible. The framework comprises three components: a doxycycline-inducible NGN2 mouse embryonic stem cell line for rapid neuronal specification, low-cost cloud microscopy for remote observation, and the validated Stem Cell Research Identity Scale (SCRIS) for quantifying educational outcomes. Implementation across a Title I high school and urban community college demonstrated significant increases in scientific identity. Students using differentiating PSCs showed broader science identity development than those using neuroblastoma cells, particularly in competence, research readiness, and recognition. High school students showed enhanced research competence gains compared to community college students despite equivalent intervention duration. Demographic analyses revealed enhanced effectiveness for Hispanic and first-generation college students. This framework provides a scalable model for broadening participation in advanced biomedical research.

scientific communication and education↗

bioRxiv2026

Scalable high-fidelity human vascularized cortical assembloids recapitulate neurovascular co-development and cell specialization

Human cortical development involves the coupling of neurogenesis and cerebrovascular growth. However, interactions between neural and vascular cells are largely missing in most brain organoids, which are crucial models for studying neurodevelopment and disease. Here, we establish vascularized cortical assembloids (vCAs) by fusing mesoderm-derived vascular organoids (VOs) with cortical organoids (COs). vCAs self-assemble lumenized networks of endothelial cells, pericytes, and perivascular fibroblasts that acquire blood-brain barrier (BBB) specialization and arteriovenous specification in vitro. Single-cell RNA-sequencing and immunofluorescence imaging revealed that vascularization improves neuroepithelial architecture, reduces hypoxia and apoptosis, expands cortical progenitor pools, and enhances neuronal maturation and connectivity compared with COs. Moreover, atlas-level integration with human neurodevelopmental tissue and cross-protocol benchmarking demonstrate superior transcriptional concordance, especially for vascular cell and glial populations, relative to existing approaches. This scalable and reproducible platform improves fidelity and throughput for modeling human neurovascular co-development and enables systematic studies across brain regions and diseases using engineered or patient-derived iPSCs.

developmental biology↗

bioRxiv2025

Incubator-Free Organoid Culture in a Sealed Recirculatory System

Discovery in human biology is pivoting toward high-dimensional computational analysis of 3D in vitro models, but this progress is limited by reliance on conventional cell culture techniques. Realism and data collection are hindered by the environmental instabilities and accessibility constraints of standard incubators. We introduce an automated, sealed recirculatory system that eliminates these barriers, enabling unconstrained instrument integration and infrastructure-independent scalability. By employing gas-tight sealing, a liquid-phase gas buffer and a non-porous plastic gas exchanger, our technology maintains biological stability without the compromises of open-air vessels. This design eliminates the need for CO2 incubators and prevents the evaporative drift that typically plagues conventional open-culture vessels. Operating on the benchtop outside the cell culture suite, we demonstrate that our system supports continuous, multi-week live imaging of vascular organoids while maintaining metabolic viability, structural fidelity and electrophysiological activity in brain organoids comparable to traditional in-incubator cultures. TeaserA sealed benchtop device enables data-rich organoid culture beyond the constraints of the cell culture laboratory.

bioengineering↗

bioRxiv2025

Self-Organizing Neural Networks in Organoids Reveal Principles of Forebrain Circuit Assembly

The mouse cortex is a canonical model for studying how functional neural networks emerge, yet it remains unclear which topological features arise from intrinsic cellular organization versus external regional cues. Mouse forebrain organoids provide a powerful system to investigate these intrinsic mechanisms. We generated dorsal (DF) and ventral (VF) forebrain organoids from mouse pluripotent stem cells and tracked their development using longitudinal electrophysiology. DF organoids showed progressively stronger network-wide correlations, while VF organoids developed more refined activity patterns, enhanced small-world topology, and increased modular organization. These differences emerged without extrinsic inputs and may be driven by the increased generation of Pvalb+ interneurons in VF organoids. Our findings demonstrate how variations in cellular composition influence the self-organization of neural circuits, establishing mouse forebrain organoids as a tractable platform to study how neuronal populations shape cortical network architecture.

neuroscience↗

bioRxiv2025

HIPPIE: A Multimodal Deep Learning Model for Electrophysiological Classification of Neurons

Extracellular electrophysiological recordings present unique computational challenges for neuronal classification due to noise, technical variability, and batch effects across experimental systems. We introduce HIPPIE (High-dimensional Interpretation of Physiological Patterns In Extracellular recordings), a deep learning framework that combines self-supervised pretraining on unlabeled datasets with supervised fine-tuning to classify neurons from extracellular recordings. Using conditional convolutional joint autoencoders, HIPPIE learns robust, technology-adjusted representations of waveforms and spiking dynamics. This model can be applied to electrophysiological classification and clustering across diverse biological cultures and technologies. We validated HIPPIE on both in vivo mouse recordings and in vitro brain slices, where it demonstrated superior performance over other unsupervised methods in cell-type discrimination and aligned closely with anatomically defined classes. Its latent space organizes neurons along electrophysiological gradients, while enabling batch and individual corrected alignment of recordings across experiments. HIPPIE establishes a general framework for systematically decoding neuronal diversity in native and engineered systems.

neuroscience↗

bioRxiv2025

Fluoxetine Delivery for Wound Treatment Through an Integrated Bioelectronic Device - Pharmacokinetic Parameters and Safety Profile in Swine

Wound infections are a significant medical challenge, often leading to chronicity or systemic infection. Selective serotonin reuptake inhibitors (SSRIs) have emerged as potential non-antibiotic candidates with demonstrated ability to limit growth and biofilm formation in Gram-negative bacteria, in addition to their pro-healing activity. Here, we compared direct delivery of the SSRI fluoxetine by topical bolus dosing to delivery from an iontophoresis bandage device with an actuator for temporally controlled drug delivery, in a porcine excisional wound model. Device delivery of fluoxetine resulted in a maximum concentration of 12.25 ng fluoxetine per mg tissue, compared to 2.926 ng/mg following bolus dosing, and tissue fluoxetine levels were higher after application using the device than after bolus dosing across the range of doses tested (p=0.0041). The half-life of fluoxetine in the wound tissue was 0.988 {+/-} 0.256 days. Fluoxetine was not detected in the pig plasma, and plasma serotonin levels were not affected by the topical application. Fluoxetine delivery using the device, but not bolus delivery, produced tissue concentrations above the minimum inhibitory concentration (MIC) for some clinically important species of bacteria. The experimental device can effectively deliver topical fluoxetine to the wound, producing higher tissue concentrations of fluoxetine at lower cumulative doses compared to bolus dosing, and with minimal risk of off-target effects. The device may simplify wound treatment by reducing the burden for daily drug application, possibly increasing adherence to a prescribed treatment regimen.

bioengineering↗

bioRxiv2024

Calcium dynamics tune developmental tempo to generate evolutionarily divergent axon tract lengths

The human brain has undergone an evolutionary expansion in size, both in terms of cell numbers and the size of cellular structures, including axon tracts. Human brain development also progresses slowly and takes particularly long. However, the functional relevance of slowed timing, and whether it is responsible for these changes in size, remains unknown. Here, we investigate this by studying axon tract development in human and mouse brain organoids. We demonstrate that human axon tracts grow [~]2x more slowly than those of mice, reflecting their slowed tempo, but that this actually leads to shorter human axons, not longer. To overcome the effect of slowed tempo, human axons have a more prolonged growth duration that enables them to project farther despite their slower growth rate. Using a combination of single-cell RNA sequencing and live imaging, we demonstrate that the prolonged duration involves a different mechanism to that controlling tempo and is driven by calcium dynamics. Human axons exhibit a reduced calcium influx compared to mouse, mediated by L-Type voltage-gated calcium channels. Stimulating this calcium influx in human neurons triggers earlier cessation of growth, leading to shorter axon tracts similar to those of mouse. We further show that increasing calcium speeds up the transition to the synaptogenesis stage. Thus, calcium regulation sets the timing of transitions to disproportionately extend developmental duration, thereby enabling evolutionary expansion of human neurons.

developmental biology↗

bioRxiv2024

Design and Validation of a Wearable Imaging System for Automated Wound Monitoring in Porcine Model

Effective wound monitoring has the potential to guide treatment regiments and improve healing outcomes, yet current clinical assessment methods remain largely subjective and labor-intensive. To address this challenge, we present a high-resolution wearable imaging system designed for continuous wound monitoring. The system integrates a 64 MP camera with a plano-convex lens housed in an enclosure measuring 73 mm in diameter and 36.1 mm in height, and features a custom printed circuit board (PCB) for programmable LED illumination. The 3D-printed device enclosure is designed to accommodate a silicone bioelectronic device and can be securely attached using a commercially available ostomy skin barrier. In porcine model validation studies, the system successfully captured daily wound progression over periods up to 7 days. The captured images were wirelessly transmitted to a processing unit where DeepMapper, a machine learning algorithm, processed z-stacked images and performed multi-level feature extraction to predict wound healing stages and indicate potential complications such as infection. This imaging system enables automated analysis of wound progression and supports the development of smart wound care platforms for personalized treatment strategies. The integrated design approach demonstrates the feasibility of creating compact, high-resolution imaging systems suitable for clinical wound monitoring applications.

bioengineering↗

bioRxiv2024

Accelerating Wound Healing Through Deep Reinforcement Learning: A Data-Driven Approach to Optimal Treatment

Advancements in bioelectronic sensors and actuators have paved the way for real-time monitoring and control of wound healing progression. Real-time monitoring allows for precise adjustment in treatment strategies that align with an individuals unique biological response. However, due to the complexities of human-drug interactions and a lack of predictive models it is challenging to determine just how one should adjust drug dosage to achieve the desired biological response. This work proposes an adaptive closed-loop control framework that integrates deep learning, optimal control, and reinforcement learning to update treatment strategies in real-time with the goal of accelerating wound closure. The proposed approach eliminates the need for mathematical modeling of complex nonlinear wound healing dynamics. We demonstrate the convergence of the controller via an in silico experimental setup, where the proposed approach successfully accelerates the wound healing process by 17.71%. Finally, we share the experimental setup and results of an in vivo implementation to highlight the translational potential of our work. Our data-driven model estimates a 40% acceleration in wound closure.

systems biology↗

bioRxiv2024

Remote-Controlled Wireless Bioelectronics for Fluoxetine Therapy to Promote Wound Healing in a Porcine Model

Wound healing presents a significant challenge in biomedical science, requiring precise therapeutic delivery and real-time monitoring. Bioelectronic systems offer a promising solution but remain largely unexplored for wound care, particularly in large animal models that reflect human healing dynamics. This study introduces a remote controlled wireless bioelectronic platform equipped with an iontophoretic pump to deliver fluoxetine, a selective serotonin reuptake inhibitor that promotes wound repair. In vitro and ex-vivo testing validated efficient on demand fluoxetine delivery. In vivo experiments in a porcine wound model demonstrated clear therapeutic efficacy over 3-day and 7-day periods. The system enhanced healing outcomes, increasing re-epithelialization by 37% (H&E staining), reducing the M1/M2 macrophage ratio by 33%, and stimulating neuronal growth at the wound site. This bioelectronic platform delivers fluoxetine in a controlled, remotely-controlled manner while allowing for wound direct wound imaging that can be used to monitor wound healing progress. Additionally, it allows precise dose and temporal delivery of treatment to enhance the outcome of future large animal wound healing studies.

bioengineering↗

bioRxiv2024

Wireless bioelectronic device for wound healing

Wireless bioelectronic actuators have been developed to deliver targeted treatments over multiple days while continuously monitoring delivery, thereby improving wound healing. Specifically, these devices can deliver charged biomolecules such as fluoxetine cations (Flx+) and electric field (EF) in freely moving pigs. Treatments can be controlled and monitored in real time via WiFi, with options for both user-specified delivery rates and durations, as well as automated closed-loop (CL) control. The devices are engineered to handle various failure scenarios that may arise in dynamic, real-world experiments--such as communication or power interruptions--ensuring that valuable experimental data is collected with minimal disruption. The ion pump features eight drug reservoirs and channels arranged around a 20 mm diameter-wound, with a central ground electrode (0 V). When voltages above 0 V are applied to the outer channels, currents flow from the reservoirs and channels into the wound, delivering Flx+ and/or EF depending on the reservoir solution. The device records applied voltages and currents locally to a microSD card at a high sampling rate, while simultaneously transmitting real-time measurements via a local WiFi network to a wound healing algorithm running on a nearby laptop. CL control of current/delivery rate is performed by an onboard microcontroller unit (MCU) and current-source microchips, based on instructions received from the wound healing algorithm. A graphical user interface (GUI) provides intuitive user control and real-time data visualization, with support for multiple devices. In vivo studies over seven days showed that Flx+-treated wounds had a 20% lower M1/M2 macrophage ratio and 41.67% greater re-epithelialization compared to controls (standard-of-care), demonstrating the actuators potential to enhance wound healing.

biochemistry↗