Machine learning-assisted optimization of food-grade spirulina cultivation in seawater-based media: From laboratory to large-scale production.

The shortage of food and freshwater sources threatens human health and environmental sustainability. Spirulina grown in seawater-based media as a healthy food is promising and environmentally friendly. This study used three machine learning techniques to identify important cultivation parameters and their hidden interrelationships and optimize the biomass yield of Spirulina grown in seawater-based media. Through optimization of hyperparameters and features, eXtreme Gradient Boosting, along with the recursive feature elimination (RFE) model demonstrated optimal performance and identified 28 important features. Among them, illumination intensity and initial pH value were critical determinants of biomass, which impacted other features. Specifically, high initial pH values (>9.0) mainly increased biomass but also increased nutrient sedimentation and ammonia (NH3) losses. Both batch and continuous additions could decrease nutrient losses by increasing their availability in the seawater-based media. When illumination intensity exceeded 200 µmol photons/m2/s, it amplified the growth of Spirulina by mitigating the light attenuation caused by a high initial inoculum level and counteracted the negative effect of low temperature (<25 °C). In large-scale cultivation, production efficiency would be reduced if illumination was not maintained at a high level. High salinity and sodium bicarbonate (NaHCO3) addition promoted carbohydrate accumulation, but suitable dilution could keep the required protein content in Spirulina with relatively low media and production costs. These findings reveal the interactive influence of cultivation parameters on biomass yield and help us determine the optimal cultivation conditions for large-scale cultivation of Spirulina-based seawater system based on a developed graphical user interface website.

An Innovative Device Based on Human-Machine Interface (HMI) for Powered Wheelchair Control for Neurodegenerative Disease: A Proof-of-Concept.

In the global context, advancements in technology and science have rendered virtual, augmented, and mixed-reality technologies capable of transforming clinical care and medical environments by offering enhanced features and improved healthcare services. This paper aims to present a mixed reality-based system to control a robotic wheelchair for people with limited mobility. The test group comprised 11 healthy subjects (six male, five female, mean age 35.2 ± 11.7 years). A novel platform that integrates a smart wheelchair and an eye-tracking-enabled head-mounted display was proposed to reduce the cognitive requirements needed for wheelchair movement and control. The approach's effectiveness was demonstrated by evaluating our system in realistic scenarios. The demonstration of the proposed AR head-mounted display user interface for controlling a smart wheelchair and the results provided in this paper could highlight the potential of the HoloLens 2-based innovative solutions and bring focus to emerging research topics, such as remote control, cognitive rehabilitation, the implementation of patient autonomy with severe disabilities, and telemedicine.

The Impact of Information Relevancy and Interactivity on Intensivists' Trust in a Machine Learning-Based Bacteremia Prediction System: Simulation Study.

Background: The exponential growth in computing power and the increasing digitization of information have substantially advanced the machine learning (ML) research field. However, ML algorithms are often considered "black boxes," and this fosters distrust. In medical domains, in which mistakes can result in fatal outcomes, practitioners may be especially reluctant to trust ML algorithms. Objective: The aim of this study is to explore the effect of user-interface design features on intensivists' trust in an ML-based clinical decision support system. Methods: A total of 47 physicians from critical care specialties were presented with 3 patient cases of bacteremia in the setting of an ML-based simulation system. Three conditions of the simulation were tested according to combinations of information relevancy and interactivity. Participants' trust in the system was assessed by their agreement with the system's prediction and a postexperiment questionnaire. Linear regression models were applied to measure the effects. Results: Participants' agreement with the system's prediction did not differ according to the experimental conditions. However, in the postexperiment questionnaire, higher information relevancy ratings and interactivity ratings were associated with higher perceived trust in the system (P<.001 for both). The explicit visual presentation of the features of the ML algorithm on the user interface resulted in lower trust among the participants (P=.05). Conclusions: Information relevancy and interactivity features should be considered in the design of the user interface of ML-based clinical decision support systems to enhance intensivists' trust. This study sheds light on the connection between information relevancy, interactivity, and trust in human-ML interaction, specifically in the intensive care unit environment.

Coupled analysis of arable land input intensity and output intensity based on sliding windows.

Sustainable intensification (SI) of agriculture can produce more food to meet the demand of a growing population while considering ecosystem health. The current SI estimation framework ignores the complex coupling between input and output intensity of arable land. A method for coupled analysis of arable land input intensity and output intensity based on sliding windows is proposed. By calculating the correlation coefficient and partial correlation coefficient between input intensity and output intensity in different value ranges as the order parameter, the phase transition and the influence process of input intensity on output intensity can be explained. Meanwhile, a python-based framework is developed. An application of the method was made to reveal the interaction process between annual provincial input intensity and output intensity in mainland China. Researchers in many fields may benefit from the method by obtaining a fast way to analysis the coupling relationship between driving and dependent variables in complex systems.•New method for SI estimation is presented.•The order parameter of the coupling relationship between input and output intensity is calculated based on sliding windows.•Analysis of coupling relationships between driving and dependent variables in complex systems.

Metis: a python-based user interface to collect expert feedback for generative chemistry models.

One challenge that current de novo drug design models face is a disparity between the user's expectations and the actual output of the model in practical applications. Tailoring models to better align with chemists' implicit knowledge, expectation and preferences is key to overcoming this obstacle effectively. While interest in preference-based and human-in-the-loop machine learning in chemistry is continuously increasing, no tool currently exists that enables the collection of standardized and chemistry-specific feedback. Metis is a Python-based open-source graphical user interface (GUI), designed to solve this and enable the collection of chemists' detailed feedback on molecular structures. The GUI enables chemists to explore and evaluate molecules, offering a user-friendly interface for annotating preferences and specifying desired or undesired structural features. By providing chemists the opportunity to give detailed feedback, allows researchers to capture more efficiently the chemist's implicit knowledge and preferences. This knowledge is crucial to align the chemist's idea with the de novo design agents. The GUI aims to enhance this collaboration between the human and the "machine" by providing an intuitive platform where chemists can interactively provide feedback on molecular structures, aiding in preference learning and refining de novo design strategies. Metis integrates with the existing de novo framework REINVENT, creating a closed-loop system where human expertise can continuously inform and refine the generative models.Scientific contributionWe introduce a novel Graphical User Interface, that allows chemists/researchers to give detailed feedback on substructures and properties of small molecules. This tool can be used to learn the preferences of chemists in order to align de novo drug design models with the chemist's ideas. The GUI can be customized to fit different needs and projects and enables direct integration into de novo REINVENT runs. We believe that Metis can facilitate the discussion and development of novel ways to integrate human feedback that goes beyond binary decisions of liking or disliking a molecule.

TooManyCellsInteractive: A visualization tool for dynamic exploration of single-cell data.

BACKGROUND: As single-cell sequencing technologies continue to advance, the growing volume and complexity of the ensuing data present new analytical challenges. Large cellular populations from single-cell atlases are more difficult to visualize and require extensive processing to identify biologically relevant subpopulations. Managing these workflows is also laborious for technical users and unintuitive for nontechnical users. RESULTS: We present TooManyCellsInteractive (TMCI), a browser-based JavaScript application for interactive exploration of cell populations. TMCI provides an intuitive interface to visualize and manipulate a radial tree representation of hierarchical cell subpopulations and allows users to easily overlay, filter, and compare biological features at multiple resolutions. Here we describe the software architecture and demonstrate how we used TMCI in a pan-cancer analysis to identify unique survival pathways among drug-tolerant persister cells. CONCLUSIONS: TMCI will facilitate exploration and visualization of large-scale sequencing data in a user-friendly way. TMCI is freely available at https://github.com/schwartzlab-methods/too-many-cells-interactive. An example tree from data within this article is available at https://tmci.schwartzlab.ca/.

Stopping criteria for neutron spectrum unfolding algorithms.

Neutron spectrum unfolding is a crucial process in radiation protection and dosimetry. Unfolding codes using iterative algorithms require a criterion to stop the iterations. One approach often relies on the Root Mean Square Error (RMSE) criterion to assess the convergence of iterative algorithms. The aim of this work is to present a new criteria: Average Ratio Scaled (AVGS) and Relative Change in AVGS (dAVGS) to address specific challenges associated with RMSE. Extensive validation tests were conducted, covering a range of scenarios with results showing high level of agreement between the unfolded spectra and the reference.

Haxe as a Swiss knife for bioinformatic applications: the SeqPHASE case story.

Haxe is a general purpose, object-oriented programming language supporting syntactic macros. The Haxe compiler is well known for its ability to translate the source code of Haxe programs into the source code of a variety of other programming languages including Java, C++, JavaScript, and Python. Although Haxe is more and more used for a variety of purposes, including games, it has not yet attracted much attention from bioinformaticians. This is surprising, as Haxe allows generating different versions of the same program (e.g. a graphical user interface version in JavaScript running in a web browser for beginners and a command-line version in C++ or Python for increased performance) while maintaining a single code, a feature that should be of interest for many bioinformatic applications. To demonstrate the usefulness of Haxe in bioinformatics, we present here the case story of the program SeqPHASE, written originally in Perl (with a CGI version running on a server) and published in 2010. As Perl+CGI is not desirable anymore for security purposes, we decided to rewrite the SeqPHASE program in Haxe and to host it at Github Pages (https://eeg-ebe.github.io/SeqPHASE), thereby alleviating the need to configure and maintain a dedicated server. Using SeqPHASE as an example, we discuss the advantages and disadvantages of Haxe's source code conversion functionality when it comes to implementing bioinformatic software.

Optimizing Technology-Based Prompts for Supporting People Living With Dementia in Completing Activities of Daily Living at Home: Experimental Approach to Prompt Modality, Task Breakdown, and Attentional Support.

BACKGROUND: Assistive technology is becoming increasingly accessible and affordable for supporting people with dementia and their care partners living at home, with strong potential for technology-based prompting to assist with initiation and tracking of complex, multistep activities of daily living. However, there is limited direct comparison of different prompt features to guide optimal technology design. OBJECTIVE: Across 3 experiments, we investigated the features of tablet-based prompts that best support people with dementia to complete activities of daily living at home, measuring prompt effectiveness and gaining feedback from people with dementia and their care partners about their experiences. METHODS: Across experiments, we developed a specialized iPad app to enable data collection with people with dementia at home over an extended experimental period. In experiment 1, we varied the prompts in a 3 (visual type: text instruction, iconic image, and photographic image) × 3 (audio type: no sound, symbolic sound, and verbal instruction) experimental design using repeated measures across multiple testing sessions involving single-step activities. In experiment 2, we tested the most effective prompt breakdown for complex multistep tasks comparing 3 conditions (1-prompt, 3-prompt, and 7-prompt conditions). In experiment 3, we compared initiation and maintenance alerts that involved either an auditory tone or an auditory tone combined with a verbal instruction. Throughout, we asked people with dementia and their care partners to reflect on the usefulness of prompting technology in their everyday lives and what could be developed to better meet their needs. RESULTS: First, our results showed that audible verbal instructions were more useful for task completion than either tone-based or visual prompts. Second, a more granular breakdown of tasks was generally more useful and increased independent use, but this varied across individuals. Third, while a voice or text maintenance alert enabled people with dementia to persist with a multistep task for longer when it was more frequent, task initiation still frequently required support from a care partner. CONCLUSIONS: These findings can help inform developers of assistive technology about the design features that promote the usefulness of home prompting systems for people with dementia as well as the preferences and insights of people with dementia and their care partners regarding assistive technology design.

User-Centered Development of Explanation User Interfaces for AI-Based CDSS: Lessons Learned from Early Phases.

This paper reports lessons learned during the early phases of the user-centered design process for an explanation user interface for an AI-based clinical decision support system for the intensive care unit. This paper focuses on identifying and verifying physicians' explanation needs in a multi-center, multi-country project. The explanation needs identified through context analysis and user requirements prioritization in an initial center differed from those identified through questionnaire responses from N= 9 physicians after a multi-center project workshop. These results highlight the caution that should be taken when eliciting explanation needs during the user-centered design process.

Visualizing Future Breast Cancer Prognosis by Generative AI.

In Sweden, 30 percent of breast cancer cases are detected between screenings, leading to later staged cancer diagnoses. Aileen Health is preventing later staged cancers by making a breast cancer prognosis with generative AI. This study investigates how breast radiologists perceive AI-generated images and their usability as cancer prognosis. Through literature review and formative usability testing, the research study emphasizes the challenges when integrating AI-generated medical images into clinical decision-making. Furthermore, our findings stress the importance of avoiding cognitive overload and following mental models. Future research should focus on radiologists' use of breast cancer prognosis at various urgency levels, as well as AI accuracy of generated images.

Creating and validating a scholarly knowledge graph using natural language processing and microtask crowdsourcing.

Due to the growing number of scholarly publications, finding relevant articles becomes increasingly difficult. Scholarly knowledge graphs can be used to organize the scholarly knowledge presented within those publications and represent them in machine-readable formats. Natural language processing (NLP) provides scalable methods to automatically extract knowledge from articles and populate scholarly knowledge graphs. However, NLP extraction is generally not sufficiently accurate and, thus, fails to generate high granularity quality data. In this work, we present TinyGenius, a methodology to validate NLP-extracted scholarly knowledge statements using microtasks performed with crowdsourcing. TinyGenius is employed to populate a paper-centric knowledge graph, using five distinct NLP methods. We extend our previous work of the TinyGenius methodology in various ways. Specifically, we discuss the NLP tasks in more detail and include an explanation of the data model. Moreover, we present a user evaluation where participants validate the generated NLP statements. The results indicate that employing microtasks for statement validation is a promising approach despite the varying participant agreement for different microtasks.

Websites, mobile apps, and social media: Premier online educational tools for radiology.

Demand for online educational tools has risen steadily as technological innovations have evolved over the past several decades. Websites were the first platform to be introduced, and eventually used for online schooling, soon after the advent of the World Wide Web. Access to information and updated content in a short period of time on a wide-screen device such as a computer made websites popular early in their development. With the technological revolution of smart phones, mobile applications have been developed on various operating systems and, through this progress, a new form of educational platform was initiated. The portable features of mobile applications represent a pioneer era of educational tools for medical professionals. Online communications have transformed into social media over the last decade and have since been adopted by much of the world. All three of these educational platforms have created a significant impact on medical education communities, specifically in radiology. We describe the relative strengths of each platform and illustrate how our experience over more than two decades guides our recommendations.

Constructing a User-Centric Visual Real-Time Surgical Dashboard.

The existing surgery timetable is very simple, not user-friendly, and has very limited information. This article outlines the creation of a surgical dynamic dashboard for users, using Microsoft Power BI. The dashboard was developed based on in-depth interviews with the surgical team members to determine their requirements. The outcome of dashboard was evaluated by 5-likert scale questionnaire to measure satisfaction of all surgical team members. All team members were very satisfied with the dashboard compared to the existing surgery timetable. In addition, with the function of auto-generated analytic reports through data criteria selection on the dashboard saved time on manually generating reports. This study supports to replace the existing surgery timetable by the user-centric and real time surgical dashboard.

Use of Electronic Health Records and Documentation Burden Among Nurses in Acute and Critical Care.

Nurses who provide the majority of hands-on care for hospitalized patients are disproportionately affected by the current state of electronic health records (EHRs), and little is known about their lived perception of EHR use. Using a mixed-methods research design, we conducted an in-depth analysis and synthesis of data from EHR usage log files, interviews, and surveys and assessed factors contributing to the nurse documentation burden in acute and critical at a large academic medical center. There remain substantial spaces where we can develop viable solutions for enhancing the usability of multi-component EHR systems.

An automation of Monte Carlo workflow for dosimetry study of an Elekta LINAC delivery system in radiotherapy.

Purpose: This study aims to automate the Monte Carlo (MC) workflow utilized for radiotherapy dosimetry, focusing on an Elekta LINAC delivery system. It addresses the challenge of integrating MC simulations into routine clinical practice, making this accurate yet complex method more accessible and efficient for radiotherapy dosimetry. Methods and Materials: We developed a user-friendly software featuring a graphical user interface (GUI) that integrates EGSnrc for MC simulations. The software streamlines the process from retrieving Digital Imaging and Communications in Medicine (DICOM) data to executing dose calculations and comparing dose distributions. To validate our proposed tool, we compared its computed doses for IMRT and VMAT plans from the Pinnacle TPS for an Elekta Versa HD linear accelerator against MC simulation results. This comparison utilized our in-house software and GUI as the tool, covering various treatment sites and prescriptions. Results: The automated MC workflow demonstrated high accuracy in dose calculations and streamlined integration with clinical workflows. The comparison between the MC-simulated and TPS-calculated doses revealed excellent agreement, highlighting the reliability of MC for independent dose verification in complex treatment scenarios. Conclusions: The automated MC workflow developed represents a substantial improvement in the practicality and efficiency of MC simulations in radiotherapy. This advancement not only simplifies the dosimetry process but also ensures high accuracy, establishing it as a valuable tool for routine patient-specific quality assurance and the development of specialized treatment procedures.

Development of "DiabetesSutra" a mobile application for lifestyle management of Type 2 Diabetes in India.

Background: Multiple mhealth (mobile health) interventions and mobile applications have been developed to support diabetes self-management. However, most of the apps are developed without the need for assessment and evaluation by experts in the field. This study aimed to design and develop a mobile application (app) supporting diabetes self-management for people with Type 2 Diabetes Mellitus (T2D) using a systematic approach. Methods: In this study mixed method design was used to develop the mobile application. The mhealth intervention was designed and developed in five steps: i) Extensive literature search, ii) Needs assessment of patients with T2D with the help of healthcare providers and patients (Interviews with 15 healthcare providers like clinicians, dietitians, and diabetes educators, and 2 focus group discussions with patients) iii) Ideation and content development of app based on outcomes of needs assessment; iv) content validation (by 10 healthcare providers) and v) App development on a hybrid platform. Evaluation of the app by users i.e., type 2 diabetes patients was done using the users' Mobile App rating scale (uMARS). The app was evaluated by 40 patients and rated on the uMARS questionnaire. Results: A patient-centric mobile app was developed for the nutritional management of diabetes with three modules: The patient module, the Evaluation module, and the Healthcare provider module. The patient module was the app that was provided to the patients with features like diet, physical activity, blood glucose log, education, etc., in addition to, a symptom checker, Stress meter blog, and FAQ. The evaluation module was integrated with the app it works when a user enters any log, it evaluates the entry against the standard cutoffs and flash prompts on the screen. The Healthcare provider module interacts with the server to provide them with patient data, comments, and feedback. Conclusions: The users found the app to be satisfactory. Incorporating additional features to enhance the user interface and streamline navigation could potentially enhance user engagement, thereby aiding in the management of T2D.

Enhancing safety in conditionally automated driving: Can more takeover request visual information make a difference in hazard scenarios with varied hazard visibility?

Autonomous driving technology has the potential to significantly reduce the number of traffic accidents. However, before achieving full automation, drivers still need to take control of the vehicle in complex and diverse scenarios that the autonomous driving system cannot handle. Therefore, appropriate takeover request (TOR) designs are necessary to enhance takeover performance and driving safety. This study focuses on takeover tasks in hazard scenarios with varied hazard visibility, which can be categorized as overt hazards and covert hazards. Through ergonomic experiments, the impact of TOR interface visual information, including takeover warning, hazard direction, and time to collision, on takeover performance is investigated, and specific analyses are conducted using eye-tracking data. The following conclusions are drawn from the experiments: (1) The visibility of hazards significantly affects takeover performance. (2) Providing more TOR visual information in hazards with different visibility has varying effects on drivers' visual attention allocation but can improve takeover performance. (3) More TOR visual information helps reduce takeover workload and increase human-machine trust. Based on these findings, this paper proposes the following TOR visual interface design strategies: (1) In overt hazard scenarios, only takeover warning is necessary, as additional visual information may distract drivers' attention. (2) In covert hazard scenarios, the TOR visual interface should better assist drivers in understanding the current hazard situation by providing information on hazard direction and time to collision to enhance takeover performance.

Development of an augmented reality system for tracheal intubation guidance of airway management.

Tracheal intubation is a crucial procedure performed in airway management to sustain life during various procedures. However, difficult airways can make intubation challenging, which is associated with increased mortality and morbidity. This is particularly important for children who undergo intubation where the situation is difficult. Improved airway management will decrease incidences of repeated attempts, decrease hypoxic injuries in patients, and decrease hospital stays, resulting in better clinical outcomes and reduced costs. Currently, 3D printed models based on CT scans and ultrasound-guided intubation are being used or tested for device fitting and procedure guidance to increase the success rate of intubation, but both have limitations. Maintaining a 3D printing facility can be logistically inconvenient, and it can be time consuming and expensive. Ultrasound-guided intubation can be hindered by operator dependence, limited two-dimensional visualization, and potential artifacts. In this study, we developed an augmented reality (AR) system that allows the overlay of intubation tools and internal airways, providing real-time guidance during the procedure. A child manikin was used to develop and test the AR system. Three-dimensional CT images were acquired from the manikin. Different tissues were segmented to generate the 3D models that were imported into Unity to build the holograms. Phantom experiments demonstrated the AR-guided system for potential applications in tracheal intubation guidance.