Future Tools Research Articles

Epidemiological and molecular characteristics of extraintestinal pathogenic escherichia coli isolated from diseased cattle and sheep in Xinjiang, China from 2015 to 2019

Escherichia coli has become a common causative agent of infections in animals, inflicting serious economic losses on livestock production and posing a threat to public health. Escherichia coli infection is common and tends to be complex in Xinjiang, a major region of cattle and sheep breeding in China. This study aims to explore the current status and molecular characteristics of Escherichia coli infection in cattle and sheep in Xinjiang, as part of the disease prevention and control strategy. Herein we isolated Extraintestinal pathogenic Escherichia coli (ExPEC) from the liver, spleen, lung, heart, and lymph nodes of infected cattle and sheep (Xinjiang, China), and phylogenetic grouping, serotyping, and multilocus sequence typing were performed to determine epidemic and molecular characteristics. We also assessed their biofilm formation ability. A total of 132 strains of ExPEC were identified from diseased cattle and sheep, belonging to 7 phylogenetic groups. A and B1 are advantageous groups. Further, 22 serogroups were found, with O101 (26/132), O154 (14/132), and O65 (8/132) being the predominant ones. Among the seven sequence types identified by multilocus sequence typing, ST10 was the most common, followed by ST23 and ST457. Of 132, 105 (79.5%) strains were able to form biofilms: 15 strains (11.4%) were strong, 28 (21.2%) were medium, and 62 (47%) were weak biofilm producers. These findings will contribute to a better understanding of the molecular epidemiology of ExPEC in Xinjiang, China, and can be applied to the development, prevention, and disease control of future diagnostic tools and vaccine.

Machine-Learning Predictions of Cochlear Implant Functional Outcomes: A Systematic Review.

Cochlear implant (CI) user functional outcomes are challenging to predict because of the variability in individual anatomy, neural health, CI device characteristics, and linguistic and listening experience. Machine learning (ML) techniques are uniquely poised for this predictive challenge because they can analyze nonlinear interactions using large amounts of multidimensional data. The objective of this article is to systematically review the literature regarding ML models that predict functional CI outcomes, defined as sound perception and production. We analyze the potential strengths and weaknesses of various ML models, identify important features for favorable outcomes, and suggest potential future directions of ML applications for CI-related clinical and research purposes. We conducted a systematic literature search with Web of Science, Scopus, MEDLINE, EMBASE, CENTRAL, and CINAHL from the date of inception through September 2024. We included studies with ML models predicting a CI functional outcome, defined as those pertaining to sound perception and production, and excluded simulation studies and those involving patients without CIs. Using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we extracted participant population, CI characteristics, ML model, and performance data. Sixteen studies examining 5058 pediatric and adult CI users (range: 4 to 2489) were included from an initial 1442 publications. Studies predicted heterogeneous outcome measures pertaining to sound production (5 studies), sound perception (12 studies), and language (2 studies). ML models use a variety of prediction features, including demographic, audiological, imaging, and subjective measures. Some studies highlighted predictors beyond traditional CI audiometric outcomes, such as anatomical and imaging characteristics (e.g., vestibulocochlear nerve area, brain regions unaffected by auditory deprivation), health system factors (e.g., wait time to referral), and patient-reported measures (e.g., dizziness and tinnitus questionnaires). Used ML models were tree-based, kernel-based, instance-based, probabilistic, or neural networks, with validation and test methods most commonly being k-fold cross-validation and train-test split. Various statistical measures were used to evaluate model performance, however, for studies reporting accuracy, the best-performing models for each study ranged from 71.0% to 98.83%. ML models demonstrate high predictive performance and illuminate factors that contribute to CI user functional outcomes. While many models showed favorable evaluation statistics, the majority were not adequately reported with regard to dataset characteristics, model creation, and validation. Furthermore, the extent of overfitting in these models is unclear and will likely result in poor generalization to new data. This suggests the need for more robust validation procedures and standardization in reporting, with the ultimate hope that the iterative improvement of these models will allow for their adoption as a future clinical tool.

Understanding the impact of racism on surgical outcomes in settler nation-states USA and Canada: a protocol for a systematic review and meta-analysis

IntroductionTo improve surgical quality and safety, health systems must prioritise equitable care for surgical patients. Racialised patients experience worse postoperative outcomes when compared with non-racialised surgical patients in settler colonial...

642. The Role of Electronic Medical Record Automation in Latent Tuberculosis Screening and Treatment in a Large Health System

Abstract Background Valley Medical Center is a tertiary county hospital in San Jose, California that also staffs 12 clinics. Our institution’s Mycobacterial and Refugee Health Clinic is the main referral center for latent tuberculosis (LTBI) and active TB cases in the county. The USPFTF recommends screening for LTBI in asymptomatic adults who are at higher risk. This is challenging in primary care settings where physicians often see a high number of patients within limited time frames. There are many points on the LTBI care cascade (Figure 1) where patients can fall off. In the age of automation, technological tools can be designed to help reduce clinician burden.Figure 1:Latent Tuberculosis Care CascadeSource: California Department of Public Health Tuberculosis Control Branch Methods An EMR tool was engineered to prompt physicians to order a QuantiFERON (QFN) if a patient meets certain criteria: 18 years or older, no prior QFN or TB skin test, and any of the following risk factors: HIV positive, in custody, drug use, homelessness, or non-English primary language (proxy for foreign birth). This EMR tool went live on 8/9/22. Using EMR data-collection tools, a retrospective analysis was done among patients who had a clinical encounter (office visit or telemedicine) between 8/9/21-8/9/22 and 8/9/22-8/9/23 (one year before and after implementation of the tool). We determined the number of patients who were eligible for screening, obtained a QFN, had a positive QFN, initiated, and completed treatment. A 2-sample Z-test was used to assess temporal trends.Table 1:LTBI Care Cascade for Patients Seen by a Primary Care Physician in 8/9/21-8/9/22 and 8/9/22-8/9/23 Results From 8/9/22-8/9/23, 34251 patients were marked eligible for LTBI screening based on the EMR tool (Table 1, Figure 2). The proportion of patients tested for LTBI increased from 45.3% to 81.2% (p-value < 2.2e-16). The proportion of positive LTBI results increased from 8.6% to 12.4% (p-value < 2.2e-16). The number of patients who initiated treatment increased from 545 (40.8%) to 1038 (30%). The number of patients who completed treatment increased from 417 (76.5%) to 770 (74.2%).Figure 2:LTBI Care Cascade for Patients Seen by a Primary Care Physician in 8/9/21-8/9/22 and 8/9/22-8/9/23 Conclusion This study highlights that a low-cost EMR intervention can lead to substantial increases in LTBI screening, and as a result higher treatment completion. EMR automation can also improve accuracy in screening, reflected by the increased proportion of positive QFN results. Future tools involving digital technology can be designed to reduce drop-offs at other points along the LTBI cascade as well. Disclosures All Authors: No reported disclosures

Analysis of small extracellular vesicles from dried blood spots

This protocol paper describes how to extract small extracellular vesicles (sEVs) from dried blood spots (DBS). The methodology is described in detail and offers further evidence that the extracted particles are sEVs using western blotting (anti-CD9, CD63 and CD81) and fluorescence nanoparticle tracking analysis (fNTA). In addition, we present evidence that approximately 40% of the sEVs were recovered from DBS compared with EVs analyzed from plasma directly. The protocol proves to be robust, reliable and displays very interesting performances even after several weeks (up to 3 weeks) of storage of the DBS when analyzing the sEVs using protein microarray for the presence of the markers CD9, CD63, CD81, EpCAM, Flotilin-1, CD62E/P, CD142 and CD235a. These findings have important implications for using sEVs as future potential diagnostic tools by supporting the validity of less-invasive methods that can be implemented within vulnerable populations or in the field.

A mixed-methods investigation of a digital mental health tool to manage posttrauma anger.

Problematic anger affects up to 30% of individuals who have experienced trauma. Digital mental health approaches, such as ecological momentary assessment (EMA) delivered via smartphone and wearabledevices (i.e., wearables), hold significant potential for the development of novel digital technology treatments. The objective of this cohort study was to examine the acceptability, feasibility, and outcomes from 10 days of usage of a digital mental health tool combining EMA and wearable use among trauma-exposed adults with problematic anger. We used mixed methods to examine feasibility and acceptability and explored quantitative changes in mental health symptoms among participants over the study period (N = 98, 80.4% women, Mage = 38 years). Quantitative and qualitative data revealed that regular EMA combined with a wearable was feasible and acceptable in the sample. We observed reductions in problem anger, p <.001, repeated-measures d (dRM) = -0.81, 95% CI [-1.04, -0.59]; and posttraumatic stress disorder symptoms, p =.025, dRM = -0.26, 95% CI [-0.55, -0.03], over the 10 days of monitoring. Qualitative findings suggest that by regularly "checking in" on anger symptoms, participants improved their self-awareness and ability to self-manage their mood. These findings provide valuable learnings for building future personalized digital mental health tools.

Is AI image generation a creative or optimize tool in product design process?—The usage of AI image generation tools for vehicle shape as an example

This study explores the shape thinking processes and decision-making factors of designers when using AI image generation tools for conceptualizing the shapes of two-wheeled vehicles through four design tasks. Eight designers were invited to create hand-drawn sketches based on a specific aesthetic direction (technological geometry), followed by a shape divergence exercise using two AI graphics tools, Stable Diffusion and Vizcom, to generate images from text prompts. After selecting the designs closest to their original concepts and their favorite designs, the designers used an iPad to explore different shape directions (technological biology) for partial shape modifications. Finally, retrospective interviews were conducted to understand whether there were differences in designers’ thinking process regarding the use of various AI tools for shape conceptualization, as well as their focal points regarding design modification and shape thinking at different stages of the process. The research findings indicate that current AI tools are more suitable for shape divergence. If designers wish to achieve shape convergence, they need to be more familiar with the various settings of AI image generation tools and understand which prompts significantly influence specific shape characteristics. Designers’ perceptions of shape modification primarily revolve around: 1. Outline contours, 2. Parting lines, 3. Variations in surface curvature, and 4. The resulting features (light and shadow effects). Furthermore, it is recommended that future AI image generation tools, if developed as professional assistive tools for product design, should provide two modes—shape divergence and convergence—focusing on both the main shape and details. Additionally, it is suggested that developing AI-3D technologies should address the four key aspects of shape manipulation presented in this study, offering adjustments for overall appearance and detailing the contour lines of parts, including the manipulation of surface curvature and shape positioning.

The potential of short-wave infrared hyperspectral imaging and deep learning for dietary assessment: a prototype on predicting closed sandwiches fillings

IntroductionAccurate measurement of dietary intake without interfering in natural eating habits is a long-standing problem in nutritional epidemiology. We explored the applicability of hyperspectral imaging and machine learning for dietary assessment of home-prepared meals, by building a proof-of-concept, which automatically detects food ingredients inside closed sandwiches.MethodsIndividual spectra were selected from 24 hyperspectral images of assembled closed sandwiches, measured in a spectral range of 1116.14 nm to 1670.62 nm over 108 bands, pre-processed with Standard Normal Variate filtering, derivatives, and subsampling, and fed into multiple algorithms, among which PLS-DA, multiple classifiers, and a simple neural network.ResultsThe resulting best performing models had an accuracy score of ~80% for predicting type of bread, ~60% for butter, and ~ 28% for filling type. We see that the main struggle in predicting the fillings lies with the spreadable fillings, meaning the model may be focusing on structural aspects and not nutritional composition.DiscussionFurther analysis on non-homogeneous mixed food items, using computer vision techniques, will contribute toward a generalizable system. While there are still significant technical challenges to overcome before such a system can be routinely implemented in studies of free-living subjects, we believe it holds promise as a future tool for nutrition research and population intake monitoring.

Shape-Dependent Structural Order of Red Blood Cells.

In this work, we show how shape matters for the ordering of red blood cells (RBCs) at a water-air interface for both artificially rigidified and sphered cells as a model system for hereditary spherocytosis. We report enhanced long-range order for spherical RBCs over disk-shaped RBCs arising from the increased local ordering of spheres relative to disks. We show that rigidity has a greater effect on the radial distribution of spherical vs disk-shaped RBCs by slightly increasing the average distance between cells. The onset of local hexatic bond order of spherical RBCs in mixed disc-sphere systems coincides with the appearance of clustering of spherical cells as the number fraction of spherocytes increases. Additionally, the radial distribution function in mixed-shape systems begins to change with the onset of local hexatic order and clustering of spherical RBCs. By analyzing the radial distribution functions of RBCs, local hexatic bond order, and clustering, we show that the structure of settled RBCs is dictated by shape. These shape-dictated structures may provide a basis for future tools for detecting RBC shape-altering diseases and disorders.

Compress and Compare: Interactively Evaluating Efficiency and Behavior Across ML Model Compression Experiments.

To deploy machine learning models on-device, practitioners use compression algorithms to shrink and speed up models while maintaining their high-quality output. A critical aspect of compression in practice is model comparison, including tracking many compression experiments, identifying subtle changes in model behavior, and negotiating complex accuracy-efficiency trade-offs. However, existing compression tools poorly support comparison, leading to tedious and, sometimes, incomplete analyses spread across disjoint tools. To support real-world comparative workflows, we develop an interactive visual system called COMPRESS AND COMPARE. Within a single interface, COMPRESS AND COMPARE surfaces promising compression strategies by visualizing provenance relationships between compressed models and reveals compression-induced behavior changes by comparing models' predictions, weights, and activations. We demonstrate how COMPRESS AND COMPARE supports common compression analysis tasks through two case studies, debugging failed compression on generative language models and identifying compression artifacts in image classification models. We further evaluate COMPRESS AND COMPARE in a user study with eight compression experts, illustrating its potential to provide structure to compression workflows, help practitioners build intuition about compression, and encourage thorough analysis of compression's effect on model behavior. Through these evaluations, we identify compression-specific challenges that future visual analytics tools should consider and COMPRESS AND COMPARE visualizations that may generalize to broader model comparison tasks..

Whole-body tumor segmentation from FDG-PET/CT: Leveraging a segmentation prior from tissue-wise projections.

Background: Accurate tumor detection and quantification are important for optimized therapy planning and evaluation. Total tumor burden is also an appealing biomarker for clinical trials. Manual examination and annotation of oncologic PET/CT is labor-intensive and demands a high level of expertise. One significant challenge is the risk for human error, leading to potential omission of especially small tumors and tumors with low FDG uptake. Purpose: In this study, we introduced an automated framework with segmentation prior, from a tissue-wise multi-channel multi-angled based approach, to enhance tumor segmentation in whole-body FDG-PET/CT. Method: The proposed framework utilized a segmentation prior generated from tumor segmentations in tissue-wise multi-channel projections of the standardized uptake value (SUV) from PET. Projections were created from various angles and the tissues were identified based on their CT Hounsfield values. The resulting segmentation masks were subsequently backprojected into a unified 3D volume for creation of the segmentation prior. Finally, the segmentation prior was provided as an additional input channel along with the CT and SUV images to three variants of 3D segmentation networks (3D UNet, dynUNet, nnUNet) to enhance the overall tumor segmentation performance. All the methods were independently evaluated using 5-fold cross-validation on the autoPET dataset and subsequently tested on the U-CAN dataset. Results: Combining the segmentation prior with the original SUV and CT images improved overall tumor segmentation performance significantly compared to a baseline network. The increase in Dice coefficient for lymphoma, lung cancer, and melanoma across different segmentation networks were: 3D UNet ( , , ), dynUNet ( , , ), and nnUNet ( , , ), respectively; *, p-value < 0.05; ns, non-significance. Conclusion: The increased segmentation accuracy could be attributed to the segmentation prior generated from tissue-wise SUV projections, revealing information from various tissues that was useful for segmentation of tumors. The results from this study highlight the potential of the proposed method as a valuable future tool for time-efficient quantification of tumor burden in oncologic FDG-PET/CT.

Future Relevance of Innovative Information Technology in External Auditing and the Gap Between the Expertise in Sri Lanka

This study addresses a critical empirical gap by investigating the perceptions of external auditors in Sri Lanka regarding their current expertise and the future relevance such expertise in terms of innovative information technology (IT) in external auditing. The research also assesses whether discrepancies exist between auditors’ current IT capabilities and the expected expertise of emerging IT technologies. Additionally, it explores how specific auditor characteristics influence IT competence and perceptions of future relevance. A structured questionnaire was administered to a sample of 200 external auditors from both Big Four and non-Big Four firms, and the data were analyzed using descriptive and inferential statistical techniques. The findings indicate that while auditors acknowledge the growing importance of IT, a significant gap persists between their existing expertise and the skills required to adopt future IT tools effectively. This gap is most pronounced for advanced technologies such as machine learning, process mining, and robotic process automation. On the other hand, regression analysis highlights that work experience and hierarchical position of external auditors significantly influence IT knowledge and perceptions of relevance. Conversely, demographic factors such as age, gender, education, and firm size did not show a significant impact. The study underscores the need for targeted professional development, particularly for senior auditors, to bridge the technological gap. The results suggest that regulatory and educational bodies should implement strategic initiatives to ensure external auditors are equipped to meet the evolving demands of the audit profession in a rapidly changing technological landscape.

Coffee Rust Severity Analysis in Agroforestry Systems Using Deep Learning in Peruvian Tropical Ecosystems

Agroforestry systems can influence the occurrence and abundance of pests and diseases because integrating crops with trees or other vegetation can create diverse microclimates that may either enhance or inhibit their development. This study analyzes the severity of coffee rust in two agroforestry systems in the provinces of Jaén and San Ignacio in the department of Cajamarca (Peru). This research used a quantitative descriptive approach, and 319 photographs were collected with a professional camera during field trips. The photographs were segmented, classified and analyzed using the deep learning MobileNet and VGG16 transfer learning models with two methods for measuring rust severity from SENASA Peru and SENASICA Mexico. The results reported that grade 1 is the most prevalent rust severity according to the SENASA methodology (1 to 5% of the leaf affected) and SENASICA Mexico (0 to 2% of the leaf affected). Moreover, the proposed MobileNet model presented the best classification accuracy rate of 94% over 50 epochs. This research demonstrates the capacity of machine learning algorithms in disease diagnosis, which could be an alternative to help experts quantify the severity of coffee rust in coffee trees and broadens the field of research for future low-cost computational tools for disease recognition and classification

From DNA’s dark corners: fresh clues about depression

Depression is a significant global health issue, affecting millions of people worldwide, and is recognized as a leading cause of disability. Researchers are exploring how genetic and environmental factors contribute to depression, noting that about 30% of the risk is genetic, and the rest is chronic stress exposure. Recent studies highlight the importance of noncoding RNA genes—previously dismissed as “junk” DNA—in regulating brain function and influencing depression. There is a whole kingdom of RNA molecules that are not templates for protein; that, as a result, are called noncoding RNA. Here, I discuss evidence showing how noncoding RNAs play an important role in the brain in depression. Such RNA levels are also altered in depression in body fluids, including within miniature compartments called extracellular vesicles. Researchers are now investigating which noncoding RNAs contribute to the development of depression in the brain and if RNA found in body fluids can be used as a diagnostic biomarker. After the COVID-19 pandemic, which gave RNA therapeutics and diagnostics a boost, the hope for future RNA-based clinical tools for depression and other mental health conditions is higher than ever.

Vection in Individuals with and without Concussion: Associations with Postural Responses and Visual Dependence

Background We investigated whether everyday situations that trigger post-concussion symptoms (i.e., dynamic visual scenes), induce vection (illusory self-motion) and/or affect postural stability. Materials &amp; Methods Concussed and control participants were moved through a virtual grocery store, and rated their vection intensity. Postural sway during visual motion was measured. Baseline tests assessed concussion symptoms and sensory functioning, including visual dependence. Results Vection ratings were higher in concussed individuals than controls, and were predicted by faster visual speeds. Vection and visual speed also predicted postural sway in the concussion group. Visual dependence was positively associated with vection intensity and all postural measures. Conclusion These findings provide valuable insights for the development of future symptom-screening tools and rehabilitation strategies.

Acceptability of interactive post-endodontic restoration decision making application among undergraduate dental students

Background/PurposeSelecting an appropriate coronal restoration after root canal treatment requires careful consideration, however, there is a lack of established web-based tool that is based on clear guidelines for suitable post-endodontic restorations. Hence, we aimed to design and determine the acceptability of an interactive application to aid the decision making process for post-endodontic restoration among the undergraduate dental students. Materials and methodsA web-based post-endodontic restoration decision making interactive application, (RestoSmart) was developed after a detailed review of available literature. Seventy undergraduate dental students of IMU University were recruited in the study. A validated questionnaire was given to the participants before and after the use of RestoSmart to answer the provided mock clinical scenarios. Obtained data were statistically analyzed using IBM SPSS version 26 software package. ResultsThe statistical analysis demonstrated that RestoSmart was acceptable among the undergraduate dental students (P < 0.05). Initially, participants recognized the potential benefits of designing an interactive application, 74% believing that it may enhance decision making and 76% emphasizing its user-friendliness. Additionally, close to half of the participants expressed the willingness to use interactive tools in future. Following the use of RestoSmart, there was a notable increase in favorability, with 94.2% finding it helpful in the decision making and 94.3% acknowledging its user-friendliness. Furthermore, 87.2% expressed intent to use RestoSmart in future. ConclusionRestoSmart was useful for decision making and user friendly. While overall response was positive, there exist some areas for improvement such as ease of use, updating evidence-based guidelines, and establishing its helpfulness in teaching and learning.

Grapefruit-derived nootkatone potentiates GABAergic signaling and acts as a dual-action mosquito repellent and insecticide.

Humanity has long battled mosquitoes and the diseases they transmit-a struggle intensified by climate change and globalization, which have expanded mosquito ranges and the spread of associated diseases.1 Additionally, widespread insecticide resistance has reduced the efficacy of current control methods, necessitating new solutions.2,3 Nootkatone, a natural compound found in grapefruit, shows promise as both a mosquito repellent and an insecticide.4,5 However, its mechanism of action remains unclear. Our study demonstrates that nootkatone acts as a potent spatial and contact repellent against multiple mosquito species. Nootkatone-induced spatial aversion, which is influenced by human odor, is in Aedes aegypti partially mediated by Orco- and ionotropic receptor (IR)-positive neurons, while contact aversion is robust and likely mediated via the proboscis and independent of TRPA1 and IRs. We further find that nootkatone potentiates γ-aminobutyric acid (GABA)-mediated signaling by modulating the broadly expressed major insect GABA-gated chloride channel resistant to dieldrin (Rdl). At low doses, the chemosensory-mediated spatial and contact repellency is likely strengthened by nootkatone's disruption of synaptic transmission in select mosquito sensory neurons. At higher doses, nootkatone induces paralysis and death, presumably through broad-range synaptic transmission disruption. These findings reveal nootkatone's unique mode of action and highlight its potential as an effective mosquito control agent. Its dual role as a repellent and an insecticide, combined with low-to-no toxicity to humans and a pleasant smell, underscores nootkatone's promise as a future tool in mosquito control efforts.

‘Endless variation on a theme’: a document analysis of international and UK major trauma triage tools

Introduction: Triage tools are used within trauma networks to identify which injured patients should be bypassed and pre-alerted to major trauma centres. Despite the importance of treating the ‘right patient in the right place at the right time’, there has been no consensus on triage tool structure or content. This study aimed to identify, collate, review, summarise and recognise patterns across established major trauma triage tools.Methods: UK and international triage tools used between 2012 and 2021 were identified through literature review and correspondence with trauma networks. A conceptual content analysis was then undertaken using an inductive codebook, comprising concepts of triage tool structure, intended population, inclusion criteria and included variables and thresholds. Thematic analysis was also performed to identify higher-level patterns within the data, with emerging patterns becoming categories for analysis. A narrative synthesis of findings was then undertaken.Results: In total, 53 major trauma tools were identified, comprising 19 UK tools and 35 published international tools. Most triage tools (n = 42/53, 80%) were developed by expert opinion, were paper based and shared a common structure of multiple domains, with constituent triage predictors assessed in parallel. A minority of tools were statistically derived prediction models, operationalised either as simple scores (n = 10, 19%) or as an electronic application (n = 1, 1%). Overall, 173 distinct triage variables were used, with the median number of constituent triage variables per triage tool being 19 (range 3‐31). Four distinct patterns of triage tools were identified during thematic analysis, which differed in terms of format, number of triage variables, thresholds, scope for clinical judgement and relative diagnostic accuracy.Conclusion: Many diverse major trauma triage tools were identified, with no consensus in format, structure or content. Quantification of constituent variables and identification of distinct categories of triage tools may guide the design of future triage tools.

Identification of cultural conversations in therapy using natural language processing models.

Researchers have historically focused on understanding therapist multicultural competency and orientation through client self-report measures and behavioral coding. While client perceptions of therapist cultural competency and multicultural orientation and behavioral coding are important, reliance on these methods limits therapists receiving systematic, scalable feedback on cultural opportunities within sessions. Prior research demonstrating the feasibility of automatically identifying topics of conversation in psychotherapy suggests that natural language processing (NLP) models could be trained to automatically identify when clients and therapists are talking about cultural concerns and could inform training and provision of rapid feedback to therapists. Utilizing 103,170 labeled talk turns from 188 psychotherapy sessions, we developed NLP models that recognized the discussion of cultural topics in psychotherapy (F-1 = 70.0; Spearman's ρ = 0.78, p < .001). We discuss implications for research and practice and applications for future NLP-based feedback tools. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

ContractMind: Trust-calibration interaction design for AI contract review tools

Traditional contract review becomes increasingly time-consuming when attorneys face a surge in the number of contracts. Artificial Intelligence (AI) offers a solution to enhance the efficiency of reviews, but its uncertainty raises concerns among legal professionals. In this paper, we explore innovative clues and interaction design for trust calibration through a scenario-centric approach. Conducting formative experiments and semi-structured interviews, we conducted a contextual investigation with 24 attorneys, uncovering mismatches and trust calibration challenges between commercial AI tools and manual review processes in practical use. Based on these findings, we collaboratively designed seven key design components with attorneys and developed the ContractMind system prototype using a Wizard-of-Oz design approach. Through evaluation with 16 attorneys, we calculated and analyzed the differences between participants’ perceived trustworthiness and AI system capabilities. Compared to commercial artificial intelligence tools, our system is more conducive to trust calibration, enabling them to smoothly review contracts and make informed decisions. This work takes the first step in exploring trust calibration design for AI contract review tools and designs and evaluates ContractMind system prototype. We also provide design considerations for future AI contract review tools and trust calibration.