5.12.1. Междисциплинарные исследования когнитивных процессов (психологические науки)

Repetitive negative thinking processes such as rumination and worry — also referred to as ‘perseverative thinking’ or ‘perseverative cognition’ — have traditionally been conceptualized as distinct processes defined, measured and studied in the clinical literature from a disorder-specific stance (for example, rumination in relation to depression, worry in relation to anxiety). However, rumination and worry are highly correlated, share phenomenological features, load onto a common factor, and predict and maintain multiple psychopathologies. In this Review, we overview the evidence base that conceptualizes repetitive negative thinking as an overarching cognitive process unconstrained by disorder-related content or temporal focus. The related theories and the empirical literature covering measurement, treatment and prevention support repetitive negative thinking as a transdiagnostic process. Adopting a transdiagnostic approach can help advance a more nuanced understanding of the role of repetitive negative thinking in clinical presentations and their treatment.

Research on social comparison as a behaviour change technique (SC-BCT) has increased substantially. We conducted a random-effects meta-analysis of randomized controlled trials investigating SC-BCTs across the behavioural sciences (PROSPERO: CRD42022343154). We searched MEDLINE, PsycINFO and Web of Science from inception to January 2024. Seventy-nine randomized controlled trials (N = 1,356,521) investigating effects on behaviours related to climate change mitigation, health, performance and service were included. In the short term (mean 3.7 months post-intervention), SC-BCTs produced small effects relative to both passive (Hedges’ g = 0.17; 95% confidence interval, 0.11–0.23; k = 37; P < 0.001) and active control conditions (g = 0.23; 95% confidence interval, 0.15–0.31; k = 42; P < 0.001). A greater number of SC-BCT sessions and emphasis on desired (versus undesired) behaviours were associated with larger effects. Moderation effects were observed in only a few analyses, highlighting the need for further testing. SC-BCTs also produced significant small effects in the long term (mean 6.2 months post-intervention). Small effects should be interpreted in the context of low cost and scalability (for example, sending one or two emails). Certainty of evidence, using GRADE criteria, ranged from low to moderate depending on the analysis. More high-quality research is needed.

Inhibitory control is essential to daily function and is a key factor in numerous psychiatric disorders. One popular measure of inhibitory control is the congruency effect, but recent research has highlighted its low reliability, limiting its use for clinical and basic research questions. Here we asked whether it is possible to obtain precise individual estimates of the congruency effect. We sampled more than 5,000 trials from nine participants across four inhibitory control tasks. This dataset, made public for the community, demonstrates that precise individual estimates are achievable but with higher numbers of trials than typically collected with common tools. Using a combination of datasets and simulations, we show that extensive sampling is necessary to reveal true individual differences and improve observations from alternative modelling approaches. We share our dataset as a resource to further understand sources of variation in inhibitory control, ultimately advancing research in this critical field.

What determines who becomes and stays friends? Many factors are linked to friendship, including physical proximity and interpersonal similarities. Recent work has leveraged neuroimaging to detect similarities among friends by capturing how people process the world around them. However, given the cross-sectional nature of past research, it is unknown if neural similarity precedes friendship or only emerges among friends following social connection. Here we show that similarities in neural responses to movie clips—acquired before participants met one another—predicted proximity in a friendship network eight months later (that is, participants with similar responses were more likely to be friends rather than several degrees of separation apart). We also examined changes in distances between participants in their shared social network, which resulted from the formation, persistence and dissolution of friendships, between two months and eight months after they met each other. Compared with people who drifted further apart, people who grew closer over this six-month period had been more neurally similar as strangers. In addition, analyses controlling for sociodemographic similarities showed that whereas these similarities appeared to drive the differences in pre-existing neural similarities between friends and dyads of a social distance of 3, they did not account for the more extensive links between pre-existing neural similarities and the tendency for people to grow closer together, rather than drift farther apart, over time. Thus, whereas some friendships may initially form due to circumstance and dissolve over time, later-emerging and longer-lasting friendships may be rooted in deeper interpersonal compatibilities that are indexed by pre-existing neural similarities. The localization of these results suggests that pre-existing similarities in how people interpret, attend to and emotionally respond to their surroundings are precursors of future friendship and increased social closeness.

Establishing a unified theory of cognition has been an important goal in psychology1,2. A first step towards such a theory is to create a computational model that can predict human behaviour in a wide range of settings. Here we introduce Centaur, a computational model that can predict and simulate human behaviour in any experiment expressible in natural language. We derived Centaur by fine-tuning a state-of-the-art language model on a large-scale dataset called Psych-101. Psych-101 has an unprecedented scale, covering trial-by-trial data from more than 60,000 participants performing in excess of 10,000,000 choices in 160 experiments. Centaur not only captures the behaviour of held-out participants better than existing cognitive models, but it also generalizes to previously unseen cover stories, structural task modifications and entirely new domains. Furthermore, the model’s internal representations become more aligned with human neural activity after fine-tuning. Taken together, our results demonstrate that it is possible to discover computational models that capture human behaviour across a wide range of domains. We believe that such models provide tremendous potential for guiding the development of cognitive theories, and we present a case study to demonstrate this.

Understanding how animals and humans learn from experience to make adaptive decisions is a fundamental goal of neuroscience and psychology. Normative modelling frameworks such as Bayesian inference1 and reinforcement learning2 provide valuable insights into the principles governing adaptive behaviour. However, the simplicity of these frameworks often limits their ability to capture realistic biological behaviour, leading to cycles of handcrafted adjustments that are prone to researcher subjectivity. Here we present a novel modelling approach that leverages recurrent neural networks to discover the cognitive algorithms governing biological decision-making. We show that neural networks with just one to four units often outperform classical cognitive models and match larger neural networks in predicting the choices of individual animals and humans, across six well-studied reward-learning tasks. Critically, we can interpret the trained networks using dynamical systems concepts, enabling a unified comparison of cognitive models and revealing detailed mechanisms underlying choice behaviour. Our approach also estimates the dimensionality of behaviour3 and offers insights into algorithms learned by meta-reinforcement learning artificial intelligence agents. Overall, we present a systematic approach for discovering interpretable cognitive strategies in decision-making, offering insights into neural mechanisms and a foundation for studying healthy and dysfunctional cognition.

Основываясь на сведениях национального опроса «China Education Panel Survey» (2017/18 учебный год) и результатах 2018/19 учебного года, авторы данной статьи представляют полученные короткие панельные данные. Исследование проводилось методом квантильной регрессии с фиксированным эффектом, не ограниченным условиями. Оценивалось влияние домашнего компьютера и интернета на когнитивные способности учащихся младших классов средней школы. Результаты показывают, что: во-первых, домашний компьютер и интернет оказывают значительное положительное воздействие на учащихся младших классов средней школы с умеренными когнитивными способностями. Во-вторых, строгое отношение родителей к использованию компьютеров и интернета учащимися младших классов оказывает обратное влияние на когнитивное развитие последних в целом. В-третьих, для учащихся младших классов средней школы с низким уровнем когнитивных способностей качество обучения в школе более важно для развития вышеуказанных способностей. В-четвертых, неоправданные ожидания учеников младших классов и их родителей в отношении образования ограничивают когнитивное развитие учащихся младших классов в целом. В соответствии с этим, с одной стороны, правительство должно обеспечить учащимся младших классов доступ к использованию домашних компьютеров и интернета. С другой стороны, родители должны играть руководящую и контролирующую роль в использовании домашних компьютеров и интернета учащимися средней школы, постоянно общаться с детьми, чтобы создать благоприятную среду для образовательной деятельности.

Актуальность. В условиях стремительного роста цифровой торговли и активного внедрения искусственного интеллекта (ИИ) в маркетинговые практики, одной из ключевых стратегий взаимодействия с потребителями становится персонализация, понимаемая в контексте электронной коммерции как индивидуализированные рекомендации товаров, ценовые предложения и рекламные сообщения, формируемые на основе пользовательских данных и алгоритмов ИИ. Это трансформирует поведение потребителей, усиливая их вовлеченность, но одновременно порождая риски манипулятивного воздействия. В этой связи чрезвычайно актуальным представляется анализ того, как персонализация влияет на формирование цифрового доверия и потребительское поведение. Цель. Исследование имело цель оценить влияние персонализации на доверие потребителей и склонность к импульсивным покупкам. Выборка. В исследовании приняли участие 140 «цифровых» потребителей (из них 100 женщин) в возрасте от 20 до 40 лет (средний возраст 28 лет). Методы. В исследовании использовался метод опроса с применением структурированного онлайн-опросника, распространенного с использованием простого метода случайной выборки. Для статистической обработки данных применялся кластерный анализ, критерий хи-квадрат Пирсона и графические визуализации для выявления тенденций. Результаты. Анализ полученных данных позволил выявить в обследованной выборке три сегмента потребителей, характеризующихся определенным потребительским поведением: 1) «ориентированные на безопасность» ценят надежные бренды и прозрачную персонализацию, 2) «чувствительные к цене» позитивно реагируют на персонализированные скидки, но менее склонны к импульсивным покупкам, 3) «импульсивные» покупатели подвержены влиянию рекомендаций ИИ и маркетинга, основанного на срочности. Выводы. Персонализация может как способствовать формированию доверия, так и провоцировать импульсивные покупки в зависимости от типа потребителя. Бизнесу рекомендуется внедрять этичную персонализацию на основе ИИ, повышать доверие за счет прозрачности и обеспечивать баланс между персонализацией и автономией потребителей.