Search

Zinc deficiency is linked to poor prognosis in COVID-19 patients while clinical trials with zinc demonstrate better clinical outcomes. The molecular targets and mechanistic details of the anti-coronaviral activity of zinc remain obscure. We show that zinc not only inhibits the SARS-CoV-2 main protease (Mpro) with nanomolar affinity, but also viral replication. We present the first crystal structure of the Mpro–Zn2+ complex at 1.9 Å and provide the structural basis of viral replication inhibition. We show that Zn2+ coordinates with the catalytic dyad at the enzyme active site along with two previously unknown water molecules in a tetrahedral geometry to form a stable inhibited Mpro–Zn2+ complex. Further, the natural ionophore quercetin increases the anti-viral potency of Zn2+. As the catalytic dyad is highly conserved across SARS-CoV, MERS-CoV and all variants of SARS-CoV-2, Zn2+ mediated inhibition of Mpro may have wider implications. [ABSTRACT FROM AUTHOR]

Plant dehydroascorbate reductases (DHARs) are only known as soluble antioxidant enzymes of the ascorbate-glutathione pathway. They recycle ascorbate from dehydroascorbate, thereby protecting plants from oxidative stress and the resulting cellular damage. DHARs share structural GST fold with human chloride intracellular channels (HsCLICs) which are dimorphic proteins that exists in soluble enzymatic and membrane integrated ion channel forms. While the soluble form of DHAR has been extensively studied, the existence of a membrane integrated form remains unknown. We demonstrate for the first time using biochemistry, immunofluorescence confocal microscopy, and bilayer electrophysiology that Pennisetum glaucum DHAR (PgDHAR) is dimorphic and is localized to the plant plasma membrane. In addition, membrane translocation increases under induced oxidative stress. Similarly, HsCLIC1 translocates more into peripheral blood mononuclear cells (PBMCs) plasma membrane under induced oxidative stress conditions. Moreover, purified soluble PgDHAR spontaneously inserts and conducts ions in reconstituted lipid bilayers, and the addition of detergent facilitates insertion. In addition to the well-known soluble enzymatic form, our data provides conclusive evidence that plant DHAR also exists in a novel membrane-integrated form. Thus, the structure of DHAR ion channel form will help gain deeper insights into its function across various life forms. [Display omitted] • Plant dehydroascorbate reductase (DHAR) is dimorphic and bifunctional. • Pennisetum glaucum DHAR (PgDHAR) moonlights as a membrane integrated channel. • PgDHAR and HsCLIC1 membrane translocation increases under induced oxidative stress. • Soluble PgDHAR spontaneously inserts and conducts ions in planar lipid bilayers. • Detergent facilitates PgDHAR channel insertion. [ABSTRACT FROM AUTHOR]

Biomedical systems have been using ontology matching as a primary technique for heterogeneity resolution. However, the natural intricacy and vastness of biomedical data have compelled biomedical ontologies to become large-scale and complex; consequently, biomedical ontology matching has become a computationally intensive task. Our parallel heterogeneity resolution system, i.e., SPHeRe, is built to cater the performance needs of ontology matching by exploiting the parallelism-enabled multicore nature of today's desktop PC and cloud infrastructure. In this paper, we present the execution and evaluation results of SPHeRe over large-scale biomedical ontologies. We evaluate our system by integrating it with the interoperability engine of a clinical decision support system (CDSS), which generates matching requests for large-scale NCI, FMA, and SNOMEDCT biomedical ontologies. Results demonstrate that our methodology provides an impressive performance speedup of 4.8 and 9.5 times over a quad-core desktop PC and a four virtual machine (VM) cloud platform, respectively. [ABSTRACT FROM AUTHOR]

The Semantic Web is dependent on extensive knowledge management by interlinking resources on the web using matching techniques. This role is played by the progressing domain of ontology matching, by introducing ontology-matching tools. The focus of these matching tools is limited to matching techniques and automation, rather than expressive formal representation of alignments. We propose Mediation Bridge Ontology (MBO), an expressive alignment representation ontology used to store correspondences between matching ontologies matched by our ontology-matching tool, System for Parallel Heterogeneity Resolution (SPHeRe). The MBO utilizes object-oriented design patterns and the proposed ontology-alignment design patterns to provide extendibility and reusability factors to SPHeRe system. We compared our proposed system with existing systems using Coupling Factor, Number of Polymorphic methods and Rate of Change metrics to support extendibility and reusability. These factors contribute to the overall objective of interoperability for knowledge management in the Semantic Web. [ABSTRACT FROM PUBLISHER]

Objective: Data interoperability among health information exchange (HIE) systems is a major concern for healthcare practitioners to enable provisioning of telemedicine-related services. Heterogeneity exists in these systems not only at the data level but also among different heterogeneous healthcare standards with which these are compliant. The relationship between healthcare organization data and different heterogeneous standards is necessary to achieve the goal of data level interoperability. We propose a personalized-detailed clinical model (P-DCM) approach for the generation of customized mappings that creates the necessary linkage between organization-conformed healthcare standards concepts and clinical model concepts to ensure data interoperability among HIE systems. Materials and Methods: We consider electronic health record (EHR) standards, open EHR, and HL7 CDA instances transformation using P-DCM. P-DCM concepts associated with open EHR and HL7 CDA help in transformation of instances among these standards. We investigated two datasets: (1) data of 100 diabetic patients, including 50 each of type 1 and type 2, from a local hospital in Korea and (2) data of a single Alzheimer's disease patient. P-DCMs were created for both scenarios, which provided the basis for deriving instances for HL7 CDA and open EHR standards. Results: For proof of concept, we present case studies of encounter information for type 2 diabetes mellitus patients and monitoring of daily routine activities of an Alzheimer's disease patient. These reflect P-DCM-based customized mappings generation with open EHR and HL7 CDA standards. Customized mappings are generated based on the relationship of P-DCM concepts with CDA and open EHR concepts. Conclusions: The objective of this work is to achieve semantic data interoperability among heterogeneous standards. This would lead to effective utilization of resources and allow timely information exchange among healthcare systems. [ABSTRACT FROM AUTHOR]

New features, design suggestions, and solutions for the next generation of cellular systems need to be tested in a variety of real-world deployments and use cases. 3GPP-compliant 3GPPcompliant system-level holistic and realistic simulators are urgently needed to evaluate the variety of AI-based network automation solutions that are being suggested in the literature. The Synthetic-NET simulator created at AI4networks Lab is presented in this publication. A python-based emulator that completely complies with 3GPP 5G standards update 15 and can be upgraded to future releases called Synthetic-NET, according to the authors. In comparison to other simulators, Synthetic-NET has a number of key advantages, including: 1) a modular design that makes it easier to cross-validate and upgrade to future releases; 2) a variety of propagation modelling options, including measurement-based, ray-tracing-based, and AI-based models; 3) the ability to import data sheets based on measurements of realistic base [5] station features, including such satellite and energy consumption patterns; and 4) sui generis support for a wide range of wireless protocols. The Synthetic-NET's ability to be utilised to test AI-based network automation solutions is another important feature of the product. Synthetic Net's built-in abilities to analyze and process large amounts of data and integrated access to Applications Of machine learning [12] contribute to this simplicity of use, which is the first python-based 5G emulator. A powerful platform for both academia and industry alike, the Synthetic-NET simulator is a powerful tool for experimenting of not only creative approaches for optimising the operation of both existing as well as starting to emerge wireless connections but also for developing AI-powered deep mechanisation in the years to come. [ABSTRACT FROM AUTHOR]

We analytically investigate a nonlinear fractional-order sine-Gordon (sG) equation. The derivatives considered herein, are taken in Caputo's sense. The Laplace transform together with the Adomian decomposition method (LADM) is applied to attain analytical approximation of the aforesaid equation. The sG equation having Caputo derivative is solved in the order of series solutions and the results are confirmed by considering two examples with appropriate initial conditions. The numerical simulations are accomplished to compare with the analytical approximations, where qualitatively better agreements are achieved. [ABSTRACT FROM AUTHOR]

Background: With the increasing use of electronic health records (EHRs), there is a growing need to expand the utilization of EHR data to support clinical research. The key challenge in achieving this goal is the unavailability of smart systems and methods to overcome the issue of data preparation, structuring, and sharing for smooth clinical research. Materials and Methods: We developed a robust analysis system called the smart extraction and analysis system (SEAS) that consists of two subsystems: (1) the information extraction system (IES), for extracting information from clinical documents, and (2) the survival analysis system (SAS), for a descriptive and predictive analysis to compile the survival statistics and predict the future chance of survivability. The IES subsystem is based on a novel permutation-based pattern recognition method that extracts information from unstructured clinical documents. Similarly, the SAS subsystem is based on a classification and regression tree (CART)-based prediction model for survival analysis. Results: SEAS is evaluated and validated on a real-world case study of head and neck cancer. The overall information extraction accuracy of the system for semistructured text is recorded at 99%, while that for unstructured text is 97%. Furthermore, the automated, unstructured information extraction has reduced the average time spent on manual data entry by 75%, without compromising the accuracy of the system. Moreover, around 88% of patients are found in a terminal or dead state for the highest clinical stage of disease (level IV). Similarly, there is an ∼36% probability of a patient being alive if at least one of the lifestyle risk factors was positive. Conclusion: We presented our work on the development of SEAS to replace costly and time-consuming manual methods with smart automatic extraction of information and survival prediction methods. SEAS has reduced the time and energy of human resources spent unnecessarily on manual tasks. [ABSTRACT FROM AUTHOR]

Recent years have witnessed a huge progress in the automatic identification of individual primitives of human behavior, such as activities or locations. However, the complex nature of human behavior demands more abstract contextual information for its analysis. This work presents an ontology-based method that combines low-level primitives of behavior, namely activity, locations and emotions, unprecedented to date, to intelligently derive more meaningful high-level context information. The paper contributes with a new open ontology describing both low-level and high-level context information, as well as their relationships. Furthermore, a framework building on the developed ontology and reasoning models is presented and evaluated. The proposed method proves to be robust while identifying high-level contexts even in the event of erroneously-detected low-level contexts. Despite reasonable inference times being obtained for a relevant set of users and instances, additional work is required to scale to long-term scenarios with a large number of users. [ABSTRACT FROM AUTHOR]

Background: The provision of health and wellness care is undergoing an enormous transformation. A key element of this revolution consists in prioritizing prevention and proactivity based on the analysis of people's conducts and the empowerment of individuals in their self-management. Digital technologies are unquestionably destined to be the main engine of this change, with an increasing number of domain-specific applications and devices commercialized every year; however, there is an apparent lack of frameworks capable of orchestrating and intelligently leveraging, all the data, information and knowledge generated through these systems.Methods: This work presents Mining Minds, a novel framework that builds on the core ideas of the digital health and wellness paradigms to enable the provision of personalized support. Mining Minds embraces some of the most prominent digital technologies, ranging from Big Data and Cloud Computing to Wearables and Internet of Things, as well as modern concepts and methods, such as context-awareness, knowledge bases or analytics, to holistically and continuously investigate on people's lifestyles and provide a variety of smart coaching and support services.Results: This paper comprehensively describes the efficient and rational combination and interoperation of these technologies and methods through Mining Minds, while meeting the essential requirements posed by a framework for personalized health and wellness support. Moreover, this work presents a realization of the key architectural components of Mining Minds, as well as various exemplary user applications and expert tools to illustrate some of the potential services supported by the proposed framework.Conclusions: Mining Minds constitutes an innovative holistic means to inspect human behavior and provide personalized health and wellness support. The principles behind this framework uncover new research ideas and may serve as a reference for similar initiatives. [ABSTRACT FROM AUTHOR]

In recent years, the focus of healthcare and wellness technologies has shown a significant shift towards personal vital signs devices. The technology has evolved from smartphone-based wellness applications to fitness bands and smartwatches. The novelty of these devices is the accumulation of activity data as their users go about their daily life routine. However, these implementations are device specific and lack the ability to incorporate multimodal data sources. Data accumulated in their usage does not offer rich contextual information that is adequate for providing a holistic view of a user’s lifelog. As a result, making decisions and generating recommendations based on this data are single dimensional. In this paper, we present our Data Curation Framework (DCF) which is device independent and accumulates a user’s sensory data from multimodal data sources in real time. DCF curates the context of this accumulated data over the user’s lifelog. DCF provides rule-based anomaly detection over this context-rich lifelog in real time. To provide computation and persistence over the large volume of sensory data, DCF utilizes the distributed and ubiquitous environment of the cloud platform. DCF has been evaluated for its performance, correctness, ability to detect complex anomalies, and management support for a large volume of sensory data. [ABSTRACT FROM AUTHOR]

With advanced technologies in hand, there exist potential applications and services built around monitoring activities of daily living (ADL) of elderly people at nursing homes. Most of the elderly people in these facilities are suffering from different chronic diseases such as dementia. Existing technologies are mainly focusing on non-medication interventions and monitoring of ADL for addressing loss of autonomy or well-being. Monitoring and managing ADL related to cognitive behaviors for non-medication intervention are very effective in improving dementia patients' conditions. However, cognitive functions of patients can be improved if appropriate recommendations of medications are delivered at a particular time. Previously we developed the Secured Wireless Sensor Network Integrated Cloud Computing for Ubiquitous-Life Care (SC3). SC3 services were limited to monitoring ADL of elderly people with Alzheimer's disease and providing non-medication recommendations to the patient. In this article, we propose a system called the Smart Clinical Decision Support System (CDSS) as an integral part of the SC3 platform. Using the Smart CDSS, patients are provided with access to medication recommendations of expert physicians. Physicians are provided with an interface to create clinical knowledge for medication recommendations and to observe the patient's condition. The clinical knowledge created by physicians as the knowledge base of the Smart CDSS produces recommendations to the caregiver for medications based on each patient's symptoms. [ABSTRACT FROM AUTHOR]

Recently, social media have been used by researchers to detect depressive symptoms in individuals using linguistic data from users' posts. In this study, we propose a framework to identify social information as a significant predictor of depression. Using the proposed framework, we develop an application called the Socially Mediated Patient Portal (SMPP), which detects depression-related markers in Facebook users by applying a data-driven approach with machine learning classification techniques. We examined a data set of 4350 users who were evaluated for depression using the Center for Epidemiological Studies Depression (CES-D) scale. From this analysis, we identified a set of features that can distinguish between individuals with and without depression. Finally, we identified the dominant features that adequately assess individuals with and without depression on social media. The model trained on these features will be helpful to physicians in diagnosing mental diseases and psychiatrists in analysing patient behaviour. [ABSTRACT FROM AUTHOR]

The lack of Interoperable healthcare data presents a major challenge, towards achieving ubiquitous health care. The plethora of diverse medical standards, rather than common standards, is widening the gap of interoperability. While many organizations are working towards a standardized solution, there is a need for an alternate strategy, which can intelligently mediate amongst a variety of medical systems, not complying with any mainstream healthcare standards while utilizing the benefits of several standard merging initiates, to eventually create digital health personas. The existence and efficiency of such a platform is dependent upon the underlying storage and processing engine, which can acquire, manage and retrieve the relevant medical data. In this paper, we present the Ubiquitous Health Profile (UHPr), a multi-dimensional data storage solution in a semi-structured data curation engine, which provides foundational support for archiving heterogeneous medical data and achieving partial data interoperability in the healthcare domain. Additionally, we present the evaluation results of this proposed platform in terms of its timeliness, accuracy, and scalability. Our results indicate that the UHPr is able to retrieve an error free comprehensive medical profile of a single patient, from a set of slightly over 116.5 million serialized medical fragments for 390,101 patients while maintaining a good scalablity ratio between amount of data and its retrieval speed. [ABSTRACT FROM AUTHOR]

The proposed Intelligent Medical Platform is a dialoguebased medical decision-making system that provides medical coaching and recommendation services, based on incremental learning methodology. The prototype demonstrates 90% accuracy for knowledge acquisition, 80% satisfaction level of user interaction with the system, and 95% accuracy for system integration with the legacy system. [ABSTRACT FROM AUTHOR]

Objective: The objective of this study is to help a team of physicians and knowledge engineers acquire clinical knowledge from existing practices datasets for treatment of head and neck cancer, to validate the knowledge against published guidelines, to create refined rules, and to incorporate these rules into clinical workflow for clinical decision support.Methods and Materials: A team of physicians (clinical domain experts) and knowledge engineers adapt an approach for modeling existing treatment practices into final executable clinical models. For initial work, the oral cavity is selected as the candidate target area for the creation of rules covering a treatment plan for cancer. The final executable model is presented in HL7 Arden Syntax, which helps the clinical knowledge be shared among organizations. We use a data-driven knowledge acquisition approach based on analysis of real patient datasets to generate a predictive model (PM). The PM is converted into a refined-clinical knowledge model (R-CKM), which follows a rigorous validation process. The validation process uses a clinical knowledge model (CKM), which provides the basis for defining underlying validation criteria. The R-CKM is converted into a set of medical logic modules (MLMs) and is evaluated using real patient data from a hospital information system.Results: We selected the oral cavity as the intended site for derivation of all related clinical rules for possible associated treatment plans. A team of physicians analyzed the National Comprehensive Cancer Network (NCCN) guidelines for the oral cavity and created a common CKM. Among the decision tree algorithms, chi-squared automatic interaction detection (CHAID) was applied to a refined dataset of 1229 patients to generate the PM. The PM was tested on a disjoint dataset of 739 patients, which gives 59.0% accuracy. Using a rigorous validation process, the R-CKM was created from the PM as the final model, after conforming to the CKM. The R-CKM was converted into four candidate MLMs, and was used to evaluate real data from 739 patients, yielding efficient performance with 53.0% accuracy.Conclusion: Data-driven knowledge acquisition and validation against published guidelines were used to help a team of physicians and knowledge engineers create executable clinical knowledge. The advantages of the R-CKM are twofold: it reflects real practices and conforms to standard guidelines, while providing optimal accuracy comparable to that of a PM. The proposed approach yields better insight into the steps of knowledge acquisition and enhances collaboration efforts of the team of physicians and knowledge engineers. [ABSTRACT FROM AUTHOR]

A huge array of personalized healthcare and wellness systems are introduced into the portfolio of digital health and quantified-self movement in recent years. These systems share common capabilities including self-tracking/monitoring and self-quantifications, based on the raw sensory data. These capabilities provide solid ground for the users to be more aware of their health; however, such measures are inefficient for changing the unhealthy habits of the users. In order to induce healthy habits in the users, a system must be capable of generating context-aware personalized recommendations. The main obstacle in this regard is the contextual interpretation of recommendations based on user's current context and contextual preferences. To resolve these issues, we propose a methodology of cross-context interpretation of recommendations (CCIR) for personalized health and wellness services. The CCIR method adds additional capabilities to the traditional reasoning methods and builds advanced form of the reasoning with the incorporation of contextual factors in the process of interpretations of the recommendations. With CCIR, the self-quantification systems can be enhanced to generate personalized recommendations in addition to tracking, quantifying, and monitoring user activities. In order to validate the proposed CCIR methodology, a set of 40 contextual scenarios and corresponding recommendations are presented for the evaluation collected from 40 different end users and 10 domain experts. Using chi-square test evaluation, the results demonstrated acceptable “goodness of fit” indices for the system developed on proposed CCIR methodology with respect to the end users’ opinion. Also from the statistical observation, it is found that there exists a higher level agreement towards the system between the participants of both end users and experts. [ABSTRACT FROM AUTHOR]

Hyperspectral image classification with a limited number of training samples without loss of accuracy is desirable, as collecting such data is often expensive and time-consuming. However, classifiers trained with limited samples usually end up with a large generalization error. To overcome the said problem, we propose a fuzziness-based active learning framework (FALF), in which we implement the idea of selecting optimal training samples to enhance generalization performance for two different kinds of classifiers, discriminative and generative (e.g. SVM and KNN). The optimal samples are selected by first estimating the boundary of each class and then calculating the fuzziness-based distance between each sample and the estimated class boundaries. Those samples that are at smaller distances from the boundaries and have higher fuzziness are chosen as target candidates for the training set. Through detailed experimentation on three publically available datasets, we showed that when trained with the proposed sample selection framework, both classifiers achieved higher classification accuracy and lower processing time with the small amount of training data as opposed to the case where the training samples were selected randomly. Our experiments demonstrate the effectiveness of our proposed method, which equates favorably with the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The emerging research on automatic identification of user's contexts from the cross-domain environment in ubiquitous and pervasive computing systems has proved to be successful. Monitoring the diversified user's contexts and behaviors can help in controlling lifestyle associated to chronic diseases using context-aware applications. However, availability of cross-domain heterogeneous contexts provides a challenging opportunity for their fusion to obtain abstract information for further analysis. This work demonstrates extension of our previous work from a single domain (i.e., physical activity) to multiple domains (physical activity, nutrition and clinical) for context-awareness. We propose multi-level Context-aware Framework (mlCAF), which fuses the multi-level cross-domain contexts in order to arbitrate richer behavioral contexts. This work explicitly focuses on key challenges linked to multi-level context modeling, reasoning and fusioning based on the mlCAF open-source ontology. More specifically, it addresses the interpretation of contexts from three different domains, their fusioning conforming to richer contextual information. This paper contributes in terms of ontology evolution with additional domains, context definitions, rules and inclusion of semantic queries. For the framework evaluation, multi-level cross-domain contexts collected from 20 users were used to ascertain abstract contexts, which served as basis for behavior modeling and lifestyle identification. The experimental results indicate a context recognition average accuracy of around 92.65% for the collected cross-domain contexts. [ABSTRACT FROM AUTHOR]

Public cloud storage services are becoming prevalent and myriad data sharing, archiving and collaborative services have emerged which harness the pay-as-you-go business model of public cloud. To ensure privacy and confidentiality often encrypted data is outsourced to such services, which further complicates the process of accessing relevant data by using search queries. Search over encrypted data schemes solve this problem by exploiting cryptographic primitives and secure indexing to identify outsourced data that satisfy the search criteria. Almost all of these schemes rely on exact matching between the encrypted data and search criteria. A few schemes which extend the notion of exact matching to similarity based search, lack realism as those schemes rely on trusted third parties or due to increase storage and computational complexity. In this paper we propose Oblivious Similarity based Search () for encrypted data. It enables authorized users to model their own encrypted search queries which are resilient to typographical errors. Unlike conventional methodologies, ranks the search results by using similarity measure offering a better search experience than exact matching. It utilizes encrypted bloom filter and probabilistic homomorphic encryption to enable authorized users to access relevant data without revealing results of search query evaluation process to the untrusted cloud service provider. Encrypted bloom filter based search enables to reduce search space to potentially relevant encrypted data avoiding unnecessary computation on public cloud. The efficacy of is evaluated on Google App Engine for various bloom filter lengths on different cloud configurations. [ABSTRACT FROM AUTHOR]

Finding appropriate evidence to support clinical practices is always challenging, and the construction of a query to retrieve such evidence is a fundamental step. Typically, evidence is found using manual or semi-automatic methods, which are time-consuming and sometimes make it difficult to construct knowledge-based complex queries. To overcome the difficulty in constructing knowledge-based complex queries, we utilized the knowledge base (KB) of the clinical decision support system (CDSS), which has the potential to provide sufficient contextual information. To automatically construct knowledge-based complex queries, we designed methods to parse rule structure in KB of CDSS in order to determine an executable path and extract the terms by parsing the control structures and logic connectives used in the logic. The automatically constructed knowledge-based complex queries were executed on the PubMed search service to evaluate the results on the reduction of retrieved citations with high relevance. The average number of citations was reduced from 56,249 citations to 330 citations with the knowledge-based query construction approach, and relevance increased from 1 term to 6 terms on average. The ability to automatically retrieve relevant evidence maximizes efficiency for clinicians in terms of time, based on feedback collected from clinicians. This approach is generally useful in evidence-based medicine, especially in ambient assisted living environments where automation is highly important. [ABSTRACT FROM AUTHOR]

Background: Diabetic Retinopathy (DR) is considered a pathology of retinal vascular complications, which stays in the top causes of vision impairment and blindness. Therefore, precisely inspecting its progression enables the ophthalmologists to set up appropriate next-visit schedule and cost-effective treatment plans. In the literature, existing work only makes use of numerical attributes in Electronic Medical Records (EMR) for acquiring such kind of DR-oriented knowledge through conventional machine learning techniques, which require an exhaustive job of engineering most impactful risk factors.Objective: In this paper, an approach of deep bimodal learning is introduced to leverage the performance of DR risk progression identification.Methods: In particular, we further involve valuable clinical information of fundus photography in addition to the aforementioned systemic attributes. Accordingly, a Trilogy of Skip-connection Deep Networks, namely Tri-SDN, is proposed to exhaustively exploit underlying relationships between the baseline and follow-up information of the fundus images and EMR-based attributes. Besides that, we adopt Skip-Connection Blocks as basis components of the Tri-SDN for making the end-to-end flow of signals more efficient during feedforward and backpropagation processes.Results: Through a 10-fold cross validation strategy on a private dataset of 96 diabetic mellitus patients, the proposed method attains superior performance over the conventional EMR-modality learning approach in terms of Accuracy (90.6%), Sensitivity (96.5%), Precision (88.7%), Specificity (82.1%), and Area Under Receiver Operating Characteristics (88.8%).Conclusions: The experimental results show that the proposed Tri-SDN can combine features of different modalities (i.e., fundus images and EMR-based numerical risk factors) smoothly and effectively during training and testing processes, respectively. As a consequence, with impressive performance of DR risk progression recognition, the proposed approach is able to help the ophthalmologists properly decide follow-up schedule and subsequent treatment plans. [ABSTRACT FROM AUTHOR]