Qualitative modeling of the decision-making process using electrooculography.

A novel method based on electrooculography (EOG) has been introduced in this work to study the decision-making process. An experiment was designed and implemented wherein subjects were asked to choose between two items from the same category that were presented within a limited time. The EOG and voice signals of the subjects were recorded during the experiment. A calibration task was performed to map the EOG signals to their corresponding gaze positions on the screen by using an artificial neural network. To analyze the data, 16 parameters were extracted from the response time and EOG signals of the subjects. Evaluation and comparison of the parameters, together with subjects' choices, revealed functional information. On the basis of this information, subjects switched their eye gazes between items about three times on average. We also found, according to statistical hypothesis testing-that is, a t test, t(10) = 71.62, SE = 1.25, p < .0001-that the correspondence rate of a subjects' gaze at the moment of selection with the selected item was significant. Ultimately, on the basis of these results, we propose a qualitative choice model for the decision-making task.

ExplaineR: an R package to explain machine learning models.

Summary: SHapley Additive exPlanations (SHAP) is a widely used method for model interpretation. However, its full potential often remains untapped due to the absence of dedicated software tools. In response, ExplaineR, an R package to facilitate interpretation of binary classification and regression models based on clustering functionality for SHAP analysis is introduced here. It additionally offers user-interactive elements in visualizations for evaluating model performance, fairness analysis, decision-curve analysis, and a diverse range of SHAP plots. It facilitates in-depth post-prediction analysis of models, enabling users to pinpoint potentially significant patterns in SHAP plots and subsequently trace them back to instances through SHAP clustering. This functionality is particularly valuable for identifying patient subgroups in clinical cohorts, thus enhancing its role as a robust profiling tool. ExplaineR empowers users to generate comprehensive reports on machine learning outcomes, ensuring consistent and thorough documentation of model performance and interpretations. Availability and implementation: ExplaineR 1.0.0 is available on GitHub (https://persimune.github.io/explainer/) and CRAN (https://cran.r-project.org/web/packages/explainer/index.html).

Development of a machine learning model for early prediction of plasma leakage in suspected dengue patients.

BACKGROUND: At least a third of dengue patients develop plasma leakage with increased risk of life-threatening complications. Predicting plasma leakage using laboratory parameters obtained in early infection as means of triaging patients for hospital admission is important for resource-limited settings. METHODS: A Sri Lankan cohort including 4,768 instances of clinical data from N = 877 patients (60.3% patients with confirmed dengue infection) recorded in the first 96 hours of fever was considered. After excluding incomplete instances, the dataset was randomly split into a development and a test set with 374 (70%) and 172 (30%) patients, respectively. From the development set, five most informative features were selected using the minimum description length (MDL) algorithm. Random forest and light gradient boosting machine (LightGBM) were used to develop a classification model using the development set based on nested cross validation. An ensemble of the learners via average stacking was used as the final model to predict plasma leakage. RESULTS: Lymphocyte count, haemoglobin, haematocrit, age, and aspartate aminotransferase were the most informative features to predict plasma leakage. The final model achieved the area under the receiver operating characteristics curve, AUC = 0.80 with positive predictive value, PPV = 76.9%, negative predictive value, NPV = 72.5%, specificity = 87.9%, and sensitivity = 54.8% on the test set. CONCLUSION: The early predictors of plasma leakage identified in this study are similar to those identified in several prior studies that used non-machine learning based methods. However, our observations strengthen the evidence base for these predictors by showing their relevance even when individual data points, missing data and non-linear associations were considered. Testing the model on different populations using these low-cost observations would identify further strengths and limitations of the presented model.

Pre-Transplant Prediction of Acute Graft-versus-Host Disease Using the Gut Microbiome.

Gut microbiota is thought to influence host responses to allogeneic hematopoietic stem cell transplantation (aHSCT). Recent evidence points to this post-transplant for acute graft-versus-host disease (aGvHD). We asked whether any such association might be found pre-transplant and conducted a metagenome-wide association study (MWAS) to explore. Microbial abundance profiles were estimated using ensembles of Kaiju, Kraken2, and DeepMicrobes calls followed by dimensionality reduction. The area under the curve (AUC) was used to evaluate classification of the samples (aGvHD vs. none) using an elastic net to test the relevance of metagenomic data. Clinical data included the underlying disease (leukemia vs. other hematological malignancies), recipient age, and sex. Among 172 aHSCT patients of whom 42 developed aGVHD post transplantation, a total of 181 pre-transplant tool samples were analyzed. The top performing model predicting risk of aGVHD included a reduced species profile (AUC = 0.672). Beta diversity (37% in Jaccard's Nestedness by mean fold change, p < 0.05) was lower in those developing aGvHD. Ten bacterial species including Prevotella and Eggerthella genera were consistently found to associate with aGvHD in indicator species analysis, as well as relief and impurity-based algorithms. The findings support the hypothesis on potential associations between gut microbiota and aGvHD based on a data-driven approach to MWAS. This highlights the need and relevance of routine stool collection for the discovery of novel biomarkers.

Eye Movements in Response to Pain-Related Feelings in the Presence of Low and High Cognitive Loads.

The affective dimension of pain contributes to pain perception. Cognitive load may influence pain-related feelings. Eye tracking has proven useful for detecting cognitive load effects objectively by using relevant eye movement characteristics. In this study, we investigated whether eye movement characteristics differ in response to pain-related feelings in the presence of low and high cognitive loads. A set of validated, control, and pain-related sounds were applied to provoke pain-related feelings. Twelve healthy young participants (six females) performed a cognitive task at two load levels, once with the control and once with pain-related sounds in a randomized order. During the tasks, eye movements and task performance were recorded. Afterwards, the participants were asked to fill out questionnaires on their pain perception in response to the applied cognitive loads. Our findings indicate that an increased cognitive load was associated with a decreased saccade peak velocity, saccade frequency, and fixation frequency, as well as an increased fixation duration and pupil dilation range. Among the oculometrics, pain-related feelings were reflected only in the pupillary responses to a low cognitive load. The performance and perceived cognitive load decreased and increased, respectively, with the task load level and were not influenced by the pain-related sounds. Pain-related feelings were lower when performing the task compared with when no task was being performed in an independent group of participants. This might be due to the cognitive engagement during the task. This study demonstrated that cognitive processing could moderate the feelings associated with pain perception.

Author Correction: Eye movement characteristics reflected fatigue development in both young and elderly individuals.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

An oculometrics-based biofeedback system to impede fatigue development during computer work: A proof-of-concept study.

A biofeedback system may objectively identify fatigue and provide an individualized timing plan for micro-breaks. We developed and implemented a biofeedback system based on oculometrics using continuous recordings of eye movements and pupil dilations to moderate fatigue development in its early stages. Twenty healthy young participants (10 males and 10 females) performed a cyclic computer task for 31-35 min over two sessions: 1) self-triggered micro-breaks (manual sessions), and 2) biofeedback-triggered micro-breaks (automatic sessions). The sessions were held with one-week inter-session interval and in a counterbalanced order across participants. Each session involved 180 cycles of the computer task and after each 20 cycles (a segment), the task paused for 5-s to acquire perceived fatigue using Karolinska Sleepiness Scale (KSS). Following the pause, a 25-s micro-break involving seated exercises was carried out whether it was triggered by the biofeedback system following the detection of fatigue (KSS≥5) in the automatic sessions or by the participants in the manual sessions. National Aeronautics and Space Administration Task Load Index (NASA-TLX) was administered after sessions. The functioning core of the biofeedback system was based on a Decision Tree Ensemble model for fatigue classification, which was developed using an oculometrics dataset previously collected during the same computer task. The biofeedback system identified fatigue with a mean accuracy of approx. 70%. Perceived workload obtained from NASA-TLX was significantly lower in the automatic sessions compared with the manual sessions, p = 0.01 Cohen's dz = 0.89. The results give support to the effectiveness of integrating oculometrics-based biofeedback in timing plan of micro-breaks to impede fatigue development during computer work.

Eye movement characteristics reflected fatigue development in both young and elderly individuals.

Fatigue can develop during prolonged computer work, particularly in elderly individuals. This study investigated eye movement characteristics in relation to fatigue development. Twenty young and 18 elderly healthy adults were recruited to perform a prolonged functional computer task while their eye movements were recorded. The task lasted 40 minutes involving 240 cycles divided into 12 segments. Each cycle consisted of a sequence involving memorization of a pattern, a washout period, and replication of the pattern using a computer mouse. The participants rated their perceived fatigue after each segment. The mean values of blink duration (BD) and frequency (BF), saccade duration (SCD) and peak velocity (SPV), pupil dilation range (PDR), and fixation duration (FD) along with the task performance based on clicking speed and accuracy, were computed for each task segment. An increased subjective evaluation of fatigue suggested the development of fatigue. BD, BF, and PDR increased whereas SPV and SCD decreased over time in the young and elderly groups. Longer FD, shorter SCD, and lower task performance were observed in the elderly compared with the young group. The present findings provide a viable approach to develop a computational model based on oculometrics to track fatigue development during computer work.