Search Results (67)

Background/Objectives: The aim of this study is to test different convolutional neural network (CNN) and Transformer-based models to detect and classify vertical misfit at the abutment-prosthesis interface on panoramic radiographs, and to develop a hybrid deep learning model enhanced with attention mechanisms. Methods: A dataset consisting of a total of 566 images, manually classified as 249 ‘fit’ and 317 ‘misfit’ cases by two experts, was created. Images were resized to 224 × 224 and divided into training, validation, and test groups. The deep learning model yielding the most successful results was determined as the backbone; a hybrid model was developed by integrating three different attention modules (SE, CBAM, and ECA) into this structure. Model performance was evaluated using accuracy, precision, sensitivity, and F1 score metrics. Results: CNN-based models (RegNetY-800MF, ConvNeXt-Tiny, EfficientNetV2-S, ResNet50) performed better than Transformer-based models (DeiT, Swin-Tiny) in all metrics. The proposed hybrid model exhibited the highest success among all tested models with a 99.12% accuracy rate. This model reached a 100% precision value in the misfit group and yielded no false positive results. The F1 scores of the hybrid model were recorded as 99.01% for the fit group and 99.21% for the misfit group. Conclusions: The findings of this study demonstrate that attention-enhancing deep learning frameworks have the potential to significantly improve the diagnostic utility of routine panoramic radiographs. It shows that panoramic imaging, when supported by advanced artificial intelligence, can provide valuable diagnostic support in detecting vertical misfit. The developed model has the potential to become a reliable clinical decision support system. Full article

The development of prostheses that accurately reproduce fine motor skills remains a key challenge for daily assistance applications. This research presents the development of a soft robotic hand prosthesis prototype inspired by the natural behavior of muscles and tendons, incorporating internal vacuum-based reinforcement and textured fingertip surfaces to enhance friction and grasp adaptability, without relying on force sensors. The prosthesis reproduces open-hand and tripod pinch movements through myoelectric signals (EMG) acquired via a wearable armband equipped with eight surface electrodes. The signals are processed in real-time and classified by a lightweight dense neural network implemented on a low-power microcontroller. Tendon-driven actuation enables biomimetic motion with smooth and compliant behavior. The proposed system was validated through laboratory-based functional tests using user-specific models, showing response times ranging from 0.49 to 2.00 s and an overall grasping effectiveness of approximately 80% when manipulating small everyday objects with different geometries. These results indicate that the prototype constitutes an accessible and functional solution for fine motor assistance, with potential applicability in low-cost and resource-constrained myoelectric prosthetic systems. Full article

Background/Objectives: It is essential to test the accuracy of artificial intelligence-assisted tools that detect dental pathologies from radiographs. This study aimed to evaluate the test characteristics of an artificial intelligence-assisted convolutional neural network-based prototype used for automated radiographic detection. Methods: A total of 300 panoramic and 100 intraoral periapical radiographs were collected between January 2020 and 2024 and then analyzed by two trained, independent specialist evaluators. The diagnostic consensus, “ground truth”, was labeled as follows: BL: bone loss; C: caries; F: filling; I: implants; IT: impacted teeth; P: prosthesis; PC: post-core; PR: periapical radiolucency; RF: root fillings; and RR: retained roots. The radiographs were uploaded to the prototype, and the results were compared. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated using Stata version 15.0 (StataCorp). Results: Overall, most of the outcomes demonstrated sensitivity greater than 82%, with values ranging from 66.41% (65.47,67.36) for BL to 100% (100.00,100.00) for I. For all outcomes, specificity was greater than 93%, with values ranging from 93.61% (93.12,94.10) for BL to 100% for I. The overall values for all the test characteristics for the periapical radiographs were above 85%. The key errors identified in the qualitative analysis were errors in tooth identification, failure to detect recurrent caries under fillings and crowns, impacted canines, and inaccurate identification of extensive fillings as crowns. Conclusions: The prototype demonstrated high sensitivity and specificity in identifying dental pathologies. Accuracy in identifying bone loss, teeth that have migrated, including impacted canines, secondary caries, and differentiating extensive fillings from crowns requires further improvement. Full article

Affordable, sensor-driven lower-limb prostheses remain scarce in middle-income health systems. We report the design, numerical justification, and bench validation of a semi-active transfemoral prosthesis featuring surface electromyography (EMG) control and inertial sensing for low-resource deployment. The mechanical architecture combines a titanium–aluminum–carbon composite frame (total mass 0.87 kg; parts cost < USD 400) with topology optimization (SIMP) to minimize weight while preserving stiffness. Finite-element analyses (critical load 2.94 kN) confirmed structural safety (yield safety factor ≥ 1.6) and favorable fatigue margins. A dual-channel sensing scheme—surface EMG from the rectus femoris and an IMU—drives a five-state gait finite state machine implemented on a low-power STM32H platform. The end-to-end EMG→PWM latency remained <200 ms (mean 185 ms). Bench tests reproduced commanded flexion within ±2.2%, with average electrical power of ~4.6 W and battery autonomy of ~5.7 h using a 1650 mAh Li-Po pack. Results demonstrate a pragmatic trade-off between functionality and cost: semi-active damping with EMG-triggered control and open, modular hardware suitable for small-lab fabrication. Meeting target metrics (mass ≤ 1 kg, latency ≤ 200 ms, autonomy ≥ 6 h, cost ≤ USD 500), the prototype indicates a viable pathway to broaden access to intelligent prostheses and provides a platform for future upgrades (e.g., neural network control and higher-efficiency actuators). Full article

Background/Objectives: This article brings forward a novel methodology for the intra-op approach of forearm amputation stumps to facilitate their subsequent wireless connection to a neural prosthesis. A neural prosthesis offers the amputee more motor functions compared to myoelectric prostheses, but the neural prosthesis must be connected to the patient’s stump nerves. Methods: An experimental animal study was conducted on 15 Wistar rats. Under anesthesia, the sciatic nerve was carefully dissected and preserved using a folding technique to maintain maximum length without tension. Nerves were repositioned with consideration for future use with biocompatible conduits. Morphometric measurements (nerve length, external diameter, fascicle count) were performed, followed by statistical analysis of length–diameter correlations. Results: The techniques show that the length of the nerves in the amputation stump can be preserved and integrated into the muscle masses with appropriate methods and biomaterials, which ensures the transmission of motor impulses to control the movements of a prosthesis. Fibrosis and mechanical injury have a lower risk of occurring with the nerves protected in the muscle mass. Through statistical analysis we find that sciatic nerve length and diameter have a positive correlation (r = 0.71, p = 0.003), supporting anatomic plausibility for human extrapolation of results. Conclusions: The amputation technique preserves much of the nerve length and viability and is simple to perform. Neural electrode implantation can be facilitated by folding the nerve within a large muscle mass and using biomaterial conduits. Better rehabilitation of the patient may occur with the use of a prosthesis equipped with more functions and superior control. Full article

Conventional hip disarticulation prostheses often require amputees to produce limited leg-lifting torque through exaggerated pelvic motion, resulting in complex control and pronounced gait abnormalities. To overcome the limitations, we present a mapping control strategy for a powered hip disarticulation prosthesis aimed at improving gait symmetry. A quaternion-based method was implemented to capture hip joint kinematics, while a gated recurrent unit (GRU) neural network was trained to model the kinematic relationship between the intact and prosthetic limbs, enabling biomimetic trajectory control. Validation experiments showed that trajectory similarity between predicted and actual motions increased with walking speed, reaching 98.12% at 3.0 km/h. Comparative walking tests revealed an 84.00% improvement in hip flexion angle with the powered prosthesis over conventional designs. Notable improvements in gait symmetry were observed: stride symmetry (measured by SI and RII) improved by 23.21% and 19.28%, respectively, while hip trajectory symmetry increased by 68.07% (SI) and 47.59% (RII). These results confirm that the GRU-based kinematic mapping model offers robust trajectory prediction and that the powered prosthesis significantly enhances gait symmetry, delivering more natural and biomimetic motion. Full article

The primary objective of the presented study is to develop a comprehensive system for the acquisition of surface electromyographic (sEMG) data and to perform time–frequency analysis aimed at extracting discriminative features for the classification of hand gestures intended for the control of a simplified bionic hand prosthesis. The proposed system is designed to facilitate precise finger gesture execution in both prosthetic and robotic hand applications. This article outlines the methodology for multi-channel sEMG signal acquisition and processing, as well as the extraction of relevant features for gesture recognition using artificial neural networks (ANNs) and other well-established machine learning (ML) algorithms. Electromyographic signals were acquired using a prototypical LPCXpresso LPC1347 ARM Cortex M3 (NXP, Eindhoven, Holland) development board in conjunction with surface EMG sensors of the Gravity OYMotion SEN0240 type (DFRobot, Shanghai, China). Signal processing and feature extraction were carried out in the MATLAB 2024b environment, utilizing both the Fourier transform and the Hilbert–Huang transform to extract selected time–frequency characteristics of the sEMG signals. An artificial neural network (ANN) was implemented and trained within the same computational framework. The experimental protocol involved 109 healthy volunteers, each performing five predefined gestures of the right hand. The first electrode was positioned on the brachioradialis (BR) muscle, with subsequent channels arranged laterally outward from the perspective of the participant. Comprehensive analyses were conducted in the time domain, frequency domain, and time–frequency domain to evaluate signal properties and identify features relevant to gesture classification. The bionic hand prototype was fabricated using 3D printing technology with a PETG filament (Spectrum, Pęcice, Poland). Actuation of the fingers was achieved using six MG996R servo motors (TowerPro, Shenzhen, China), each with an angular range of ${180}^{\circ }$, controlled via a PCA9685 driver board (Adafruit, New York, NY, USA) connected to the main control unit. Full article

Trigeminal somatosensory-evoked potentials (TSEPs) provide valuable insight into neural responses to oral stimuli. This study investigates TSEP recording methods and their impact on interpreting results in clinical settings to improve the development process of neurostimulation-based therapies. The experiments and results presented here aim at identifying appropriate stimulation characteristics to design an active dental prosthesis capable of contributing to restoring the lost neurosensitive connection between the teeth and the brain. Two methods of TSEP acquisition, traditional and occluded, were used, each conducted by a different volunteer. Traditional TSEP acquisition involves stimulation at different sites with varying parameters to achieve a control base. In contrast, occluded TSEPs examine responses acquired under low- and high-force bite conditions to assess the influence of periodontal mechanoreceptors and muscle activation on measurements. Traditional TSEPs demonstrated methodological feasibility with satisfactory results despite a limited subject pool. However, occluded TSEPs presented challenges in interpreting results, with responses deviating from expected norms, particularly under high force conditions, due to the simultaneous occurrence of stimulation and dental occlusion. While traditional TSEPs highlight methodological feasibility, the occluded approach highlights complexities in outcome interpretation and urges caution in clinical application. Previously unreported results were achieved, which underscores the importance of conducting further research with larger sample sizes and refined protocols in order to strengthen the reliability and validity of TSEP assessments. Full article

To improve the adaptability of the hand prosthesis, we propose a new smart control for a physiological finger prosthesis using the advantages of the deep deterministic policy gradient (DDPG) algorithm. A rigid body model was developed to represent the finger as a training environment. The geometric characteristics and physiological physical properties of the finger available in the literature were assumed, but the joint’s stiffness and damping were neglected. The standard DDPG algorithm was modified to train an artificial neural network (ANN) to perform two predetermined trajectories: linear and sinusoidal. The ANN was evaluated through the use of a computational model that simulated the functionality of the finger prosthesis. The model demonstrated the capacity to successfully execute both sinusoidal and linear trajectories, exhibiting a mean error of $3.984\pm 2.899$ mm for the sinusoidal trajectory and $3.220\pm 1.419$ mm for the linear trajectory. Observing the torques, it was found that the ANN used different strategies to control the movement in order to adapt to the different trajectories. Allowing the ANN to use a combination of both trajectories, our model was able to perform trajectories that differed from purely linear and sinusoidal, showing its ability to adapt to the movement of the physiological finger. The results showed that it was possible to develop a controller for multiple trajectories, which is essential to provide more integrated and personalized prostheses. Full article

Cervical disc arthroplasty (CDA) is a motion-preserving alternative to anterior cervical discectomy and fusion (ACDF) for cervical degenerative disease, reducing adjacent segment degenerative disease (ASD). Despite its benefits, some patients experience CDA failure due to prosthesis-related complications, heterotopic ossification, segmental kyphosis, ASD, or facet joint degeneration, necessitating revision surgery. Reoperation strategies depend on the failure mechanism, instability, sagittal malalignment, and neural compression. Anterior revision is suited for prosthesis failure, recurrent disc herniation, or ASD, enabling prosthesis removal, decompression, and fusion. In select cases, reimplantation may restore motion. Posterior approaches are preferred for facet degeneration, multilevel stenosis, or posterior hypertrophy, with options including foraminotomy, laminoplasty, or laminectomy and fusion. Complex cases may require combined anterior and posterior surgery for optimal decompression and stability. This narrative review outlines revision strategies, emphasizing biomechanical assessment, radiographic evaluation, and patient-specific considerations. Despite surgical challenges, meticulous planning and execution can optimize outcomes. Full article

This paper presents an integrated approach for a retinal prosthesis that overcomes the scalability challenges and limitations of conventional systems that use external cameras. Silicon nanowires (SiNWs) are utilized as photonic sensors due to their nanoscale dimensions and high surface-to-volume ratio. To enhance these properties and achieve high photoresponsivity, our research team developed a vertically stacked SiNW structure using a fabrication method entirely based on dry etching. The fabricated SiNW photodetector demonstrated excellent electrical and optical characteristics, including linear I–V characteristics that confirmed ohmic contact formation and high photoresponsivity exceeding 10⁵ A/W across the 400–800 nm wavelength range. The SiNW photodetector, following its integration with a switched capacitor stimulator circuit, exhibited a proportional increase in stimulation current in response to higher light intensity and increased SiNW density. In vitro experiments confirmed the efficacy of the integrated system in inducing neural responses from retinal cells, as indicated by an increased number of neural spikes observed at higher light intensities and SiNW densities. This study contributes to sensor technology by demonstrating an approach to integrating nanostructures and electronic components, which enhances control and functionality. Full article

Traditional myoelectric controls of trans-humeral prostheses fail to provide intuitive coordination of the necessary degrees of freedom. We previously showed that by using artificial neural network predictions to reconstruct distal joints, based on the shoulder posture and movement goals (i.e., position and orientation of the targeted object), participants were able to position and orient an avatar hand to grasp objects with natural arm performances. However, this control involved rapid and unintended prosthesis movements at each modification of the movement goal, impractical for real-life scenarios. Here, we eliminate this abrupt change using novel methods based on an angular trajectory, determined from the speed of stump movement and the gap between the current and the ‘goal’ distal configurations. These new controls are tested offline and online (i.e., involving participants-in-the-loop) and compared to performances obtained with a natural control. Despite a slight increase in movement time, the new controls allowed twelve valid participants and six participants with trans-humeral limb loss to reach objects at various positions and orientations without prior training. Furthermore, no usability or workload degradation was perceived by participants with upper limb disabilities. The good performances achieved highlight the potential acceptability and effectiveness of those controls for our target population. Full article

This paper demonstrates, for the first time, the stability of synthetic diamond as a passive layer within neural implants. Leveraging the exceptional biocompatibility of intrinsic nanocrystalline diamond, a comprehensive review of material aging analysis in the context of in-vivo implants is provided. This work is based on electric impedance monitoring through the formulation of an analytical model that scrutinizes essential parameters such as the deposited metal resistivity, insulation between conductors, changes in electrode geometry, and leakage currents. The evolution of these parameters takes place over an equivalent period of approximately 10 years. The analytical model, focusing on a fractional capacitor, provides nuanced insights into the surface conductivity variation. A comparative study is performed between a classical polymer material (SU8) and synthetic diamond. Samples subjected to dynamic impedance analysis reveal distinctive patterns over time, characterized by their physical degradation. The results highlight the very high stability of diamond, suggesting promise for the electrode’s enduring viability. To support this analysis, microscopic and optical measurements conclude the paper and confirm the high stability of diamond and its strong potential as a material for neural implants with long-life use. Full article

Objective: To train a machine-learning (ML) algorithm to classify stumbling in transfemoral amputee gait. Methods: Three subjects completed gait trials in which they were induced to stumble via three different means. Several iterations of ML algorithms were developed to ultimately classify whether individual steps were stumbles or normal gait using leave-one-out methodology. Data cleaning and hyperparameter tuning were applied. Results: One hundred thirty individual stumbles were marked and collected during the trials. Single-layer networks including Long-Short Term Memory (LSTM), Simple Recurrent Neural Network (SimpleRNN), and Gradient Recurrent Unit (GRU) were evaluated at 76% accuracy (LSTM and GRU). A four-layer LSTM achieved an 88.7% classic accuracy, with 66.9% step-specific accuracy. Conclusion: This initial trial demonstrated the ML capabilities of the gathered dataset. Though further data collection and exploration would likely improve results, the initial findings demonstrate that three forms of induced stumble can be learned with some accuracy. Significance: Other datasets and studies, such as that of Chereshnev et al. with HuGaDB, demonstrate the cataloging of human gait activities and classifying them for activity prediction. This study suggests that the integration of stumble data with such datasets would allow a knee prosthesis to detect stumbles and adapt to gait activities with some accuracy without depending on state-based recognition. Full article

Powered ankle prostheses have been proven to improve the walking economy of people with transtibial amputation. All commercial powered ankle prostheses that are currently available can only perform one-degree-of-freedom motion in a limited range. However, studies have shown that the frontal plane motion during ambulation is associated with balancing. In addition, as more advanced neural interfaces have become available for people with amputation, it is possible to fully recover ankle function by combining neural signals and a robotic ankle. Accordingly, there is a need for a powered ankle prosthesis that can have active control on not only plantarflexion and dorsiflexion but also eversion and inversion. We designed, built, and evaluated a two-degree-of-freedom (2-DoF) powered ankle–foot prosthesis that is untethered and can support level-ground walking. Benchtop tests were conducted to characterize the dynamics of the system. Walking trials were performed with a 77 kg subject that has unilateral transtibial amputation to evaluate system performance under realistic conditions. Benchtop tests demonstrated a step response rise time of less than 50 milliseconds for a torque of 40 N·m on each actuator. The closed-loop torque bandwidth of the actuator is 9.74 Hz. Walking trials demonstrated torque tracking errors (root mean square) of less than 7 N·m. These results suggested that the device can perform adequate torque control and support level-ground walking. This prosthesis can serve as a platform for studying biomechanics related to balance and has the possibility of further recovering the biological function of the ankle–subtalar–foot complex beyond the existing powered ankles. Full article