Smart Health最新文献

Devices characterizing diabetic foot ulcers and other wounds currently fall into two categories. Expensive clinically-oriented devices that use mature technologies such as X-ray CT and hyperspectral imaging or low-cost solutions that leverage deep learning to infer wound characterization from conventional smartphone camera images or simple surrogate markers. Mature medical-grade devices are too expensive for primary care and assisted living facilities. Low-cost solutions rely too much on indirect statistical inference to be clinically suitable. Therefore, we propose a device that leverages mature, clinically suitable optical technologies to provide a solution for these facilities. Recognizing that individual combinations of 1–2 bands of active illumination are used individually to capture pulsation, vascular, and oxygenation images. We combine all these bands into a single multispectral lighting source to create a multi-functional, reliable device for wound assessment. We selected these bands to leverage CMOS cameras near orthogonality between the RGB channels and leverage that CMOS cameras can also sense near IR light if a filter is not present, reducing overall system complexity and needed bands. For each function, the necessary lights are turned on, and the captured raw video is then fed to the corresponding sequence of image processing steps. No deep learning models are used, so large training datasets are not required. Our device is also small, lightweight, and handheld.

The proactive identification of behavioral health incidents concerns from police reports is a critical yet underexplored area. The law enforcement officers provide follow-up services to improve community life by manually analyzing and identifying generated public narrative reports after the 911 incident calls. Therefore, automatically identifying these behavioral health calls from public narrative reports helps reduce the current manual labor-intensive identification process for law enforcement officers. In this work, we introduce a novel, multi-faceted approach that combines manual expert annotations, natural language processing (NLP), and cutting-edge machine learning strategies to classify and understand these incidents within police narratives efficiently. Our proposed method retrieves relevant narrative reports utilizing domain knowledge as behavioral health cues as terms/keywords. Our approach automatically extracts different categories of behavioral health cues by utilizing the limited domain knowledge enabled behavioral health terms using a cosine similarity-based thresholding approach. The behavioral health classification model employs an automatic attention-aware feature representation of terms/keywords, categories, and narrative reports to identify behavioral health cases with an accuracy of 85%. Extensive evaluation shows that our proposed model outperforms all the state-of-the-art models by approximately 4%.

Emotional states, as indicators of affect, are pivotal to overall health, making their accurate prediction before onset crucial. Current studies are primarily centered on immediate short-term affect detection using data from wearable and mobile devices. These studies typically focus on objective sensory measures, often neglecting other forms of self-reported information like diaries and notes. In this paper, we propose a multimodal deep learning model for affect status forecasting. This model combines a transformer encoder with a pre-trained language model, facilitating the integrated analysis of objective metrics and self-reported diaries. To validate our model, we conduct a longitudinal study, enrolling college students and monitoring them over a year, to collect an extensive dataset including physiological, environmental, sleep, metabolic, and physical activity parameters, alongside open-ended textual diaries provided by the participants. Our results demonstrate that the proposed model achieves predictive accuracy of 82.50% for positive affect and 82.76% for negative affect, a full week in advance. The effectiveness of our model is further elevated by its explainability.

The profound impact of food on health necessitates advanced nutrition-oriented food recommendation services. Conventional methods often lack the crucial elements of personalization, explainability, and interactivity. While Large Language Models (LLMs) bring interpretability and explainability, their standalone use falls short of achieving true personalization. In this paper, we introduce ChatDiet, a novel LLM-powered framework designed specifically for personalized nutrition-oriented food recommendation chatbots. ChatDiet integrates personal and population models, complemented by an orchestrator, to seamlessly retrieve and process pertinent information. The personal model leverages causal discovery and inference techniques to assess personalized nutritional effects for a specific user, whereas the population model provides generalized information on food nutritional content. The orchestrator retrieves, synergizes and delivers the output of both models to the LLM, providing tailored food recommendations designed to support targeted health outcomes. The result is a dynamic delivery of personalized and explainable food recommendations, tailored to individual user preferences. Our evaluation of ChatDiet includes a compelling case study, where we establish a causal personal model to estimate individual nutrition effects. Our assessments, including a food recommendation test showcasing a 92$\text{\%}$ effectiveness rate, coupled with illustrative dialogue examples, underscore ChatDiet’s strengths in explainability, personalization, and interactivity.

Decoding eye movements from non-invasive electroencephalography (EEG) data is a challenging yet vital task for both scientific and practical purposes, especially for identifying neurodegenerative disorders like Alzheimer’s disease (AD). Our research tackles this complexity by adapting inverse reinforcement learning (IRL), a machine learning method, to infer decision-making strategies from observed behaviors. We implement this to understand the processes driving eye direction and movements during diverse cognitive tasks, providing new insights into this field. Our paper begins with a detailed description of the procedures for collecting and preprocessing EEG data related to gaze behavior. We then elaborate on the development of an IRL framework designed to predict the spatial and temporal dynamics of eye movements (scanpaths) in participants engaged in cognitive tasks of varying complexity. Our model is tailored to accommodate the complexities inherent in neural signals and the stochastic nature of human gaze patterns. Our research findings underscore IRL’s effectiveness in precisely forecasting gaze patterns based on a combination of EEG and image data. The correlation between the model’s predictions and the actual gaze behavior observed in controlled experiments reinforces the utility of IRL in cognitive neuroscience research. Notably, our IRL-EEG models demonstrated superior performance, especially in more complex cognitive tasks. We further delve into the implications of our results for enhancing the understanding of neural mechanisms that govern gaze behavior.

Numerous studies have been conducted on the connection between music and the brain, and it has been established that listening to music directly affects brain activity and stimulation. The potential benefits of music therapy, which uses music as a tool for healing and fostering well-being, have come to light in a number of circumstances. However, there is a gap in understanding the effects of Indian classical music (ICM) on the brain and its therapeutic applications. Yaman and Puria Dhanashree were the two chosen ragas, which share same notes (swaras) and differs in two of their counter notes. The brain responses are captured from five volunteers through 24 channel Electroencephalogram (EEG) cap using a smartphone, which is utilized to allocate electrodes to different regions of the brain. In this work, different automated approaches for identifying brain regions evoked to live ICM stimuli are proposed, considering input and output uncertainties. These approaches are based on automated energy and Mahalanobis distance measurements, and, a region-specific real-time algorithm based on eigen perturbation, which provides a measure to capture the time evolution of brain activity. This identification is relevant in understanding dynamic changes in brain responses during musical experiences providing a more comprehensive perception and processing of ICM in the human brain. Also, significant change in beta band power reduction was observed after music. These approaches can help integrate it into evidence-based music therapy for cognitive, emotional, and psychological conditions. The findings of this study provide evidence indicating ragas activate different brain regions based on listener’s musical knowledge and is a first step for mhealth based applications.

The Internet of Things (IoT) applied to the health area is in significant growth, with companies putting effort into developing specialized devices. Remote patient healthcare monitoring, in particular, benefits society as it unburdens hospitals and helps patients with chronic diseases. Analyzing the health status with IoT and Artificial Intelligence (AI) is new in the digital community. The literature yet presents a limited number of references explicitly concerning the topic of qualified data acquisition. In this sense, the present literature review aims to update the joint subject of IoT and vital signs, seeking to understand the state-of-the-art and future directions. We have analyzed 78 articles and IoT manufacturer websites that address vital signs collection to answer a group of primary and specific questions. In particular, we revisited architectures, communication protocols, data acquisition mechanisms, evaluation metrics, and how to efficiently transfer data through the lens of sensors, actuators, and healthcare. Currently, two themes are considered as promising directions for studies in the joint area of IoT and vital sign-based healthcare monitoring. The first is the connection promotion between third-party applications and IoT devices to collect and process time-critical data with the support of edge, fog, and cloud infrastructures. The second theme again brings the focus to data compression methodologies since monitoring vital signs in a smart city geographical area naturally requires strategies to optimize network bandwidth consumption and data storage on computational resources. Moreover, both themes are directly linked to energy-saving approaches and quality of service (QoS) for efficient patient healthcare checking.

Healthcare-associated infections (HAIs), or nosocomial infections, refer to patients getting new infections while getting treatment for an existing condition in a healthcare facility. HAI poses a significant challenge in healthcare delivery that results in higher rates of mortality and morbidity as well as a longer duration of hospital stay. While the real cause of HAI in a hospital varies widely and in most cases untraceable, it is popularly believed that patient flow in a hospital—which hospital units patients visit and where they spend the most time since their admission into the hospital—can trace back to HAI incidence in the hospital. Based on this observation, we, in this paper, model and simulate patient flow in an emergency department of a hospital and then utilize the developed model to study HAI incidence therein. We obtain (a) a flowchart of patient movement (admission to discharge) and (b) anonymous patient data from University Health Medical Center for a duration of 11 months (Aug 2022–June 2023). Based on these data, we develop and validate the patient flow model. Our model captures patient movement in different areas of a typical emergency department, such as triage, waiting room, and minor procedure rooms. We employ the discrete-event simulation (DES) technique to model patient flow and associated HAI infections using the simulation software, Anylogic. Our simulation results show that the rates of HAI incidence are proportional to both the specific areas patients occupy and the duration of their stay. By utilizing our model, hospital administrators and infection control teams can implement targeted strategies to reduce the incidence of HAI and enhance patient safety, ultimately leading to improved healthcare outcomes and more efficient resource allocation.

Human Body Communication (HBC) is an emerging technology that uses the human body as a communication channel. It offers significant advantages over traditional RF techniques in terms of power consumption and security. In recent developments, Electro-quasistatic HBC (EQS-HBC) in the frequency band below 1 MHz has been employed to enable communication without signal radiation beyond the body, effectively turning the body into a wired communication medium. This paper delves into the application of the EQS band for HBC. Experimental results show the determinantal effect of intermittent noise that sporadically disrupts communications across the band of interest. To address this challenge, we introduce an innovative frequency-hopping transceiver system, which allows the transmitter to seamlessly adapt to different frequencies. In addition, we present a miniature transmitter design, incorporating a simplified micro-controller unit (MCU) to facilitate the implementation of HBC. Furthermore, to validate this proposed design, we present a fully functional prototype of an HBC system that effectively employs frequency hopping techniques for practical applications.

Monitoring blood pressure (BP) is an essential component of evaluating cardiovascular health, aiding in the early detection and management of hypertension-related complications. Traditional methods of BP measurement often involve invasive or cumbersome devices, leading to discomfort and reduced compliance. We propose a framework to monitor BP non-invasively, analyzing the face video captured by a webcam or smartphone camera leveraging the relationship of image-based Pulse Transit Time (iPTT) and Heart Rate Variability (HRV) with BP. We have built a dataset of 90 sets of collected videos using a mobile phone front camera and BP data from a standard digital BP monitor from 12 individuals from an approved Institutional Review Board (IRB) to evaluate our system. We have got a Mean Absolute Error (MAE) of $10.35+/-2.5$ mmHg for systolic BP (SBP) and $7.8+/-1.5$ mmHg for diastolic BP (DBP) while using the HRV representation RMSSD. On the other hand, an MAE of $8.25+/-3.5$ mmHg for SBP and $7.7+/-2.5$ mmHg for DBP while using the HRV representation SDRR. Finally, we have developed a framework and built a real-time system to monitor BP as a mobile and web-based application that can facilitate early detection of trends and anomalies, allowing healthcare providers to intervene promptly and personalize treatment plans.