pith. sign in

arxiv: 2606.15121 · v2 · pith:HUUYUA2Hnew · submitted 2026-06-13 · 💻 cs.CL

When Cognitive Graphs Meet LLMs: BDEI Cognitive Pathways for Panic Emotional Arousal Prediction

Mengzhu Liu , Long Qin , Chuan Ai , Zhengqiu Zhu , Hongru Liang , Chen Gao , Yong Li , Xin Lu
show 1 more author
Quanjun Yin
This is my paper

Pith reviewed 2026-06-27 04:43 UTC · model grok-4.3

classification 💻 cs.CL
keywords panic emotional arousalcognitive pathwaysBDEI modelappraisal emotion theoryLLM integrationpsychological safety distanceemergency predictionarousal timing
0
0 comments X

The pith

PCP framework predicts panic emotional arousal timing by fusing four-domain signals into a risk metric and routing through a BDEI pathway with LLM limited to one transition.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that predicting when individuals will experience panic emotional arousal requires explicitly modeling the appraisal process from threat evaluation to emotion, which existing methods fail to do directly. It identifies gaps in fusing multiple threat dimensions into risk perception, decoupling emotion from cognitive models, and over-relying on LLMs which can hallucinate. PCP addresses these by using a Psychological Safety Distance model to create a unified risk metric from four domains as an entry condition, adding an Emotion node to form the BDEI pathway, and restricting the LLM to parameter estimation only in the Belief-to-Desire step. This setup couples appraisal directly to arousal and confines errors. On Hurricane Sandy data, it improves arousal timing accuracy by 10.68% over baselines and reduces peak count error to 7.07%.

Core claim

The paper claims that the PCP framework solves the three identified problems by mapping four-domain signals via the Psychological Safety Distance model into a risk metric that triggers entry into the BDEI pathway, where an explicit Emotion node grounded in appraisal emotion theory directly links threat appraisal to emotional arousal, all state transitions are governed by BDEI, and the LLM is confined to the Belief-to-Desire transition to prevent hallucination propagation, resulting in more accurate prediction of panic emotional arousal timing.

What carries the argument

The BDEI pathway, an extension of BDI that inserts an explicit Emotion node to couple appraisal directly to arousal, with the LLM restricted to Belief-to-Desire parameter estimation.

If this is right

  • Agents whose risk metric exceeds the PSD threshold enter the BDEI pathway for direct coupling of appraisal to emotional arousal.
  • The LLM is confined to a single transition step, preventing error propagation through the cognitive graph.
  • All state transitions in the model are governed by the BDEI pathway rather than LLM outputs.
  • Experiments demonstrate a 10.68% gain in arousal timing accuracy and reduction of peak count error to 7.07% on Hurricane Sandy data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The hybrid structure could be adapted to predict timing of other emotions in non-emergency settings such as mental health monitoring.
  • Testing the PSD fusion step independently on additional signal sources would clarify whether the risk metric generalizes beyond the four domains used here.
  • Real-time deployment in emergency systems might allow earlier interventions if the threshold reliably precedes observable arousal peaks.

Load-bearing premise

The Psychological Safety Distance model correctly fuses four-domain signals into a risk metric that serves as a valid entry condition for the BDEI pathway, and the BDEI structure with LLM limited to Belief-to-Desire accurately models the direct link from appraisal to emotional arousal.

What would settle it

If applying the PCP framework to datasets from other panic-inducing events yields no improvement in arousal timing accuracy or peak count error over the same baselines, the performance advantage would be falsified.

Figures

Figures reproduced from arXiv: 2606.15121 by Chen Gao, Chuan Ai, Hongru Liang, Long Qin, Mengzhu Liu, Quanjun Yin, Xin Lu, Yong Li, Zhengqiu Zhu.

Figure 1
Figure 1. Figure 1: Multi-domain coupling for panic arousal timing and macro outbreak prediction. (a) Task formulation. At the time step ti , physical, social, cognitive and information inputs drive PCP reasoning to predict each individual’s panic arousal timing and the macro peak panic scale and its time step. (b) PCP reasoning at step ti . User A activates all three desires and panics; User C perceives lower risk, activates… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the PCP framework. Four-domain inputs are coupled via a PSD model into a unified risk metric; agents below the safety threshold enter a BDEI pathway where the cognitive graph governs all transitions and the LLM serves only Belief-to-Desire parameter inference. Outputs include individual panic state and first arousal timestep (micro), and peak panic count with corresponding timestep (macro). the… view at source ↗
Figure 3
Figure 3. Figure 3: BDEI cognitive pathway graph for panic arousal. The graph formalizes appraisal emotion theory as a causal pathway of five node types (O, B, D, E, I) and sub-nodes, with directed edges annotated by computational models. from objective risk, and is better suited for quantifying individ￾ual risk perception in sudden disasters. Tailored to hurricane￾evacuation panic arousal, we adapt its input variables and pa… view at source ↗
Figure 4
Figure 4. Figure 4: Spatial prediction error of peak panic arousal at Step 61 during Hurricane Sandy. Each panel shows kernel density estimation (KDE) difference (simulated vs. real panic density). Blue: under-prediction; Red: over-prediction; White: accurate match. All KDE heatmaps normalized to [0,1]. Macro F1: Equal-weighted average of per-class F1 scores, mitigating class-imbalance bias and emphasizing minority class (del… view at source ↗
Figure 5
Figure 5. Figure 5: Robustness analysis of LLM scoring consistency. (a) [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: PCP framework predicts counterfactual panic in a low-neuroticism user (Case 2) by accounting for situational uncertainty [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
read the original abstract

Predicting individual panic emotional arousal timing before manifestation is essential for proactive emergency intervention. Existing methods incorporate cognitive elements but none explicitly model the emotional arousal process, making them ill-suited for emotional arousal timing prediction. We argue that grounding prediction in appraisal emotion theory is necessary because it explicitly models this process, but three problems must be solved. (1) Appraisal theory posits that emotion arises from simultaneous evaluation across multiple threat dimensions, yet no prior work fuses these inputs into risk perception. (2) Existing cognitive models lack an Emotion node, decoupling threat appraisal from emotional arousal and forcing emotions to be inferred indirectly from behaviors. (3) Given their generalizable cognitive reasoning, current approaches adopt LLMs as the primary decision-maker, yet overlook the fragility and hallucination-proneness of their outputs. To address these issues, we introduce PanicCognitivePath (PCP), a framework that addresses all three. A Psychological Safety Distance (PSD) model, grounded in psychological distance theory, maps four-domain signals into a unified risk metric as the entry condition for subsequent cognitive reasoning. An explicit Emotion node grounded in appraisal emotion theory is introduced into BDI, forming a Belief-Desire-Emotion-Intention (BDEI) pathway. Agents whose risk metric exceeds the PSD threshold enter this pathway, coupling threat appraisal directly to emotional arousal. The BDEI pathway governs all state transitions while the LLM is confined to parameter estimation for the Belief-to-Desire transition, confining hallucinations to a single step and preventing error propagation. Experiments on Hurricane Sandy show PCP improves arousal timing accuracy by 10.68% over baselines, reduces peak count error to 7.07%.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript introduces the PanicCognitivePath (PCP) framework for predicting individual panic emotional arousal timing. It proposes a Psychological Safety Distance (PSD) model, grounded in psychological distance theory, to fuse four-domain signals into a unified risk metric that gates entry into a Belief-Desire-Emotion-Intention (BDEI) pathway; an explicit Emotion node is added to the classic BDI structure so that threat appraisal couples directly to emotional arousal. The LLM is restricted to parameter estimation only in the Belief-to-Desire transition. Experiments on the Hurricane Sandy dataset report that PCP improves arousal timing accuracy by 10.68% over baselines and reduces peak-count error to 7.07%.

Significance. If the PSD risk metric can be shown to function as a psychologically grounded gate rather than a post-hoc threshold and if the reported gains can be attributed to the explicit Emotion node and the restricted LLM role, the work would supply a concrete mechanism for embedding appraisal theory inside cognitive-agent architectures that use LLMs. This could be relevant for proactive intervention systems where the timing of emotional arousal must be anticipated from multi-domain signals.

major comments (3)
  1. [Abstract] Abstract: the headline claims of a 10.68% gain in arousal timing accuracy and a 7.07% peak-count error are presented without any definition of the baselines, the exact metric used for “arousal timing accuracy,” statistical significance tests, confidence intervals, or error analysis. Because these numbers are the sole empirical support for the central claim that the BDEI pathway improves prediction, their unverifiability is load-bearing.
  2. [PSD model description] PSD model: the manuscript states that the PSD model “maps four-domain signals into a unified risk metric as the entry condition” for the BDEI pathway, yet supplies neither the fusion equation, the definitions of the four domains, nor any separate validation (e.g., correlation between threshold crossings and observed arousal onsets). Without this, it is impossible to confirm that the reported performance originates from the claimed appraisal-to-arousal coupling rather than from an arbitrary detector.
  3. [Experiments] Experiments section: evaluation is performed on a single dataset (Hurricane Sandy) with no cross-validation, no sensitivity analysis on the PSD threshold (listed as a free parameter), and no comparison against alternative threshold-based or non-cognitive baselines. This leaves open the possibility that the 10.68% figure is not independent of post-hoc choices.
minor comments (1)
  1. [BDEI pathway] The precise state-transition rules inside the BDEI pathway (how the Emotion node alters Desire and Intention) are described at a high level but lack the formal specification or pseudocode that would be needed for replication.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major comment point by point below, indicating where revisions will be made to improve clarity and verifiability.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the headline claims of a 10.68% gain in arousal timing accuracy and a 7.07% peak-count error are presented without any definition of the baselines, the exact metric used for “arousal timing accuracy,” statistical significance tests, confidence intervals, or error analysis. Because these numbers are the sole empirical support for the central claim that the BDEI pathway improves prediction, their unverifiability is load-bearing.

    Authors: We agree that the abstract would benefit from greater specificity on these elements to make the claims immediately verifiable. In the revised manuscript we will expand the abstract to define arousal timing accuracy (as the proportion of correctly predicted arousal onsets within a tolerance window), name the primary baselines, and note that statistical significance was assessed via paired t-tests with reported p-values. Full confidence intervals and error breakdowns will remain in the Experiments section due to length constraints, but the abstract will now direct readers to them. revision: yes

  2. Referee: [PSD model description] PSD model: the manuscript states that the PSD model “maps four-domain signals into a unified risk metric as the entry condition” for the BDEI pathway, yet supplies neither the fusion equation, the definitions of the four domains, nor any separate validation (e.g., correlation between threshold crossings and observed arousal onsets). Without this, it is impossible to confirm that the reported performance originates from the claimed appraisal-to-arousal coupling rather than from an arbitrary detector.

    Authors: This observation is correct; the current description is high-level. We will add an explicit subsection that (1) defines the four domains drawn from psychological distance theory, (2) presents the fusion equation used to compute the unified PSD risk metric, and (3) reports a validation analysis correlating PSD threshold crossings with observed arousal onsets on the Hurricane Sandy data. These additions will allow readers to evaluate whether the performance gain stems from the intended appraisal mechanism. revision: yes

  3. Referee: [Experiments] Experiments section: evaluation is performed on a single dataset (Hurricane Sandy) with no cross-validation, no sensitivity analysis on the PSD threshold (listed as a free parameter), and no comparison against alternative threshold-based or non-cognitive baselines. This leaves open the possibility that the 10.68% figure is not independent of post-hoc choices.

    Authors: We acknowledge the limitations of the reported experimental design. In the revision we will (1) add k-fold cross-validation results, (2) include a sensitivity analysis sweeping the PSD threshold and showing that performance remains stable across a reasonable range, and (3) introduce comparisons against both simple threshold detectors and standard non-cognitive baselines (e.g., LSTM and random forest on the same features). These changes will better isolate the contribution of the explicit Emotion node and restricted LLM role. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper presents a descriptive framework (PSD model + BDEI pathway) grounded in existing psychological theories and reports empirical results on the Hurricane Sandy dataset. No equations, parameter-fitting steps, or self-citations are visible in the provided text that would reduce any claimed prediction or result to an input by construction. The performance metrics are presented as experimental outcomes rather than derived quantities, and the central claims rest on the introduction of new nodes and thresholds without evidence of self-referential definition or load-bearing self-citation chains. This is a standard case of a self-contained empirical proposal against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 2 invented entities

The central claim rests on the unproven effectiveness of the newly introduced PSD mapping and BDEI structure; these components are postulated rather than derived from first principles or independently validated data.

free parameters (1)
  • PSD threshold
    The risk-metric value that triggers entry into the BDEI pathway is a tunable cutoff whose value is not derived in the abstract.
axioms (1)
  • domain assumption Appraisal emotion theory provides an accurate simultaneous multi-dimension evaluation that produces emotional arousal
    The framework is explicitly grounded in this theory to justify the Emotion node and risk fusion.
invented entities (2)
  • Psychological Safety Distance (PSD) model no independent evidence
    purpose: Maps four-domain signals into a single risk metric that serves as entry condition for cognitive reasoning
    New construct introduced to solve the fusion problem stated in the abstract.
  • Emotion node in BDEI pathway no independent evidence
    purpose: Couples threat appraisal directly to emotional arousal inside the cognitive graph
    Explicit addition to standard BDI to address the missing emotion step.

pith-pipeline@v0.9.1-grok · 5858 in / 1406 out tokens · 47046 ms · 2026-06-27T04:43:56.461282+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

46 extracted references · 2 canonical work pages · 1 internal anchor

  1. [1]

    Looting and antisocial behavior after disasters: a systematic review,

    R. A. Sheikhi, R. Javanbakhtian, and M. Heidari, “Looting and antisocial behavior after disasters: a systematic review,”BMC Public Health, vol. 25, no. 1, p. 309, 2025

    2025

  2. [2]

    Exposure to rapid succession disasters: a study of residents at the epicenter of the chilean b ´ıo b´ıo earthquake.,

    D. R. Garfin, R. C. Silver, F. J. Ugalde, H. Linn, and M. Inostroza, “Exposure to rapid succession disasters: a study of residents at the epicenter of the chilean b ´ıo b´ıo earthquake.,”Journal of abnormal psychology, vol. 123, no. 3, p. 545, 2014

    2014

  3. [3]

    A novel agent- based approach for dynamic emotion modeling in social networks,

    X. Wu, L. Lu, M. Dotoli, G. Fortino, and M. Chen, “A novel agent- based approach for dynamic emotion modeling in social networks,” IEEE Transactions on Cybernetics, 2025

    2025

  4. [4]

    Bdi agents in social simulations: a survey,

    C. Adam and B. Gaudou, “Bdi agents in social simulations: a survey,” The Knowledge Engineering Review, vol. 31, no. 3, pp. 207–238, 2016

    2016

  5. [5]

    Purple: Making a large language model a better sql writer,

    T. Ren, Y . Fan, Z. He, R. Huang, J. Dai, C. Huang, Y . Jing, K. Zhang, Y . Yang, and X. S. Wang, “Purple: Making a large language model a better sql writer,” in2024 IEEE 40th International Conference on Data Engineering (ICDE), pp. 15–28, IEEE, 2024

    2024

  6. [6]

    Lego-graphrag: Modularizing graph-based retrieval-augmented generation for design space exploration,

    Y . Cao, Z. Gao, Z. Li, X. Xie, S. K. Zhou, and J. Xu, “Lego-graphrag: Modularizing graph-based retrieval-augmented generation for design space exploration,”Proceedings of the VLDB Endowment, vol. 18, p. 3269–3283, June 2025

    2025

  7. [7]

    In-depth analysis of graph-based rag in a unified framework,

    Y . Zhou, Y . Su, Y . Sun, S. Wang, T. Wang, R. He, Y . Zhang, S. Liang, X. Liu, Y . Ma, and Y . Fang, “In-depth analysis of graph-based rag in a unified framework,”Proc. VLDB Endow., vol. 18, p. 5623–5637, Sept. 2025

    2025

  8. [8]

    Multirag: a knowledge-guided framework for mitigating hallucination in multi- source retrieval augmented generation,

    W. Wu, H. Wang, B. Li, P. Huang, X. Zhao, and L. Liang, “Multirag: a knowledge-guided framework for mitigating hallucination in multi- source retrieval augmented generation,” in2025 IEEE 41st International Conference on Data Engineering (ICDE), pp. 3070–3083, IEEE, 2025

    2025

  9. [9]

    Computational models of emotion inference in theory of mind: A review and roadmap,

    D. C. Ong, J. Zaki, and N. D. Goodman, “Computational models of emotion inference in theory of mind: A review and roadmap,”Topics in cognitive science, vol. 11, no. 2, pp. 338–357, 2019

    2019

  10. [10]

    Some: A realistic benchmark for llm-based social media agents,

    D. Xue, J. Cui, S. Qian, C. Hu, and C. Xu, “Some: A realistic benchmark for llm-based social media agents,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 40, pp. 1391–1399, 2026

    2026

  11. [11]

    Computational models and optimal control strategies for emotion contagion in the human population in emergencies,

    X. Wang, L. Zhang, Y . Lin, Y . Zhao, and X. Hu, “Computational models and optimal control strategies for emotion contagion in the human population in emergencies,”Knowledge-Based Systems, vol. 109, pp. 35– 47, 2016

    2016

  12. [12]

    Modeling the spread of negative emotions in social networks during sudden public crisis events: Dual mechanisms of social reinforcement and individual regulation,

    T. W ANG, T. LIU, and C. LI, “Modeling the spread of negative emotions in social networks during sudden public crisis events: Dual mechanisms of social reinforcement and individual regulation,”Journal of Tsinghua University (Science and Technology), vol. 65, no. 6, pp. 1040–1049, 2025

    2025

  13. [13]

    Emotional contagion on social media and the simulation of intervention strategies after a disaster event: A modeling study,

    M. Chu, W. Song, Z. Zhao, T. Chen, and Y .-c. Chiang, “Emotional contagion on social media and the simulation of intervention strategies after a disaster event: A modeling study,”Humanities and Social Sciences Communications, vol. 11, no. 1, 2024

    2024

  14. [14]

    Improved crowd dynamics analysis considering physical contact force and panic emotional propagation,

    R. Zhao, B. Wei, C. Han, P. Jia, W. Zhu, C. Li, and Y . Ma, “Improved crowd dynamics analysis considering physical contact force and panic emotional propagation,”IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 2, pp. 1840–1851, 2024

    2024

  15. [15]

    Generative agents: Interactive simulacra of human behavior,

    J. S. Park, J. O’Brien, C. J. Cai, M. R. Morris, P. Liang, and M. S. Bernstein, “Generative agents: Interactive simulacra of human behavior,” inProceedings of the 36th annual acm symposium on user interface software and technology, pp. 1–22, 2023

    2023

  16. [16]

    Agentsociety: Large-scale simulation of llm- driven generative agents advances understanding of human behaviors and society,

    J. Piao, Y . Yan, J. Zhang, N. Li, J. Yan, X. Lan, Z. Lu, Z. Zheng, J. Y . Wang, D. Zhou,et al., “Agentsociety: Large-scale simulation of llm- driven generative agents advances understanding of human behaviors and society,” 2025

    2025

  17. [17]

    An llm-enhanced agent-based simulation tool for information propagation.,

    Y . Hu, G. Sherpa, L. Zhang, W. Li, Q. Bai, Y . Wang, and X. Wang, “An llm-enhanced agent-based simulation tool for information propagation.,” inIJCAI, pp. 8679–8682, 2024

    2024

  18. [18]

    Yulan-onesim: Towards the next generation of social simulator with large language models,

    L. Wang, H. Gao, X. Bo, X. Chen, and J.-R. Wen, “Yulan-onesim: Towards the next generation of social simulator with large language models,” inWorkshop on Scaling Environments for Agents, 2025

    2025

  19. [19]

    Coe: A clue of emotion framework for emotion recognition in conversations,

    Z. Shen, Y . Pang, Y . Rao, and J. Yu, “Coe: A clue of emotion framework for emotion recognition in conversations,” inProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 23548–23563, 2025

    2025

  20. [20]

    Ortony, G

    A. Ortony, G. L. Clore, and A. Collins,The cognitive structure of emotions. Cambridge university press, 2022

    2022

  21. [21]

    On the logic of agent’s emotions,

    Y . Wang, Z. Liu, T. Liu, A. V . Samsonovich, and V . V . Klimov, “On the logic of agent’s emotions,”Cognitive Systems Research, vol. 88, p. 101281, 2024

    2024

  22. [22]

    Modeling Multi-Dimensional Cognitive States in Large Language Models under Cognitive Crowding

    L. Zhong, S. Zhu, Z. Yuan, J. Cui, X. Zhao, L. Wang, H. Chen, and Q. Liao, “Modeling multi-dimensional cognitive states in large language models under cognitive crowding,”arXiv preprint arXiv:2604.17174, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  23. [23]

    Bdi agents: from theory to practice.,

    A. S. Rao, M. P. Georgeff,et al., “Bdi agents: from theory to practice.,” inIcmas, vol. 95, pp. 312–319, 1995

    1995

  24. [24]

    Bdi agent architectures: A survey,

    L. De Silva, F. Meneguzzi, and B. Logan, “Bdi agent architectures: A survey,” in29th International Joint Conference on Artificial Intelligence, IJCAI 2020, International Joint Conferences on Artificial Intelligence, 2020

    2020

  25. [25]

    Emotional contagion,

    E. Hatfield, J. T. Cacioppo, and R. L. Rapson, “Emotional contagion,” Current directions in psychological science, vol. 2, no. 3, pp. 96–100, 1993

    1993

  26. [26]

    Conceptualising and measuring psychological preparedness for disaster: The psychological preparedness for disaster threat scale,

    J. McLennan, M. D. Marques, and D. Every, “Conceptualising and measuring psychological preparedness for disaster: The psychological preparedness for disaster threat scale,”Natural Hazards, vol. 101, no. 1, pp. 297–307, 2020

    2020

  27. [27]

    Psychical distance

    E. Bullogh, “Psychical distance” as a factor in art and an aesthetic principle,”Journal of Psychology, vol. 5, no. 2, pp. 87–118, 1912

    1912

  28. [28]

    Construal-level theory of psychological distance.,

    Y . Trope and N. Liberman, “Construal-level theory of psychological distance.,”Psychological review, vol. 117, no. 2, p. 440, 2010

    2010

  29. [29]

    Construal levels and psycho- logical distance: Effects on representation, prediction, evaluation, and behavior,

    Y . Trope, N. Liberman, and C. Wakslak, “Construal levels and psycho- logical distance: Effects on representation, prediction, evaluation, and behavior,”Journal of consumer psychology, vol. 17, no. 2, pp. 83–95, 2007

    2007

  30. [30]

    Recipient syringe sharing and its relation- ship to social proximity, perception of risk and preparedness to share,

    B. P. Smyth and A. Roche, “Recipient syringe sharing and its relation- ship to social proximity, perception of risk and preparedness to share,” Addictive Behaviors, vol. 32, no. 9, pp. 1943–1948, 2007

    1943

  31. [31]

    Judgmental discounting and environmen- tal risk perception: Dimensional similarities, domain differences, and implications for sustainability,

    A. Gattig and L. Hendrickx, “Judgmental discounting and environmen- tal risk perception: Dimensional similarities, domain differences, and implications for sustainability,”Journal of Social Issues, vol. 63, no. 1, pp. 21–39, 2007

    2007

  32. [32]

    Psychological distance model for environmental risk communication,

    S. She, M. Chao-Qun, L. Qiang, and C. Xie, “Psychological distance model for environmental risk communication,”Syst. Eng, vol. 9, 2012

    2012

  33. [33]

    Psychological distance,

    N. Liberman, Y . Trope, E. Stephan,et al., “Psychological distance,” Social psychology: Handbook of basic principles, vol. 2, no. 2, pp. 353– 383, 2007

    2007

  34. [34]

    Psychological stress and the coping process,

    R. S. Lazarus, “Psychological stress and the coping process,”The American Journal of Psychology, vol. 83, no. 4, 1966

    1966

  35. [35]

    Motivated closing of the mind:

    A. W. Kruglanski and D. M. Webster, “Motivated closing of the mind:” seizing” and” freezing.”,”Psychological review, vol. 103, no. 2, p. 263, 1996

    1996

  36. [36]

    Pychoagent: Psychology-driven llm agents for explainable panic prediction on social media during sudden disaster events,

    M. Liu, Z. Zhu, C. Ai, C. Gao, X. Li, L. He, K. Lai, Y . Chen, X. Lu, Y . Li,et al., “Pychoagent: Psychology-driven llm agents for explainable panic prediction on social media during sudden disaster events,” in Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pp. 17127–17145, 2025

    2025

  37. [37]

    Performance of social network sensors during hurricane sandy,

    Y . Kryvasheyeu, H. Chen, E. Moro, P. Van Hentenryck, and M. Cebrian, “Performance of social network sensors during hurricane sandy,”PLoS one, vol. 10, no. 2, p. e0117288, 2015

    2015

  38. [38]

    A new product growth for model consumer durables,

    F. M. Bass, “A new product growth for model consumer durables,” Management science, vol. 15, no. 5, pp. 215–227, 1969

    1969

  39. [39]

    Mass media and its impact on opinion dynamics of the nonlinear q-voter model,

    R. Muslim, R. A. Nqz, and M. A. Khalif, “Mass media and its impact on opinion dynamics of the nonlinear q-voter model,”Physica A: Statistical Mechanics and its Applications, vol. 633, p. 129358, 2024

    2024

  40. [40]

    Agent-based modeling of covid-19 vaccine uptake in new york state: Information diffusion in hy- brid spaces,

    F. Yin, N. Jiang, A. Crooks, and L. Laurian, “Agent-based modeling of covid-19 vaccine uptake in new york state: Information diffusion in hy- brid spaces,” inProceedings of the 7th ACM SIGSPATIAL International Workshop on GeoSpatial Simulation, pp. 11–20, 2024

    2024

  41. [41]

    Crowd: A social network simulation framework,

    A. N. N. Rende, T. Yilmaz, and O. Ulusoy, “Crowd: A social network simulation framework,”IEEE Transactions on Computational Social Systems, 2025

    2025

  42. [42]

    Modelling internet users’ negative emotion based on soar model,

    P. Wu, S. Qiang, and Q. Gao, “Modelling internet users’ negative emotion based on soar model,”Chin J Manag Sci, vol. 26, no. 3, pp. 126–138, 2018

    2018

  43. [43]

    Predicting emotion dynamics sequence on twitter via deep learning approach,

    D. Naskar, E. Onaindia, M. Rebollo, and S. R. Singh, “Predicting emotion dynamics sequence on twitter via deep learning approach,” in Proceedings of the 18th International Conference on Advances in Mobile Computing & Multimedia, pp. 20–24, 2020

    2020

  44. [44]

    Information diffusion prediction with graph neural ordinary differential equation network,

    D. Wang, W. Zhou, and S. Hu, “Information diffusion prediction with graph neural ordinary differential equation network,” inProceedings of the 32nd ACM International Conference on Multimedia, pp. 9699–9708, 2024

    2024

  45. [45]

    Ecr-chain: Advancing generative language models to better emotion-cause reasoners through reasoning chains,

    Z. Huang, J. Zhao, and Q. Jin, “Ecr-chain: Advancing generative language models to better emotion-cause reasoners through reasoning chains,”arXiv preprint arXiv:2405.10860, 2024

    work page arXiv 2024

  46. [46]

    Modeling cognitive-affective processes with appraisal and reinforcement learning,

    J. E. Zhang, J. Broekens, and J. P. Jokinen, “Modeling cognitive-affective processes with appraisal and reinforcement learning,”IEEE Transactions on Affective Computing, vol. 16, no. 2, pp. 771–782, 2024

    2024