Signal analysis is a domain which is an amalgamation of different processes coming together to form robust pipelines for the automation of data analysis. When applied to the medical world, physiological signals are used. It is becoming increasingly common in today's day and age to be working with very large datasets, on the scale of having thousands of features. This is largely due to the fact that the acquisition of biomedical signals can be taken over multi-hour timeframes, which is another challenge to solve in and of itself. This paper will focus on the electrocardiogram (ECG) signal specifically, and common feature extraction techniques used for digital health and artificial intelligence (AI) applications. Feature extraction is a vital step of biomedical signal analysis. The basic goal of feature extraction is for signal dimensionality reduction and data compaction. In simple terms, this would allow one to represent data with a smaller subset of features; these features could then later be leveraged to be used more efficiently for machine learning and deep learning models for applications, such as classification, detection, and automated applications. In addition, the redundant data in the overall dataset is filtered out as the data is reduced during feature extraction. In this review, we cover ECG signal processing and feature extraction in the time domain, frequency domain, time-frequency domain, decomposition, and sparse domain. We also provide pseudocode for the methods discussed so that they can be replicated by practitioners and researchers in their specific areas of biomedical work. Furthermore, we discuss deep features, and machine learning integration, to complete the overall pipeline design for signal analysis. Finally, we discuss future work that can be innovated upon in the feature extraction domain for ECG signal analysis.

BACKGROUND: Isolated ACL reconstructions (ACLR) demonstrate limitations in restoring native knee kinematics. This study investigates the knee mechanics of ACLR plus various anterolateral augmentations using a patient-specific musculoskeletal knee model. MATERIALS AND METHODS: A patient-specific knee model was developed in OpenSim using contact surfaces and ligament details derived from MRI and CT data. The contact geometry and ligament parameters were varied until the predicted knee angles for intact and ACL-sectioned models were validated against cadaveric test data for that same specimen. Musculoskeletal models of the ACLR combined with various anterolateral augmentations were then simulated. Knee angles were compared between these reconstruction models to determine which technique best matched the intact kinematics. Also, ligament strains calculated by the validated knee model were compared to those of the OpenSim model driven by experimental data. The accuracy of the results was assessed by calculating the normalised RMS error (NRMSE); an NRMSE < 30% was considered acceptable. RESULTS: All rotations and translations predicted by the knee model were acceptable when compared to the cadaveric data (NRMSE < 30%), except for the anterior/posterior translation (NRMSE > 60%). Similar errors were observed between ACL strain results (NRMSE > 60%). Other ligament comparisons were acceptable. All ACLR plus anterolateral augmentation models restored kinematics toward the intact state, with ACLR plus anterolateral ligament reconstruction (ACLR + ALLR) achieving the best match and the greatest strain reduction in ACL, PCL, MCL, and DMCL. CONCLUSION: The intact and ACL-sectioned models were validated against cadaveric experimental results for all rotations. It is acknowledged that the validation criteria are very lenient; further refinement is required for improved validation. The results indicate that anterolateral augmentation moves the kinematics closer to the intact knee state; combined ACLR and ALLR provide the best outcome for this specimen.

Major trauma is a condition that can result in severe bone damage. Customised orthopaedic reconstruction allows for limb salvage surgery and helps to restore joint alignment. For the best possible outcome three dimensional (3D) medical imaging is necessary, but its availability and access, especially in developing countries, can be challenging. In this study, 3D bone shapes of the femur reconstructed from planar radiographs representing bone defects were evaluated for use in orthopaedic surgery. Statistical shape and appearance models generated from 40 cadaveric X-ray computed tomography (CT) images were used to reconstruct 3D bone shapes. The reconstruction simulated bone defects of between 0% and 50% of the whole bone, and the prediction accuracy using anterior-posterior (AP) and anterior-posterior/medial-lateral (AP/ML) X-rays were compared. As error metrics for the comparison, measures evaluating the distance between contour lines of the projections as well as a measure comparing similarities in image intensities were used. The results were evaluated using the root-mean-square distance for surface error as well as differences in commonly used anatomical measures, including bow, femoral neck, diaphyseal-condylar and version angles between reconstructed surfaces from the shape model and the intact shape reconstructed from the CT image. The reconstructions had average surface errors between 1.59 and 3.59 mm with reconstructions using the contour error metric from the AP/ML directions being the most accurate. Predictions of bow and femoral neck angles were well below the clinical threshold accuracy of 3°, diaphyseal-condylar angles were around the threshold of 3° and only version angle predictions of between 5.3° and 9.3° were above the clinical threshold, but below the range reported in clinical practice using computer navigation (i.e., 17° internal to 15° external rotation). This study shows that the reconstructions from partly available planar images using statistical shape and appearance models had an accuracy which would support their potential use in orthopaedic reconstruction.

BACKGROUND: Electroencephalogram (EEG) signals record electrical activity on the scalp. Measured signals, especially EEG motor imagery signals, are often inconsistent or distorted, which compromises their classification accuracy. Achieving a reliable classification of motor imagery EEG signals opens the door to possibilities such as the assessment of consciousness, brain computer interfaces or diagnostic tools. We seek a method that works with a reduced number of variables, in order to avoid overfitting and to improve interpretability. This work aims to enhance EEG signal classification accuracy by using methods based on time series analysis. Previous work on this line, usually took a univariate approach, thus losing the possibility to take advantage of the correlation information existing within the time series provided by the different electrodes. To overcome this problem, we propose a multivariate approach that can fully capture the relationships among the different time series included in the EEG data. To perform the multivariate time series analysis, we use a multi-resolution analysis approach based on the discrete wavelet transform, together with a stepwise discriminant that selects the most discriminant variables provided by the discrete wavelet transform analysis RESULTS: Applying this methodology to EEG data to differentiate between the motor imagery tasks of moving either hands or feet has yielded very good classification results, achieving in some cases up to 100% of accuracy for this 2-class pre-processed dataset. Besides, the fact that these results were achieved using a reduced number of variables (55 out of 22,176) can shed light on the relevance and impact of those variables. CONCLUSIONS: This work has a potentially large impact, as it enables classification of EEG data based on multivariate time series analysis in an interpretable way with high accuracy. The method allows a model with a reduced number of features, facilitating its interpretability and improving overfitting. Future work will extend the application of this classification method to help in diagnosis procedures for detecting brain pathologies and for its use in brain computer interfaces. In addition, the results presented here suggest that this method could be applied to other fields for the successful analysis of multivariate temporal data.

BACKGROUND: Monitoring the body temperature of premature infants is vital, as it allows optimal temperature control and may provide early warning signs for severe diseases such as sepsis. Thermography may be a non-contact and wireless alternative to state-of-the-art, cable-based methods. For monitoring use in clinical practice, automatic segmentation of the different body regions is necessary due to the movement of the infant. METHODS: This work presents and evaluates algorithms for automatic segmentation of infant body parts using deep learning methods. Based on a U-Net architecture, three neural networks were developed and compared. While the first two only used one imaging modality (visible light or thermography), the third applied a feature fusion of both. For training and evaluation, a dataset containing 600 visible light and 600 thermography images from 20 recordings of infants was created and manually labeled. In addition, we used transfer learning on publicly available datasets of adults in combination with data augmentation to improve the segmentation results. RESULTS: Individual optimization of the three deep learning models revealed that transfer learning and data augmentation improved segmentation regardless of the imaging modality. The fusion model achieved the best results during the final evaluation with a mean Intersection-over-Union (mIoU) of 0.85, closely followed by the RGB model. Only the thermography model achieved a lower accuracy (mIoU of 0.75). The results of the individual classes showed that all body parts were well-segmented, only the accuracy on the torso is inferior since the models struggle when only small areas of the skin are visible. CONCLUSION: The presented multi-modal neural networks represent a new approach to the problem of infant body segmentation with limited available data. Robust results were obtained by applying feature fusion, cross-modality transfer learning and classical augmentation strategies.

BACKGROUND: The development and innovation of biomechanical measurement methods provide a solution to the problems in ski jumping research. At present, research on ski jumping mostly focuses on the local technical characteristics of different phases, but studies on the technology transition process are less. OBJECTIVES: This study aims to evaluate a measurement system (i.e. the merging of 2D video recording, inertial measurement unit and wireless pressure insole) that can capture a wide range of sport performance and focus on the key transition technical characteristics. METHODS: The application validity of the Xsens motion capture system in ski jumping was verified under field conditions by comparing the lower limb joint angles of eight professional ski jumpers during the takeoff phase collected by different motion capture systems (Xsens and Simi high-speed camera). Subsequently, the key transition technical characteristics of eight ski jumpers were captured on the basis of the aforementioned measurement system. RESULTS: Validation results indicated that the joint angle point-by-point curve during the takeoff phase was highly correlated and had excellent agreement (0.966 ≤ r ≤ 0.998, P < 0.001). Joint root-mean-square error (RMSE) differences between model calculations were 5.967° for hip, 6.856° for knee and 4.009° for ankle. CONCLUSIONS: Compared with 2D video recording, the Xsens system shows excellent agreement to ski jumping. Furthermore, the established measurement system can effectively capture the key transition technical characteristics of athletes, particularly in the dynamic changes of straight turn into arc in inrun, the adjustment of body posture and ski movement during early flight and landing preparation.

OBJECTIVE: The purpose of this study was to compare the biomechanical and clinical results of two surgical methods for the treatment of vertical femoral neck fractures: Femoral neck system (FNS) and traditional three cannulated cancellous screws (CCS). METHODS: First, we developed three different vertical femoral neck fracture models for the finite element analysis, with angles of 55°, 65°, and 75°, respectively. Two experimental groups were set up: the FNS group and the CCS group. Each fracture group was tested under axial loads of 2100 N to measure the femur's displacement, Von Mises stress (VMS), and its internal fixation components. Secondly, we retrospectively included the cases of vertical femoral neck fractures with FNS and CCS in our hospital from May 2019 to May 2021. In this study, we compared the duration of intraoperative fluoroscopy, operative time, hospital stay, fracture healing time, Hemoglobin loss, Harris score of hip joint function, and postoperative complications among patients undergoing hip joint replacement. RESULTS: In terms of finite element analysis, FNS has better anti-displacement stability than CCS at 55°and 65°, while FNS is greater than CCS in Von Mises stress. Clinically, we followed up on 87 patients for an average of 12 months. FNS was superior to traditional CCS in fracture healing time, operation time, fluoroscopy duration, fracture healing time, and Harris hip function score. CONCLUSION: FNS is superior to traditional CCS in biomechanical and clinical aspects of treating vertical femoral neck fractures. There is potential for FNS to become a new treatment option for vertical femoral neck fractures.

BACKGROUND: Percutaneous coronary interventions (PCIs) within left main coronary arteries are high-risk procedures that require optimization of interactions between stent(s) and diseased vessels. Optical Coherence Tomography (OCT) is a widely accepted tool that enhances physicians' ability to assess proper stent appositions during clinical procedures. The primary aim of this study was to develop complementary post-procedure imaging methodologies to better assess and interpret outcomes of left main PCI procedures, utilizing both reanimated and perfusion-fixed human hearts. METHODS: PCIs were performed while obtaining OCT scans within the left main anatomies of six human hearts. Subsequently, each heart was scanned with a micro-CT scanner with optimized parameters to achieve resolutions up to 20 µm. Scans were reconstructed and imported into a DICOM segmentation software to generate computational models of implanted stents and associated coronary vessels. 2D images from OCT that were obtained during PCIs were compared to the 3D models generated from micro-CT reconstructions. In addition, the 3D models were utilized to create virtual reality scenes and enlarged 3D prints for development of "mixed reality" tools relative to bifurcation stenting within human left main coronary arteries. RESULTS: We developed reproducible methodologies for post-implant analyses of coronary artery stenting procedures. In addition, we generated high-resolution 3D computational models, with ~ 20-micron resolutions, of PCIs performed within reanimated and perfusion-fixed heart specimens. CONCLUSIONS: Generated computational models of left main PCIs performed in isolated human hearts can be used to obtain detailed measurements that provide further clinical insights on procedural outcomes. The 3D models from these procedures are useful for generating virtual reality scenes and 3D prints for physician training and education.

BACKGROUND: Infusion Pumps (IP) are medical devices that were developed in the 1960s and generate fluid flow at pressures higher than that of normal blood pressure. Various hospital sectors make use of them, and they have become indispensable in therapies requiring continuity and precision in the administration of medication and/or food. As they are classified Class III (high risk) equipment, their maintenance is crucial for proper performance of the device, as well as patient and operator safety. The principal consideration of the pump is the volume infused, and the device demands great attention to detail when being calibrated. A lack of necessary care with this equipment can lead to uncertainty in volume and precision during the administration of substances. Because of this, it is essential to evaluate its reliability, to prevent possible failures at time of execution. This control aims at the quality of the intended infusion result, becoming an indication of quality. METHODS: This systematic review summarizes studies done over the last 10 years (2011 to December 2021) that address the reliability and accuracy of hospital infusion pumps, in order to identify planning of maintenance and/or other techniques used in management of the equipment. The Prisma method was applied and the databases utilized were Embase, MEDLINE/Pubmed, Web of Science, Scopus, IEEE Xplore, and Science Direct. In addition, similar reviews were studied in Prospero and the Cochrane Library. For data analysis, softwares such as Mendeley, Excel, RStudio, and VOSviewer were used, and Robvis helped in plotting risk of bias results for studies performed with Cochrane tools. RESULTS: The six databases selected produced 824 studies. After applying eligibility criteria (inclusion and exclusion), removing duplicates, and applying filters 1 and 2, 15 studies were included in the present review. It was found that the most relevant sources came from the Institute of Electrical and Electronics Engineers (IEEE) and that the most relevant keywords revolved around the terms ("device failure", "infusion pumps", "adverse effects", "complications", etc.). These results made clear that there remains substantial room for improvement as it relates to the study of accuracy and reliability of infusion. CONCLUSIONS: We verified that the reliability and precision analysis of hospital infusion pumps need to be performed in a more detailed and consistent way. New developments, considering the model and IP specification, are intended, clearly explaining the adopted methodology.

Core body temperature (CBT) is a key vital sign and fever is an important indicator of disease. In the past decade, there has been growing interest for vital sign monitoring technology that may be embedded in wearable devices, and the COVID-19 pandemic has highlighted the need for remote patient monitoring systems. While wrist-worn sensors allow continuous assessment of heart rate and oxygen saturation, reliable measurement of CBT at the wrist remains challenging. In this study, CBT was measured continuously in a free-living setting using a novel technology worn at the wrist and compared to reference core body temperature measurements, i.e., CBT values acquired with an ingestible temperature-sensing pill. Fifty individuals who received the COVID-19 booster vaccination were included. The datasets of 33 individuals were used to develop the CBT prediction algorithm, and the algorithm was then validated on the datasets of 17 participants. Mean observation time was 26.4 h and CBT > 38.0 °C occurred in 66% of the participants. CBT predicted by the wrist-worn sensor showed good correlation to the reference CBT (r = 0.72). Bland-Altman statistics showed an average bias of 0.11 °C of CBT predicted by the wrist-worn device compared to reference CBT, and limits of agreement were - 0.67 to + 0.93 °C, which is comparable to the bias and limits of agreement of commonly used tympanic membrane thermometers. The small size of the components needed for this technology would allow its integration into a variety of wearable monitoring systems assessing other vital signs and at the same time allowing maximal freedom of movement to the user.

BACKGROUND: Heart auscultation is an easy and inexpensive tool for early diagnosis of congenital heart defects. In this regard, a simple device which can be used easily by physicians for heart murmur detection will be very useful. The current study was conducted to evaluate the validity of a Doppler-based device named "Doppler Phonolyser" for the diagnosis of structural heart diseases in pediatric patients. In this cross-sectional study, 1272 patients under 16 years who were referred between April 2021 and February 2022, to a pediatric cardiology clinic in Mofid Children Hospital, Tehran, Iran, were enrolled. All the patients were examined by a single experienced pediatric cardiologist using a conventional stethoscope at the first step and a Doppler Phonolyser device at the second step. Afterward, the patient underwent trans-thoracic echocardiography, and the echocardiogram results were compared with the conventional stethoscope as well as the Doppler Phonolyser findings. RESULTS: Sensitivity of the Doppler Phonolyser for detecting congenital heart defects was 90.5%. The specificity of the Doppler Phonolyser in detecting heart disease was 68.9% in compared with the specificity of the conventional stethoscope, which was 94.8%. Among the most common congenital heart defects in our study population, the sensitivity of the Doppler Phonolyser was 100% for detection of tetralogy of Fallot (TOF); In contrast, sensitivity of both the conventional stethoscope and the Doppler Phonolyser was relatively low for detecting atrial septal defect. CONCLUSIONS: Doppler Phonolyser could be useful as a diagnostic tool for the detection of congenital heart defects. The main advantages of the Doppler Phonolyser over the conventional stethoscope are no need for operator experience, the ability to distinguish innocent murmurs from the pathologic ones and no effect of environmental sounds on the performance of the device.

BACKGROUND: Exercise electrocardiography (ECG) is a noninvasive test aiming at producing ischemic changes. However, resting ECG cannot be adopted in diagnosing myocardial ischemia till ST-segment depressions. Therefore, this study aimed to detect myocardial energy defects in resting ECG using the Hilbert-Huang transformation (HHT) in patients with angina pectoris. METHODS: Electrocardiographic recordings of positive exercise ECG by performing coronary imaging test (n = 26) and negative exercise ECG (n = 47) were collected. Based on the coronary stenoses severity, patients were divided into three categories: normal, < 50%, and ≥ 50%. During the resting phase of the exercise ECG, all 10-s ECG signals are decomposed by HHT. The RT intensity index, composed of the power spectral density of the P, QRS, and T components, is used to estimate the myocardial energy defect. RESULTS: After analyzing the resting ECG using HHT, the RT intensity index was significantly higher in patients with positive exercise ECG (27.96%) than in those with negative exercise ECG (22.30%) (p < 0.001). In patients with positive exercise ECG, the RT intensity index was gradually increasing with the severity of coronary stenoses: 25.25% (normal, n = 4), 27.14% (stenoses < 50%, n = 14), and 30.75% (stenoses ≥ 50%, n = 8). The RT intensity index of different coronary stenoses was significantly higher in patients with negative exercise ECG, except for the normal coronary imaging test. CONCLUSIONS: Patients with coronary stenoses had a higher RT index at the resting stage of exercise ECG. Resting ECG analyzed using HHT could be a method for the early detection of myocardial ischemia.