Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
David Le and colleagues report a deep learning approach to construct optical coherence tomography angiography (OCTA) images by analyzing spatial variance in the structural images within retinal vasculature. This method has the potential to eliminate the need for repeated acquisitions at identical locations, thereby enhancing the clinical utility of OCTA.
Zhang, Pan, Liu, and their colleagues propose a self-locking kirigami pattern that utilizes in-plane tension and release to fold 3D configurations. This method has been applied to construct a series of modular surfaces.
Hanwen Luo and colleagues report a device-compatible ultra-high-order quantum noise stream cipher encrypted method. The approach flexibly regulates the data transmission rate and security performance of high-speed long-distance communications networks.
Wang, Nolen and colleagues present an iteration of the existing visible light optical coherence tomography system, which improves both imaging depth, resolution and suppresses noise. The system was demonstrated in human retinal imaging.
Maximilian Gießler and colleagues introduce a wearable sensor to track human movement by directly measuring the angular acceleration. This approach eliminates the noise associated with numerical differentiation. The authors show that the improved tracking accuracy leads to benefits when assessing balance recovery responses to perturbed locomotion.
High-resolution fiber shape sensors face limited application due to high costs. Manavi et al. proposed a solution employing deep learning for shape prediction directly from the fiber sensor’s spectrum. This approach eliminates the need for expensive measurements and complex post-processing, providing a cost-effective yet accurate method for detecting complex multi-bend deformations.
Xingyu Zhao and colleagues report a Bayesian learning framework for runtime self-verification of robotics systems. This framework allows robots to autonomously evaluate and reconfigure themselves after both regular and singular events, using only imprecise and partial prior knowledge.
Dickmann and colleagues present an interferometric sensing scheme that delivers improved accuracy with wider dynamic range and high speed. They use the multiple modes of a Fabry-Perot laser diode with the weighted power to measure with moderate coherence and thus break up the periodicity of the sensor signal.
Yao and Ning present a method for sensing the folding and deployment behavior of tape spring hinges using skin-like strain and motion sensors. The integration of these soft sensors does not affect the structural deformations, thereby providing a solution to monitor the mechanical behavior of high-strain components under large deformations.
Hyun Kum and colleagues report on the heterogeneous integration of high-k single-crystalline strontium titanium oxide within gallium nitride high-electron-mobility transistors. As the dielectric material is epitaxially grown and transferred, they achieve low defect density between the gate oxide and transistor heterostructures.
Large-scale deployment of direct air carbon capture and storage (DACS) is required to offset CO2 emissions. Here, Moritz Gutsch and Jens Leker present a cost model and life cycle assessment for several combinations of off-grid DACSs, powered by photovoltaic energy and heat pumps combined with battery storage. They find a cost optimal energy system layout for implementation in Nevada USA.
Transcranial focused ultrasound has had limited applications in humans due to the unpredictable distortions of ultrasound by the human head. Thomas Riis and colleagues report an approach which enables direct correction for the attenuation of ultrasound by the skull and hair, thus enabling controlled ultrasound therapies in humans.
Fushimi and colleagues report a digital twin based method to correct the discrepancies between simulation and experimental states for acoustic hologram optimization. This method reduces the number of experimental measurements needed, thus more efficient as compared to classical finite difference based approaches.
The fixed and narrow working frequency bandwidth limits the use of most acoustic absorbers in a dynamic acoustic environment. Shrestha and colleagues designed a tunable acoustic absorber capable of absorbing sound with varying frequency bands using dielectric elastomer bending actuators. The triangular petal-like structure actively reconfigures its shape by voltage activation to shift the noisy absorption spectrum, offering a practical solution for real-time outdoor noise absorption.
Tomoki Inoue and colleagues report a time-series data clustering algorithm using a quantum-inspired digital annealer technology to improve the clustering performance. The algorithm was implemented to cluster time-series data derived from benchmark problems and flow measurement images.
Chomiak and Hu introduce a versatile time-series data prediction algorithm using recurring local topological patterning. This algorithm reduces computational complexity and cost, and its feasibility is demonstrated across various dynamic systems including macroeconomic, wearable sensor, and dynamic population systems.
Medium-range wireless power transfer using magnetically coupled resonators requires a wireless data link between the contactless coils to regulate power. Here, Xiangning He and colleagues report a wireless power transfer system which efficiently multiplexes the information modulation and signal channel with the power conversion and transfer channel.
Frequency dependent impairment degrades the transmission performance but its accurate calibration is still challenging due to the complex polarization. Dai, Zhang and colleagues report a calibration method able to simultaneously calibrate various frequency-dependent impairments in coherent optical transceivers, with a measured bit error rate improvement from 1.61e^-1 to 4.41e^-2 after calibration.
Data harmonization of MRI scans can improve consistency when analysing MRI data from different instruments in different locations. Liu and Yap report an efficient deep neural network method to disentangle site-specific information from site-invariant anatomical information in MRI images. The approach allows data in a wide range of existing studies, conducted via different imaging protocols, to be harmonized.
Yimo Yan and colleagues combined droplet-based microfluidics and controlled photopolymerization in an approach to mass produce soft microrobots with programmable structural and magnetic anisotropy. This large-scale production enabled the formation of microrobot swarms and observations of their emergent structures and behaviours.