IntroductionFlexible use of memory involves both the ability to form detailed memories of individual experiences (specificity) and to generalize across related experiences (generalization). Memory specificity and generalization have been attributed to distinct neocortical regions, such as ventrolateral prefrontal cortex (PFC) and ventromedial PFC respectively. The hippocampus has been traditionally associated with memory specificity, but more recent work highlights additional role in generalization. Here, we tested the hypothesis that the hippocampus supports both memory specificity and generalization, but through interactions with distinct cortical regions.MethodsFifty-two adults learned to categorize blended face stimuli, enabling both extraction of category structure (generalization) and encoding of item-specific features (specificity). Background functional connectivity was measured using fMRI during passive viewing of the same faces before and after learning.ResultsParticipants showed robust category learning, above-chance recognition of studied faces from similar lures, and successful category generalization to novel category members. Recognition and categorization performance were not highly correlated, suggesting distinct processes supporting each memory function. In the brain, we found distinct connectivity profiles of anterior hippocampus, presumed to preferentially contribute to generalization, and posterior hippocampus, presumed to preferentially contribute to specificity. Learning led to increased anterior hippocampal connectivity with default mode regions including ventromedial PFC, and posterior hippocampal connectivity with visual cortex. Increased anterior hippocampal connectivity with ventromedial PFC, somatomotor cortex, and visual cortex predicted better category generalization, whereas increased posterior hippocampal connectivity with ventrolateral PFC predicted more accurate face recognition. Exploratory analyses revealed widespread learning-related changes in cortico-cortical interactions, with changes in connectivity among visual, somatomotor, and default mode networks predicting categorization.DiscussionTogether, these findings support the notion that the hippocampus supports both memory specificity and generalization through interactions with distinct cortical networks. These results advance mechanistic accounts of how the hippocampus and cortex coordinate to balance competing memory demands.
Pedro Antonio Alarcon Granadeno, Arturo Miguel Bernal Russell, Sofia Nelson
et al.
Cyber-physical systems increasingly rely on foundational models, such as Large Language Models (LLMs) and Vision-Language Models (VLMs) to increase autonomy through enhanced perception, inference, and planning. However, these models also introduce new types of errors, such as hallucinations, over-generalizations, and context misalignments, resulting in incorrect and flawed decisions. To address this, we introduce the concept of Cognition Envelopes, designed to establish reasoning boundaries that constrain AI-generated decisions while complementing the use of meta-cognition and traditional safety envelopes. As with safety envelopes, Cognition Envelopes require practical guidelines and systematic processes for their definition, validation, and assurance. In this paper we describe an LLM/VLM-supported pipeline for dynamic clue analysis within the domain of small autonomous Uncrewed Aerial Systems deployed on Search and Rescue (SAR) missions, and a Cognition Envelope based on probabilistic reasoning and resource analysis. We evaluate the approach through assessing decisions made by our Clue Analysis Pipeline in a series of SAR missions. Finally, we identify key software engineering challenges for systematically designing, implementing, and validating Cognition Envelopes for AI-supported decisions in cyber-physical systems.
Spatial cognition is fundamental to real-world multimodal intelligence, allowing models to effectively interact with the physical environment. While multimodal large language models (MLLMs) have made significant strides, existing benchmarks often oversimplify spatial cognition, reducing it to a single-dimensional metric, which fails to capture the hierarchical structure and interdependence of spatial abilities. To address this gap, we propose a hierarchical spatial cognition framework that decomposes spatial intelligence into five progressively complex levels from basic observation to high-level planning. Building upon this taxonomy, we construct SpatialBench, a large-scale, fine-grained benchmark covering 15 tasks aligned with these cognitive levels. To provide a unified evaluation across heterogeneous tasks, we further introduce a high-level capability-oriented metric that reliably assesses a model's overall spatial reasoning ability. Extensive experiments over massive MLLMs reveal distinct performance stratification across cognitive levels: models exhibit strong perceptual grounding yet remain limited in symbolic reasoning, causal inference, and planning. Additional human tests demonstrate that humans perform selective, goal-directed abstraction, while MLLMs tend to over-attend to surface details without coherent spatial intent. Our work establishes the first systematic framework for measuring hierarchical spatial cognition in MLLMs, laying the foundation for future spatially intelligent systems.
Explaining individual differences in cognitive abilities requires both identifying brain parameters that vary across individuals and understanding how brain networks are recruited for specific tasks. Typically, task performance relies on the integration and segregation of functional subnetworks, often captured by parameters like regional excitability and connectivity. Yet, the high dimensionality of these parameters hinders pinpointing their functional relevance. Here, we apply stiff-sloppy analysis to human brain data, revealing that certain subtle parameter combinations ("stiff dimensions") powerfully influence neural activity during task processing, whereas others ("sloppy dimensions") vary more extensively but exert minimal impact. Using a pairwise maximum entropy model of task fMRI, we show that even small deviations in stiff dimensions-derived through Fisher Information Matrix analysis-govern the dynamic interplay of segregation and integration between the default mode network (DMN) and a working memory network (WMN). Crucially, separating a 0-back task (vigilant attention) from a 2-back task (working memory updating) uncovers partially distinct stiff dimensions predicting performance in each condition, along with a global DMN-WMN segregation shared across both tasks. Altogether, stiff-sloppy analysis challenges the conventional focus on large parameter variability by highlighting these subtle yet functionally decisive parameter combinations.
LLMs trained for logical reasoning excel at step-by-step deduction to reach verifiable answers. However, this paradigm is ill-suited for navigating social situations, which induce an interpretive process of analyzing ambiguous cues that rarely yield a definitive outcome. To bridge this gap, we introduce Cognitive Reasoning, a paradigm modeled on human social cognition. It formulates the interpretive process into a structured cognitive flow of interconnected cognitive units (e.g., observation or attribution), which combine adaptively to enable effective social thinking and responses. We then propose CogFlow, a complete framework that instills this capability in LLMs. CogFlow first curates a dataset of cognitive flows by simulating the associative and progressive nature of human thought via tree-structured planning. After instilling the basic cognitive reasoning capability via supervised fine-tuning, CogFlow adopts reinforcement learning to enable the model to improve itself via trial and error, guided by a multi-objective reward that optimizes both cognitive flow and response quality. Extensive experiments show that CogFlow effectively enhances the social cognitive capabilities of LLMs, and even humans, leading to more effective social decision-making.
Abstract There have been shifts toward more systematic and standardized methods for studying non-human primate facial signals, thanks to advancements like animalFACS. Additionally, there have been calls to better integrate the study of both facial and gestural communication in terms of theory and methodology. However, few studies have taken this important integrative step. By doing so, researchers could gain greater insight into how the physical flexibility of facial signals affects social flexibility. Our study combines both approaches to examine the relationship between the flexibility of physical form and the social function of chimpanzee facial “gestures”. We used chimpFACS along with established gestural ethograms that provide insights into four key gesture properties and their associated variables documented in chimpanzee gestures. We specifically investigated how the combinatorics (i.e., the different combinations of facial muscle movements) and complexity (measured by the number of discrete facial muscle movements) of chimpanzee facial signals varied based on: (1) how many gesture variables they exhibit; (2) the presence of a specific goal; and (3) the context in which they were produced. Our findings indicate that facial signals produced with vocalizations exhibit fewer gesture variables, rarely align with specific goals, and exhibit reduced contextual flexibility. Furthermore, facial signals that include additional visual movements (such as those of the head) and other visual signals (like manual gestures) exhibit more gestural variables, are frequently aligned with specific goals, and exhibit greater contextual flexibility. Finally, we discovered that facial signals become more morphologically complex when they exhibit a greater number of gesture variables. Our findings indicate that facial “gesturing” significantly enhanced the facial signaling repertoire of chimpanzees, offering insights into the evolution of complex communication systems like human language.
Can one shift attention among voices at a cocktail party during a silent pause? Researchers have required participants to attend to one of two simultaneous voices – cued by its gender or location. Switching the target gender or location has resulted in a performance ‘switch cost’ – which was recently shown to reduce with preparation when a gender cue was presented in advance. The current study asks if preparation for a switch is also effective when a voice is selected by location. We displayed a word or image 50/800/1400 ms before the onset of two simultaneous dichotic (male and female) voices to indicate whether participants should classify as odd/even the number spoken by the voice on the left or on the right; in another condition, we used gender cues. Preparation reduced the switch cost in both spatial-and gender-cueing conditions. Performance was better when each voice was heard on the same side as on the preceding trial, suggesting ‘binding’ of non-spatial and spatial voice features – but this did not materially influence the reduction in switch cost with preparation, indicating that preparatory attentional shifts can be effective within a single (task-relevant) dimension. We also asked whether words or pictures are more effective for cueing a voice. Picture cues resulted in better performance than word cues, especially when the interval between the cue and the stimulus was short, suggesting that (presumably phonological) processes involved in the recognition of the word cue interfered with the (near) concurrent encoding of the target voice’s speech.
Alexandra C. Pike, Katrina H. T. Tan, Hoda Tromblee
et al.
Affective biases are commonly seen in disorders such as depression and anxiety, where individuals may show attention towards and preferential processing of negative or threatening stimuli. Affective biases have been shown to change with effective intervention: randomized controlled trials into these biases and the mechanisms that underpin them may allow greater understanding of how interventions can be improved and their success be maximized. For such trials to be informative, we must have reliable ways of measuring affective bias over time, so we can detect how and whether they are altered by interventions: the test-retest reliability of our measures puts an upper bound on our ability to detect any changes. In this online study we therefore examined the test-retest reliability of two behavioural affective bias tasks (an ‘Ambiguous Midpoint’ and a ‘Go-Nogo’ task). 58 individuals recruited from the general population completed the tasks twice, with at least 14 days in between sessions. We analysed the reliability of both summary statistics and parameters from computational models using Pearson’s correlations and intra-class correlations. Standard summary statistic measures from these affective bias tasks had reliabilities ranging from 0.18 (poor) to 0.49 (moderate). Parameters from computational modelling of these tasks were in many cases less reliable than summary statistics. However, embedding the covariance between sessions within the generative modelling framework resulted in higher estimates of stability. We conclude that measures from these affective bias tasks are moderately reliable, but further work to improve the reliability of these tasks would improve still further our ability to draw inferences in randomized trials.
Computer applications to medicine. Medical informatics, Psychiatry
Empathy is an important characteristic to be considered when building a more intelligent and humanized dialogue agent. However, existing methods did not fully comprehend empathy as a complex process involving three aspects: cognition, affection and behavior. In this paper, we propose CAB, a novel framework that takes a comprehensive perspective of cognition, affection and behavior to generate empathetic responses. For cognition, we build paths between critical keywords in the dialogue by leveraging external knowledge. This is because keywords in a dialogue are the core of sentences. Building the logic relationship between keywords, which is overlooked by the majority of existing works, can improve the understanding of keywords and contextual logic, thus enhance the cognitive ability. For affection, we capture the emotional dependencies with dual latent variables that contain both interlocutors' emotions. The reason is that considering both interlocutors' emotions simultaneously helps to learn the emotional dependencies. For behavior, we use appropriate dialogue acts to guide the dialogue generation to enhance the empathy expression. Extensive experiments demonstrate that our multi-perspective model outperforms the state-of-the-art models in both automatic and manual evaluation.
When people are asked to recall their social networks, theoretical and empirical work tells us that they rely on shortcuts, or heuristics. Cognitive Social Structures (CSS) are multilayer social networks where each layer corresponds to an individual's perception of the network. With multiple perceptions of the same network, CSSs contain rich information about how these heuristics manifest, motivating the question, Can we identify people who share the same heuristics? In this work, we propose a method for identifying cognitive structure across multiple network perceptions, analogous to how community detection aims to identify social structure in a network. To simultaneously model the joint latent social and cognitive structure, we study CSSs as three-dimensional tensors, employing low-rank nonnegative Tucker decompositions (NNTuck) to approximate the CSS--a procedure closely related to estimating a multilayer stochastic block model (SBM) from such data. We propose the resulting latent cognitive space as an operationalization of the sociological theory of social cognition by identifying individuals who share relational schema. In addition to modeling cognitively independent, dependent, and redundant networks, we propose a specific model instance and related statistical test for testing when there is social-cognitive agreement in a network: when the social and cognitive structures are equivalent. We use our approach to analyze four different CSSs and give insights into the latent cognitive structures of those networks.
John Stewart Fabila-Carrasco, Avalon Campbell-Cousins, Mario A. Parra-Rodriguez
et al.
Permutation Entropy ($PE$) is a powerful nonlinear analysis technique for univariate time series. Recently, Permutation Entropy for Graph signals ($PEG$) has been proposed to extend PE to data residing on irregular domains. However, $PEG$ is limited as it provides a single value to characterise a whole graph signal. Here, we introduce a novel approach to evaluate graph signals \emph{at the vertex level}: graph-based permutation patterns. Synthetic datasets show the efficacy of our method. We reveal that dynamics in graph signals, undetectable with $PEG$, can be discerned using our graph-based patterns. These are then validated in DTI and fMRI data acquired during a working memory task in mild cognitive impairment, where we explore functional brain signals on structural white matter networks. Our findings suggest that graph-based permutation patterns in individual brain regions change as the disease progresses, demonstrating potential as a method of analyzing graph-signals at a granular scale.
Kekerasan dalam rumah tangga pada individu tertentu dapat menimbulkan gejala trauma, seperti kecenderungan menghindari situasi pencetus ingatan buruk hingga perasaan waspada terhadap lingkungan. Korban kekerasan berpotensi memiliki gangguan psikologis, salah satunya yaitu post traumatic stress disorder (PTSD). Tujuan penelitian ini adalah mengetahui efektivitas Cognitive Behavior Therapy pada individu dengan Post Traumatic Stress Disorder (PTSD) korban kekerasan sebelum dan sesudah diberikan intervensi. Desain penelitian ini menggunakan single subject design. Partisipan dalam penelitian ini merupakan seorang perempuan yang pernah mengalami kekerasan dari suaminya. Teknik pengambilan data yang digunakan yaitu observasi, wawancara, dan tes psikologi. Hasil penelitian menunjukkan bahwa cognitive behaviour therapy berpengaruh terhadap penurunan rasa cemas dan takut serta mengubah distorsi kognitif mengenai sumber ketakutan. Hasil observasi, wawancara, dan worksheet yang telah dikerjakan menunjukkan partisipan berhasil mengatasi rasa cemas dan takut yang dialaminya.
Scott H. Fraundorf, Zachary A. Caddick, Timothy J. Nokes-Malach
et al.
Abstract Although tests and assessments—such as those used to maintain a physician’s Board certification—are often viewed merely as tools for decision-making about one’s performance level, strong evidence now indicates that the experience of being tested is a powerful learning experience in its own right: The act of retrieving targeted information from memory strengthens the ability to use it again in the future, known as the testing effect. We review meta-analytic evidence for the learning benefits of testing, including in the domain of medicine, and discuss theoretical accounts of its mechanism(s). We also review key moderators—including the timing, frequency, order, and format of testing and the content of feedback—and what they indicate about how to most effectively use testing for learning. We also identify open questions for the optimal use of testing, such as the timing of feedback and the sequencing of complex knowledge domains. Lastly, we consider how to facilitate adoption of this powerful study strategy by physicians and other learners.
Abstract Many concepts are defined by their relationships to one another. However, instructors might teach these concepts individually, neglecting their interconnections. For instance, students learning about statistical power might learn how to define alpha and beta, but not how they are related. We report two experiments that examine whether there is a benefit to training subjects on relations among concepts. In Experiment 1, all subjects studied material on statistical hypothesis testing, half were subsequently quizzed on relationships among these concepts, and the other half were quizzed on their individual definitions; quizzing was used to highlight the information that was being trained in each condition (i.e., relations or definitions). Experiment 2 also included a mixed training condition that quizzed both relations and definitions, and a control condition that only included study. Subjects were then tested on both types of questions and on three conceptually related question types. In Experiment 1, subjects trained on relations performed numerically better on relational test questions than subjects trained on definitions (nonsignificant trend), whereas definitional test questions showed the reverse pattern; no performance differences were found between the groups on the other question types. In Experiment 2, relational training benefitted performance on relational test questions and on some question types that were not quizzed, whereas definitional training only benefited performance on test questions on the trained definitions. In contrast, mixed training did not aid learning above and beyond studying. Relational training thus seems to facilitate transfer of learning, whereas definitional training seems to produce training specificity effects.
Abstract The response time concealed information test (RT-CIT) can reveal that a person recognizes a relevant item (probe) among other, irrelevant items, based on slower responding to the probe compared to the irrelevant items. Thereby, if this person is concealing knowledge about the relevance of this item (e.g., recognizing it as a murder weapon), this deception can be unveiled. In the present paper, we examined the impact of a speed versus accuracy instruction: Examinees (N = 235) were either presented with instructions emphasizing a focus on speed, with instructions emphasizing a focus on accuracy, or with no particular speed or accuracy instructions at all. We found that although participants responded to the probe and the irrelevants marginally faster when they had received instructions emphasizing speed, there was no significant difference between RTs of the different experimental groups and crucially no significant difference between the probe–irrelevant RT differences either. This means that such instructions are unlikely to benefit the RT-CIT, but it also suggests that related deliberate manipulation (focusing on speed on or accuracy) is unlikely to decrease the efficiency of the RT-CIT—contributing further evidence to the RT-CIT’s resistance to faking.
Cristina Salles, Miguel Meira e Cruz, Igor Cardoso Freire
et al.
Introduction: Insomnia is a common sleep disorder in elderly. Although the HIV-positive population have a similar life expectancy when compared to the general population, some factors may interact with immunity conditions and therefore contribute to a worse prognosis.
Objective: This is a review of literature that aims to identify prevalence of insomnia in older HIV-positive patients.
Material and Methods: This is a review of literature conducted by using MEDLINE-PubMed, Embase, Cochrane Library, CINAHL, Web of Science, Scopus, SciELO, LILACS, and VHL databases, in addition to conducting manual searches. The terms used for the search were related to prevalence, HIV, insomnia, and advanced age. Inclusion criteria were: cross-sectional, cohort, and longitudinal studies, patients with a previous diagnosis of HIV in old age, studies reporting the frequency of insomnia or insomnia symptoms. The criteria for exclusion were: clinical trials, animal studies, letters, abstracts, conference proceedings, studies with other sleep scales that did not include insomnia.
Results: There were 2,805 publications found in the database and a further 10 articles were included manually. Of this total, four were included in this review, resulting in a total of 2,227 participants. The prevalence of insomnia in HIV-positive patients over 50 years varied from 12.5% to 76.5%.
Conclusion: The frequency of insomnia was higher in the profile of the population studied than in the general population. This should be clinically relevant in order to adequately treat and impact on the prognosis of those patient.
Zahra Banafsheh Alemohammad, Khosro Sadeghniiat-Haghighi, Mohammad Mehraban Parisa Fazlipanah
et al.
Objective: Several studies confirmed a positive association between obstructive sleep apnea (OSA) and metabolic syndrome. Continuous positive airway pressure (CPAP) is the main treatment for patients with moderate and severe OSA. CPAP therapy in adults with OSA results in reduction in sleepiness, blood pressure and improvement of metabolic profile. In this study, we aimed to evaluate the effects of CPAP therapy on various components of metabolic syndrome and subjective sleep parameters in patients with OSA.
Material and Methods: In this prospective trial study, 28 patients with moderate and severe OSA enrolled. Patients were asked to fill out the validated Persian version of questionnaires including Epworth sleepiness scale, insomnia severity index, STOP-BANG and Beck depression inventory - II, before and after treatment with CPAP. Weight and blood pressure were recorded before and after treatment. Only 14 patients agreed to blood sampling before and after CPAP therapy (at least 3 months of treatment). Fasting blood samples were analyzed for measuring the levels of FBS (fasting blood sugar), TG (triglyceride), total cholesterol, HDL, LDL, AST, and ALT.
Results: Diastolic blood pressure, ISI and STOP-BANG score significantly decreased after treatment (p-value: 0.008, 0.022 and 0.004, respectively). FBS and TG levels decreased after treatment, but only TG levels had significant difference (p-value: 0.46 and 0.016, respectively).
Discussion: CPAP therapy had positive effects on diastolic blood pressure, TG levels and ISI score. More studies with larger sample size and longer follow-up periods are warranted to investigate the effects of CPAP therapy on blood pressure, and metabolic parameters.