Classifying Alzheimer's disease and normal subjects using machine learning techniques and genetic-environmental features.

BACKGROUND: Alzheimer's disease (AD) is complicated by multiple environmental and polygenetic factors. The accuracy of artificial neural networks (ANNs) incorporating the common factors for identifying AD has not been evaluated. METHODS: A total of 184 probable AD patients and 3773 healthy individuals aged 65 and over were enrolled. AD-related genes (51 SNPs) and 8 environmental factors were selected as features for multilayer ANN modeling. Random Forest (RF) and Support Vector Machine with RBF kernel (SVM) were also employed for comparison. Model results were verified using traditional statistics. RESULTS: The ANN achieved high accuracy (0.98), sensitivity (0.95), and specificity (0.96) in the intrinsic test for AD classification. Excluding age and genetic data still yielded favorable results (accuracy: 0.97, sensitivity: 0.94, specificity: 0.96). The assigned weights to ANN features highlighted the importance of mental evaluation, years of education, and specific genetic variations (CASS4 rs7274581, PICALM rs3851179, and TOMM40 rs2075650) for AD classification. Receiver operating characteristic analysis revealed AUC values of 0.99 (intrinsic test), 0.60 (TWB-GWA), and 0.72 (CG-WGS), with slightly lower AUC values (0.96, 0.80, 0.52) when excluding age in ANN. The performance of the ANN model in AD classification was comparable to RF, SVM (linear kernel), and SVM (RBF kernel). CONCLUSIONS: The ANN model demonstrated good sensitivity, specificity, and accuracy in AD classification. The top-weighted SNPs for AD prediction were CASS4 rs7274581, PICALM rs3851179, and TOMM40 rs2075650. The ANN model performed similarly to RF and SVM, indicating its capability to handle the complexity of AD as a disease entity.

Serial electrocardiogram recordings revealed a high prevalence of QT interval prolongation in patients with tuberculosis receiving fluoroquinolones.

BACKGROUND: Fluoroquinolones, crucial components of treatment regimens for drug-resistant tuberculosis (TB), are associated with QT interval prolongation and risks of fatal cardiac arrhythmias. However, few studies have explored dynamic changes in the QT interval in patients receiving QT-prolonging agents. METHODS: This prospective cohort study recruited hospitalized patients with TB who received fluoroquinolones. The study investigated the variability of the QT interval by using serial electrocardiograms (ECGs) recorded four times daily. This study analyzed the accuracy of intermittent and single-lead ECG monitoring in detecting QT interval prolongation. RESULTS: This study included 32 patients. The mean age was 68.6 ± 13.2 years. The results revealed mild-to-moderate and severe QT interval prolongation in 13 (41%) and 5 (16%) patients, respectively. The incremental yields in sensitivity of one to four daily ECG recordings were 61.0%, 26.1%, 5.6%, and 7.3% in detecting mild-to-moderate QT interval prolongation, and 66.7%, 20.0%, 6.7%, and 6.7% in detecting severe QT interval prolongation. The sensitivity levels of lead II and V5 ECGs in detecting mild-to-moderate and severe QT interval prolongation exceeded 80%, and their specificity levels exceeded 95%. CONCLUSION: This study revealed a high prevalence of QT interval prolongation in older patients with TB who receive fluoroquinolones, particularly those with multiple cardiovascular risk factors. Sparsely intermittent ECG monitoring, the prevailing strategy in active drug safety monitoring programs, is inadequate owing to multifactorial and circadian QT interval variability. Additional studies performing serial ECG monitoring are warranted to enhance the understanding of dynamic QT interval changes in patients receiving QT-prolonging anti-TB agents.

A Deep Learning Quantification Algorithm for HER2 Scoring of Gastric Cancer.

Gastric cancer is the third most common cause of cancer-related death in the world. Human epidermal growth factor receptor 2 (HER2) positive is an important subtype of gastric cancer, which can provide significant diagnostic information for gastric cancer pathologists. However, pathologists usually use a semi-quantitative assessment method to assign HER2 scores for gastric cancer by repeatedly comparing hematoxylin and eosin (H&E) whole slide images (WSIs) with their HER2 immunohistochemical WSIs one by one under the microscope. It is a repetitive, tedious, and highly subjective process. Additionally, WSIs have billions of pixels in an image, which poses computational challenges to Computer-Aided Diagnosis (CAD) systems. This study proposed a deep learning algorithm for HER2 quantification evaluation of gastric cancer. Different from other studies that use convolutional neural networks for extracting feature maps or pre-processing on WSIs, we proposed a novel automatic HER2 scoring framework in this study. In order to accelerate the computational process, we proposed to use the re-parameterization scheme to separate the training model from the deployment model, which significantly speedup the inference process. To the best of our knowledge, this is the first study to provide a deep learning quantification algorithm for HER2 scoring of gastric cancer to assist the pathologist's diagnosis. Experiment results have demonstrated the effectiveness of our proposed method with an accuracy of 0.94 for the HER2 scoring prediction.

O-Net: A Novel Framework With Deep Fusion of CNN and Transformer for Simultaneous Segmentation and Classification.

The application of deep learning in the medical field has continuously made huge breakthroughs in recent years. Based on convolutional neural network (CNN), the U-Net framework has become the benchmark of the medical image segmentation task. However, this framework cannot fully learn global information and remote semantic information. The transformer structure has been demonstrated to capture global information relatively better than the U-Net, but the ability to learn local information is not as good as CNN. Therefore, we propose a novel network referred to as the O-Net, which combines the advantages of CNN and transformer to fully use both the global and the local information for improving medical image segmentation and classification. In the encoder part of our proposed O-Net framework, we combine the CNN and the Swin Transformer to acquire both global and local contextual features. In the decoder part, the results of the Swin Transformer and the CNN blocks are fused to get the final results. We have evaluated the proposed network on the synapse multi-organ CT dataset and the ISIC 2017 challenge dataset for the segmentation task. The classification network is simultaneously trained by using the encoder weights of the segmentation network. The experimental results show that our proposed O-Net achieves superior segmentation performance than state-of-the-art approaches, and the segmentation results are beneficial for improving the accuracy of the classification task. The codes and models of this study are available at https://github.com/ortonwang/O-Net.

Artificial neural networks-based classification of emotions using wristband heart rate monitor data.

Heart rate variability (HRV) is an objective measure of emotional regulation. This study aimed to estimate the accuracy with which an artificial neural network (ANN) algorithm could classify emotions using HRV data that were obtained using wristband heart rate monitors.Four emotions were evoked during gameplay: pleasure, happiness, fear, and anger. Seven normalized HRV features (i.e., 3 time-domain features, 3 frequency-domain features, and heart rate), which yielded 29,727 segments during gameplay, were collected and analyzed first by statistics and then classified by the trained ANN model.General linear model adjusted for individual differences in HRV showed that all HRV features significantly differed across emotions, despite disparities in their magnitudes and associations. When compared to neutral status (i.e., no emotion evoked), the mean of R-R interval was significantly higher for pleasure and fear but lower for happiness and anger. In addition, pleasure evidenced the HRV features that suggested a superior parasympathetic to sympathetic activation. Happiness was associated with a prominent sympathetic activation. These statistical findings suggest that HRV features significantly differ across emotions evoked by gameplay. When further utilizing ANN-based emotion classification, the accuracy rates for prediction were above 75.0% across the 4 emotions with accuracy rates for classification of paired emotions ranging from 82.0% to 93.4%.For classifying emotion in an individual person, the trained ANN model utilizing HRV features yielded a high accuracy rate in our study. ANN is a time-efficient and accurate means to classify emotions using HRV data obtained from wristband heart rate monitors. Thus, this integrated platform can help monitor and quantify human emotions and physiological biometrics.

Development of Home-Based Frailty Detection Device Using Wireless Sensor Networks.

This study develops a home-based frailty detection device that uses embedded systems and wireless sensing technology. This system helps monitor the impact of aging among the elderly through wireless automatic detection. The detection system consists of four devices. The first device, called eScale, simulates the traditional falling ruler test to measure reaction time. Another device, called eChair, measures the pressure exerted by a test subject through a pressure sensor. It is used to test three symptoms of frailty, namely slowness of movement, physical weakness, and body weight. The third device, called ePad, consists of a soft membrane switch placed on the ground to detect footsteps and is used to test balance. The fourth device, called eReach, measures displacement through ultrasound sensors. It is used to carry out the functional reach test. The sampling rate of each device is the main factor that determines system performance. When the test distance was set to 5 m for Home-Gateway, a 1-Hz sampling rate showed the best performance (98 %). Up to eight devices can be connected simultaneously to the gateway. The proposed system was compared with conventional approaches through testing with test subjects (n = 8). The results of the five tests were as follows: standing forward bend (r = 0.929), balance (r = 0.996), slowness of movement (r = 0.976), and physical weakness (r = 0.991), with p < 0.01. In the reaction time test, r = 0.871, with p < 0.1. All results suggest high correlations. Tests of aging symptoms were performed on 309 people aged over 65 years. Among males, degradation of over 20 % was found in the areas of physical weakness, slowness of movement, and functional reach. Among females, a degradation of 75 % was found in the balance test.

Wireless sensor networks for indoor air quality monitoring.

The purpose of this study is to build an indoor air quality monitoring system based on wireless sensor networks (WSNs) technology. The main functions of the system include (1) remote parameter adjustment and firmware update mechanism for the sensors to enhance the flexibility and convenience of the system, (2) sensor nodes are designed by referring to the IEEE 1451.4 standard. This way, sensor nodes can automatically adjust and be plug and play, and (3) calibration method to strength the measurement value's sensitivity and accuracy. The experimental results show that transmission speed improves 30% than Trickle, transmission volume reduced to 42% of the original volume, updating task in 5*5 network topology can be executed 1.79 times and power consumption reduced to 30%. When baseline drifts, we can use the firmware update mechanism to adjust the reference value. The way can reduce error percentage from 15% to 7%.

eFurniture for home-based frailty detection using artificial neural networks and wireless sensors.

The purpose of this study is to integrate wireless sensor technologies and artificial neural networks to develop a system to manage personal frailty information automatically. The system consists of five parts: (1) an eScale to measure the subject's reaction time; (2) an eChair to detect slowness in movement, weakness and weight loss; (3) an ePad to measure the subject's balancing ability; (4) an eReach to measure body extension; and (5) a Home-based Information Gateway, which collects all the data and predicts the subject's frailty. Using a furniture-based measuring device to provide home-based measurement means that health checks are not confined to health institutions. We designed two experiments to obtain optimum frailty prediction model and test overall system performance: (1) We developed a three-step process to adjust different parameters to obtain an optimized neural identification network whose parameters include initialization, L.R. dec and L.R. inc. The post-process identification rate increased from 77.85% to 83.22%. (2) We used 149 cases to evaluate the sensitivity and specificity of our frailty prediction algorithm. The sensitivity and specificity of this system are 79.71% and 86.25% respectively. These results show that our system is a high specificity prediction tool that can be used to assess frailty.

A healthcare integration system for disease assessment and safety monitoring of dementia patients.

Elderly dementia patients often get lost due to lack of a sense of direction. This may put them at risk and cause their families much worry. The goal of this study is to use information technology to enhance the professional judgment of caregivers, strengthen internal safety monitoring at care organizations, and improve the quality of care for dementia patients. An eXtensible-Markup-Language-based dementia assessment system combining program code and assessment content is used to provide caregivers with better flexibility and real-time response ability. Beyond establishing long-term case files, the system can also perform data consistency analysis, strengthen caregivers' continuing education, improve caregivers' case judgment skills, and reduce the incidence of accidents due to neglect. This study also applies radio frequency identification (RFID) technology to the development of an indoor and outdoor active safety monitoring mechanism. The system can automatically remind caregivers whenever an elderly person approaches a dangerous area or strays too far. Apart from the use of different size tags, the realization of the system also employs the tame transformation signatures (TTS) algorithm to encrypt tag IDs and protect personal privacy. Clinical testing of the system showed that the indoor RFID reader has a response time of 0.5 s when sensing 40 tags, while the outdoor reader has a sensing time of approximately 5 s due to the need to save power. In the latter case, the system can ensure that elderly patients stay less than 15 m away from their caregivers. Patients were relatively willing to wear light tags. We also found that irritable patients with strong mobility were less compliant and often removed their own tags. Caregivers must provide active care and adopt various safety measures to protect the type of patients.

A pervasive health monitoring service system based on ubiquitous network technology.

OBJECTIVES: The phenomenon of aging society has derived problems such as shortage of medical resources and reduction of quality in healthcare services. METHOD: This paper presents a system infrastructure for pervasive and long-term healthcare applications, i.e. a ubiquitous network composed of wireless local area network (WLAN) and cable television (CATV) network serving as a platform for monitoring physiological signals. Users can record vital signs including heart rate, blood pressure, and body temperature anytime either at home or at frequently visited public places in order to create a personal health file. RESULTS: The whole system was formally implemented in December 2004. Analysis of 2000 questionnaires indicates that 85% of users were satisfied with the provided community-wide healthcare services. Among the services provided by our system, health consultation services offered by family doctors was rated the most important service by 17.9% of respondents, and was followed by control of one's own health condition (16.4% of respondents). Convenience of data access was rated most important by roughly 14.3% of respondents. DISCUSSION/CONCLUSION: We proposed and implemented a long-term healthcare system integrating WLAN and CATV networks in the form of a ubiquitous network providing a service platform for physiological monitoring. This system can classify the health levels of the resident according to the variation tendency of his or her physiological signal for important reference of health management.

A mobile care system with alert mechanism.

Hypertension and arrhythmia are chronic diseases, which can be effectively prevented and controlled only if the physiological parameters of the patient are constantly monitored, along with the full support of the health education and professional medical care. In this paper, a role-based intelligent mobile care system with alert mechanism in chronic care environment is proposed and implemented. The roles in our system include patients, physicians, nurses, and healthcare providers. Each of the roles represents a person that uses a mobile device such as a mobile phone to communicate with the server setup in the care center such that he or she can go around without restrictions. For commercial mobile phones with Bluetooth communication capability attached to chronic patients, we have developed physiological signal recognition algorithms that were implemented and built-in in the mobile phone without affecting its original communication functions. It is thus possible to integrate several front-end mobile care devices with Bluetooth communication capability to extract patients' various physiological parameters [such as blood pressure, pulse, saturation of haemoglobin (SpO2), and electrocardiogram (ECG)], to monitor multiple physiological signals without space limit, and to upload important or abnormal physiological information to healthcare center for storage and analysis or transmit the information to physicians and healthcare providers for further processing. Thus, the physiological signal extraction devices only have to deal with signal extraction and wireless transmission. Since they do not have to do signal processing, their form factor can be further reduced to reach the goal of microminiaturization and power saving. An alert management mechanism has been included in back-end healthcare center to initiate various strategies for automatic emergency alerts after receiving emergency messages or after automatically recognizing emergency messages. Within the time intervals in system setting, according to the medical history of a specific patient, our prototype system can inform various healthcare providers in sequence to provide healthcare service with their reply to ensure the accuracy of alert information and the completeness of early warning notification to further improve the healthcare quality. In the end, with the testing results and performance evaluation of our implemented system prototype, we conclude that it is possible to set up a complete intelligent healt care chain with mobile monitoring and healthcare service via the assistance of our system.

An automatic analysis method for detecting and eliminating ECG artifacts in EEG.

An automated method for detecting and eliminating electrocardiograph (ECG) artifacts from electroencephalography (EEG) without an additional synchronous ECG channel is proposed in this paper. Considering the properties of wavelet filters and the relationship between wavelet basis and characteristics of ECG artifacts, the concepts for selecting a suitable wavelet basis and scales used in the process are developed. The analysis via the selected basis is without suffering time shift for decomposition and detection/elimination procedures after wavelet transformation. The detection rates, above 97.5% for MIT/BIH and NTUH recordings, show a pretty good performance in ECG artifact detection and elimination.

Wireless health care service system for elderly with dementia.

The purpose of this paper is to integrate the technologies of radio frequency identification, global positioning system, global system for mobile communications, and geographic information system (GIS) to construct a stray prevention system for elderly persons suffering from dementia without interfering with their activities of daily livings. We also aim to improve the passive and manpowered way of searching the missing patient with the help of the information technology. Our system provides four monitoring schemes, including indoor residence monitoring, outdoor activity area monitoring, emergency rescue, and remote monitoring modes, and we have developed a service platform to implement these monitoring schemes. The platform consists of a web service server, a database server, a message controller server, and a health-GIS (H-GIS) server. Family members or volunteer workers can identify the real-time positions of missing elderly using mobile phone, PDA, Notebook PC, and various mobile devices through the service platform. System performance and reliability is analyzed. Experiments performed on four different time slots, from three locations, through three mobile telecommunication companies show that the overall transaction time is 34 s and the average deviation of the geographical location is about 8 m. A questionnaire surveyed by 11 users show that eight users are satisfied with the system stability and 10 users would like to carry the locating device themselves, or recommend it to their family members.

Design and implementation of a mobile-care system over wireless sensor network for home healthcare applications.

According to home healthcare requirement of chronic patients, this paper proposes a mobile-care system integrated with a variety of vital-sign monitoring, where all the front-end vital-sign measuring devices are portable and have the ability of short-range wireless communication. In order to make the system more suitable for home applications, the technology of wireless sensor network is introduced to transmit the captured vital signs to the residential gateway by means of multi-hop relay. Then the residential gateway uploads data to the care server via Internet to carry out patient's condition monitoring and the management of pathological data. Furthermore, the system is added in the alarm mechanism, which the portable care device is able to immediately perceive the critical condition of the patient and to send a warning message to medical and nursing personnels in order to achieve the goal of prompt rescue.