Heart failure (HF) is a critical condition in which the accurate prediction of mortality plays a vital role in guiding patient management decisions. However, clinical datasets used for mortality prediction in HF often suffer from an imbalanced distribution of classes, posing significant challenges. In this paper, we explore preprocessing methods for enhancing one-month mortality prediction in HF patients. We present a comprehensive preprocessing framework including scaling, outliers processing and resampling as key techniques. We also employed an aware encoding approach to effectively handle missing values in clinical datasets. Our study utilizes a comprehensive dataset from the Persian Registry Of cardio Vascular disease (PROVE) with a significant class imbalance. By leveraging appropriate preprocessing techniques and Machine Learning (ML) algorithms, we aim to improve mortality prediction performance for HF patients. The results reveal an average enhancement of approximately 3.6% in F1 score and 2.7% in MCC for tree-based models, specifically Random Forest (RF) and XGBoost (XGB). This demonstrates the efficiency of our preprocessing approach in effectively handling Imbalanced Clinical Datasets (ICD). Our findings hold promise in guiding healthcare professionals to make informed decisions and improve patient outcomes in HF management.
The shoulder complex consists of various joints which correspond to specific degrees of freedom (DOF). Musculoskeletal modeling is a method to reconstruct the real motions that involved a big challenge to simulation accuracy. Different shoulder models with various DOF and constraint definitions are available although the accuracy of the simulation is still debated. Thus, the objective of this study was to evaluate DOF and constraint definitions of shoulder models for inverse kinematics (IK) simulation during the middle direct punch (MDP). The experimental marker data in twenty elite martial arts players (65.4±5.8 kg, 172±7.8 cm, 29.5±8.5 years) were collected during the MDP. Four models were chosen as base models (M3, three-DOF between humerus and trunk Glenohumeral joint; M9, three-DOF for Scapulothoracic joint, three-DOF for Acromioclavicular joint, and three-DOF for Glenohumeral joint; Mst, coupled motions for scapula, clavicle, and humerus; Mel, an ellipsoid mobilizer for the Sternoclavicular joint). The subject-specific models were generated for each subject and model by marker data in static posture and scaling tools of OpenSim. The inverse kinematics tool of OpenSim was used to reconstruct MDP motion via models. RMS of marker error was used to compare models as indicators. Results represented significant differences in RMS of marker error for various models during the MDP tasks (P<0.05). Results illustrated the RMS of marker error for model Mel was minimum (12.87±0.09mm), whereas the RMS of marker error for all subject-specific models was lower than generic models. Based on the results the best model is model Mel which indicated this musculoskeletal model of the shoulder complex can reconstruct the MDP motion better than other represented models. In addition, our results indicated using the subject-specific model instead of generic models is vital to access reliable results. The results of this study would be utilized by sport science researchers, sports medicine doctors, coaches, trainers, and players to simulate shoulder motion by musculoskeletal modeling.
Image registration is the process of matching the coordinate systems of two or more images. Medical image registration has been used in a variety of applications such as segmentation, motion tracking, etc. Recently, the use of deep neural networks has been demonstrated as a useful approach to registration problems. In this article, we propose two separate novel Convolutional Neural Network (CNN) architectures for multi-modal rigid and affine registration of the CT-MRI images of the brain. A dataset consisting of CT-MRI images of 37 subjects was used for training and evaluation of the networks. For both networks, the proposed models achieved a high mutual information value between predicted CT images and their corresponding MRIs and a mean dice score of 0.984 for rigid registration.
In the past, many people lost their lives due to superficial wound infections. Over time, the importance of wound healing as the reconstruction of the body's first defense barrier, the skin, became more important. Today, traditional medicine has given way to reconstructive medicine. Tissue engineering has always been a pioneer in the development and application of new methods as the main branch of reconstructive medicine. In this research, a Nano-Scaffold was designed and fabricated for skin Tissue engineering with unique properties using electrospinning method. In this project, to increase the function of the scaffold, two nanoparticles containing two types of drugs have been used, which is considered as a hybrid drug delivery system. Concomitant use of layered double hydroxide hybrid (LDH) nanoparticles containing curcumin and imidazole zeolite (ZIF-8) containing aspirin in poly-lactic acid scaffolding will accelerate wound healing and reduce inflammation. After processing two layered double hydroxide nanoparticles containing curcumin and imidazole zeolite containing aspirin, these two nanoparticles were loaded on poly-lactic acid nanofibers. By adding 3% by weight of nanoparticles to poly-lactic acid, the tensile strength increased from 1.31 MPa to 1.6 MPa, the contact angle increased from 59 ° to 120 ° and cell viability (within 72 hours) increased from 49% to 88%. Tests performed on the scaffold confirmed its biocompatibility.
Gait analysis is one of the major topics in rehabilitation and sport. Tracking and determining gait phases can be done using various sensors and methods. In this paper, a fuzzy logic method is proposed to analyze and detect the five phases of a gait cycle using ground reaction force (GRF) and its gradient. The proposed method enables better detection adaptability at different walking speeds and body weights compared with the traditional threshold algorithms. In this algorithm, the GRF, measured by an insole equipped with force sensing resistors (FSR) and GRF gradient, which represent the plantar pressure transmission during a cycle, is passed through a set of fuzzy rules to detect the five gaits. A genetic algorithm (GA) is also applied for optimizing the fuzzy logic membership functions to reach minimum detection delay. A cost function is defined based on the difference between the normal reference gait and the output of the fuzzy logic gait phases. Detected phases are IC (initial contact), LR (loading response), MS (mid-stance), PS (pre-swing), and SW (swing). It is shown that the proposed method reaches a highly reliable performance of phase detection, especially for the initial contact (IC) and toe-off (TO). The average detection delays for the IC and TO phases, using the fuzzy-based method for three walking speeds of 0.4, 0.85, and 1.3 m/s, were -14.3±16.9ms and 1.24±17.0ms, respectively, and the average duration of stance and swing phases are 61.42% and 38.58%, respectively.
