Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Influence of prior and task generated emotions on XAI explanation retention and understanding

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Emotional arousal triggered during a feature's explanation is significantly associated with lower understanding of that feature's role in the AI's decision.

desk verdict A plausible but measurement-sensitive result: explanation-triggered arousal (via an unvalidated EmoNet-on-faces detector) may reduce XAI understanding; worth reviewing as a preliminary study, not yet strong evidence. read the letter →

arxiv 2505.10427 v2 pith:ZGJ45EZP submitted 2025-05-15 cs.HC

Birte Richter , Christian Schütze , Anna Aksonova , Britta Wrede This is my paper · ORCID
classification cs.HC
keywords explainableAIemotionalarousalexplanationretentionunderstandingdecisionsupportsystemfacialexpressionanalysisemotioninductionhuman-AIinteraction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper asks whether emotions—both those a user brings into an interaction and those triggered by the explanation itself—change what people take away from an AI's feature-relevance explanations. In a study with 24 participants receiving risk-advice from an embodied agent, the authors find that task-unrelated induced emotions (fear vs happiness) did not affect how many explained features people could recall, and at most marginally affected their understanding of feature influence. The central result is that emotional arousal detected during a feature's explanation is significantly negatively associated with understanding of that feature's contribution: in a mixed model, $\beta = -0.45$, $SE = 0.19$, $p = .015$. The authors interpret this as arousal during explanation hindering comprehension, and argue that XAI systems should seek the right level of arousal rather than assuming rational, emotion-free processing.

What carries the argument

The load-bearing mechanism is a thresholded anomaly detector on arousal. For each facial frame, EmoNet supplies an arousal value; a rolling z-score with $k = 2.5$ over a 500 ms window (Equations 1 and 2) flags an emotional reaction whenever the current arousal exceeds the recent mean by 2.5 standard deviations. Each explained feature segment is then coded as having caused an emotional reaction or not, and that binary predictor enters a binomial GLMM (logit link) with emotion induction and explained feature as fixed effects and participant ID as random intercept. Retention is measured by verbal recall mapped to variable names; understanding is measured by whether the participant correctly sorts each feature as increasing or decreasing risk. This arrangement lets the authors separate prior emotion, content salience, and momentary arousal as predictors of the two outcomes.

What would settle it

Re-run the analysis on the same 240 feature-explanation observations with arousal thresholds varied (for example, $k = 1.5$, $2.0$, $3.0$; window 250–1000 ms), or replace the facial arousal detector with HRV-derived or self-reported arousal for each feature. If the significant negative association between emotional reaction and understanding disappears, changes sign, or fails to replicate under any reasonable alternative detection setting, the claim that task-generated arousal hinders understanding would not survive.

Watch

Extended reading notes

Core claim

The paper's central claim is that task-generated emotional reactions, measured as brief arousal peaks during each feature's explanation, reduce the explainee's understanding of that feature's influence on the system's decision, while neither prior induced emotion nor arousal predicts whether the feature is verbally recalled. The evidence is a generalized linear mixed model on 240 feature-explanation observations: emotional reaction has a significant negative fixed effect on understanding ($\beta = -0.45$, $SE = 0.19$, $p = .015$), with a positive intercept indicating generally good baseline understanding. The authors also report that some features (gender, current health status, political orientation) significantly predict retention, that feature categories differ in how often they trigger arousal, and that the wrongly explained 'attitude towards future' feature produced a suggestive emotion-congruent pattern in understanding: 80% of fearful participants accepted the system's wrong explanation versus 44% of happy participants.

Load-bearing premise

The core result depends on treating a brief spike above a fixed threshold in a facial-expression arousal score as an emotional reaction; the paper does not validate that threshold or that the model, originally built for speech, works on faces, so the measured 'emotional reactions' could be an artifact of the detector.

Editorial extensions

If this is right

  • If the central result is correct, XAI evaluation should treat recall and understanding as separate constructs, because arousal can lower comprehension without lowering memory.
  • Explanation interfaces that monitor arousal could adapt timing or presentation to keep the explainee in a comprehension-friendly arousal range.
  • Designers should expect some features (gender, health, political orientation) to be inherently more memorable and more arousal-inducing, so per-feature statistics are needed rather than aggregate explanation metrics.
  • Prior emotions may shape how users judge a feature's influence in an emotion-congruent way, so systems should verify understanding rather than relying on users' self-reports.
  • The absence of a retention effect suggests emotionally salient explanations do not improve memory, contrary to a common assumption that arousal aids encoding.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: the single wrongly explained feature functions as an accidental experiment—comparing understanding for correct versus erroneous explanations under fear versus happiness could isolate error detection from arousal, a contrast the paper notes but does not model.
  • Beyond the paper: because the arousal detector threshold ($k = 2.5$, 500 ms) was not varied, the reported coefficient could partly reflect measurement noise; testing the same GLMM with $k$ from 1.5 to 3.5 and with HRV-based or self-reported arousal would show whether the negative association is robust.
  • Beyond the paper: if the effect replicates, the practical target is not removing emotion from XAI but regulating arousal—explanations could be paced, chunked, or preceded by a calming step, with user state fed back into explanation generation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper presents an empirical HCI study (N=24) in which an embodied decision support system explains ten feature contributions to a risk estimate, and the authors test whether prior task-unrelated emotions (fear vs. happiness) and task-generated emotional reactions influence retention and understanding of those explanations. Emotional reactions are detected from EmoNet arousal using a rolling z-score threshold (k = 2.5, 500 ms window), while retention and understanding are measured via verbal recall and a direction-sorting task. The main reported findings are that prior emotion did not significantly affect retention (and only marginally affected understanding), and that detected emotional reactions were significantly negatively associated with understanding in a GLMM (Section 4.6, β = -0.45, SE = 0.19, p = .015), interpreted as arousal during explanation reducing comprehension. The paper also reports feature-level variation in arousal and acknowledges a systematic explanation error for one feature.

Significance. If the RQ3b effect is reliable, the paper makes a useful contribution to XAI evaluation by connecting affective state during explanation to comprehension outcomes, with a concrete implication that explanation systems may need to monitor and pace content in response to arousal. The study has genuine strengths: it separates prior task-unrelated emotions from task-related reactions, uses a mixed-effects model with a participant-level random intercept, and collects multimodal measures (HRV, facial data, self-report). However, the central result currently rests on an unvalidated arousal detector, so the contribution is conditional on resolving the measurement concerns below.

major comments (4)
  1. [Section 3.1.1 / 4.6] The task-related emotion detector is not validated. EmoNet, as cited in reference [6], is a speech emotion recognition model, but it is applied here to facial frames; no check is reported that its arousal output is meaningful for faces in this setup. Equations (1)-(2) introduce an arbitrary threshold k = 2.5 and a 500 ms rolling window, and no sensitivity analysis is provided. Since the headline RQ3b result (Section 4.6, β = -0.45, SE = 0.19, p = .015) is computed from this detector, plausible changes in k or the window, or contamination from blinks, head movements, or speech-related facial motion, could turn the association into a measurement artifact. The manuscript itself states in Section 4.4 that the causes of the detected arousal remain unclear. I request a sensitivity analysis over thresholds and window lengths, a validity check against concurrently recorded HRV, SAM, or human-coded video data, and a careful statement of what the binary reaction variable actually captures.
  2. [Section 4.1] The emotion manipulation check is descriptive only. Section 4.1 reports median SAM valence and arousal per condition and a 'trend' in the expected direction, but no inferential test is reported. The conclusions on RQ1a (null retention effect) and RQ1b (marginal understanding effect) depend on the induction having actually worked; without an inferential manipulation check (e.g., Mann-Whitney U or mixed ANOVA with effect sizes), a null or marginal result could simply reflect a failed or weak manipulation. Please add formal tests and effect sizes for the SAM data, and ideally for the mDES and HRV measures as well.
  3. [Sections 4.2 / 4.4 / 4.6] The systematically erroneous explanation of the feature 'Einstellung bzgl. Zukunft' is acknowledged in Section 4.2 and shown to produce differential arousal between the Fear and Happy groups in Section 4.4, but this feature is retained in the GLMMs of Sections 4.5 and 4.6 without any control or exclusion. Because the error plausibly drives both emotional reactions and failures of understanding, the significant negative association between arousal and understanding may be confounded by this known stimulus artifact. I ask for a robustness analysis excluding this feature or including an error-flag covariate, and for reporting whether the arousal-understanding effect survives that analysis.
  4. [Sections 4.4 / 4.5 / 4.6] The definition of the emotional-reaction variable is inconsistent across sections. Section 4.4 states that a feature segment with one or more arousal bouts is counted as an emotional reaction, which is a binary variable; however, Figures 10 and 11 and the phrase 'number of emotional reactions' imply a count variable, and the model descriptions in Sections 4.5 and 4.6 do not state whether the fixed effect is binary or a count. The interpretation in Section 4.6 ('higher levels of positive emotional reactions') is also misleading because the detector captures arousal, not positive valence, and the variable as defined is binary. Please clarify the exact coding, provide the model formula, and align the wording with the actual variable definition.
minor comments (6)
  1. [Abstract] The abstract contains typos: 'ratantion' should be 'retention', and 'do not affected' should be 'did not affect'.
  2. [Section 3.1.3] 'Hear Rate Variability' should be 'Heart Rate Variability'.
  3. [Section 3.4] The phrase 'cf. Fig. 3.4' should refer to the correct figure number (likely Figure 2).
  4. [Figure 9] The caption placeholder 'Enter Caption' must be replaced with an actual descriptive caption.
  5. [Section 4.6] The coefficient 'β = ˘2.31' appears to be a rendering error for 'β = -2.31'; please correct it.
  6. [Section 4.2] There are typographical issues such as 'ANOV A' and 'numberemotionalreactions' with missing spacing; a careful proofread is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the RQ3b association is an empirical GLMM estimate with independently measured predictors and outcomes.

full rationale

This paper is an empirical interaction study, not a derivation. The central RQ3b result (Section 4.6) is a generalized linear mixed model predicting the binary outcome 'understanding' from emotional reactions, emotion induction, and explained feature, with a random intercept for participant. The predictor 'emotional reaction' is defined in Section 3.1.1 via EmoNet arousal with a rolling z-score threshold (k = 2.5, 500 ms window), and the outcome 'understanding' is measured separately by participants sorting features according to their remembered influence on the risk estimate (Section 3.1.2). These are independently operationalized measurements; neither quantity is defined in terms of the other, and no parameter of the arousal detector is fitted to maximize the reported association. The self-citations in the paper are methodological: the decision task and advice scheme are reused from a prior study [10], and the emotion induction is similar to [11]. These citations import experimental procedures, not the target result, and are therefore not load-bearing circularity. The absence of a sensitivity analysis for the arousal threshold, and the use of a speech-trained EmoNet model on facial frames, are legitimate threats to construct validity and robustness, but they are not circularity: a flawed or arbitrary measurement does not make the statistical claim equivalent to its inputs by construction. The paper itself concedes in Section 4.4 that 'What the underlying reasons for the arousal are, remains unclear, so far,' which further confirms that the detected events are not assumed to be the same thing as understanding. Accordingly, no circular step can be exhibited from the paper's own equations or citations, and the appropriate circularity score is 0.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central quantitative result depends on the arousal detector's threshold and the transfer of EmoNet from speech to faces; both are assumptions rather than validated measurement. The free parameters (k and window) are the main hand-set choices. No new entities are postulated.

free parameters (2)
  • Arousal z-score threshold k = 2.5
    Threshold in Equation (2) for classifying an emotional reaction; set by hand, no sensitivity analysis.
  • Rolling window length = 500 ms
    Window for rolling mean and SD in Equation (1); arbitrary choice.
assumptions (4)
  • domain assumption EmoNet arousal output from facial frames is a valid index of emotional arousal
    EmoNet is cited as a speech emotion recognition model [6]; using it for facial expression frames is an unvalidated transfer.
  • domain assumption Rolling z-score outliers correspond to meaningful emotional reactions
    The operationalization assumes that a 2.5 SD deviation over 500 ms reflects an emotional response rather than noise or movement artifact.
  • domain assumption The understanding task measures comprehension of the explanation
    Participants' binary higher/lower judgments are taken as understanding; with a faulty explanation, this conflates agreement with understanding.
  • standard math GLMM assumptions (binomial, logit, random intercept by participant)
    Standard mixed model assumptions used in Sections 4.5 and 4.6.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Influence of prior and task generated emotions on XAI explanation retention and understanding." pith.science (2026). https://pith.science/paper/ZGJ45EZP

@misc{pith2026250510427,
  author       = {Pith},
  title        = {Pith review of: Influence of prior and task generated emotions on XAI explanation retention and understanding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZGJ45EZP}},
  note         = {Machine review of arXiv:2505.10427}
}
read the original abstract

The explanation of AI results and how they are received by users is an increasingly active research field. However, there is a surprising lack of knowledge about how social factors such as emotions affect the process of explanation by a decision support system (DSS). While previous research has shown effects of emotions on DSS supported decision-making, it remains unknown in how far emotions affect cognitive processing during an explanation. In this study, we, therefore, investigated the influence of prior emotions and task-related arousal on the retention and understanding of explained feature relevance. To investigate the influence of prior emotions, we induced happiness and fear prior to the decision support interaction. Before emotion induction, user characteristics to assess their risk type were collected via a questionnaire. To identify emotional reactions to the explanations of the relevance of different features, we observed heart rate variability (HRV), facial expressions, and self-reported emotions of the explainee while observing and listening to the explanation and assessed their retention of the features as well as their influence on the outcome of the decision task. Results indicate that (1) task-unrelated prior emotions do not affected the ratantion but may affect the understanding of the relevance of certain features in the sense of an emotion-induced confirmation bias, (2) certain features related to personal attitudes yielded arousal in individual participants, (3) this arousal affected the understanding of these variables.

Figures

Figures reproduced from arXiv: 2505.10427 by the authors.

Figure 1
Figure 1. Study Setup during the first Hold and Laury decission (staged) experts (i.e., medical experts) have been identified as an important target group. This shift in focus has lead to a re-evaluation of existing research, identifying the need for more interactive approaches [8, 9]. However, the underlying assumption in this research has mostly been that interaction takes place with a rational decision maker who fol￾lows p… view at source ↗
Figure 2
Figure 2. Experimental procedure with six phases. 3.4.1 Questionnaire During the first phase of the interaction, the participants were asked to fill out an online questionnaire, with the virtual robot Flobi being present and asking the questions. A total of ten questions were asked. We refer to these questions as the “variables” or “features” that the system uses to compute and explain the risk type of the participant. Based … view at source ↗
Figure 3
Figure 3. Visual representation while Flobi was explaining which variables contributed to which risk type classification of the explainee. The red arrows indicated that the explainee’s value of this variable contributed to an estima￾tion of a lower risk type whereas a blue arrow indicated evidence for a higher risk type. 3.4.5 Assessment of user’s retention and understanding of the explanations and the task In the last phase,… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Visualization of the user interface to answer the question what effect the user’s value of the presented variable had on the system’s estimation of the user’s risk type. This task was used to measure the explainee’s retention of each variable. 4 Results 4.1 Emotion Man…
Figure 6
Figure 6. Figure 6: (left) Mean feature retention divided for task-unrelated prior emo￾tion. (right) Mean retention of ’wrong’ feature divided for task-unrelated prior emotion [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]
Figure 5
Figure 5. Figure 5: SAM scores for Valence (left) and Arousal (right), indicating higher (= more positive) valence for participants in the Happy condition, and a trend to higher self-reported arousal for participants in the Fear condition. 4.2 Influence of task-unrelated prior emotion on …
Figure 7
Figure 7. Figure 7: A comparison of the understanding of each presented feature, as tested by the recall task shown in [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 8
Figure 8. Figure 8: Arousal over the Feature Presentaion Time with emotional reac￾tion detection If a feature explanation segment contained one (or more) bouts of arousal – as indicated by the black lines – it was counted as a feature explanation causing an emotional reaction (or arousal)…
Figure 10
Figure 10. Figure 10: visualizes the mean retention as a function of the number of emotional reactions that a feature explanation evoked. The size of the bullet visualizes the number of occurrences. Here, the explanations of all 10 features for all 24 participants (10 x 24 = 240) have been…
Figure 11
Figure 11. Figure 11: The mean understanding of the feature divided for number of emotional reactions during the feature presentation for all (left) and divided by condition (right). 5 Discussion In the following, we will discuss our results regarding our initial re￾search questions [PITH…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Emotion-sensitive Explanation Model

    cs.HC 2025-05 conditional novelty 6.0 of 10

    The paper introduces a three-stage emotion-sensitive grounding model (arousal, understanding, agreement) for XAI systems to adapt explanations to users' emotional states.

Reference graph

Works this paper leans on

12 extracted references · 11 canonical work pages · cited by 1 Pith paper

  1. [6]

    Gerczuk, S

    M. Gerczuk, S. Amiriparian, S. Ottl, and B. W. Schuller. Emonet: A transfer learning framework for multi-corpus speech emotion recog- nition. IEEE Transactions on Affective Computing , 14(2):1472–1487, 2021

    work page 2021

  2. [1]

    Battefeld, S

    D. Battefeld, S. Mues, T. Wehner, P. House, C. Kellinghaus, J. Wellmer, and S. Kopp. Revealing the dynamics of medical diagnostic reasoning as step-by-step cognitive process trajectories. In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 46, 2024

    work page 2024

  3. [2]

    Bernardo and R

    E. Bernardo and R. Seva. Exploration of emotions developed in the lnteraction with explainable ai. In 2022 15th International Symposium on Computational Intelligence and Design (ISCID) , pages 143–146. IEEE, 2022

    work page 2022

  4. [3]

    M. M. Bradley and P. J. Lang. Measuring emotion: the self-assessment manikin and the semantic differential. Journal of behavior therapy and experimental psychiatry, 25(1):49–59, 1994

    work page 1994

  5. [4]

    Galanakis, A

    M. Galanakis, A. Stalikas, C. Pezirkianidis, I. Karakasidou, et al. Re- liability and validity of the modified differential emotions scale (mdes) in a greek sample. Psychology, 7(01):101, 2016

    work page 2016

  6. [5]

    J. M. George and E. Dane. Affect, emotion, and decision making. Or- ganizational behavior and human decision processes, 136:47–55, 2016

    work page 2016

  7. [7]

    O. Lammert. Can ai regulate your emotions? an empirical investigation of the influence of ai explanations and emotion regulation on human decision-making factors. In World Conference on Explainable Artificial Intelligence. Springer, forthcoming, 2025

    work page 2025

  8. [8]

    K. J. Rohlfing, P. Cimiano, I. Scharlau, T. Matzner, H. M. Buhl, H. Buschmeier, E. Esposito, A. Grimminger, B. Hammer, R. Häb- Umbach, et al. Explanation as a social practice: Toward a conceptual framework for the social design of ai systems. IEEE Transactions on Cognitive and Developmental Systems, 13(3):717–728, 2020

    work page 2020

Show all 12 references
  1. [9]

    Schmid and B

    U. Schmid and B. Wrede. What is missing in xai so far? An interdisci- plinary perspective. KI-Künstliche Intelligenz, 36(3):303–315, 2022

    2022

  2. [10]

    Schütze, O

    C. Schütze, O. Lammert, B. Richter, K. Thommes, and B. Wrede. Emo- tional debiasing explanations for decisions in hci. In H. Degen and S. Ntoa, editors, Artificial Intelligence in HCI. HCII 2023. Lecture Notes in Computer Science , volume 14050, pages 318–336. Springer, Cham, ...

    2023 doi

  3. [11]

    Thommes, O

    K. Thommes, O. Lammert, C. Schütze, B. Richter, and B. Wrede. Hu- man emotions in ai explanations. In L. Longo, S. Lapuschkin, and C. Seifert, editors, Explainable Artificial Intelligence , pages 270–293. Springer, 2024. doi: 10.1007/978-3-031-63803-9_15

    2024 doi

  4. [12]

    D. Wang, Q. Yang, A. Abdul, and B. Y . Lim. Designing theory-driven user-centric explainable ai. In Proceedings of the 2019 CHI conference on human factors in computing systems, pages 1–15, 2019

    2019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.