Anna Alexeenko

The frequency characteristic of biological systems is a sensitive tool for describing physiological dynamics, yet it has been virtually unused in pupillometry. A novel eye-tracking feature is proposed—the amplitude–frequency characteristic (AFC) of the “stimulus–pupil size” system—and evaluated as a candidate diagnostic marker on an open ADHD dataset. After quality control, 61 sessions (18 controls, 27 unmedicated ADHD, 16 on-medication ADHD) were analyzed. The AFC of the raw diameter signal reveals a systematic elevation of AFC amplitude in ADHD across the 0.5–30 Hz range, strongest in the 20–30 Hz band (Cohen’s d = +0.995, p = 0.0034). This is a pilot and methodology- and reproducibility-focused study. At the individual diagnostic level, a normative composite z-score of three bands (10–15, 15–20, 20–30 Hz), computed relative to the control group with fixed weights and no training, reaches an ROC-AUC = 0.788 (Ctrl vs. off-ADHD: p = 0.0012). At the group level, a linear discriminant projection of the eight-band profile reaches an ROC-AUC = 0.854. The inter-individual variability of within-subject AFC stability is reduced by a factor of 2.97 in unmedicated ADHD (p = 0.0085) and normalizes under methylphenidate (paired Wilcoxon p = 0.0067, n = 17). The widely used ratio of high- to low-frequency AFC amplitude yields an inflated ROC-AUC of 0.865 on the pre-filtered signal, but only 0.63 on the raw signal. All diagnostic claims are exploratory. The proposed normative feature is simple, interpretable, requires no training set, and can be computed from a single pupil recording.

Many time series forecasting applications use ranges rather than point forecasts. Producing forecasts in the form of Prediction Intervals (PIs) is natural, since intervals are an important component of many mathematical models. The LUBE (Lower Upper Bound Estimation) method is aimed at finding ranges based on solving optimization problems taking into account interval width and coverage. Using the Particle Swarm Training of simple neural network, we look for a solution to the optimization problem of the Coverage Width-Based Criterion (CWC), which is the exponential convolution of conflicting criteria PICP (Prediction Interval Coverage Probability) and PINRW (Prediction Interval Normalized Root-mean-square Width). Based on the concept of the Pareto compromise, it is introduced as a Pareto front in the space of specified criteria. The Pareto compromise is constructed as a relationship between conflicting criteria based on the found solution to the optimization problem. The data under consideration are the financial time series of the MOEX closing prices. Our findings reveal that a relatively simple neural network, comprising eight neurons and their corresponding 26 parameters (weights of neuron connections and neuron signal biases), is sufficient to yield reliable PIs for the investigated financial time series. The novelty of our approach lies in the use of a simple network structure (containing fewer than 100 parameters) to construct PIs for a financial time series. Additionally, we offer an experimental construction of the Pareto frontier, formed by the PICP and PINRW criteria.

Contemporary digital platforms provide a large number of web services for learning and professional growth. In most cases, educational web services only control access when connecting to resources and platforms. However, for educational and similar resources (internet surveys, online research), which are characterized by interactive interaction with the platform, it is important to assess user engagement in the learning process. A fairly large body of research is devoted to assessing learner engagement based on automatic, semi-automatic, and manual methods. Those methods include self-observation, observation checklists, engagement tracing based on learner reaction time and accuracy, computer vision methods (analysis of facial expressions, gestures, and postures, eye movements), methods for analyzing body sensor data, etc. Computer vision and body sensor methods for assessing engagement give a more complete objective picture of the learner’s state for further analysis in comparison with the methods of engagement tracing based on learner’s reaction time, however, they require the presence of appropriate sensors, which may often not be applicable in a particular context. Sensory observation is explicit to the learner and is an additional stressor, such as knowing the learner is being captured by the webcam while solving a problem. Thus, the further development of the hidden engagement assessment methods is relevant, while new computationally efficient techniques of converting the initial signal about the learner’s reaction time to assess engagement can be applied. On the basis of the hypothesis about the randomness of the dynamics of the time series, the largest Lyapunov exponent can be calculated for the time series formed from the reaction time of learners during prolonged work with web interfaces to assess the learner’s engagement. A feature of the proposed engagement assessment method is the relatively high computational efficiency, absence of high traffic loads in comparison with computer vision as well as secrecy from the learner coupled with no processing of learner’s personal or physical data except the reaction time to questions displayed on the screen. The results of experimental studies on a large amount of data are presented, demonstrating the applicability of the selected technique for learner’s engagement assessment.

Web surveys are an integral part of the feedback of Internet services, a research tool for respondents, including in the field of health and psychology. Web technologies allow conducting research on large samples. For mental health, an important metric is reaction time in cognitive tests and in answering questions. The use of mobile devices such as smartphones and tablets has increased markedly in web surveys, so the impact of device types and operating systems needs to be investigated. This article proposes an architectural solution aimed at reducing the effect of device variability on the results of cognitive psychological experiments. An experiment was carried out to formulate the requirements for software and hardware. Three groups of 1000 respondents were considered, corresponding to three types of computers and operating systems: Mobile Device, Legacy PC, and Modern PC. The results obtained showed a slight bias in the estimates for each group. It is noticed that the error for a group of devices differs both upward and downward for various tasks in a psychological experiment. Thus, for cognitive tests, in which the reaction time is critical, an architectural solution was synthesized for conducting psychological research in a web browser. The proposed architectural solution considers the characteristics of the device used by participants to undergo research in the web platform and allows to restrict access from devices that do not meet the specified criteria.

Web-surveys are one of the most popular forms of primary data collection used for various researches. However, mass surveys involve some challenges. It is required to consider different platforms and browsers, as well as different data transfer rates using connections in different regions of the country. Ensuring guaranteed data delivery in these conditions should determine the right choice of technologies for implementing web-surveys. The paper describes the solution to transfer a questionnaire to the client side in the form of an archive. This technological solution ensures independence from the data transfer rate and the stability of the communication connection with significant survey filling time. The conducted survey benefited the service of education psychologists under the federal Ministry of Education. School psychologists consciously took part in the survey, realizing the importance of their opinion for organizing and improving their professional activities. The desire to answer open-ended questions in detail created a part of the answers in the dataset, where there were several sentences about different aspects of professional activity. An important challenge of the problem is the Russian language, for which there are not as many tools as for the languages more widespread in the world. The survey involved 20,443 school psychologists from all regions of the Russian Federation, both from urban and rural areas. The answers did not contain spam, runaround answers, and so on as evidenced by the average response time. For the surveys, an authoring development tool DigitalPsyTools.ru was used.