Hang Lu

HYPER-Net enables physics-conditioned self-supervised reconstruction for Fourier light-field microscopy, improving volumetric imaging across diverse biological contexts.

This study develops a microfluidic device to simultaneously image neural activity and behavior in chemically stimulated, head-fixed C. elegans, showing that greater movement freedom leads to more naturalistic responses.

The paper uses network modularity to analyze time-varying functional connectivity in single-cell neural recordings, showing how context and network state influence module organization in C. elegans and whole-brain activity.

This study shows that aversive learning in C. elegans reorganizes neural dynamics globally, altering activity only during specific tasks while preserving essential functions.

ASCENT tracks neurons in 3D microscopy data without manual annotation, achieving state-of-the-art accuracy on C. elegans datasets.

This paper introduces a microfluidics-compatible open-top light-sheet microscope using model predictive control of a tunable lens to enable high-speed 3D imaging of C. elegans.

EZ-FRCNN provides a user-friendly, offline deep learning tool for object detection in biological datasets, enabling applications from cell biology to ecological studies.

This study measures neuromechanical phase lags in C. elegans and shows they are controlled by body elasticity and environmental forces, shaping gait.

WormID-Bench evaluates methods for neuron detection, identification, and tracking in C. elegans using a multi-laboratory dataset, highlighting the need for improved computational approaches.

This paper develops a mesofluidic bioreactor platform with high-content imaging for non-invasive quality control of brain organoids, improving their development and reducing cell death.

The paper harmonizes whole-brain imaging datasets from multiple labs to enable automated neuron identification and identifies factors influencing neuron positioning in C. elegans.

The paper introduces WormPsyQi, an image analysis pipeline for high-throughput quantification of synaptically localized fluorescent signals, along with 30 transgenic C. elegans strains for synaptic labeling.

The paper introduces TARDIS, a high-throughput transgenesis method in C. elegans that generates thousands of genetically unique transgenic individuals by splitting the transgenesis process into two steps.

This paper develops and provides a set of optimized GCaMP reporters for C. elegans functional imaging, offering improved dynamic range, varied kinetics, and subcellular localization.

This paper introduces a deep learning framework for neuro-imaging denoising that improves speed, memory efficiency, and accuracy while requiring minimal training data.

CRF_ID automates cell identity annotation in dense images, achieving higher accuracy and robustness than existing methods across various experimental conditions and organisms.

This study shows how C. elegans maintains food-responsiveness across temperature changes through a multicellular gene network involving TGF-β and serotonin signaling.

The paper describes a low-cost device for precise temperature control during specimen manipulation under stereomicroscopes, demonstrating improved characterization of temperature-sensitive alleles in C. elegans and other Caenorhabditis species.

The paper develops a semi-supervised learning framework to integrate heterogeneous biological data and demonstrates its application in analyzing Drosophila pattern formation.