Medical Engineering & Physics最新文献

Epilepsy is one of the most common brain diseases, characterised by repeated seizures that occur on a regular basis. During a seizure, a patient's muscles flex uncontrollably, causing a loss of mobility and balance, which can be harmful or even fatal. Developing an automatic approach for warning patients of oncoming seizures necessitates substantial research. Analyzing the electroencephalogram (EEG) output from the human brain's scalp region can help predict seizures. EEG data were analyzed to extract time domain features such as Hurst exponent (Hur), Tsallis entropy (TsEn), enhanced permutation entropy (impe), and amplitude-aware permutation entropy (AAPE). In order to automatically diagnose epileptic seizure in children from normal children, this study conducted two sessions. In the first session, the extracted features from the EEG dataset were classified using three machine learning (ML)-based models, including support vector machine (SVM), K nearest neighbor (KNN), or decision tree (DT), and in the second session, the dataset was classified using three deep learning (DL)-based recurrent neural network (RNN) classifiers in The EEG dataset was obtained from the Neurology Clinic of the Ibn Rushd Training Hospital. In this regard, extensive explanations and research from the time domain and entropy characteristics demonstrate that employing GRU, LSTM, and BiLSTM RNN deep learning classifiers on the $All-time-entropy$ fusion feature improves the final classification results.

Statistical shape models (SSMs) are useful tools in evaluating variation in bony anatomy to assess pathology, plan surgical interventions, and inform the design of orthopaedic implants and instrumentation. Recently, by considering multiple bones spanning a joint or the whole lower extremity, SSMs can support studies investigating articular conformity and joint mechanics. The objective of this study was to assess tradeoffs in accuracy between SSMs of the femur or tibia individually versus a combined joint-level model. Three statistical shape models were developed (femur-only, tibia-only, and joint-level) for a training set of 179 total knee arthroplasty (TKA) patients with osteoarthritis representing both genders and several ethnicities. Bone geometries were segmented from preoperative CT scans, meshed with triangular elements, and registered to a template for each SSM. Principal component analysis was performed to determine modes of variation. The statistical shape models were compared using measures of compactness, accuracy, generalization, and specificity. The generalization evaluation, assessing the ability to describe an unseen instance in a leave-one-out analysis, showed that errors were consistently smaller for the individual femur and tibia SSMs than for the joint-level model. However, when additional modes were included in the joint-level model, the errors were comparable to the individual bone results, with minimal additional computational expense. When developing more complex SSMs at the joint, lower limb, or whole-body level, the use of an error threshold to inform the number of included modes, instead of 95 % of the variation explained, can help to ensure accurate representations of anatomy.

Model-based Roentgen Stereophotogrammetric Analysis (RSA) is able to measure the migration of metallic prostheses with submillimeter accuracy through contour-detection and 3D surface model matching techniques. However, contour-detection is only possible if the prosthesis is clearly visible in the radiograph; consequently Model-based RSA cannot be directly used for polymeric materials due to their limited X-ray attenuation; this is especially clinically relevant for all-polyethylene implants. In this study the radiopacity of unicompartmental Ultra-High Molecular Weight Polyethylene (UHMWPE) knee bearings was increased by diffusing an oil-based contrast agent into the surface to create three different levels of surface radiopacity. Model-based RSA was performed on the bearings alone, the bearings alongside a metallic component held in position using a phantom, the bearings cemented into a Sawbone tibia, and the bearings at different distances from the femoral component. For each condition the precision and accuracy of zero motion of Model-based RSA were assessed. The radiopaque bearings could be located in the stereo-radiographs using Model-based RSA an accuracy comparable to metallic parts for translational movements (0.03 mm to 0.50 mm). For rotational movements, the accuracy was lower (0.1^∘ to 3.0^∘). The measurement accuracy was compared for all the radiopacity levels and no significant difference was found (p=0.08). This study demonstrates that contrast enhanced radiopaque polyethylene can be used for Model-based RSA studies and has equivalent translational measurement precision to metallic parts in the superior-inferior direction.

Objectives

Intimal tears caused by aortic dissection can weaken the arterial wall and lead to aortic aneurysms. However, the effect of different tear states on the blood flow behaviour remains complex. This study uses a novel approach that combines numerical haemodynamic simulation with in vitro experiments to elucidate the effect of arterial dissection rupture on the complex blood flow state within the abdominal aneurysm and the endogenous causes of end-organ malperfusion.

Materials and methods

Based on the CT imaging data and clinical physiological parameters, the overall arterial models including aortic dissection and aneurysm with single tear and double tear were established, and the turbulence behaviours and haemodynamic characteristics of arterial dissection and aneurysm under different blood pressures were simulated by using non-Newtonian flow fluids with the pulsatile blood flow rate of the clinical patients as a cycle, and the results of the numerical simulation were verified by in vitro simulation experiments.

Results

Hemodynamic simulations revealed that the aneurysm and single-tear false lumen generated a maximum pressure of 320.591 mmHg, 267 % over the 120 mmHg criterion. The pressure differential generates reflux, leading to a WSS of 2247.9 Pa at the TL inlet and blood flow velocities of up to 6.41 m/s inducing extend of the inlet. DTD Medium FL instantaneous WP above 120 mmHg Standard 151 % Additionally, there was 82.5 % higher flow in the right iliac aorta than in the left iliac aorta, which triggered malperfusion. Thrombus was accumulated distal to the tear and turbulence. These results are consistent with the findings of the in vitro experiments.

Conclusions

This study reveals the haemodynamic mechanisms by which aortic dissection induces aortic aneurysms to produce different risk states. This will contribute to in vitro simulation studies as a new fulcrum in the process of moving from numerical simulation to clinical trials.

Measuring the kyphotic angle (KA) and lordotic angle (LA) on lateral radiographs is important to truly diagnose children with adolescent idiopathic scoliosis. However, it is a time-consuming process to measure the KA because the endplate of the upper thoracic vertebra is normally difficult to identify. To save time and improve measurement accuracy, a machine learning algorithm was developed to automatically extract the KA and LA. The accuracy and reliability of the T1-T12 KA, T5-T12 KA, and L1-L5 LA were reported. A convolutional neural network was trained using 100 radiographs with data augmentation to segment the T1-L5 vertebrae. Sixty radiographs were used to test the method. Accuracy and reliability were reported using the percentage of measurements within clinical acceptance (≤9°), standard error of measurement (SEM), and inter-method intraclass correlation coefficient (ICC_2,1). The automatic method detected 95 % (57/60), 100 %, and 100 % for T1-T12 KA, T5-T12 KA, and L1-L5 LA, respectively. The clinical acceptance rate, SEM, and ICC_2,1 for T1-T12 KA, T5-T12 KA, and L1-L5 LA were (98 %, 0.80°, 0.91), (75 %, 4.08°, 0.60), and (97 %, 1.38°, 0.88), respectively. The automatic method measured quickly with an average of 4 ± 2 s per radiograph and illustrated how measurements were made on the image, allowing verifications by clinicians.

The objective of this work was to determine the effects of using simplified finite element (FE) mesh geometry in the process of performing reverse iterative fitting to estimate cartilage material parameters from in situ indentation testing. Six bovine tibial osteochondral explants were indented with sequential 5 % step-strains followed by a 600 s hold while relaxation force was measured. Three sets of porous viscohyperelastic material parameters were estimated for each specimen using reverse iterative fitting of the indentation test with (1) 2D axisymmetric, (2) 3D idealized, and (3) 3D specimen-specific FE meshes. Variable material parameters were identified using the three different meshes, and there were no systematic differences, correlation to basic geometric features, nor distinct patterns of variation based on the type of mesh used. Implementing the three material parameter sets in a separate 3D FE model of 40 % compressive strain produced differences in von Mises stresses and pore pressures up to 25 % and 50 %, respectively. Accurate material parameters are crucial in any FE model, and parameter differences influenced by idealized assumptions in initial material property determination have the potential to alter subsequent FE models in unpredictable ways and hinder the interpretation of their results.

Quantifying the mechanical behavior of skin has been foundational in applications of cosmetics, surgical techniques, forensic science, and protective clothing development. However, previous puncture studies have lacked consistent and physiological boundary conditions of skin. To determine natural skin tension, excision of in situ porcine skin resulted in significantly different diameter reduction (shrinkage) in leg (19.5 %) and abdominal skin (38.4 %) compared to flank skin (28.5 %) (p = 0.047). To examine effects of initial tension and pre-conditioning, five conditions of initial tension (as percentage of diameter increase) and pre-conditioning were tested in quasistatic puncture with a 5 mm spherical impactor using an electrohydraulic load frame and custom clamping apparatus. Samples with less than 5 % initial tension resulted in significantly greater (p = 0.011) force at failure (279.2 N) compared to samples with greater than 25 % initial tension (195.1 N). Eight pre-conditioning cycles of 15 mm displacement reduced hysteresis by 45 %. The coefficient of variance was substantially reduced for force, force normalized by cutis thickness, displacement, stiffness, and strain energy up to 46 %. Pre-conditioned samples at physiological initial tension (14–25 %) resulted in significantly greater (p = 0.03) normalized forces at failure (278.3 N/mm) compared to non-conditioned samples of the same initial tension (234.4 N/mm). Pre-conditioned samples with 14–25 % initial tension, representing physiological boundary conditions, resulted in the most appropriate failure thresholds with the least variation. For in vitro puncture studies, the magnitude of applied initial tension should be defined based on anatomical location, through a shrinkage experimentation, to match natural tension of skin. Characterizing the biological behavior and tolerances of skin may be utilized in finite element models to aid in protective clothing development and forensic science analyses.

Intention detection of the reaching movement is considerable for myoelectric human and machine collaboration applications. A comprehensive set of handcrafted features was mined from windows of electromyogram (EMG) of the upper-limb muscles while reaching nine nearby targets like activities of daily living. The feature selection-based scoring method, neighborhood component analysis (NCA), selected the relevant feature subset. Finally, the target was recognized by the support vector machine (SVM) model. The classification performance was generalized by a nested cross-validation structure that selected the optimal feature subset in the inner loop. According to the low spatial resolution of the target location on display and following the slight discrimination of signals between targets, the best classification accuracy of 77.11 % was achieved for concatenating the features of two segments with a length of 2 and 0.25 s. Due to the lack of subtle variation in EMG, while reaching different targets, a wide range of features was applied to consider additional aspects of the knowledge contained in EMG signals. Furthermore, since NCA selected features that provided more discriminant power, it became achievable to employ various combinations of features and even concatenated features extracted from different movement parts to improve classification performance.

The 12-lead electrocardiogram (ECG) is widely used for diagnosing cardiovascular diseases in clinical practice. Recently, deep learning methods have become increasingly effective for automatically classifying ECG signals. However, most current research simply combines the 12-lead ECG signals into a matrix without fully considering the intrinsic relationships between the leads and the heart's structure. To better utilize medical domain knowledge, we propose a multi-branch network for multi-label ECG classification and introduce an intuitive and effective lead grouping strategy. Correspondingly, we design multi-branch networks where each branch employs a multi-scale convolutional network structure to extract more comprehensive features, with each branch corresponding to a lead combination. To better integrate features from different leads, we propose a feature weighting fusion module. We evaluate our method on the PTB-XL dataset for classifying 4 arrhythmia types and normal rhythm, and on the China Physiological Signal Challenge 2018 (CPSC2018) database for classifying 8 arrhythmia types and normal rhythm. Experimental results on multiple multi-label datasets demonstrate that our proposed multi-branch network outperforms state-of-the-art networks in multi-label classification tasks