Criticality in the fracture of silica glass: Insights from molecular mechanics.

The universality of avalanches characterizing the inelastic response of disordered materials has the potential to bridge the gap from micro to macroscale. In this study, we explore the statistics and the scaling behavior of avalanches occurring during the fracture process in silica glass using molecular mechanics. We introduce a robust method for capturing and quantifying these avalanches, allowing us to perform rigorous statistical analyses, revealing universal power laws associated with critical phenomena. The influence of an initial crack is explored, observing deviations from mean-field predictions while maintaining the property of criticality. However, the avalanche exponents in the unnotched samples are predicted correctly by the mean-field depinning model. Furthermore, we investigate the strain-dependent probability density function, its cutoff function, and the interrelation between the critical exponents. Finally, we unveil distinct scaling behavior for small and large avalanches of the crack growth, shedding light on the underlying fracture mechanisms in silica glass.

Tether pre-tension within vertebral body tethering reduces motion of the spine and influences coupled motion: a finite element analysis.

Anterior Vertebral Body Tethering (VBT) is a novel fusionless treatment option for selected adolescent idiopathic scoliosis patients which is gaining widespread interest. The primary objective of this study is to investigate the effects of tether pre-tension within VBT on the biomechanics of the spine including sagittal and transverse parameters as well as primary motion, coupled motion, and stresses acting on the L2 superior endplate. For that purpose, we used a calibrated and validated Finite Element model of the L1-L2 spine. The VBT instrumentation was inserted on the left side of the L1-L2 segment with different cord pre-tensions and submitted to an external pure moment of 6 Nm in different directions. The range of motion (ROM) for the instrumented spine was measured from the initial post-VBT position. The magnitudes of the ROM of the native spine and VBT-instrumented with pre-tensions of 100 N, 200 N, and 300 N were, respectively, 3.29°, 2.35°, 1.90° and 1.61° in extension, 3.30°, 3.46°, 2.79°, and 2.17° in flexion, 2.11°, 1.67°, 1.33° and 1.06° in right axial rotation, and 2.10°, 1.88°, 1.48° and 1.16° in left axial rotation. During flexion-extension, an insignificant coupled lateral bending motion was observed in the native spine. However, VBT instrumentation with pre-tensions of 100 N, 200 N, and 300 N generated coupled right lateral bending of 0.85°, 0.81°, and 0.71° during extension and coupled left lateral bending of 0.32°, 0.24°, and 0.19° during flexion, respectively. During lateral bending, a coupled extension motion of 0.33-0.40° is observed in the native spine, but VBT instrumentation with pre-tensions of 100 N, 200 N, and 300 N generates coupled flexion of 0.67°, 0.58°, and 0.42° during left (side of the implant) lateral bending and coupled extension of 1.28°, 1.07°, and 0.87° during right lateral bending, respectively. Therefore, vertebral body tethering generates coupled motion. Tether pre-tension within vertebral body tethering reduces the motion of the spine.

Prediction of Temperature and Loading History Dependent Lumbar Spine Biomechanics Under Cyclic Loading Using Recurrent Neural Networks.

Extended-duration cyclic loading of the spine is known to be correlated to lower back pain (LBP). Therefore, it is important to understand how the loading history affects the entire structural behavior of the spine, including the viscoelastic effects. Six human spinal segments (L4L5) were loaded with pure moments up to 7.5 Nm cyclically for half an hour, kept unloaded for 15 min, and loaded with three cycles. This procedure was performed in flexion-extension (FE), axial rotation (AR), and lateral bending (LB) and repeated six times per direction for a total of 18 h of testing per segment. A Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) was trained to predict the change in the biomechanical response under cyclic loading. A strong positive correlation between the total testing time and the ratio of the third cycle to the last cycle of the loading sequence was found (BT: [Formula: see text] =  0.3469, p = 0.0003, RT: [Formula: see text] =0.1988, p  =   0.0377). The moment-range of motion (RoM) curves could be very well predicted with an RNN ([Formula: see text]=0.988), including the correlation between testing time and testing temperature as inputs. This study shows successfully the feasibility of using RNNs to predict changing moment-RoM curves under cyclic moment loading.

Rapid diagnosis of Covid-19 infections by a progressively growing GAN and CNN optimisation.

BACKGROUND AND OBJECTIVE: Covid-19 infections are spreading around the globe since December 2019. Several diagnostic methods were developed based on biological investigations and the success of each method depends on the accuracy of identifying Covid infections. However, access to diagnostic tools can be limited, depending on geographic region and the diagnosis duration plays an important role in treating Covid-19. Since the virus causes pneumonia, its presence can also be detected using medical imaging by Radiologists. Hospitals with X-ray capabilities are widely distributed all over the world, so a method for diagnosing Covid-19 from chest X-rays would present itself. Studies have shown promising results in automatically detecting Covid-19 from medical images using supervised Artificial neural network (ANN) algorithms. The major drawback of supervised learning algorithms is that they require huge amounts of data to train. Also, the radiology equipment is not computationally efficient for deep neural networks. Therefore, we aim to develop a Generative Adversarial Network (GAN) based image augmentation to optimize the performance of custom, light, Convolutional networks used for the classification of Chest X-rays (CXR). METHODS: A Progressively Growing Generative Adversarial Network (PGGAN) is used to generate synthetic and augmented data to supplement the dataset. We propose two novel CNN architectures to perform the Multi-class classification of Covid-19, healthy and pneumonia affected Chest X-rays. Comparisons have been drawn to the state of the art models and transfer learning methods to evaluate the superiority of the networks. All the models are trained using enhanced and augmented X-ray images and are compared based on classification metrics. RESULTS: The proposed models had extremely high classification metrics with proposed Architectures having test accuracy of 98.78% and 99.2% respectively while having 40% lesser training parameters than their state of the art counterpart. CONCLUSION: In the present study, a method based on artificial intelligence is proposed, leading to a rapid diagnostic tool for Covid infections based on Generative Adversarial Network (GAN) and Convolutional Neural Networks (CNN). The benefit will be a high accuracy of detection with up to 99% hit rate, a rapid diagnosis, and an accessible Covid identification method by chest X-ray images.

Generative adversarial network based data augmentation for CNN based detection of Covid-19.

Covid-19 has been a global concern since 2019, crippling the world economy and health. Biological diagnostic tools have since been developed to identify the virus from bodily fluids and since the virus causes pneumonia, which results in lung inflammation, the presence of the virus can also be detected using medical imaging by expert radiologists. The success of each diagnostic method is measured by the hit rate for identifying Covid infections. However, the access for people to each diagnosis tool can be limited, depending on the geographic region and, since Covid treatment denotes a race against time, the diagnosis duration plays an important role. Hospitals with X-ray opportunities are widely distributed all over the world, so a method investigating lung X-ray images for possible Covid-19 infections would offer itself. Promising results have been achieved in the literature in automatically detecting the virus using medical images like CT scans and X-rays using supervised artificial neural network algorithms. One of the major drawbacks of supervised learning models is that they require enormous amounts of data to train, and generalize on new data. In this study, we develop a Swish activated, Instance and Batch normalized Residual U-Net GAN with dense blocks and skip connections to create synthetic and augmented data for training. The proposed GAN architecture, due to the presence of instance normalization and swish activation, can deal with the randomness of luminosity, that arises due to different sources of X-ray images better than the classical architecture and generate realistic-looking synthetic data. Also, the radiology equipment is not generally computationally efficient. They cannot efficiently run state-of-the-art deep neural networks such as DenseNet and ResNet effectively. Hence, we propose a novel CNN architecture that is 40% lighter and more accurate than state-of-the-art CNN networks. Multi-class classification of the three classes of chest X-rays (CXR), ie Covid-19, healthy and Pneumonia, is performed using the proposed model which had an extremely high test accuracy of 99.2% which has not been achieved in any previous studies in the literature. Based on the mentioned criteria for developing Corona infection diagnosis, in the present study, an Artificial Intelligence based method is proposed, resulting in a rapid diagnostic tool for Covid infections based on generative adversarial and convolutional neural networks. The benefit will be a high accuracy of lung infection identification with 99% accuracy. This could lead to a support tool that helps in rapid diagnosis, and an accessible Covid identification method using CXR images.

An experimental-numerical method for the calibration of finite element models of the lumbar spine.

We present a systematic and automated stepwise method to calibrate computational models of the spine. For that purpose, a sequential resection study on one lumbar specimen (L2-L5) was performed to obtain the individual contribution of the IVD, the facet joints and the ligaments to the kinematics of the spine. The experimental data was prepared for the calibration procedure in such manner that the FE model could reproduce the average motion of the 10 native spines from former cadaveric studies as well as replicate the proportional change in ROM after removal of a spinal structure obtained in this resection study. A Genetic Algorithm was developed to calibrate the properties of the intervertebral discs and facet joints. The calibration of each ligament was performed by a simple and novel technique that requires only one simulation to obtain its mechanical property. After calibration, the model was capable of reproducing the experimental results in all loading directions and resections.

Motion preservation surgery for scoliosis with a vertebral body tethering system: a biomechanical study.

PURPOSE: There is a paucity of studies on new vertebral body tethering (VBT) surgical constructs especially regarding their potentially motion-preserving ability. This study analyses their effects on the ROM of the spine. METHODS: Human spines (T10-L3) were tested under pure moment in four different conditions: (1) native, (2) instrumented with one tether continuously connected in all vertebrae from T10 to L3, (3) additional instrumented with a second tether continuously connected in all vertebrae from T11 to L3, and (4) instrumented with one tether and one titanium rod (hybrid) attached to T12, L1 and L2. The instrumentation was inserted in the left lateral side. The intersegmental ROM was evaluated using a magnetic tracking system, and the medians were analysed. Please check and confirm the author names and initials are correct. Also, kindly confirm the details in the metadata are correct. The mentioned information is correct RESULTS: Compared to the native spine, the instrumented spine presented a reduction of less than 13% in global ROM considering flexion-extension and axial rotation. For left lateral bending, the median global ROM of the native spine (100%) significantly reduced to 74.6%, 66.4%, and 68.1% after testing one tether, two tethers and the hybrid construction, respectively. In these cases, the L1-L2 ROM was reduced to 68.3%, 58.5%, and 38.3%, respectively. In right lateral bending, the normalized global ROM of the spine with one tether, two tethers and the hybrid construction was 58.9%, 54.0%, and 56.6%, respectively. Considering the same order, the normalized L1-L2 ROM was 64.3%, 49.9%, and 35.3%, respectively. CONCLUSION: The investigated VBT techniques preserved global ROM of the spine in flexion-extension and axial rotation while reduced the ROM in lateral bending.

Biomechanical In Vitro and Finite Element Study on Different Sagittal Alignment Postures of the Lumbar Spine During Multiaxial Daily Motion.

Lumbar lordotic correction (LLC), the gold standard treatment for sagittal spinal malalignment (SMA), and its effect on sagittal balance have been critically discussed in recent studies. This paper assesses the biomechanical response of the spinal components to LLC as an additional factor for the evaluation of LLC. Human lumbar spines (L2L5) were loaded with combined bending moments in flexion (Flex)/extension (Ex) or lateral bending (LatBend) and axial rotation (AxRot) in a physiological environment. We examined the dependency of AxRot range of motion (RoM) on the applied bending moment. The results were used to validate a finite element (FE) model of the lumbar spine. With this model, the biomechanical response of the intervertebral discs (IVD) and facet joints under daily motion was studied for different sagittal alignment postures, simulated by a motion in Flex/Ex direction. Applied bending moments decreased AxRot RoM significantly (all P < 0.001). A stronger decline of AxRot RoM for Ex than for Flex direction was observed (all P < 0.0001). Our simulated results largely agreed with the experimental data (all R2 > 0.79). During the daily motion, the IVD was loaded higher with increasing lumbar lordosis (LL) for all evaluated values at L2L3 and L3L4 and posterior annulus stress (AS) at L4L5 (all P < 0.0476). The results of this study indicate that LLC with large extensions of LL may not always be advantageous regarding the biomechanical loading of the IVD. This finding may be used to improve the planning process of LLC treatments.

From mine to mind and mobiles - Lithium contamination and its risk management.

With the ever-increasing demand for lithium (Li) for portable energy storage devices, there is a global concern associated with environmental contamination of Li, via the production, use, and disposal of Li-containing products, including mobile phones and mood-stabilizing drugs. While geogenic Li is sparingly soluble, Li added to soil is one of the most mobile cations in soil, which can leach to groundwater and reach surface water through runoff. Lithium is readily taken up by plants and has relatively high plant accumulation coefficient, albeit the underlying mechanisms have not been well described. Therefore, soil contamination with Li could reach the food chain due to its mobility in surface- and ground-waters and uptake into plants. High environmental Li levels adversely affect the health of humans, animals, and plants. Lithium toxicity can be considerably managed through various remediation approaches such as immobilization using clay-like amendments and/or chelate-enhanced phytoremediation. This review integrates fundamental aspects of Li distribution and behaviour in terrestrial and aquatic environments in an effort to efficiently remediate Li-contaminated ecosystems. As research to date has not provided a clear picture of how the increased production and disposal of Li-based products adversely impact human and ecosystem health, there is an urgent need for further studies on this field.

Development of convolutional neural networks for recognition of tenogenic differentiation based on cellular morphology.

BACKGROUND AND OBJECTIVE: The use of automated systems for image recognition is highly preferred for regenerative medicine applications to evaluate stem cell differentiation early in the culturing state with non-invasive methodologies instead of invasive counterparts. Bone marrow-derived mesenchymal stem cells (BMSCs) are able to differentiate into desired cell phenotypes, and thereby promise a proper cell source for tendon regeneration. The therapeutic success of stem cell therapy requires cellular characterization prior to the implantation of cells. The foremost problem is that traditional characterization techniques require cellular material which would be more useful for cell therapy, complex laboratory procedures, and human expertise. Convolutional neural networks (CNNs), a class of deep neural networks, have recently made great improvements in image-based classifications, recognition, and detection tasks. We, therefore, aim to develop a potential CNN model in order to recognize differentiated stem cells by learning features directly from image data of unlabelled cells. METHODS: The differentiation of bone marrow mesenchymal stem cells (BMSCs) into tenocytes was induced with the treatment of bone morphogenetic protein-12 (BMP-12). Following the treatment and incubation step, the phase-contrast images of cells were obtained and immunofluorescence staining has been applied to characterize the differentiated state of BMSCs. CNN models were developed and trained with the phase-contrast cell images. The comparison of CNN models was performed with respect to prediction performance and training time. Moreover, we have evaluated the effect of image enhancement method, data augmentation, and fine-tuning training strategy to increase classification accuracy of CNN models. The best model was integrated into a mobile application. RESULTS: All the CNN models can fit the biological data extracted from immunofluorescence characterization. CNN models enable the cell classification with satisfactory accuracies. The best result in terms of accuracy and training time is achieved by the model proposed based on Inception-ResNet V2 trained from scratch using image enhancement and data augmentation strategies (96.80%, 434.55 sec). CONCLUSION: Our study reveals that the CNN models show good performance by identifying stem cell differentiation. Importantly this technique provides a faster and real-time tool in comparison to traditional methods enabling the adjustment of culture conditions during cultivation to improve the yield of therapeutic stem cells.

A Comparison of Three Neural Network Approaches for Estimating Joint Angles and Moments from Inertial Measurement Units.

The application of artificial intelligence techniques to wearable sensor data may facilitate accurate analysis outside of controlled laboratory settings-the holy grail for gait clinicians and sports scientists looking to bridge the lab to field divide. Using these techniques, parameters that are difficult to directly measure in-the-wild, may be predicted using surrogate lower resolution inputs. One example is the prediction of joint kinematics and kinetics based on inputs from inertial measurement unit (IMU) sensors. Despite increased research, there is a paucity of information examining the most suitable artificial neural network (ANN) for predicting gait kinematics and kinetics from IMUs. This paper compares the performance of three commonly employed ANNs used to predict gait kinematics and kinetics: multilayer perceptron (MLP); long short-term memory (LSTM); and convolutional neural networks (CNN). Overall high correlations between ground truth and predicted kinematic and kinetic data were found across all investigated ANNs. However, the optimal ANN should be based on the prediction task and the intended use-case application. For the prediction of joint angles, CNNs appear favourable, however these ANNs do not show an advantage over an MLP network for the prediction of joint moments. If real-time joint angle and joint moment prediction is desirable an LSTM network should be utilised.

Design of a vertical Loss-in-Weight feeder prototype with experimental proof of concept validation.

Loss-in-Weights (LiW) feeders are commonly oriented in a horizontal way. In this work, an experimental proof of concept, including mechanical and electrical design, construction, and operation, of a vertical LiW feeder prototype is performed. In a systematic design process, based on functional design specifications, the semi-automated vertical LiW feeder for dosing a wide range of powders, especially cohesive ones, is developed. The new dosing machine is assessed with regard to a number of key features such as high dosing accuracy, first-in-first-out powder discharge, easily interchange of the powder container, and flexibility in controlling the speed of the auger and stirrer motors independently. An experimental sensitivity analysis to study the functionality of the dosing machine and to investigate the weight variability of the weighing platform, i.e. mass flow rate, and quantity of dosed mass, is carried out. The results of the sensitivity analysis and the powder dosing tests of five diverse powders using different auger and stirrer geometries verified the proof of concept prototype.HighlightsA systematic design approach for validating a proof of concept of a vertical loss in weight feeder is appliedA full mechanical CAD design and implementation along with electric installation and software programming are executedSensitivity analysis approach is performed to validate the functionality of the semi-automated machine and successfully dispense dissimilar powders tested with different process parametersThe machine is characterized with a number of key features: first-in-first-out powder discharge, high dosing accuracy, flexible and modular concept design, flexibility in controlling the speed of the auger and the stirrer independently, lightweight and user-friendly design.

Prediction of ground reaction force and joint moments based on optical motion capture data during gait.

The standard camera- and force plate-based set-up for motion analysis suffers from the disadvantage of being limited to laboratory settings. Since adaptive algorithms are able to learn the connection between known inputs and outputs and generalise this knowledge to unknown data, these algorithms can be used to leverage motion analysis outside the laboratory. In most biomechanical applications, feedforward neural networks are used, although these networks can only work on time normalised data, while recurrent neural networks can be used for real time applications. Therefore, this study compares the performance of these two kinds of neural networks on the prediction of ground reaction force and joint moments of the lower limbs during gait based on joint angles determined by optical motion capture as input data. The accuracy of both networks when generalising to new data was assessed using the normalised root-mean-squared error, the root-mean-squared error and the correlation coefficient as evaluation metrics. Both neural networks demonstrated a high performance and good capabilities to generalise to new data. The mean prediction accuracy over all parameters applying a feedforward network was higher (r = 0.963) than using a recurrent long short-term memory network (r = 0.935).

Origin of reversible and irreversible atomic-scale rearrangements in a model two-dimensional network glass.

In this contribution, we investigate the fundamental mechanism of plasticity in a model two-dimensional network glass. The glass is generated by using a Monte Carlo bond-switching algorithm and subjected to athermal simple shear deformation, followed by subsequent unloading at selected deformation states. This enables us to investigate the topological origin of reversible and irreversible atomic-scale rearrangements. It is shown that some events that are triggered during loading recover during unloading, while some do not. Thus, two kinds of elementary plastic events are observed, which can be linked to the network topology of the model glass.

Artificial Neural Networks in Motion Analysis-Applications of Unsupervised and Heuristic Feature Selection Techniques.

The use of machine learning to estimate joint angles from inertial sensors is a promising approach to in-field motion analysis. In this context, the simplification of the measurements by using a small number of sensors is of great interest. Neural networks have the opportunity to estimate joint angles from a sparse dataset, which enables the reduction of sensors necessary for the determination of all three-dimensional lower limb joint angles. Additionally, the dimensions of the problem can be simplified using principal component analysis. Training a long short-term memory neural network on the prediction of 3D lower limb joint angles based on inertial data showed that three sensors placed on the pelvis and both shanks are sufficient. The application of principal component analysis to the data of five sensors did not reveal improved results. The use of longer motion sequences compared to time-normalised gait cycles seems to be advantageous for the prediction accuracy, which bridges the gap to real-time applications of long short-term memory neural networks in the future.

Frequency-related viscoelastic properties of the human incisor periodontal ligament under dynamic compressive loading.

Studies concerning the mechanical properties of the human periodontal ligament under dynamic compression are rare. This study aimed to determine the viscoelastic properties of the human periodontal ligament under dynamic compressive loading. Ten human incisor specimens containing 5 maxillary central incisors and 5 maxillary lateral incisors were used in a dynamic mechanical analysis. Frequency sweep tests were performed under the selected frequencies between 0.05 Hz and 5 Hz with a compression amplitude that was 2% of the PDL's initial width. The compressive strain varied over a range of 4%-8% of the PDL's initial width. The storage modulus, ranging from 28.61 MPa to 250.21 MPa, increased with the increase in frequency. The loss modulus (from 6.00 MPa to 49.28 MPa) also increased with frequency from 0.05 Hz- 0.5 Hz but remained constant when the frequency was higher than 0.5 Hz. The tanδ showed a negative logarithmic correlation with frequency. The dynamic moduli and the loss tangent of the central incisor were higher than those of the lateral incisor. This study concluded that the human PDL exhibits viscoelastic behavior under compressive loadings within the range of the used frequency, 0.05 Hz- 5 Hz. The tooth position and testing frequency may have effects on the viscoelastic properties of PDL.

Molecular Investigation of Mechanical Properties and Fracture Behavior of Graphene Aerogel.

An unusual combination of ultralight weight and outstanding mechanical properties of graphene aerogel made it popular for a wide range of applications in the fields of material science, energy, and technology. In the present work, the mechanical properties and fracture behavior of graphene aerogels, which are highly brittle in nature, are investigated using molecular dynamics (MD) simulations. In tensile tests, elastic modulus and tensile strength exhibit a power law dependence on the density with their exponents predicted to be 2.95 ± 0.05 and 1.61 ± 0.04, respectively, which are in an excellent agreement with the reported works in the literature. In the compression simulations, Lennard-Jones contribution in the AIREBO potential is vital to predicting early densification. Moreover, in the compression loading-unloading simulations, as the density decreases, the dissipation energy increases. At the onset of the crack propagation, as the crack length to height ratio increases, the fracture strength decreased. However, for a considered range of the ratios, the fracture toughness values were nearly constant for all densities. Furthermore, the fracture toughness shows a power law dependence on the density, with the exponent estimated to be 1.41 ± 0.04. The outcome of this work is a vital step in the in-depth understanding of nanomechanics while designing advanced, highly porous materials.

Active source localization in wave guides based on machine learning.

In the present work, an active source localization strategy is proposed. The presence of active sources in a waveguide can have several reasons, such as crack initiation or internal friction. In this study, the active source is represented by an impact event. A steel ball is dropped on an aluminum plate at different positions. Elastic waves are excited and will propagate through the plate. The wave response is acquired by a piezoelectric sensor network, which is attached to the plate. After performing numerical and physical experiments, enough data are collected in order to train an artificial neural network and a support vector machine. Those machine learning algorithms will predict the impact position based on the wave response of each sensor, while only numerical data from the finite element simulations are used to train both methods. After the training process is completed, the algorithms are applied to experimental data. A good agreement between reference and predicted results proves that the wave responses at the piezoelectric transducers contain sufficient information in order to localize the impact position precisely.

Estimation of Gait Mechanics Based on Simulated and Measured IMU Data Using an Artificial Neural Network.

Enhancement of activity is one major topic related to the aging society. Therefore, it is necessary to understand people's motion and identify possible risk factors during activity. Technology can be used to monitor motion patterns during daily life. Especially the use of artificial intelligence combined with wearable sensors can simplify measurement systems and might at some point replace the standard motion capturing using optical measurement technologies. Therefore, this study aims to analyze the estimation of 3D joint angles and joint moments of the lower limbs based on IMU data using a feedforward neural network. The dataset summarizes optical motion capture data of former studies and additional newly collected IMU data. Based on the optical data, the acceleration and angular rate of inertial sensors was simulated. The data was augmented by simulating different sensor positions and orientations. In this study, gait analysis was undertaken with 30 participants using a conventional motion capture set-up based on an optoelectronic system and force plates in parallel with a custom IMU system consisting of five sensors. A mean correlation coefficient of 0.85 for the joint angles and 0.95 for the joint moments was achieved. The RMSE for the joint angle prediction was smaller than 4.8° and the nRMSE for the joint moment prediction was below 13.0%. Especially in the sagittal motion plane good results could be achieved. As the measured dataset is rather small, data was synthesized to complement the measured data. The enlargement of the dataset improved the prediction of the joint angles. While size did not affect the joint moment prediction, the addition of noise to the dataset resulted in an improved prediction accuracy. This indicates that research on appropriate augmentation techniques for biomechanical data is useful to further improve machine learning applications.

Validation of two hybrid approaches for clustering age-related groups based on gait kinematics data.

Age-associated changes in walking parameters are relevant to recognize functional capacity and physical performance. However, the sensible nuances of slightly different gait patterns are hardly noticeable by inexperienced observers. Due to the complexity of this evaluation, we aimed at verifying the efficiency of applied hybrid-adaptive algorithms to cluster groups with similar gait patterns. Based on self-organizing maps (SOM), k-means clustering (KM), and fuzzy c-means (FCM), we compared the hybrid algorithms to a conventional FCM approach to cluster accordingly age-related groups. Additionally, we performed a relevance analysis to identify the principal gait characteristics. Our experiments, based on inertial-sensors data, comprised a sample of 180 healthy subjects, divided into age-related groups. The outcomes suggest that our methods outperformed the FCM algorithm, demonstrating a high accuracy (88%) and consistent sensitivity also to distinguish groups that presented a significant difference (p < .05) only in one of the six observed gait features. The applied algorithms showed a compatible performance, but the SOM + KM required less computation cost and, therefore, was more efficient. Furthermore, the results indicate the overall importance of cadence, as a measurement of physical performance, especially when clustering subjects by their age. Such output provides valuable information to healthcare professionals, concerning the subject's physical performance related to his age, supporting and guiding the physical evaluation.