Cyborg and bionic systems (Washington, D.C.)最新文献

Growing evidence highlights the importance of body composition (BC), including bone, muscle, and adipose tissue (AT), as a critical biomarker for cardiometabolic risk stratification. However, conventional methods for quantifying BC components using medical images are hindered by labor-intensive workflows and limited anatomical coverage. This study developed BioCompNet-an end-to-end deep learning workflow that integrates dual-parametric magnetic resonance imaging (MRI) sequences (water/fat) with a hierarchical U-Net architecture to enable fully automated quantification of 15 biomechanically critical BC components. BioCompNet targets 10 abdominal compartments (vertebral bone, psoas muscles, core muscles, subcutaneous AT [SAT], superficial SAT, deep SAT, intraperitoneal AT, retroperitoneal AT, visceral AT, and intermuscular AT [IMAT]) and 5 thigh compartments (femur, muscle, SAT, IMAT, and vessels). The workflow was developed on 8,048 MRI slices from a community-based cohort (n = 503) and independently validated on 240 MRI slices from a tertiary hospital (n = 30). The model's performance was benchmarked against expert annotations. On internal and external validation datasets, BioCompNet achieved average Dice similarity coefficients of 0.944 and 0.938 for abdominal compartments and 0.961 and 0.936 for thigh compartments, respectively. Excellent interreader reliability was observed (intraclass correlation coefficient ≥ 0.881) across all quantified features, and IMAT quantification showed a strong linear trend (P _trend < 0.001) compared to physician-rated assessments. The workflow substantially reduced processing time from 128.8 ± 5.6 to 0.12 ± 0.001 min per case. By enabling rapid, accurate, and comprehensive volumetric analysis of BC components, BioCompNet establishes a scalable framework for precision cardiometabolic risk assessment and clinical decision support.

C-tactile afferents are low-threshold mechanoreceptors that innervate the hairy skin of mammals, essential for emotional interactions. Replication of such a mechanism could facilitate emotional interactions between humans and embodied intelligence robotic systems. Herein, we demonstrate a monolithic synaptic device that replicates and integrates tactile sensing and neuromorphic processing functions for in-sensor computing. The device is operable by both mechanical and electrical inputs, with the mechanoelectrical operation mechanism stemming from the synergistic effect of dynamic ionic migration and injection. As a proof of concept, the device effectively converts spatiotemporal tactile stimuli into distinct electrical signals, which are subsequently encoded to enable the microcomputer to classify multiple discrete emotional states, such as happiness, calmness, and excitement. This monolithic integrated device, which converges mild tactile perception with neuromorphic processing, with high tactile sensitivity and low-energy consumption, establishes an approach for emotional interaction between intelligent robots and human beings.

Motivated by the agility of animal and human locomotion, highly dynamic bionic legged robots have been extensively applied across various domains. Legged robotics represents a multidisciplinary field that integrates manufacturing, materials science, electronics, and biology, and other disciplines. Among its core subsystems, the lower limbs are particularly critical, necessitating the integration of structural optimization, advanced modeling techniques, and sophisticated control strategies to fully exploit robots' dynamic performance potential. This paper presents a comprehensive review of recent developments in the structural design of single-legged robots and systematically summarizes prevailing modeling approaches and control strategies. Key challenges and potential future directions are also discussed, serving as a reference for the future application of state-of-the-art manufacturing and control methodologies in legged robotic systems.

Flexible continuum robots exhibit excellent adaptability to a wide range of tasks and environments. However, accurate and efficient modeling and control remain challenging due to their inherent nonlinearities. In this article, a hybrid model-based and online data-driven control method is proposed for a tendon-driven continuum robot, which requires no prior dataset collection or training. The method incorporates the Jacobian derived from the piecewise constant curvature model with online Jacobian error compensation using a Kalman filter. Consecutive Jacobian estimates are constrained to reduce fluctuations and improve stability in real-time estimation. Experimental results validate the effectiveness of the proposed hybrid approach in enhancing tracking accuracy and demonstrate its robustness against external disturbances.

Advancements in health wearable technology hold the potential to prevent critical health issues such as hyponatremia and other hydration-related conditions often triggered by intense physical activities. Approaches to address this issue include the development of thin-film wearable sensors incorporating carbon nanotubes (CNTs), which offer scalability, lightweight design, and exceptional electrical properties. CNT paper serves as an ideal substrate for electrochemical sensors like ion-selective membranes (ISMs), enabling effective on-skin electrolyte monitoring. However, current on-skin devices often face limitations in maintaining performance during human motion. This study introduces a bioinspired surface texturing technique that mimics the microstructures of rose petals to enhance wettability, self-cleaning, and ISM sensitivity. By replicating the mechanical properties of the surface texture found on rose petals, the newly developed ISM achieves accurate measurements across a 2-mm air gap, offering an improved interfacing solution that promotes better sweat recirculation and comfort. This advancement overcomes the constraints of traditional sensors, paving the way for more reliable and effective noninvasive health monitoring in real-world conditions.

In this study, we investigate the complex network dynamics of in vitro neural systems using DishBrain, which integrates live neural cultures with high-density multi-electrode arrays in real-time, closed-loop game environments. By embedding spiking activity into lower-dimensional spaces, we distinguish between spontaneous activity (Rest) and Gameplay conditions, revealing underlying patterns crucial for real-time monitoring and manipulation. Our analysis highlights dynamic changes in connectivity during Gameplay, underscoring the highly sample efficient plasticity of these networks in response to stimuli. To explore whether this was meaningful in a broader context, we compared the learning efficiency of these biological systems with state-of-the-art deep reinforcement learning (RL) algorithms (Deep Q Network, Advantage Actor-Critic, and Proximal Policy Optimization) in a simplified Pong simulation. Through this, we introduce a meaningful comparison between biological neural systems and deep RL. We find that when samples are limited to a real-world time course, even these very simple biological cultures outperformed deep RL algorithms across various game performance characteristics, implying a higher sample efficiency.

Centimeter-scale robots have unique advances such as small size, light weight, and flexible motions, which exhibit great application potential in many fields. Notably, high integration and robustness are 2 key factors determining the locomotion characteristics and practical applications. Here, we propose a novel centimeter-scale quadruped piezo robot. The robot's locomotion is generated by multi-dimensional vibration trajectories at the feet, which are produced through a novel built-in actuation method. The robot achieves high locomotion speed (47.38 body length per second), high carrying capability (28.96 times self-weight), and high-resolution motion (minimum step size of 0.33 μm). Benefiting from the built-in integration method, the robot realizes the built-in integration of actuation, control, communication, and power supply, enabling untethered movement and strong robustness. It has a low startup voltage (10 V _0-p) and an endurance time of 32 min. Furthermore, after enduring 3 consecutive drops, 2 kicks, and being stepped on by an adult (over 3,500 times its own weight), the system remains functional and continues to move afterward. The robot utilizes modular expansion to achieve image sensing applications, including multi-object image capture and object detection. This work provides inspiration for the balance between high-integration design and robustness in centimeter-scale robots.

Epidural electrical stimulation (EES) has emerged as a promising treatment for spinal cord injury (SCI). However, the therapeutic potential of EES in functional recovery following incomplete SCI remains limited, with few studies of a large sample size. This study included 11 patients who received EES combined with physical therapy (PT) and 10 who received only PT. Follow-ups were conducted pre-surgery, post-surgery, and at 19 to 25 months postoperatively. After the surgery, patients in the EES + PT group showed significant improvements in sensory function (P < 0.001) and muscle spasticity (P < 0.001). Long-term follow-up indicated that the EES + PT group had significant improvements in sensory function (P < 0.001), muscle spasticity (P < 0.01), and urinary function (P < 0.05). Among them, all 11 patients had improvements in sensory function and muscle spasticity, and 6 of 11 reported an improvement in urinary function. Moreover, of the 5 patients with neuropathic pain, 4 exhibited reduced pain scores. Compared with the PT-only group, the EES + PT group had significantly better recovery in sensory function (P < 0.01), muscle spasticity (P < 0.0001), muscle strength (P < 0.01), and bowel function (P < 0.01). Further analysis suggested that patients with less severe SCIs in the EES + PT group tend to achieve better functional recovery. With a relatively large sample size compared to those in previous studies, this study confirms the promising therapeutic effects of EES in SCI. EES combined with PT provides a potential approach for functional recovery in patients with incomplete SCI.

Small extracellular vesicles (sEVs) have emerged as powerful vectors for liquid biopsy, offering a noninvasive window into the dynamic physiological and pathological states of the body. However, to fully leverage the clinical potential of sEV biomarkers, it is imperative to develop robust and efficient technologies for their isolation and analysis. In this study, we introduce a novel sharp-edge acoustofluidic platform designed for rapid and effective sample preparation, coupled with sensitive detection of specific sEV populations based on their surface markers. Our approach utilizes acoustically activated sharp-edge microstructures to concentrate bead-bound sEVs within the microfluidic device, facilitating immediate visualization by fluorescence microscopy. As a proof of principle, we demonstrate the capability of this portable acoustofluidic chip to selectively isolate and detect epidermal growth factor receptor (EGFR)-expressing vesicles, achieving nearly a 6-fold signal enhancement in EGFR-positive sEVs compared to EGFR-negative populations from sample volumes as small as 50 μl. This advancement not only underscores the potential of our platform for high-sensitivity biomarker detection but also paves the way for its application in isolating organ-specific sEVs. Such capability could be transformative for real-time monitoring of organ function and the simultaneous detection of multiple sEV markers, thereby broadening the scope of diagnostic precision and therapeutic decision-making in clinical practice.

Abnormalities in the ossicular chain, a key middle-ear component that is crucial for sound transmission, can lead to conductive hearing loss; reconstruction offers an effective treatment. Accurate preoperative ossicular-chain measurements are essential for creating prostheses; however, current methods rely on cadaver studies or manual measurements from 2-dimensional images, which are time-intensive and laborious and depend heavily on radiologist expertise. To improve efficiency, we aimed to develop a systematic approach for automated ossicular-chain segmentation and measurement using ultra-high-resolution computed tomography (U-HRCT). One hundred forty patients (226 ears) with normal ear anatomy underwent U-HRCT. Twelve parameters were defined to measure ossicular-chain components. Automated measurements based on automated segmentation of 226 ear images were verified through manual measurements. We analyzed variations by ear side, sex, and age group. Stapes analysis was limited by segmentation accuracy. Complete segmentation of the malleus, incus, and stapes was achieved in 47 ears. Automated measurements of 8 parameters showed no significant differences compared to manual measurements in 47 cases. Significant sex-based differences emerged in all parameters except stapes footplate length, incudostapedial joint angle, and stapes volume (P = 0.205, P = 0.560, and P = 0.170, respectively). Notable side-specific differences were observed in female incus height and male malleus volume (P = 0.017 and P = 0.037, respectively). No statistically significant differences were found in other parameters across different age groups, except for malleus and incus volumes (P = 0.015 and P = 0.031). The proposed algorithm effectively automated ossicular-chain segmentation and measurement, establishing a normative range for ossicular parameters and providing a valuable reference for detecting abnormalities.