PaperlayerView exact revision on arXiv

Source-linked AI summary

Designing for emotion regulation interventions: an agenda for HCI theory and research

Petr Slovak, Alissa N. Antle, Nikki Theofanopoulou, Claudia Daudén Roquet, James J Gross, Katherine Isbister

arXiv:2204.00118v2cs.HC

TL;DR

HCI has explored technology-enabled emotion-regulation support, but has not fully combined innovative designs with established psychological interventions. The paper synthesises HCI and psychology into a three-part design framework, finding that interventions are most likely to be effective when they support a range of emotion-regulatory processes.

  • Problem

    HCI work has explored emotion-regulation support, but the potential to combine innovative approaches with state-of-the-art psychological interventions remains unrealised.

    Source: §1 · p. 2

  • Method

    The authors synthesise psychological models and HCI intervention work into a framework covering intervention targets, delivery mechanisms, and practical design components.

    Source: Practical part summary (Section 4) · p. 5, §4.1 · p. 18, §6 · p. 34

  • Results

    Existing clinical-programme evidence suggests interventions are most likely to be effective when they support a range of emotion-regulatory processes.

    Source: Theory part summary (Section 2) · p. 4

  • Takeaways & Limitations

    The framework can help identify gaps, inform projects, and support synthesis across HCI and psychology.

    Source: §6 · p. 34

    Prior HCI work is largely early-stage, with little substantive data on long-term efficacy or effects on everyday emotion regulation and well-being.

    Source: How are these combined into… · p. 6

Abstract

from arXiv · show

There is a growing interest in HCI to envision, design, and evaluate technology-enabled interventions that support users' emotion regulation. This interest stems in part from increased recognition that the ability to regulate emotions is critical to mental health, and that a lack of effective emotion regulation is a transdiagnostic factor for mental illness. However, the potential to combine innovative HCI designs with the theoretical grounding and state-of-art interventions from psychology has yet to be fully realised. In this paper, we synthesise HCI work on emotion regulation interventions and propose a three-part framework to guide technology designers in making: (i) theory-informed decisions about intervention targets; (ii) strategic decisions regarding the technology-enabled intervention mechanisms to be included in the system; and (iii) practical decisions around previous implementations of the selected intervention components. We show how this framework can both systematise HCI work to date and suggest a research agenda for future work.

1 INTRODUCTION

HCI interest in technology-enabled emotion regulation is growing, but combining innovative HCI approaches with established psychological interventions remains incomplete. The paper proposes a three-part framework to guide intervention design and future research.

1 INTRODUCTION

  • HCI work has explored digital emotion-regulation support, but has not fully combined innovative approaches with state-of-the-art psychological interventions.The authors frame this as a step toward more robust systems and evaluation of real-world impacts.Source: §1 · p. 2
  • The paper takes stock of HCI work, synthesises psychological research, and identifies directions for future technology-enabled emotion-regulation interventions.Source: §1 · p. 2
  • The framework separates intervention design into theory-informed targets, strategic delivery mechanisms, and practical implementations.These are decisions about what to affect, how and when or where to offer support, and which design components to use.Source: §1 · p. 3
  • Its theory, strategic, and practical aspects bring psychological theory to HCI, distinguish delivery choices, and synthesise existing design components.The synthesis also helps identify gaps and potential research areas.Source: §1 · p. 3

EXECUTIVE SUMMARY

The executive summary links psychological models of emotion regulation to technology design, then synthesises HCI implementations and gaps. It argues that interventions are most likely to be effective when they support a range of emotion-regulatory processes.

EXECUTIVE SUMMARY

  • The Process Model offers intervention targets by distinguishing emotion generation from regulation and identifying processes designers could aim to affect.Emotion generation involves encountering, attending to, evaluating, and responding to situations; regulation involves evaluating one’s emotional state relative to goals and deciding whether to modify it.Source: Theory part summary (Section 2) · p. 4
  • The summary frames emotion regulation as four stages and five strategy families that cover different points in the regulation process.The stages are identification, strategy selection, implementation, and monitoring; the families are situation selection, situation modification, attentional deployment, cognitive change, and response modulation.Source: Theory part summary (Section 2) · p. 4
  • Existing clinical-programme evidence suggests interventions are most likely to be effective when they support a range of emotion-regulatory processes.The authors suggest combining multiple technology-enabled components into a broader intervention package when real-world impact on emotion-regulation capabilities is a concern.Source: Theory part summary (Section 2) · p. 4
  • Technology can support a specific strategy or an overarching regulation stage, while learning to regulate emotions is framed as embodied skill development rather than information delivery.Interactive experiences that elicit target emotions or support users during everyday emotional situations are presented as ways to support skill development and transfer.Source: Theory part summary (Section 2) · p. 4

Strategic part summary (Section 3)

The strategic framework asks designers to consider how, when, and where technology provides support, while the practical synthesis reviews HCI implementations and their gaps. Existing designs cluster around a limited set of approaches and often assume continuing technological support.

Strategic part summary (Section 3)

  • The strategic framework distinguishes didactic from experiential learning and offline from on-the-spot support, focusing on how, when, and where technology contributes.A game-based biofeedback intervention can combine offline and experiential mechanisms, and may also include didactic support.Source: Strategic part summary (Section 3) · pp. 4–5
  • Existing HCI work is relatively coherent, with only a small number of design approaches supporting each learning or delivery type.The practical synthesis catalogues components, combinations, and the emotion-regulation targets they support, based on a review of work published from 2009 through December 8, 2021.Source: Design components · pp. 5–6, Practical part summary (Section 4) · p. 5
  • Experiential designs commonly use biofeedback or implicit feedback, while didactic designs use reminders, recommenders, emotion-awareness prompts, or cognitive restructuring.Offline components include emotion-generating games, relaxation training, and visualisations for reflection; on-the-spot components use triggers and single-point or ongoing support.Source: Design components · p. 6
  • Interventions cluster into three largely independent areas: implicit on-the-spot support, biofeedback games, and reminder-recommenders or awareness systems.These clusters draw on different background literatures, goals, technologies, and recurring component combinations.Source: How are these combined into… · p. 6
  • Most systems target response modulation, while reminder-recommenders and awareness systems often support identification or situation selection rather than scaffolding specific strategies.Additional strategies are commonly introduced through psychoeducation without added technology support.Source: Common ER targets · pp. 6–7
  • 31/36 papers assume continuing system support, and the field remains fragmented with little substantive evidence on long-term efficacy or everyday impact.The authors describe much prior work as early-stage exploration rather than large-scale deployment or trials of changes in emotion-regulation skills.Source: How are these combined into… · p. 6, Common ER targets · p. 7

Research agenda and envisioned framework use (Section 5)

The framework can be used separately or together to identify gaps, inform projects, and support synthesis across disciplines. The research agenda highlights gaps in emotion-regulation targets, delivery mechanisms, and interdisciplinary progression.

Research agenda and envisioned framework use (Section 5)

  • The framework’s theory, strategic, and practical lenses can identify gaps, inform new projects, and support ongoing cross-disciplinary synthesis.Source: Research agenda and envisioned… · p. 7
  • The agenda identifies limited support for transfer, uneven coverage of emotion-regulation strategies, and a lack of efficacy data.Few systems aim to develop users’ emotion-regulation competence; response modulation predominates, and psychological impact is poorly understood.Source: ER targets gaps · p. 7
  • For each of the four delivery mechanisms, the authors propose research questions about how to develop effective support.Examples include supporting performative aspects of adaptive strategies and using technology to scaffold or guide emotion regulation in situ.Source: Intervention mechanism gaps · p. 7
  • The authors call for supporting research at different scales, from in-lab proof-of-concept innovation to large-scale real-world evaluations of efficacy.They offer a possible progression and examples from other domains.Source: Connecting interdisciplinary communities · p. 7

2 THEORY COMPONENT – EMOTION REGULATION MODELS (THE ‘WHATS’)

The theory component uses Gross’ Process Model to connect emotion generation, regulation processes, and strategy families, giving designers a systematic way to identify intervention targets. It emphasizes that strategies depend on where regulation intervenes in emotion generation and may be adaptive or maladaptive depending on context and use.

2.1 Psychological models of emotion regulation

  • Emotions arise through person–situation transactions involving attention, appraisal, and responses, while regulation modifies emotion generation in pursuit of a goal.Emotions are transient and situation-dependent, making them difficult to elicit in training without reinstating the situation or supporting in-situ strategy use.Source: §2.1 · pp. 8–9

2.2 What are the basic emotion regulation processes?

  • Emotion regulation is a dynamic process that can be intrinsic or extrinsic, explicit or implicit, and directed toward increasing, maintaining, or decreasing emotion.People identify a need to regulate, select a strategy, implement context-specific tactics, and monitor whether to continue, stop, or switch.Source: §2.2 · p. 9, §2.2.1 · pp. 9–10

2.2.1 Overarching emotion regulation process.

  • The overarching process comprises identification, strategy selection, implementation, and monitoring, with difficulties possible at each stage.For example, limited emotional awareness can affect identification, while inability to execute a tactic can affect implementation.Source: §2.2.1 · p. 10

2.2.2 Emotion regulation strategies and tactics.

  • The five strategy families target different points in emotion generation: situation selection, situation modification, attentional deployment, cognitive change, and response modulation.Examples include avoidance or problem solving, distraction, reappraisal or acceptance, and suppression; a strategy’s adaptiveness depends on context and short- and long-term effects.Source: §2.2.2 · pp. 10–11

2.3 Articulating the theory framework component – the ‘whats’

  • The model provides a systematic structure for identifying technology-relevant intervention targets by linking emotion generation, regulation processes, and strategies.The framework maps five strategy families to stages in emotion generation and distinguishes the four stages of regulation.Source: §2.3 · pp. 11–12

3 STRATEGY COMPONENTS: HOWS, WHENS, AND WHERES OF EMOTION REGULATION SUPPORT

Emotion regulation requires both knowledge of strategies and the ability to enact them in context, so the paper frames it as a skill with both intellectual and performative aspects. It develops a strategic framework around how, when, and where technology can support skill learning or application.

3.1 Emotion regulation as a ‘skill development’

  • Emotion regulation combines symbolic cognitive processes with a nonsymbolic, embodied goal, so skill acquisition can require both intellectual and performative learning.The paper uses this dual nature to motivate consideration of different kinds of technology-enabled support.Source: §3.1 · pp. 12–13

3.2 Exemplar ER interventions – didactic to experiential

  • ERT and UP predominantly teach skills didactically through psychoeducation, then complement it with practice in sessions and expected practice in daily life.Transfer from in-session practice to everyday use is a recognized challenge; ERT distinguishes offline training from on-the-spot application without therapist support.Source: §3.2 · pp. 13–14

3.3 Exemplar ER interventions – experiential to didactic

  • MBSR and MABT emphasize experiential, body-oriented practice to build awareness and tolerance of bodily sensations, often with coaching and staged exercises.Internal states are difficult to observe and provide feedback on, and learners may lack support transferring embodied skills into emotionally intense everyday moments.Source: §3.3 · pp. 14–15

3.4 Articulating the strategy framework component – hows, whens, wheres

  • The strategic framework asks designers to choose whether technology supports didactic or experiential learning and offline training or on-the-spot application.Its modular structure connects these delivery mechanisms to practical design components and is proposed to support empirically grounded understanding of what works.Source: §3.4 · pp. 15–16

4 PRACTICAL COMPONENT: SCOPING REVIEW OF HCI WORK

The practical component synthesizes HCI emotion-regulation technology designs into commonly used components organized by delivery mechanism. It aims to help designers consider implementations and identify gaps for improvement or innovation.

4 PRACTICAL COMPONENT: SCOPING REVIEW OF HCI WORK

  • The practical framework organizes existing HCI design components across experiential, didactic, offline, and on-the-spot delivery mechanisms.It is intended to help designers consider established approaches and where innovation may be possible.Source: §4 · p. 17
  • The scoping review develops a dictionary of design components and maps which regulation stages and strategy families existing HCI work supports or leaves understudied.It also examines how components have been combined into interventions and identifies remaining gaps.Source: §4 · p. 18
  • The practical and strategic framework parts connect delivery mechanisms with the intervention design components identified for each mechanism.Figure 5 provides an overview of those connections.Source: §4 · p. 18

4.1 Scoping review methodology

The authors conducted a scoping review of HCI work on interactive technology for emotion regulation, using defined inclusion criteria and iterative coding. The review’s final set comprised 36 papers from an initial search of 5,574.

4.1 Scoping review methodology

  • The search across ACM, IEEE, and Google Scholar covered January 1, 2009, through December 8, 2021, and yielded 5,574 papers, of which 36 met the review criteria.Included works were peer reviewed, addressed human emotion regulation, involved interactive technology and a user study, and aimed at an emotion-regulation intervention.Source: §4.1 · p. 18
  • Two researchers iteratively developed coding categories for experiential versus didactic learning and offline versus on-the-spot training, then independently coded about 25% of papers and reconciled differences.The codebook was finalized after they had read and analyzed about half the papers.Source: §4.1 · p. 18
  • The researchers also coded systems across six a priori dimensions, including technology type, research stage, intervention mechanism, specificity, and intended use case.These dimensions drew on prior research and theoretical grounding from psychology.Source: §4.1 · pp. 18–19

4.2 Overview of HCI work

The reviewed HCI systems span multiple technology platforms and research stages, but many remain exploratory and provide limited specification of how their design is expected to improve emotion regulation.

4.2 Overview of HCI work

  • Mobile devices (16 papers) and PCs/laptops (11 papers) were the most common platforms, alongside bespoke embedded prototypes (10) and smart watches (4).Many systems also used biosensors or non-traditional outputs such as haptics and ambient lighting.Source: §4.2 · p. 19
  • 13 papers were early-stage, 14 mid-stage, and 9 late-stage; about half had low (15 papers) or medium (12 papers) theory-of-change specificity.Late-stage work involved high-fidelity systems ready for wider deployment and summative evaluation.Source: §4.2 · pp. 19–20
  • Biofeedback (14 papers), physiological synchronisation (7), strategy reminders (6), and emotion-awareness systems (5) were the most common interaction approaches.Other bespoke systems appeared in 4 papers.Source: §4.2 · p. 20
  • 31 systems assumed ongoing support was needed, while only 5 aimed to reduce dependence so users could apply emotion-regulation strategies independently.In the former systems, the effects were expected to disappear if the technology was removed.Source: §4.2 · p. 20

4.3 Which intervention design components are used in each of the delivery mechanisms?

The review organizes intervention components by how they develop emotion-regulation skills and when or where support is delivered, then describes the interaction patterns and targets associated with each mechanism.

4.3 Which intervention design components are used in each of the delivery mechanisms?

  • The review characterizes components across four delivery mechanisms—experiential, didactic, offline, and on-the-spot—and describes their targets, use in HCI, and interaction patterns.Figures 6 and 7 overview the experiential/didactic and offline/on-the-spot contrasts.Source: §4.3 · p. 20

4.3.1 Experiential intervention components.

  • 23 papers used experiential components, mainly structured practice feedback or implicit cues; these approaches largely targeted response modulation, often breathing, and often did not build skills for use without the system.Real-time tailored biofeedback supported practice through a closed loop, while implicit target feedback nudged users toward a target state, often through vibrotactile cues.Source: §4.3.1 · pp. 20–22

4.3.2 Didactic intervention components.

  • Didactic components provide strategy suggestions, emotion-awareness and reflection support, or psychoeducation; their targets vary by mechanism, with CBT-based teaching spanning several strategy families.Unlike experiential components, they do not scaffold the application of strategies in the emotional context, though they may describe steps for later use.Source: §4.3.2 · pp. 22–23

4.3.3 Offline intervention components.

  • Offline components train skills in deliberate settings for later everyday use, through emotion-eliciting games, relaxation training, or reflection on emotion data; psychoeducation was usually delivered outside the technology.Game-based interventions mostly trained response modulation, while awareness interventions relied on the assumption that greater awareness improves regulation.Source: §4.3.3 · pp. 23–25

4.3.4 On-the-spot intervention components.

  • On-the-spot components support regulation in naturally occurring emotional situations, varying by trigger and duration; their interaction pattern is not intrinsically tied to a specific emotion-regulation strategy family.Triggers may be user-initiated, system-initiated, or situation-specific, while support may be discrete or ongoing.Source: §4.3.4 · pp. 25–26

4.4 How are the identified design components combined in existing interventions?

The reviewed interventions cluster into three mostly separate research directions, leaving combinations of delivery mechanisms unexplored and motivating cross-cluster connections.

4.4 How are the identified design components combined in existing interventions?

  • Existing systems form three largely independent clusters with consistent component combinations, and several combinations of delivery mechanisms remain unexplored.The clusters draw on different background literatures, intervention goals, and technologies.Source: §4.4 · pp. 26–27

4.4.1 Three clusters in the HCI work so far.

  • The clusters are implicit on-the-spot support (n = 12), biofeedback in interactive games (n = 12), and reminder-recommender and awareness systems (n = 11).The first centers on implicit experiential support; the second on experiential biofeedback with offline training; and the third on didactic systems delivered offline or on-the-spot.Source: §4.4.1 · pp. 27–28

4.4.2 Cluster X: identifying missing design component combinations.

  • Three combinations were absent: all four component types together, or experiential plus on-the-spot components combined with either didactic or offline components.The authors suggest combining game-based training with on-the-spot recommenders to address transfer, and pairing on-the-spot breathing support with awareness or psychoeducation.Source: §4.4.2 · p. 28

5 DISCUSSION AND RESEARCH AGENDA

The discussion uses the framework to identify gaps in ER targets and delivery mechanisms, and to show how it can guide research across HCI and psychology. It highlights limited transfer support, uneven strategy coverage, and sparse efficacy and process evidence, alongside opportunities for each delivery mechanism.

5.1.1 Lack of transfer support.

  • Fewer than 15% of interventions were designed to reduce dependence on the system; 31 implicitly assumed ongoing support, leaving transfer to everyday use a central gap.The authors suggest combining offline learning with on-the-spot scaffolding, or didactic instruction with experiential support, as possible approaches.Source: §5.1.1 · p. 30

5.1.2 Uneven support for ER strategies.

  • Most interactive components target response modulation, while cognitive reappraisal and attentional deployment are virtually unsupported and other ER stages receive little attention.The passage notes that response modulation strategies are theoretically known to be less effective in the long term.Source: §5.1.2 · p. 30

5.1.3 Lack of efficacy data on (multi-)component systems.

  • Efficacy evidence remains limited: only four field-based experimental studies and two RCTs were identified, and process measures of learning and real-world strategy use were largely absent.The review also reports few robust, high-fidelity systems ready for widespread deployment and few outcome measures tied to specific ER targets.Source: §5.1.3 · pp. 30–31

5.2.1 Didactic.

  • Didactic systems commonly deliver information, but fewer approaches scaffold practice or embed learning into daily life rather than relying on modules and reminders.The discussion raises step-by-step practice, personalized bite-sized learning, and role-play-inspired approaches as research directions.Source: §5.2.1 · p. 31

5.2.2 Experiential.

  • Experiential support has yet to address performative practice of adaptive strategies such as cognitive change, attentional deployment, or situation modification, or to guide users through emotional trajectories.The authors contrast these opportunities with feedback-only approaches and identify narrative, reflection, games, arts, and mixed reality as possible directions.Source: §5.2.2 · pp. 31–32

5.2.3 Offline.

  • Offline training contexts remain limited to stress-inducing games or relaxation support, with opportunities to broaden contexts and better couple biofeedback with psychoeducation.The discussion also calls for training under-supported ER stages such as identification or monitoring.Source: §5.2.3 · p. 32

5.2.4 On-the-spot.

  • On-the-spot designs span many possible techniques, but none is explored in depth; current approaches mainly use reminders or on-body haptic feedback and support few emotional processes.The discussion asks how designs might embed support in heightened everyday emotional moments and account for social interaction.Source: §5.2.4 · pp. 32–33

5.3 Guiding the interdisciplinary community development

  • The framework’s shared structure can orient researchers across HCI and psychology, support reuse of interaction components, and help promising prototypes progress toward real-world deployments.The proposed progression ranges from theory-informed component development to efficacy and effectiveness studies, including large-scale trials.Source: §5.3 · pp. 33–34

6 CONCLUSIONS

The paper synthesizes psychological foundations, plausible technology-enabled delivery mechanisms, and existing HCI work into a three-part framework. It presents the framework as a common structure for identifying gaps, informing projects, and connecting research across disciplines and stages of design readiness.

6 CONCLUSIONS

  • The framework combines psychology’s ER fundamentals, technology-enabled delivery mechanisms, and a synthesis of HCI implementations to shape future research at their intersection.Its three parts are theory, strategic, and practical.Source: §6 · p. 34
  • The framework aims to provide a common language and conceptual structure for connecting work across technological and design readiness and supporting interdisciplinary discussion of research questions.Researchers may use the practical component to identify gaps and the theory and strategic components to position their work.Source: §6 · p. 34
  • At study level, the framework can sharpen research questions and help teams reuse techniques, combining theory-informed targets, delivery mechanisms, and specific design choices.The example combines offline, experiential, and didactic components with psychoeducation, biofeedback, and narrative experiences in VR.Source: §6 · p. 34
  • The authors hope work across these levels will help turn interdisciplinary interest into a coherent research body with potential for substantial real-world impact.The paper invites researchers from HCI and psychology to contribute to that effort.Source: §6 · p. 34

APPENDIX 1 – SCOPING REVIEW OF EXISTING HCI SYSTEMS/INTERVENTION TARGETING ER

The appendix describes a synthesis of HCI research on technology-based interventions for learning, developing, and practising emotion regulation. It focuses on psychological mechanisms as well as technology and design dimensions, including how interventions support skill development or ongoing use.

APPENDIX 1 – SCOPING REVIEW OF EXISTING HCI SYSTEMS/INTERVENTION TARGETING ER

  • The synthesis covers HCI research on technology-based interventions that support emotion-regulation learning, development, and practice for adults and children.The review codes psychological mechanisms where possible alongside more commonly studied aspects such as technology type.Source: APPENDIX 1 – SCOPING REVIEW OF… · p. 41
  • The analysis used six dimensions: technology type, research stage, HCI focus, intervention mechanism, mechanism specificity, and ongoing versus skill-development support.The dimensions were derived from prior research, HCI reviews, and psychological theory concerning intervention change and learning.Source: APPENDIX 1 – SCOPING REVIEW OF… · p. 41
  • The review also coded whether systems supported offline versus on-the-spot training and didactic versus experiential learning.These additional dimensions followed the theoretical review and are addressed in the results section.Source: APPENDIX 1 – SCOPING REVIEW OF… · p. 41

6.1 Methodology

The authors used a scoping review to identify evidence and knowledge gaps in HCI research on technology-enabled ER interventions. They searched and filtered literature, then inductively developed and applied a coding framework to the included papers.

6.1 Methodology

  • The scoping review was chosen to map available evidence across fields, examine research methods, and identify knowledge gaps; the coded database and analysis code were to be made available through OSF.The authors describe the scoping procedure as systematic and replicable.Source: §6.1 · p. 41
  • The search returned 5,574 papers, and successive screening stages produced a final dataset of 36 papers whose research aimed to create ER interventions.The review targeted HCI publications from January 1, 2009 through December 8, 2021, using ACM, IEEE, and Google Scholar.Source: §6.1 · pp. 41–42
  • Researchers used inductive interpretive analysis to develop and iteratively refine categories and a codebook for coding the selected papers.They derived six dimensions after reviewing a subset of papers, then coded the full dataset using the finalized categories.Source: §6.1 · pp. 42–43
  • The six coding dimensions captured technology, research stage, HCI focus, intervention mechanism and specificity, and ongoing support versus skill development.The authors treated the last dimension as relevant to long-term effectiveness and the learning theories underpinning intervention design.Source: §6.1 · pp. 42–43

6.2 Results

The 36-paper review maps intervention technologies, research stages and foci, mechanisms, and mechanisms’ specificity. Most mechanisms are weakly specified, and few interventions explicitly support skills transfer or reduced dependence on ongoing system support.

6.2 Results

  • The dataset covers 36 peer-reviewed papers; mobile devices (16) and desktop or laptop computers (11) were common platforms, alongside smart watches (4) and bespoke embedded prototypes (10).The papers were sourced from ACM (63.9%), IEEE (19.4%), and Google Scholar (16.7%).Source: §6.2 · p. 43, §6.2.1 · p. 43

6.2.2 Stage of research.

  • The review classified 13 papers as early-stage, 14 as mid-stage, and 9 as late-stage research, based on research aims, evaluation design, and prototype robustness.Late-stage work involved robust, high-fidelity systems and summative lab or field evaluations.Source: §6.2.2 · p. 44

6.2.3 HCI research focus.

  • Intervention Efficacy Research had n=25; the dataset contained only one example of these co-design and social-interaction directions, and social factors were under-represented as an HCI research focus.Modality Research had n=9.Source: §6.2.3 · pp. 44–45
  • Efficacy studies spanned mechanisms, but often measured system usability or understanding rather than ER-skill change; biofeedback evaluations were more rigorous, yet mixed results and limited follow-up constrained evidence.The review notes that only two studies measured everyday-life transfer or followed up to assess maintenance.Source: §6.2.3 · pp. 44–45

6.2.4 Intervention mechanisms.

  • Biofeedback was the most commonly used mechanism (n=14); physiological-synchronization systems (n=7) had mixed results and no proposed mechanism for developing ER competency that persists without the system.Biofeedback systems aim for users to modify behavior to bring physiological or neurological states closer to a target, but most rely on users interpreting feedback; few explain how skills transfer.Source: §6.2.4 · pp. 46–47
  • Awareness systems (n=5) did not explain how displayed information leads to ER strategies; reminders and recommendations (n=6) faced mistimed sensor-triggered prompts, while other approaches included relaxation exercises and experiential experiences.Self-reported data combined with customized, just-in-time content was associated with positive effects over time; no reminder system described reducing reliance on prompts.Source: §6.2.4 · pp. 47–48

6.2.5 Specificity (of the intervention mechanism).

  • Most mechanisms were specified at low (n=15) or medium (n=12) levels; only a few successful interventions clearly linked multiple mechanisms to design features and ER-skill development.The review’s specificity coding considered how explicitly mechanisms and their links to intervention elements were described.Source: §6.2.5 · p. 48

6.2.6 Ongoing support vs Skills development.

  • Thirty-one papers assumed ongoing system support was needed, while five interventions included scaffolding or an explicit transfer mechanism; examples combined biofeedback with coaching or a shared parent-child experience.The latter examples aimed to support applying ER strategies beyond the intervention.Source: §6.2.6 · pp. 48–49

APPENDIX 2 – HOW ARE THE IDENTIFIED DESIGN COMPONENTS COMBINED IN EXISTING INTERVENTIONS?

The reviewed systems cluster into implicit on-the-spot support, biofeedback games, and reminder and awareness systems, with clusters corresponding to different combinations of delivery mechanisms. Cross-cluster combinations remain underexplored.

APPENDIX 2 – HOW ARE THE IDENTIFIED DESIGN COMPONENTS COMBINED IN EXISTING INTERVENTIONS?

  • The review identifies three mostly independent areas—implicit on-the-spot support, biofeedback games, and reminder and awareness systems—with distinct backgrounds, goals, and technology components.The areas align with recurring combinations of delivery mechanisms and design components.Source: APPENDIX 2 – HOW ARE THE IDENTIFIED… · p. 51
  • Several combinations remain unexplored, including systems mixing all four component types and combinations of experiential, on-the-spot components with didactic or offline components.The review frames these gaps as opportunities for future work.Source: APPENDIX 2 – HOW ARE THE IDENTIFIED… · p. 51

6.2.7 Cluster 1 – implicit on-the-spot support (experiential + on-the-spot) – n = 11.

  • Cluster 1 systems mainly provide experiential, implicit support alongside other activities, often using haptic or audio cues to guide breathing or heart rate in situation-specific contexts.This work nearly exclusively targets response modulation in Gross’ Process Model.Source: §6.2.7 · p. 52
  • Alternative Cluster 1 designs include biofeedback visualisation during tasks and user-initiated objects for respiration training or broader emotion-regulation approaches.The interactive toy combines attention deployment, in-situ response modulation, and potential cognitive change.Source: §6.2.7 · p. 52

6.2.8 Cluster 2 – bio-feedback in interactive games (experiential + offline +- didactic) – n = 12.

  • Cluster 2 mainly uses experiential biofeedback games (n=10), combining emotion-eliciting gameplay with real-time feedback; some add psychoeducation, while most lack direct support for transfer to everyday settings.Some games embed therapeutic techniques, whereas others overlay biofeedback on existing games; relaxation-focused systems pair biofeedback with psychoeducation.Source: §6.2.8 · pp. 52–53

6.2.9 Cluster 3 – recommender & awareness systems (didactic + offline || on-the-spot) – n = 13.

  • Cluster 3 combines awareness and reflection systems with recommenders, but has less established theories of change; only two of twelve papers were rated high in intervention-model specificity.Awareness systems pair emotional-data capture with later pattern reflection, while recommenders offer contextualized ER strategies, sometimes linked to offline psychoeducation.Source: §6.2.9 · p. 53
  • Awareness-and-reflection systems primarily target self-awareness and potentially situation selection; recommenders span several ER targets, but may cover strategies without sufficient depth absent prior psychoeducation.Source: §6.2.9 · p. 54

Related papers

Loading 2204.00118v2…