Biology subjects

Ventura, J.

Publications and source records attributed to Ventura, J..

10 recordsLinked to original sources

bioRxiv2026

Low-latency neuromorphic closed-loop control of hippocampal ripples in vivo

Real-time closed-loop neuromodulation, in which stimulation is precisely timed to ongoing brain dynamics, holds transformative potential for treating neurological disorders and probing neural circuit function. However, it requires low-latency, energy-efficient processing of high-bandwidth neural signals that conventional computing architectures struggle to deliver. Neuromorphic computing, which emulates the event-driven and massively parallel operation of biological neural circuits, offers a compelling alternative. Yet, its integration into closed-loop frameworks validated in vivo for fast, transient oscillations has not been demonstrated. Here, we present a fully integrated neuromorphic framework for real-time detection and manipulation of hippocampal ripples: brief (30-100 ms), high-frequency (100-250 Hz) oscillations that are critical for memory consolidation and implicated in neurological disorders. We train compact spiking neural networks comprising 41 neurons and 530 parameters using surrogate-gradient backpropagation, achieving detection performance competitive with deep learning models across 23 recording sessions while consuming up to 200-fold less energy when deployed on SpiNNaker neuromorphic hardware. Integration with the open-source Open Ephys platform yields total closed-loop latencies of approximately 50 ms, enabling intra-event stimulation in up to 80% of ripples. Validating the complete sensing-processing-stimulation pipeline in awake, head-fixed mice, we demonstrate that neuromorphic-triggered optogenetic inhibition significantly alters ripple dynamics and reduces oscillatory energy. This work establishes a practical and accessible neuromorphic framework for low-latency closed-loop control of fast brain dynamics in vivo.

neuroscience↗

bioRxiv2026

Mapping California's Urban Forest at Scale: An Error-Adjusted Canopy Time Series for Monitoring Change

Statewide tracking of urban tree canopy change is essential for evaluating progress toward policy targets, but detecting real change requires both high-resolution mapping and rigorous uncertainty estimation. We produced a four-year canopy cover time series for all California census-designated places using 60-cm NAIP aerial imagery and a U-Net deep learning model trained with semi-automated LiDAR-derived labels and manually annotated tiles. Canopy cover and change were estimated using stratified, error-adjusted area estimation, enabling comparisons across years. Statewide canopy cover showed a modest negative trend from 2016 to 2022 (Sens slope: -0.60% per year), but confidence intervals included zero across all groups and climate zones, indicating that trends were not statistically distinguishable from no change. Urban canopy cover was consistently lower than non-urban canopy by approximately six percentage points, and canopy cover was highest in the Northern California Coast and lowest in the Southwest Desert. Residential parcels accounted for 55-56% of canopy within incorporated urban areas across all years, indicating that statewide canopy increase goals will require engagement with private landowners. Error adjustment substantially altered canopy estimates relative to raw pixel-count totals, with direct implications for AB 2251 canopy tracking where baselines and targets drawn from unadjusted maps may not reflect true canopy extent. This open-source workflow is transferable to future NAIP acquisition years and other U.S. states, providing a scalable framework for long-term urban forest monitoring.

ecology↗

bioRxiv2026

Comparison of Deep Learning Approaches for Extreme Low-SNR Image Restoration

BackgroundLive-cell fluorescence microscopy enables the study of dynamic cellular processes. However, fluorescence microscopy can damage cells and disrupt these dynamic processes through photobleaching and phototoxicity. Reducing light exposure mitigates the effects of photobleaching and phototoxicity but results in low signal-to-noise ratio (SNR) images. Deep learning provides a solution for restoring these low-SNR images. However, these deep learning methods require large, representative datasets for training, testing, and benchmarking, as well as substantial GPU memory, particularly for denoising large images. ResultsWe present a new fluorescence microscopy dataset designed to expand the range of imaging conditions and specimens currently available for evaluating denoising methods. The dataset contains 324 paired high/low-SNR images ranging from four to 282 megapixels across 12 sub-datasets that vary in specimen, objective used, staining type, excitation wavelength, and exposure time. The dataset also includes spinning disk confocal microscopy examples and extreme-noise cases. We evaluated three state-of-the-art deep learning denoising models on the dataset: a supervised transformer-based model, a supervised CNN model, and an unsupervised single image model. We also developed an image stitching method that enables large images to be processed in smaller crops and reconstructed. ConclusionsOur dataset provides a diverse benchmark for evaluating deep learning denoising methods, and our stitching method provides a solution to GPU memory constraints encountered when processing large images. Among the evaluated deep learning models, the supervised transformer-based model had the highest denoising performance but required the longest training time.

bioinformatics↗

bioRxiv2024

Stress-actuated Flexible Microelectrode Arrays for Activity Recording in 3D Neuronal Cultures

Microelectrode arrays (MEAs) are instrumental in monitoring electrogenic cell populations, such as neuronal cultures, allowing high precision measurements of electrical activity. Although three-dimensional neuronal cultures replicate the behavior of in vivo systems better than two-dimensional models, conventional planar MEAs are not well suited to capture activity within such networks. Novel MEA geometries can overcome this difficulty, but often at the cost of increased fabrication complexity. Here, we used the stress mismatch between thin film layers to fabricate MEAs with vertical electrodes, using methods compatible with established microfabrication protocols. A micrometric SiO2 hinge enables control over the bending angle of flexible polyimide structures with embedded electrodes. The performance of the patterned electrodes was assessed before and after stress actuation, through impedance measurements, voltage noise mapping, and neuronal activity recordings. 3D MEAs with 30x30 {micro}m2 electrodes showed an impedance of 0.96 {+/-} 0.07 M{Omega} per electrode and detected neuronal activity spikes with amplitudes as high as 400 {micro}V. These results demonstrate the potential of the developed methods to provide a scalable approach to fabricate 3D MEAs, enabling enhanced recording capabilities for in vitro neuronal cultures.

bioengineering↗

bioRxiv2024

Visually Induced Involuntary Movements

Looking at a virtual 3D environment with structural features rotating at 60{degrees}/s in a head-mounted display soon elicits an illusion of self-rotation and displacement in the opposite direction. We explored in 75 s long trials the effects of visually induced self-rotation on the head, torso, and horizontally extended right arm of standing subjects. The degree of body and limb movement was contingent on whether the arm was extended out freely or pointing at a briefly proprioceptively specified target position, but did not depend on whether the hand held a rod or not. Most subjects in the Free condition showed significant unintentional arm deviations, which averaged approximately 55{degrees} in the direction opposite the induced illusory self-motion, and were more than 150{degrees} in some cases. In contrast, on average, the deviations in the Pointing condition were a quarter of those in the Free condition. Deviations of head and torso positions also occurred in all conditions. Total arm and head deviations were the sum of deviations of the arm and head with respect to the torso plus deviations of the torso with respect to space. When given a pointing target, subjects were largely able to detect and correct for arm and head deviations with respect to the torso but not for the parts of arm and head deviation that were due to deviations of the torso with respect to space. In all conditions, the arm, head, and torso deviations occurred before subjects began to experience compelling self-rotation and displacement. This is contrasted with the compensations for expected but absent Coriolis forces that are made when stationary subjects make reaching movements to targets during exposure to structured moving visual scenes. These compensations do not occur until subjects experience self-rotation and spatial displacement. These results have implications for vehicle control and maneuvering in environments that induce illusory motion and displacement, and in situations where there is motion in a large area of the visual field. The impact of these effects on joystick control is described and discussed. We also describe the subjective sense of ownership attributed to hand-held objects when experiencing illusory self-motion and displacement.

neuroscience↗

bioRxiv2022

Rapid Loss of CD4 T Cells by Pyroptosis During Acute SIV Infection in Rhesus Macaques

The mechanisms underlying depletion of CD4 T cells during acute HIV-1 infection are not well understood. Here we show that caspase-1-induced pyroptosis, a highly inflammatory programmed cell death pathway, is the dominant mechanism responsible for the rapid depletion of CD4 T cells in gut-associated lymphatic tissue (GALT), spleen, and lymph nodes during acute simian immunodeficiency virus (SIV) infection in rhesus macaques. Upregulation of interferon-gamma inducible factor 16 (IFI16), a host DNA sensor that triggers pyroptosis, was also observed in tissue-resident CD4 T cells and correlated with viral loads and CD4 T cell loss. In contrast, caspase-3-mediated apoptosis and viral cytotoxicity only accounted for a small fraction of CD4 T cell death. Other programmed cell death mechanisms, including mitochondria-induced caspase-independent cell death, necroptosis, and autophagy, did not significantly contribute to CD4 T cell depletion. These data support a model in which caspase-1-mediated pyroptosis is the principal mechanism that results in CD4 T cell loss in the GALT and lymphoid organs and release of proinflammatory cytokines. These findings contribute to our understanding of the pathogenesis of acute SIV infection and have important implications for the development of therapeutic strategies. ImportanceDifferent mechanisms for CD4 T cell depletion during acute HIV-1 infection have been proposed. In this study, we should in SIV infected rhesus macaques that depletion of CD4 T cells is primarily due to pyroptosis. Other pathophysiologic mechanisms may also contribute in a minor way to CD4 T cell depletion.

microbiology↗

bioRxiv2021

A memristor-based neuromodulation device for real-time monitoring and adaptive control of in vitro neuronal populations

Neurons are specialized cells in information transmission and information processing. Following this, many neurologic disorders are directly linked not to cellular viability/homeostasis issues but rather to specific anomalies in electrical activity dynamics. Acknowledging this fact, therapeutic strategies based on direct modulation of neuronal electrical activity have been producing remarkable results, with successful examples ranging from cochlear implants to deep brain stimulation. Development on these implantable devices are hindered, however, by important challenges: power requirements, size factor, signal transduction, and adaptability/computational capabilities. Memristors, nanoscale electronic components able to emulate natural synapses, provide unique properties to address these constraints and their use in neuroprosthetic devices is being actively explored. Here we demonstrate for the first time the use of memristive devices in a clinically relevant setting where communication between two neuronal populations is conditioned to specific activity patterns in the source population. In our approach, the memristor device performs a simple pattern detection computation and acts as a synapstor capable of reversible short-term plasticity. Using in vitro hippocampal neuronal cultures, we show real-time adaptive control with a high degree of reproducibility using our monitor-compute-actuate paradigm. We envision very similar systems being used for automatic detection and suppression of seizures in epileptic patients.

neuroscience↗

bioRxiv2021

Substantial Differences in SARS-CoV-2 Antibody Responses Elicited by Natural Infection and mRNA Vaccination

We analyzed data from two ongoing COVID-19 longitudinal serological surveys in Orange County, CA., between April 2020 and March 2021. A total of 8,476 finger stick blood specimens were collected before and after an aggressive mRNA vaccination campaign. IgG levels were determined using a multiplex antigen microarray containing 10 SARS-CoV-2 antigens, 4 SARS, 3 MERS, 12 Common CoV, and 8 Influenza antigens. Twenty-six percent of 3,347 specimens from unvaccinated Orange County residents in December 2020 were SARS-CoV-2 seropositive. The Ab response was predominantly against nucleocapsid (NP), full length spike and the spike S2 domain. Anti-receptor binding domain (RBD) reactivity was low and there was no cross-reactivity against SARS S1 or SARS RBD. An aggressive mRNA vaccination campaign at the UCI Medical Center started on December 16, 2020 and 6,724 healthcare workers were vaccinated within 3 weeks. Seroprevalence increased from 13% in December to 79% in January, 93% in February and 99% in March. mRNA vaccination induced much higher Ab levels especially against the RBD domain and significant cross-reactivity against SARS RBD and S1 was also observed. Nucleocapsid protein Abs can be used to distinguish individuals in a population of vaccinees to classify those who have been previously infected and those who have not, because nucleocapsid is not in the vaccine. Previously infected individuals developed higher Ab titers to the vaccine than those who have not been previously exposed. These results indicate that mRNA vaccination rapidly induces a much stronger and broader Ab response than SARS-CoV-2 infection.

immunology↗

bioRxiv2020

Fluorescence Microscopy Datasets for Training Deep Neural Networks

BackgroundFluorescence microscopy is an important technique in many areas of biological research. Two factors which limit the usefulness and performance of fluorescence microscopy are photobleaching of fluorescent probes during imaging, and, when imaging live cells, phototoxicity caused by light exposure. Recently developed methods in machine learning are able to greatly improve the signal to noise ratio of acquired images. This allows researchers to record images with much shorter exposure times, which in turn minimizes photobleaching and phototoxicity by reducing the dose of light reaching the sample. FindingsTo employ deep learning methods, a large amount of data is needed to train the underlying convolutional neural network. One way to do this involves use of pairs of fluorescence microscopy images acquired with long and short exposure times. We provide high quality data sets which can be used to train and evaluate deep learning methods under development. ConclusionThe availability of high quality data is vital for training convolutional neural networks which are used in current machine learning approaches.

bioinformatics↗

bioRxiv2019

Chromosomal length reference assembly for Diaphorina citri using single-molecule sequencing and Hi-C proximity ligation with manually curated genes in developmental, structural and immune pathways

BackgroundDiaphorina citri is a vector of "Candidatus Liberibacter asiaticus" (CLas), the gram-negative bacterial pathogen associated with citrus greening disease. Control measures rely on pesticides with negative impacts on the environment, natural ecosystems and human and animal health. In contrast, gene-targeting methods have the potential to specifically target the vector species and/or reduce pathogen transmission. ResultsTo improve the genomic resources needed for targeted pest control, we assembled a D. citri genome based on PacBio long reads followed by proximity ligation-based scaffolding. The 474 Mb genome has 13 chromosomal-length scaffolds. 1,036 genes were manually curated as part of a community annotation project, composed primarily of undergraduate students. We also computationally identified a total of 1,015 putative transcription factors (TFs) and were able to infer motifs for 337 TFs (33 %). In addition, we produced a genome-independent transcriptome and genomes for D. citri endosymbionts. ConclusionsManual annotation provided more accurate gene models for use by researchers and also provided an excellent training opportunity for students from multiple institutions. All resources are available on CitrusGreening.org and NCBI. The chromosomal-length D. citri genome assembly serves as a blueprint for the development of collaborative genomics projects for other medically and agriculturally significant insect vectors.

genomics↗