Review of machine learning solutions for eating disorders.

BACKGROUND: Eating Disorders (EDs) are one of the most complex psychiatric disorders, with significant impairment of psychological and physical health, and psychosocial functioning, and are associated with low rates of early detection, low recovery and high relapse rates. This underscores the need for better diagnostic and treatment methods. OBJECTIVE: This narrative review explores current Machine Learning (ML) and Artificial Intelligence (AI) applications in the domain of EDs, with a specific emphasis on clinical management in treatment settings. The primary objective are to (i) decrease the knowledge gap between ED researchers and AI-practitioners, by presenting the current state-of-the-art AI applications (including models for causality) in different ED use-cases; (ii) identify limitations of these existing AI interventions and how to address them. RESULTS: AI/ML methods have been applied in different ED use-cases, including ED risk factor identification and incidence prediction (including the analysis of social media content in the general population), diagnosis, monitoring patients and treatment response and prognosis in clinical populations. A comparative analysis of AI-techniques deployed in these use-cases have been performed, considering factors such as complexity, flexibility, functionality, explainability and adaptability to healthcare constraints. CONCLUSION: Multiple restrictions have been identified in the existing methods in ML and Causality in terms of achieving actionable healthcare for ED, like lack of good quality and quantity of data for models to train on, while requiring models to be flexible, high-performing, yet being explainable and producing counterfactual explanations, for ensuring the fairness and trustworthiness of its decisions. We conclude that to overcome these limitations and for future AI research and application in clinical management of ED, (i) careful considerations are required with regards to AI-model selection, and (ii) joint efforts from ED researcher and patient community are essential in building better quality and quantity of dedicated ED datasets and secure AI-solution framework.

Vitamin C and epigenetics: A short physiological overview.

In recent years, ascorbic acid (vitamin C) has acquired great interest due to its multiple functions, which results in homeostasis of normal tissues and organs. On the other hand, it has been shown that epigenetic modifications may have an important role in various diseases and therefore are a focus of the extraordinary investigation. Ascorbic acid serves as a cofactor for ten-eleven translocation dioxygenases, which are responsible for deoxyribonucleic acid methylation. Also, vitamin C is required for histone demethylation, since it acts as a cofactor of Jumonji C-domain-containing histone demethylases. It seems that vitamin C may be a mediator between the environment and the genome. The precise and multistep mechanism of ascorbic acid in epigenetic control is still not definitely determined. This article intends to provide the basic and newly discovered functions of vitamin C that are related to epigenetic control. Also, this article will help us to better understand the functions of ascorbic acid and will provide the possible implications of this vitamin in the regulation of epigenetic modifications.

Tracking of Nutritional Intake Using Artificial Intelligence.

In this short communication paper, we present the results we achieved for automated calorie intake measurement for patients with obesity or eating disorders. We demonstrate feasibility of applying deep learning based image analysis to a single picture of a food dish to recognize food types and make a volume estimation.

Generating High-Resolution 3D Faces and Bodies Using VQ-VAE-2 with PixelSNAIL Networks on 2D Representations.

Modeling and representing 3D shapes of the human body and face is a prominent field due to its applications in the healthcare, clothes, and movie industry. In our work, we tackled the problem of 3D face and body synthesis by reducing 3D meshes to 2D image representations. We show that the face can naturally be modeled on a 2D grid. At the same time, for more challenging 3D body geometries, we proposed a novel non-bijective 3D-2D conversion method representing the 3D body mesh as a plurality of rendered projections on the 2D grid. Then, we trained a state-of-the-art vector-quantized variational autoencoder (VQ-VAE-2) to learn a latent representation of 2D images and fit a PixelSNAIL autoregressive model to sample novel synthetic meshes. We evaluated our method versus a classical one based on principal component analysis (PCA) by sampling from the empirical cumulative distribution of the PCA scores. We used the empirical distributions of two commonly used metrics, specificity and diversity, to quantitatively demonstrate that the synthetic faces generated with our method are statistically closer to real faces when compared with the PCA ones. Our experiment on the 3D body geometry requires further research to match the test set statistics but shows promising results.

Decreased Activity of Erythrocyte Catalase and Glutathione Peroxidase in Patients with Schizophrenia.

Background and Objectives: Catalase and glutathione peroxidase (GPx) are important antioxidant enzymes that break down hydrogen peroxide (H2O2) in order to control its intracellular concentration, thus enabling its physiological role and preventing toxic effects. A lack or disruption of their function leads to the accumulation of hydrogen peroxide and the occurrence of oxidative stress. Accumulating studies have shown that the activities of key antioxidant enzymes are impaired in patients with schizophrenia. Since the published results are contradictory, and our previous studies found significantly higher erythrocyte superoxide dismutase (SOD) activity in patients with schizophrenia, the aim of this study was to determine the activity of enzymes that degrade hydrogen peroxide in the same group of patients, as well as to examine their dependence on clinical symptoms, therapy, and parameters associated with this disease. Materials and Methods: Catalase and GPx activities were determined in the erythrocytes of 68 inpatients with schizophrenia and 59 age- and gender-matched healthy controls. The clinical assessment of patients was performed by using the Positive and Negative Syndrome Scale (PANSS). The catalase activity was measured by the kinetic spectrophotometric method, while the GPx activity was determined by the commercially available Ransel test. Results: Erythrocyte catalase and GPx activities were significantly lower (p < 0.001 and p < 0.01, respectively) in subjects with schizophrenia than they were in healthy individuals. Lower catalase activity does not depend on heredity, disease onset, the number of episodes, or disease duration, while GPx activity showed significant changes in patients who had more than one episode and in those who had been suffering from the disease for over a year. Significantly lower catalase activity was noted in the PANSS(+/−) group in comparison with the PANSS(+) and PANSS(−) groups. The lowest catalase activity was found in subjects who were simultaneously treated with first- and second-generation antipsychotics; this was significantly lower than it was in those who received only one class of antipsychotics. Conclusion: These results indicate the presence of oxidative stress in the first years of clinically manifested schizophrenia and its dependence on the number of psychotic episodes, illness duration, predominant symptomatology, and antipsychotic medication.

Development of Novel Therapeutics for Schizophrenia Treatment Based on a Selective Positive Allosteric Modulation of α1-Containing GABAARs-In Silico Approach.

For the development of atypical antipsychotics, the selective positive allosteric modulation of the ionotropic GABAA receptor (GABAAR) has emerged as a promising approach. In the presented research, two unrelated methods were used for the development of QSAR models for selective positive allosteric modulation of 1-containing GABAARs with derivatives of imidazo [1,2-a]-pyridine. The development of conformation-independent QSAR models, based on descriptors derived from local molecular graph invariants and SMILES notation, was achieved with the Monte Carlo optimization method. From the vast pool of 0D, 1D, and 2D molecule descriptors, the GA-MLR method developed additional QSAR models. Various statistical methods were utilised for the determination of the developed models' robustness, predictability, and overall quality, and according to the obtained results, all QSAR models are considered good. The molecular fragments that have a positive or negative impact on the studied activity were obtained from the studied molecules' SMILES notations, and according to the obtained results, nine novel compounds were designed. The binding affinities to GABAAR of designed compounds were assessed with the application of molecular docking studies and the obtained results showed a high correlation with results obtained from QSAR modeling. To assess all designed molecules' "drug-likeness", their physicochemical descriptors were computed and utilised for the prediction of medicinal chemistry friendliness, pharmacokinetic properties, ADME parameters, and druglike nature.

Towards Effective Patient Simulators.

In this paper we give an overview of the field of patient simulators and provide qualitative and quantitative comparison of different modeling and simulation approaches. Simulators can be used to train human caregivers but also to develop and optimize algorithms for clinical decision support applications and test and validate interventions. In this paper we introduce three novel patient simulators with different levels of representational accuracy: HeartPole, a simplistic transparent rule-based system, GraphSim, a graph-based model trained on intensive care data, and Auto-ALS-an adjusted version of an educational software package used for training junior healthcare professionals. We provide a qualitative and quantitative comparison of the previously existing as well as proposed simulators.

Impact of polymer type, ASD loading and polymer-drug ratio on ASD tablet disintegration and drug release.

Amorphous solid dispersion (ASD) has become an attractive strategy to enhance solubility and bioavailability of poorly water-soluble drugs. To facilitate oral administration, ASDs are commonly incorporated into tablets. Disintegration and drug release from ASD tablets are thus critical for achieving the inherent solubility advantage of amorphous drugs. In this work, the impact of polymer type, ASD loading in tablet and polymer-drug ratio in ASD on disintegration and drug release of ASD tablets was systematically studied. Two hydrophilic polymers PVPVA and HPMC and one relatively hydrophobic polymer HPMCAS were evaluated. Dissolution testing was performed, and disintegration time was recorded during dissolution testing. As ASD loading increased, tablet disintegration time increased for all three polymer-based ASD tablets, and this effect was more pronounced for hydrophilic polymer-based ASD tablets. As polymer-drug ratio increased, tablet disintegration time increased for hydrophilic polymer-based ASD tablets, however, it remained short and largely unchanged for HPMCAS-based ASD tablets. Consequently, at high ASD loadings or high polymer-drug ratios, HPMCAS-based ASD tablets showed faster drug release than PVPVA- or HPMC-based ASD tablets. These results were attributed to the differences between polymer hydrophilicities and viscosities of polymer aqueous solutions. This work is valuable for understanding the disintegration and drug release of ASD tablets and provides insight to ASD composition selection from downstream tablet formulation perspective.

Artificial Intelligence in Clinical Health Care Applications: Viewpoint.

The idea of artificial intelligence (AI) has a long history. It turned out, however, that reaching intelligence at human levels is more complicated than originally anticipated. Currently, we are experiencing a renewed interest in AI, fueled by an enormous increase in computing power and an even larger increase in data, in combination with improved AI technologies like deep learning. Healthcare is considered the next domain to be revolutionized by artificial intelligence. While AI approaches are excellently suited to develop certain algorithms, for biomedical applications there are specific challenges. We propose six recommendations-the 6Rs-to improve AI projects in the biomedical space, especially clinical health care, and to facilitate communication between AI scientists and medical doctors: (1) Relevant and well-defined clinical question first; (2) Right data (ie, representative and of good quality); (3) Ratio between number of patients and their variables should fit the AI method; (4) Relationship between data and ground truth should be as direct and causal as possible; (5) Regulatory ready; enabling validation; and (6) Right AI method.

Nanoliposome-encapsulated ellagic acid prevents cyclophosphamide-induced rat liver damage.

In this study, we aimed to evaluate whether the encapsulation of ellagic acid (EA) into nanoliposomes would improve its potential in preventing cyclophosphamide-induced liver damage. Stability and antioxidative potential of free and encapsulated EA were determined. Experimental study conducted in vivo included ten groups of rats treated with cyclophosphamide and ellagic acid in its free and encapsulated form during 5 days. The protective effect of EA in its free and encapsulated form was determined based on serum liver function, liver tissue antioxidative capacities, and oxidative tissue damage parameters. Also, tissue morphological changes following cyclophosphamide administration were studied using standard histopathological and immunohistochemical analyses. The encapsulation of EA significantly prevented its degradation and improved its antioxidant properties in in vitro conditions. In in vivo experiments in both forms of EA were found to prevent rat liver damage induced by cyclophosphamide estimated through the changes in serum liver-damage parameters and tissue antioxidant capacities, as well as based on oxidatively modified lipids and proteins. Also, changes in morphology of liver cells and the expressions of Bcl-2, HIF-1α, and CD15 molecules in livers of animals of different experimental groups are in accordance with the obtained biochemical parameters. Thus, the encapsulation process might be effective in preventing EA from different environmental influences and could significantly increase its hepatoprotective potential. The encapsulation could prevent ellagic acid degradation and might deliver this potent compound to its target tissue in significantly larger quantities than when it is administered in its free form.

Making sense of big data in health research: Towards an EU action plan.

Medicine and healthcare are undergoing profound changes. Whole-genome sequencing and high-resolution imaging technologies are key drivers of this rapid and crucial transformation. Technological innovation combined with automation and miniaturization has triggered an explosion in data production that will soon reach exabyte proportions. How are we going to deal with this exponential increase in data production? The potential of "big data" for improving health is enormous but, at the same time, we face a wide range of challenges to overcome urgently. Europe is very proud of its cultural diversity; however, exploitation of the data made available through advances in genomic medicine, imaging, and a wide range of mobile health applications or connected devices is hampered by numerous historical, technical, legal, and political barriers. European health systems and databases are diverse and fragmented. There is a lack of harmonization of data formats, processing, analysis, and data transfer, which leads to incompatibilities and lost opportunities. Legal frameworks for data sharing are evolving. Clinicians, researchers, and citizens need improved methods, tools, and training to generate, analyze, and query data effectively. Addressing these barriers will contribute to creating the European Single Market for health, which will improve health and healthcare for all Europeans.

Personal health data: patient consent in information age.

In this paper we report on findings related to treatment of patient consent in various circumstances and geographic domains; explore transfer of health data between custodians and geo-political entities; and emphasize importance of educating general public about issues related to handling health data. A specific set of questions about consent/legislation and related issues in the Canada, the USA and the EU are addressed in an attempt to answer them systematically. This comparison identifies similarities and differences along a set of dimensions.

End-to-end security for personal telehealth.

Personal telehealth is in rapid development with innovative emerging applications like disease management. With personal telehealth people participate in their own care supported by an open distributed system with health services. This poses new end-to-end security and privacy challenges. In this paper we introduce new end-to-end security requirements and present a design for consent management in the context of the Continua Health Alliance architecture. Thus, we empower patients to control how their health information is shared and used in a personal telehealth eco-system.

Emergency access to protected health records.

Digital Rights Management (DRM) schemes are receiving increased attention in the healthcare domain for the protection of sensitive health records as they offer security against insider attacks and advance protection features such as usage control. However, to be accepted by health care providers, a DRM solution has to fulfill specific healthcare requirements including emergency access. In this paper, we propose such DRM solution that can be deployed in highly distributed environments of electronic or personal health record infrastructures.

Reliable personal health records.

A number of applications based on personal health records (PHRs) are emerging in the field of health care and wellness. PHRs empower patients by giving them control over their health data. Health data for PHRs can be supplied by patients, wellness providers and health care providers. Health care providers may use the PHRs to provide medical care. Unfortunately, the quality of the health data in PHRs cannot be guaranteed in all cases. For example, consider cases where non-professionals such as patients and wellness providers supply data. To address this problem, we present in this paper a system that provides health care professionals with an indication of the quality of health data in a PHR. This indication is based on the reputation of the supplier and on metadata provided by measurement devices. The proposed reputation system mimics the way in which trust in health data and their suppliers is built in the real world. The system introduces minimal overhead for health care providers and patients.