NPNicolas C. Pegard

Nicolas C. Pegard

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Nicolas C. Pegard has a total of 105 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    3D Segmentation of Pathological Muscle with a Physics-Informed Latent-Regularized U-Net

    by Negar Mehrabi, Nicolas C. Pégard, Geoffrey Gale Handsfield

    Objective: We developed a data-efficient deep learning framework for three-dimensional segmentation of pathological musculoskeletal anatomy from magnetic resonance imaging (MRI) when only limited manual annotations are available. Methods: We developed a Physics-Informed Latent-Regularized U-Net (PILR-U-Net) that combines transfer learning from healthy MRI, latent-space anatomical regularization, and elasticity-based physics-informed constraints. The physics-informed loss enforces mechanical equilibrium, near-incompressibility, and spatial smoothness of predicted deformation fields. The framework was evaluated on MRI datasets from 50 participants with cerebral palsy across 15 lower-limb musculoskeletal structures using sparse manual annotations. Performance was assessed using volumetric overlap, boundary accuracy, volume error, sensitivity, and precision. Ablation experiments evaluated the individual contributions of latent-space and physics-informed regularization. Results: Our experimental results show accurate segmentations with three-dimensional Dice coefficients ranging from 0.750 to 0.943 across evaluated structures, while most structures exhibited low surface-distance errors. Predicted deformation fields maintained positive Jacobian determinants near unity and smooth strain-energy distributions. Ablation analysis showed that both regularization components improved performance, with removal of physics-informed regularization producing the largest reductions in Dice accuracy and increases in boundary error. Conclusion: PILR-U-Net enables accurate and anatomically plausible segmentation of pathological musculoskeletal MRI under sparse supervision. Significance: Incorporating anatomical priors and biomechanical constraints into deep segmentation networks may reduce dependence on extensive pathological annotations and support patient-specific musculoskeletal modeling and clinical analysis.

  • Open Access

    Chromatix: a differentiable, GPU-accelerated wave-optics library

    by Diptodip Deb, Gert-Jan Both, Eric Bezzam, Amit Kohli, Siqi Yang, Amey Chaware, Cédric Allier, Changjia Cai, Geneva Anderberg, M. Hossein Eybposh, et al.

    Many current microscopy methods incorporate computational modeling as an integral part of the imaging process, either to solve inverse problems or optimize the optical system design itself. These methods often depend on differentiable optics simulations, yet no standardized framework exists, forcing computational optics researchers to repeatedly and independently implement simulations with limited reusability and performance. These common problems limit the potential impact of computational optics as a field. Here we present Chromatix: an open-source, graphics processing unit (GPU)-accelerated, differentiable wave-optics simulation library. Chromatix builds on JAX to democratize fast, parallelized simulation of diverse optical systems and expand the design space in computational optics. Chromatix standardizes a growing collection of optical elements and propagation methods allowing a broad range of applications, which we demonstrate here for snapshot microscopy, holography and phase retrieval. We demonstrate speed improvements of 2–6 times on a single GPU and up to 22 times on 8 GPUs.


  • Open Access

    Motor Assisted Commutator to Harness Electronics in Tethered Experiments

    by Jose Rodriguez-Romaguera, Jovan Tormes-Vaquerano, Ellora M. McTaggart, Maria M. Ortiz-Juza, Noah W. Miller, Kameron Thomas, Antonio Florido, Nicolas C. Pégard

    Research that combines advanced technological devices with complex behavioral tasks has enabled investigations into the neural mechanisms underlying brain and behavioral states. Freely moving rodent experiments often require a tether—a wired connection between an implanted device and an external power supply or data acquisition system. Traditionally, these experiments have used passive commutators to manage tethers, but such setups are often inadequate for reducing twisting and mechanical strain during behavioral tasks. Existing motorized commutators have extended the range of motion for these experiments but generally rely on stepper motors that produce auditory noise, potentially interfering with behavior. To address these limitations, we developed the Motor Assisted Commutator to Harness Electronics in Tethered Experiments (MACHETE), a motor-assisted commutator featuring a low-noise brushless motor. MACHETE dynamically adjusts tethers based on mouse movement, reducing torque and mechanical strain, and minimizing the animal's physical exertion during behavioral assays. Its onboard microcontroller provides customizable controls and seamless integration with custom electronic devices. The design includes a central through-hole to accommodate wires or fibers from external devices such as head-mounted miniature microscopes, electrophysiology probes, and optogenetics systems. We validated MACHETE across standard behavioral assays, including the open field test, the splash test, and the three-chamber social test. Our results showed no significant changes in mobility or behavior compared with untethered controls. By combining precise motor control, low auditory noise equipment, and accessibility, MACHETE can be used to support research that aims to adapt tools for use in freely moving behavior experiments.


  • Open Access

    Chromatix: a differentiable, GPU-accelerated wave-optics library

    by Diptodip Deb, Gert-Jan Both, Eric Bezzam, Amit Kohli, Siqi Yang, Amey Chaware, Cédric Allier, Changjia Cai, Geneva Anderberg, M. Hossein Eybposh, et al.

    Modern microscopy methods incorporate computational modeling as an integral part of the imaging process, either to solve inverse problems or optimize the optical system design itself. These methods often depend on differentiable optics simulations, yet no standardized framework exists—forcing computational optics researchers to repeatedly and independently implement simulations with limited reusability and performance. These common problems limit the potential impact of computational optics as a field. Here we present Chromatix: an open-source, GPU-accelerated, differentiable wave optics simulation library. Chromatix builds on JAX to democratize fast, parallelized simulation of diverse optical systems and expand the design space in computational optics. Chromatix standardizes a growing collection of optical elements and propagation methods allowing a broad range of applications, which we demonstrate here for snapshot microscopy, holography, and phase retrieval. We demonstrate speed improvements of 2-6× on a single GPU and up to 22× on 8 GPUs.


  • Open Access

    Compressive streak microscopy for fast sampling of fluorescent reporters of neural activity

    by Changjia Cai, Owen Traubert, Jovan Tormes-Vaquerano, M. Hossein Eybposh, Srinivas C. Turaga, Jose Rodriguez-Romaguera, Eva A. Naumann, Nicolas C. Pégard

    SignificanceIn vivo one-photon fluorescence imaging of calcium and voltage indicators expressed in neurons enables noninvasive recordings of neural activity with submillisecond precision. However, data acquisition speed is limited by the frame rate of cameras.AimWe developed a compressive streak fluorescence microscope to record fluorescence in individual neurons at high speeds (≥200 frames per second) exceeding the nominal frame rate of the camera by trading off spatial pixels for temporal resolution.ApproachOur microscope leverages a digital micromirror device for targeted illumination, a galvo mirror for temporal scanning, and a ridge regression algorithm for fast computational reconstruction of fluorescence traces with high temporal resolution.ResultsIn simulations, the ridge regression algorithm reconstructs traces of high temporal resolution with limited signal loss. Validation experiments with fluorescent beads and experiments in larval zebrafish demonstrate accurate reconstruction with a data compression ratio of 10 and accurate recordings of neural activity with 200- to 400-Hz sampling speeds.ConclusionsOur compressive microscopy enables new experimental capabilities to monitor activity at a sampling speed that outpaces the nominal frame rate of the camera.


  • Disentangling the Neural Circuits of Arousal and Anxiety-Like Behavior

    by Antonio Florido, Vincent R. Curtis, Nicolas C. Pégard, Jose Rodriguez-Romaguera

    Anxiety disorders are prevalent and debilitating conditions characterized by excessive concern and fear, affecting thoughts, behaviors, and sensations. A critical component of anxiety is arousal, a complex process involving alertness regulation and stimulus salience modulation. While arousal is adaptive in normal circumstances, dysregulation can lead to hypoarousal or hyperarousal, affecting response selection and threat perception. This chapter reviews challenges in studying arousal in preclinical anxiety models, emphasizing the need for multicomponent measurement and analysis. Novel methodologies integrating physiological measurement with activity tracking of neurons with single-cell resolution in awake animals are discussed, with emphasis in current challenges. Understanding these mechanisms is crucial for developing effective treatments for anxiety disorders.


  • Open Access

    A fully automated social interaction chamber for studying social threat learning in mice

    by Ellora M. McTaggart, Noah W. Miller, Maria M. Ortiz-Juza, Nicolas C. Pégard, Jose Rodriguez-Romaguera

    Social interactions are fundamental for our survival as a predominately social species. We need and seek positive social interactions to navigate the world. However, when social interactions are negative, and occur in the presence of an aversive event, learning occurs to associate such social interactions as threatening. Gaining insight into the neural circuits that drive social threat learning is crucial for understanding social interactions. Animal models can be leveraged to employ technologies that allow us to track neuronal processes with very high resolution to obtain a better understanding of the neural circuits involved. To accomplish this, we need robust behavioral models that are replicable and high throughput. Here, we present an open-source social interaction chamber that detects social interaction and automatically pairs it with foot shock. The social interaction chamber is designed to easily integrate into modular chambers commonly used for auditory and context threat learning. It contains an array of optical gates that precisely track mouse-to-mouse interactions in real time with digital triggers that can communicate with external devices to deliver a foot shock. We find that pairing social interactions with electric foot shock using our fully automated social interaction chamber is optimal for social threat associations. We further demonstrate that timing of social contact with foot shock produces optimal learning.


  • Open Access

    Protocol for an open-source system to integrate calcium imaging, pupillometry, and locomotion-estimated tracking in head-fixed mice

    by Maria M. Ortiz-Juza, Jovan Tormes-Vaquerano, Sophia M. Hegel, Vincent R. Curtis, Rizk A. Alghorazi, Noah W. Miller, Ellora M. McTaggart, Nicolas C. Pégard, Jose Rodriguez-Romaguera


  • Open Access

    ConIQA: A deep learning method for perceptual image quality assessment with limited data

    by M. Hossein Eybposh, Changjia Cai, Aram Moossavi, Jose Rodriguez-Romaguera, Nicolas C. Pégard

    Effectively assessing the realism and naturalness of images in virtual (VR) and augmented (AR) reality applications requires Full Reference Image Quality Assessment (FR-IQA) metrics that closely align with human perception. Deep learning-based IQAs that are trained on human-labeled data have recently shown promise in generic computer vision tasks. However, their performance decreases in applications where perfect matches between the reference and the distorted images should not be expected, or whenever distortion patterns are restricted to specific domains. Tackling this issue necessitates training a task-specific neural network, yet generating human-labeled FR-IQAs is costly, and deep learning typically demands substantial labeled data. To address these challenges, we developed ConIQA, a deep learning-based IQA that leverages consistency training and a novel data augmentation method to learn from both labeled and unlabeled data. This makes ConIQA well-suited for contexts with scarce labeled data. To validate ConIQA, we considered the example application of Computer-Generated Holography (CGH) where specific artifacts such as ringing, speckle, and quantization errors routinely occur, yet are not explicitly accounted for by existing IQAs. We developed a new dataset, HQA1k, that comprises 1000 natural images each paired with an image rendered using various popular CGH algorithms, and quality-rated by thirteen human participants. Our results show that ConIQA achieves superior Pearson (0.98), Spearman (0.965), and Kendall’s tau (0.86) correlations over fifteen FR-IQA metrics by up to 5%, showcasing significant improvements in aligning with human perception on the HQA1k dataset.


  • Open Access

    High-Speed All-Optical Neural Interfaces with 3D Temporally Focused Holography

    by Ian Antón Oldenburg, Hayley Anne Bounds, Nicolas C. Pégard