Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

The AI-Therapist Duo: Exploring the Potential of Human-AI Collaboration in Personalized Art Therapy for PICS Intervention

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper tries to establish that a human-in-the-loop AI art recommender can make art therapy for post-intensive care syndrome at least as effective as expert curation while cutting therapist preparation time by more than half.

desk verdict Useful efficiency result, but the headline claim of enhanced therapeutic outcomes is contradicted by the paper's own null results. read the letter →

arxiv 2502.09757 v1 pith:6DIOEYCH submitted 2025-02-13 cs.HC cs.AI

Bereket A. Yilma , Chan Mi Kim , Geke Ludden , Thomas van Rompay , Luis A. Leiva This is my paper · ORCID
classification cs.HCcs.AI
keywords post-intensivecaresyndromearttherapyhuman-in-the-loopvisualrecommendationhuman-AIcollaborationpersonalizationPICSrecommendersystems
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

Post-intensive care syndrome (PICS) is the cluster of physical, cognitive, and psychological problems that can follow an ICU stay, and art therapy by visual exposure is a promising but labor-intensive intervention. This paper tries to establish that a human-in-the-loop AI system—a visual art recommender that proposes candidate paintings for an art therapist to approve—can deliver therapy that is at least as effective as expert-only selection while saving the therapist more than half of their preparation time. The system works by asking the patient to choose one painting that resonates with their healing journey, computing similarity to thousands of paintings in a learned embedding space, and presenting a shortlist to the therapist for final curation. A 150-participant study found mood improvement after guided sessions in all conditions, and the AI-assisted shortlists matched expert-only choices across most quality ratings, with multimodal recommendations rated higher in novelty. The practical claim, if true, is that AI can handle scale and therapist judgment can handle safety.

What carries the argument

The load-bearing object is the human-in-the-loop visual art recommender (VA RecSys) pipeline. It takes the patient's chosen painting, represents all paintings as embeddings from a pre-trained model—ResNet-50 for visual-only and BLIP for visual-plus-text—computes cosine similarity between embeddings, and returns a ranked top-200 list. The art therapist then filters, modifies, or regenerates the list, turning an algorithmic shortlist into a curated set of nine paintings embedded in a guided session built on narrative-therapy questions. The mechanism does two jobs at once: the embedding-space ranking supplies scale and personalization from a single patient choice, and the therapist's gate keeps unsafe or irrelevant content out, which the paper shows is still necessary because even top-50 AI lists contained distressing imagery.

What would settle it

A randomized trial with clinically confirmed PICS patients, randomized to expert-only, HITL, and standard care, with standard anxiety, depression, and cognitive symptom scales at one and three months, would settle the claim. If the HITL arm shows no better symptom trajectories than expert-only, the paper's 'enhances therapeutic outcomes' claim would be unsupported despite the time savings.

Watch

Extended reading notes

Core claim

The central claim is stated in the discussion: a collaborative human-AI approach to art therapy indeed enhances both personalization and therapeutic outcomes for PICS patients. The quantitative backbone is the human-in-the-loop pipeline: a patient picks a seed painting; a ResNet-50-based visual recommender or a BLIP-based multimodal recommender ranks 2,325 gallery paintings by cosine similarity to the seed; the therapist selects the final nine paintings. In an expert evaluation, curation time dropped from 37.5 minutes to 15.5 minutes for visual HITL and 17.3 minutes for multimodal HITL. In the main study, 150 post-COVID hospital survivors in three between-subject groups completed a guided narrative-therapy session; all groups shifted from predominantly negative to predominantly positive moods, and the HITL groups were not statistically worse than the expert group on accuracy, diversity, immersion, or engagement, while multimodal HITL was rated more novel than visual HITL. The paper also identifies three new healing themes—togetherness versus solitude, awe, and escape and refuge—in participants' reflections on the recommended paintings.

Load-bearing premise

The load-bearing premise is that a single 24-minute online session—with COVID-hospitalized participants, not all of whom were in the ICU or formally diagnosed with PICS, reporting mood changes after viewing three paintings—measures therapeutic effectiveness for PICS.

Editorial extensions

If this is right

  • A therapist can prepare a three-painting art-therapy set in about 15-17 minutes instead of 37, so more patients could receive personalized sessions per therapist.
  • Guided art therapy with AI-assisted selections produces a large immediate mood lift, with the majority of participants moving from negative to positive affect after one session.
  • HITL recommendations are not inferior to expert-only selections on user-rated accuracy, diversity, immersion, or engagement, and multimodal HITL is more novel than visual HITL.
  • An expert remains essential even with strong AI recommendations, because automatic shortlists include paintings that are irrelevant or potentially distressing.
  • The same human-in-the-loop structure can be reused for other emotionally supportive domains where professional judgment must gate algorithmic suggestions.

Reading between the lines

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

  • Beyond the paper's stated scope, the efficiency gain is likely to compound on larger art collections: a top-200 shortlist filters a fixed amount of noise, so the time saved versus full manual browsing should grow as the corpus grows.
  • The three newly identified healing themes suggest a testable extension: instead of ranking by visual similarity alone, a recommender could also model the narrative affordances of paintings, such as 'escape and refuge,' and use those as explicit personalization dimensions.
  • If the mood effects replicate in a clinical sample, the same human-in-the-loop pattern could transfer to music, virtual nature scenes, or other sensory therapies that currently rely on therapist time for personalization.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes a human-in-the-loop (HITL) visual art recommender system for personalized art therapy aimed at post-intensive care syndrome (PICS), combining a therapist's curation with two AI backbones (ResNet-50 visual, BLIP multimodal). It reports an expert time-efficiency evaluation and a between-subjects user study (N=150) comparing Expert, HITL Visual, and HITL Multimodal recommendation sets on self-reported recommendation quality, mood change (Pick-A-Mood), PANAS scores, and qualitative reflections. The manuscript's central claim is that the HITL approach enhances personalization and therapeutic outcomes for PICS patients while also saving therapist time. The time-saving result is statistically supported, but the user-study results show essentially no significant differences between the HITL arms and the Expert arm on any recommendation-quality or affect measure, so the paper's affirmative answer to its research question is not supported by its own reported evidence.

Significance. If the HITL approach had been shown to improve therapeutic outcomes, this would be a meaningful contribution to HCI and digital health, with useful implications for therapist workload and scalable art-based interventions. The study has genuine strengths: a reasonably large between-subjects user study, quantitative measurement of expert curation time, and a qualitative thematic analysis that extends the authors' prior work on healing themes. However, the load-bearing claim of enhanced personalization and therapeutic effectiveness is contradicted by the null results in Sections 5.5.1 and 5.5.2; at best the data establish non-inferiority of HITL against expert curation plus an efficiency gain measured with a single expert. This overreach affects the abstract, the introduction, and the discussion, and it is not a local presentation issue.

major comments (5)
  1. [Abstract; §6; §7] The central claim is not supported by the paper's own results. The abstract states that the collaboration "enhances the personalization and effectiveness of art therapy," and §6 states that "a collaborative Human-AI approach in art therapy indeed enhances both personalization and therapeutic outcomes." However, §5.5.1 reports no statistically significant differences between either HITL arm and the Expert arm on accuracy, diversity, novelty, serendipity, immersion, or engagement; the only significant comparison is HITL Multimodal versus HITL Visual on novelty (p=.0472). §5.5.2 reports no significant group difference in mood change (χ²(2,N=38)=0.68, p=.710) and no significant group difference on any PANAS item. The data therefore support at most non-inferiority plus an efficiency benefit, not superiority in personalization or therapeutic outcomes. The research question should be answered negatively or the claims must be substantially weakened and reframed as an efficiency/non-inferiority result.
  2. [§5.2; §6; §7] The sample does not validate the PICS population the paper claims to study. PICS is defined as a syndrome following intensive care unit (ICU) stays, but the inclusion criterion in §5.2 is only "having been hospitalized post-COVID," and only 69 of the 150 participants reported having been in the ICU. No diagnostic confirmation of PICS is reported; the PHQ-4 items measure anxiety and depression symptoms, which are not equivalent to a PICS diagnosis. Thus statements about effects "for PICS patients" in the abstract and §6 overstate the population to which the results can be generalized. The authors should either recruit a verified PICS/ICU-survivor cohort or explicitly restrict all claims to a post-COVID hospitalization sample.
  3. [§5.4; §5.5; §6] The outcome measures do not support the phrase "therapeutic outcomes." The study measures immediate self-reported mood (Pick-A-Mood) and PANAS changes around a single, up-to-24-minute online session in which participants viewed three paintings. These are proximal affective/self-report proxies, not clinical or therapeutic outcomes, and no follow-up or longitudinal assessment is reported. Moreover, the within-participant mood improvement appears in all three arms, so it cannot distinguish the HITL conditions from Expert curation. The discussion's claim that the intervention "amplifies the effectiveness of therapy" requires either a clinical outcome measure or a demonstration of between-group superiority on the proximal measures, neither of which is present.
  4. [§4.1; Table 1] The efficiency result is based on a single expert selecting three paintings per condition, yielding only three time observations per approach. The ANOVA F(2,6)=57.56 has very few degrees of freedom, and with one expert the result cannot support broad claims about reducing "therapists'" workload. The expert's own quoted comments also note that even top-50 recommendations contained irrelevant or potentially distressing paintings. The time-saving claim should be presented as a preliminary single-expert finding, or additional experts and a larger selection task should be used before generalizing.
  5. [§5.5.1; §5.5.2] Several reported statistics are internally inconsistent and need correction. In §5.5.1 the significant novelty contrast is reported as "χ²(129) = −2.447, p=.0472," but a chi-square statistic cannot be negative and the degrees of freedom do not correspond to a pairwise contrast from the described LME model; this is likely a z or t statistic mislabeled as χ². In §5.5.2 the mood-change test is reported as "χ²(2, N=38)" while each group had 50 participants and Figure 5 displays counts totaling well above 38; the sample size used in this test is unexplained. These discrepancies make it difficult for a reader to verify the only significant user-facing result and the null mood result.
minor comments (5)
  1. [§3; §4] The notation is inconsistent: a painting is introduced as p_i and its embedding as bold p_i, but Equation (1) uses the same symbol for both the painting and its embedding; the text also contains a duplicated "and" in the embedding definition and the phrase "visual + tex VA RecSys" appears to be missing "tual." These should be cleaned up.
  2. [§2.2; References] The text attributes "BLIP [46]" in the related-work discussion of multimodal VA RecSys, but reference [46] is Yilma and Leiva's UMAP paper; the actual BLIP model is reference [20] (Li et al.). Please correct the citation or provide the original BLIP source.
  3. [§5.5.3; Figure 9] The t-SNE figure caption and text mention "non-linear projection t-SNE" and sample sentences, but the figure does not clearly indicate how the projection relates to the quantified sentiment percentages; a brief description of how sentence embeddings were mapped and how Figure 9 was generated would improve interpretability.
  4. [§4; Table 2] The table of healing themes lists ten themes but the coding process is described only in passing ("a random selection of participants responses... comprising 40%"); it would be helpful to state how many responses were coded, whether coding was done by one or multiple researchers, and whether inter-rater reliability was assessed.
  5. [Abstract; §1] The phrase "large-scale user study (N=150)" is defensible for an online study, but given that each of the three conditions contains only 50 participants and that the key comparisons are null, calling the sample "large-scale" in the abstract may overstate the statistical power; at minimum the power limitations should be acknowledged in the limitations section.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the user study is an independent empirical test, and the self-citations to prior work provide materials and architecture choices rather than the evidential basis for the central claim.

full rationale

This paper makes no formal derivation of its central claim; it reports a between-subjects user study (N=150) comparing Expert, HITL Visual, and HITL Multimodal recommendation arms. The recommendation scores are produced by cosine similarity to a user-chosen seed painting (Eq. 1), but these scores are not fitted to the outcome variables (PAM, PANAS, recommendation-quality ratings); user ratings are collected independently after exposure. The HITL arms use ResNet-50 and BLIP from prior work by the same group [45], and the stimuli follow that prior setup; those citations provide materials and architecture choices, not the evidential basis for the current outcome claims. The user study is self-contained against the Expert baseline, and the claim that Human-AI collaboration enhances personalization and therapeutic outcomes is an interpretation of the between-group comparisons, not a quantity derived from the model's own inputs. Therefore no step in the paper's argument reduces to its own definition, and no fitted parameter is renamed as a prediction. The fact that the reported statistics do not show significant HITL-over-Expert differences is a correctness or support concern, not circularity, and is outside this pass's scope.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

There is no formal derivation; the study's conclusions rest on the validity of the sample, the pre-trained embeddings, and self-report mood measures. One design parameter (top-200) is hand-chosen.

free parameters (1)
  • top-k recommendation list size = 200
    Hand-chosen cutoff for the recommendation shortlist presented to the therapist; no sensitivity analysis is provided, and it directly affects the time-saving result.
assumptions (5)
  • domain assumption Nature-themed visual art exposure improves mood and psychological states in ICU or PICS patients.
    Sections 2.1 and 4 cite prior art therapy and nature exposure evidence; the study does not re-test this assumption.
  • domain assumption Prolific participants hospitalized for COVID are a representative proxy for PICS patients.
    Section 5.2 screening only required COVID hospitalization; only 69 of 150 report ICU stay, and no PICS diagnosis is required.
  • domain assumption Pre-trained ResNet-50 and BLIP embeddings are appropriate for therapeutic art similarity.
    Section 3 uses these models; there is no validation that embedding similarity tracks therapeutic relevance beyond expert curation.
  • domain assumption Single-session, 24-minute online mood change (PAM and PANAS) is a meaningful proxy for therapeutic effectiveness.
    Section 5.4 measures immediate mood before and after; no longitudinal or clinical outcomes are assessed.
  • standard math Inferential statistics (LME, chi-square, pairwise t-tests) are valid for the study design.
    Statistical methods are used in Sections 4.2 and 5.5 without formal proofs, which is standard practice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The AI-Therapist Duo: Exploring the Potential of Human-AI Collaboration in Personalized Art Therapy for PICS Intervention." pith.science (2026). https://pith.science/paper/6DIOEYCH

@misc{pith2026250209757,
  author       = {Pith},
  title        = {Pith review of: The AI-Therapist Duo: Exploring the Potential of Human-AI Collaboration in Personalized Art Therapy for PICS Intervention},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6DIOEYCH}},
  note         = {Machine review of arXiv:2502.09757}
}
read the original abstract

Post-intensive care syndrome (PICS) is a multifaceted condition that arises from prolonged stays in an intensive care unit (ICU). While preventing PICS among ICU patients is becoming increasingly important, interventions remain limited. Building on evidence supporting the effectiveness of art exposure in addressing the psychological aspects of PICS, we propose a novel art therapy solution through a collaborative Human-AI approach that enhances personalized therapeutic interventions using state-of-the-art Visual Art Recommendation Systems. We developed two Human-in-the-Loop (HITL) personalization methods and assessed their impact through a large-scale user study (N=150). Our findings demonstrate that this Human-AI collaboration not only enhances the personalization and effectiveness of art therapy but also supports therapists by streamlining their workload. While our study centres on PICS intervention, the results suggest that human-AI collaborative Art therapy could potentially benefit other areas where emotional support is critical, such as cases of anxiety and depression.

Figures

Figures reproduced from arXiv: 2502.09757 by the authors.

Figure 1
Figure 1. Overview of our art therapy approach for PICS prevention and treatment. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Diagram of the VA RecSys pipeline, from automatic feature extraction to automatic painting selection. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Overview of the painting selection process using three different approaches: Expert, HITL Visual, and HITL Multimodal. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Distribution (box plots) of user ratings for the user-centric dependent variables of recommendation quality. Dots denote mean [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Mood improvement comparison before and after art therapy. [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Individual Positive Affect Negative Affect Schedule (PANAS) scores per dimension. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Emotion score changes according to the Positive Affect Negative Affect Schedule (PANAS) scale. [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Sentiment analysis of user reflections per group. [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Non-linear projection (t-SNE) of sentence embeddings from user reflections. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

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. Affect-aware Cross-Domain Recommendation for Art Therapy via Music Preference Elicitation

    cs.IR 2025-07 reject novelty 5.0 of 10

    A 200-person study of music-driven cross-domain recommendation for art therapy shows music-based and visual-based engines perform equally, contradicting the paper's 'outperforming' claim.

Reference graph

Works this paper leans on

46 extracted references · 36 canonical work pages · cited by 1 Pith paper

  1. [1]

    J Appleton. 1975. The Experience of Landscape John Wiley. New York (1975)

    work page 1975

  2. [2]

    David M Blei, Andrew Y Ng, and Michael I Jordan. 2003. Latent dirichlet allocation. Journal of machine Learning research 3, Jan (2003), 993–1022

    2003

  3. [3]

    Martin Daniel Cooney and Maria Luiza Recena Menezes. 2018. Design for an art therapy robot: An explorative review of the theoretical foundations for engaging in emotional and creative painting with a robot. Multimodal Technologies and Interaction 2, 3 (2018), 52

    work page 2018

  4. [4]

    Shane Cross, Imogen Bell, Jennifer Nicholas, Lee Valentine, Shaminka Mangelsdorf, Simon Baker, Nick Titov, Mario Alvarez-Jimenez, et al. 2024. Use of AI in Mental Health Care: Community and Mental Health Professionals Survey. JMIR Mental Health 11, 1 (2024), e60589

    work page 2024

  5. [5]

    BH Cuthbertson, J Rattray, Marion Kay Campbell, M Gager, Susanna Roughton, A Smith, A Hull, S Breeman, J Norrie, David Jenkinson, et al. 2009. The PRaCTICaL study of nurse led, intensive care follow-up programmes for improving long term outcomes from critical illness: a pragmatic randomised controlled trial. Bmj 339 (2009)

    work page 2009

  6. [6]

    Dimitry S Davydow, Douglas Zatzick, Catherine L Hough, and Wayne J Katon. 2013. In-hospital acute stress symptoms are associated with impairment in cognition 1 year after intensive care unit admission. Annals of the American Thoracic Society 10, 5 (2013), 450–457

    work page 2013

  7. [7]

    Pieter MA Desmet, Martijn H Vastenburg, and Natalia Romero. 2016. Mood measurement with Pick-A-Mood: review of current methods and design of a pictorial self-report scale. Journal of Design Research 14, 3 (2016), 241–279

    2016

  8. [8]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)

    arXiv 2018

Show all 46 references
  1. [9]

    Gregory B Diette, Noah Lechtzin, Edward Haponik, Aline Devrotes, and Haya R Rubin. 2003. Distraction therapy with nature sights and sounds reduces pain during flexible bronchoscopy: A complementary approach to routine analgesia. Chest 123, 3 (2003), 941–948

    2003

  2. [10]

    David W Dowdy, Mark P Eid, Artyom Sedrakyan, Pedro A Mendez-Tellez, Peter J Pronovost, Margaret S Herridge, and Dale M Needham. 2005. Quality of life in adult survivors of critical illness: a systematic review of the literature. Intensive care medicine 31 (2005), 611–620

    2005

  3. [11]

    Amelia Fiske, Peter Henningsen, and Alena Buyx. 2019. Your robot therapist will see you now: ethical implications of embodied artificial intelligence in psychiatry, psychology, and psychotherapy. Journal of medical Internet research 21, 5 (2019), e13216

    2019

  4. [12]

    John Griffiths, Robert A Hatch, Judith Bishop, Kayleigh Morgan, Crispin Jenkinson, Brian H Cuthbertson, and Stephen J Brett. 2013. An exploration of social and economic outcome and associated health-related quality of life after critical illness in general intensive care unit ...

    2013

  5. [13]

    Suzanne Haeyen, Farid Chakhssi, and Susan Van Hooren. 2020. Benefits of art therapy in people diagnosed with personality disorders: a quantitative survey. Frontiers in Psychology 11 (2020), 686

    2020

  6. [14]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778

    2016

  7. [15]

    Shigeaki Inoue, Junji Hatakeyama, Yutaka Kondo, Toru Hifumi, Hideaki Sakuramoto, Tatsuya Kawasaki, Shunsuke Taito, Kensuke Nakamura, Takeshi Unoki, Yusuke Kawai, et al. 2019. Post-intensive care syndrome: its pathophysiology, prevention, and future directions. Acute medicine &...

    2019

  8. [16]

    Dacher Keltner and Jonathan Haidt. 2003. Approaching awe, a moral, spiritual, and aesthetic emotion. Cognition and emotion 17, 2 (2003), 297–314

    2003

  9. [17]

    Chan Mi Kim. 2024. Towards healing through digital nature in critical care and beyond: An integrated approach to design nature-based aesthetic experiences to promote delirium prevention and patient wellbeing . PhD Thesis - Research UT, graduation UT. University of Twente, Neth...

    2024 doi

  10. [18]

    Chan Mi Kim, Thomas Van Rompay, and Geke Ludden. 2024. Outside In: Creating Digital Nature Tailored To The Needs of Intensive Care Unit Patients. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems . 1–7

    2024

  11. [19]

    Chan Mi Kim, Thomas JL van Rompay, Gijs LM Louwers, Jungkyoon Yoon, and Geke DS Ludden. 2023. From a Morning Forest to a Sunset Beach: Understanding Visual Experiences and the Roles of Personal Characteristics for Designing Relaxing Digital Nature. International Journal of Hum...

    2023

  12. [20]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. arXiv preprint arXiv:2201.12086 (2022)

    2022 arXiv

  13. [21]

    When He Feels Cold, He Goes to the Seahorse

    Di Liu, Hanqing Zhou, and Pengcheng An. 2024. " When He Feels Cold, He Goes to the Seahorse"—Blending Generative AI into Multimaterial Storymaking for Family Expressive Arts Therapy. In Proceedings of the CHI Conference on Human Factors in Computing Systems . 1–21

    2024

  14. [22]

    Bernd Löwe, Inka Wahl, Matthias Rose, Carsten Spitzer, Heide Glaesmer, Katja Wingenfeld, Antonius Schneider, and Elmar Brähler. 2010. A 4-item measure of depression and anxiety: validation and standardization of the Patient Health Questionnaire-4 (PHQ-4) in the general populat...

    2010

  15. [23]

    Xuexing Luo, Aijia Zhang, Yu Li, Zheyu Zhang, Fangtian Ying, Runqing Lin, Qianxu Yang, Jue Wang, and Guanghui Huang. 2024. Emergence of Artificial Intelligence Art Therapies (AIATs) in Mental Health Care: A Systematic Review. International Journal of Mental Health Nursing (2024)

    2024

  16. [24]

    Stephen Madigan. 2011. Narrative therapy. American Psychological Association

    2011

  17. [25]

    M Marć, A Bartosiewicz, J Burzyńska, Z Chmiel, and P Januszewicz. 2019. A nursing shortage–a prospect of global and local policies. International nursing review 66, 1 (2019), 9–16

    2019

  18. [26]

    Upali Nanda, S Eisen, Rana S Zadeh, and Deborah Owen. 2011. Effect of visual art on patient anxiety and agitation in a mental health facility and implications for the business case. Journal of psychiatric and mental health nursing 18, 5 (2011), 386–393

    2011

  19. [27]

    Upali Nanda, Sarajane L Eisen, and Veerabhadran Baladandayuthapani. 2008. Undertaking an art survey to compare patient versus student art preferences. Environment and Behavior 40, 2 (2008), 269–301

    2008

  20. [28]

    Dale M Needham, Victor D Dinglas, Peter E Morris, James C Jackson, Catherine L Hough, Pedro A Mendez-Tellez, Amy W Wozniak, Elizabeth Colantuoni, E Wesley Ely, Todd W Rice, et al. 2013. Physical and cognitive performance of patients with acute lung injury 1 year after initial ...

    2013

  21. [29]

    Kathy O’Reilly. 2019. Philips introduces VitalMinds, new non-pharmacological approach to help reduce delirium in the ICU . https://www.philips.com/a- w/about/news/archive/standard/news/press/2019/20190425-philips-introduces-vitalminds-new-non-pharmacological-approach-to-help-r...

    2019

  22. [30]

    Pratik P Pandharipande, Timothy D Girard, James C Jackson, Alessandro Morandi, Jennifer L Thompson, Brenda T Pun, Nathan E Brummel, Christopher G Hughes, Eduard E Vasilevskis, Ayumi K Shintani, et al. 2013. Long-term cognitive impairment after critical illness. New England Jou...

    2013

  23. [31]

    Pearl Pu, Li Chen, and Rong Hu. 2011. A user-centric evaluation framework for recommender systems. In Proceedings of the fifth ACM conference on Recommender systems. 157–164

    2011

  24. [32]

    Dharmanand Ramnarain, Emily Aupers, Brenda den Oudsten, Annemarie Oldenbeuving, Jolanda de Vries, and Sjaak Pouwels. 2021. Post Intensive Care Syndrome (PICS): an overview of the definition, etiology, risk factors, and possible counseling and treatment strategies. Expert Revie...

    2021

  25. [33]

    Gautam Rawal, Sankalp Yadav, and Raj Kumar. 2017. Post-intensive care syndrome: an overview. Journal of translational internal medicine 5, 2 (2017), 90–92

    2017

  26. [34]

    Dafna Regev and Liat Cohen-Yatziv. 2018. Effectiveness of art therapy with adult clients in 2018—what progress has been made? Frontiers in psychology 9 (2018), 1531

    2018

  27. [35]

    Fujiko Robledo Yamamoto, Amy Voida, and Stephen Voida. 2021. From therapy to teletherapy: Relocating mental health services online. Proceedings of the ACM on Human-Computer Interaction 5, CSCW2 (2021), 1–30

    2021

  28. [36]

    Karin Alice Schouten, Gerrit J de Niet, Jeroen W Knipscheer, Rolf J Kleber, and Giel JM Hutschemaekers. 2015. The effectiveness of art therapy in the treatment of traumatized adults: A systematic review on art therapy and trauma. Trauma, violence, & abuse 16, 2 (2015), 220–228...

    2015

  29. [37]

    Fereshtehossadat Shojaei, John Osorio Torres, and Patrick C Shih. 2024. Exploring the integration of technology in art therapy: insights from interviews with art therapists. Art Therapy (2024), 1–7

    2024

  30. [38]

    R Ulrich and L Gilpin. 2003. HealingArts-NutritionforTheSoul. Charmel, PM, Frampton, SB, Gilpin, L (2003), 89–104

    2003

  31. [39]

    Roger S Ulrich. 1984. View through a window may influence recovery from surgery. science 224, 4647 (1984), 420–421

    1984

  32. [40]

    Roger S Ulrich, Outi Lunden, and JL Eltinge. 1993. Effects of exposure to nature and abstract pictures on patients recovering from heart surgery. Psychophysiology 30, 7 (1993)

    1993

  33. [41]

    van der Maaten and G.E

    L.J.P. van der Maaten and G.E. Hinton. 2008. Visualizing Data Using t-SNE. J. Mach. Learn. Res. (2008), 2579–2605

    2008

  34. [42]

    Fahui Wang and Wei Luo. 2005. Assessing spatial and nonspatial factors for healthcare access: towards an integrated approach to defining health professional shortage areas. Health & place 11, 2 (2005), 131–146

    2005

  35. [43]

    David Watson, Lee Anna Clark, and Auke Tellegen. 1988. Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of personality and social psychology 54, 6 (1988), 1063

    1988

  36. [44]

    Edward O Wilson. 1986. Biophilia. Harvard university press

    1986

  37. [45]

    Yilma, Chan Mi Kim, Gerald C

    Bereket A. Yilma, Chan Mi Kim, Gerald C. Cupchik, and Luis A. Leiva. 2024. Artful Path to Healing: Using Machine Learning for Visual Art Recommendation to Prevent and Reduce Post-Intensive Care Syndrome (PICS). In Proceedings of the CHI Conference on Human Factors in Computing...

    2024

  38. [46]

    Yilma and Luis A

    Bereket A. Yilma and Luis A. Leiva. 2023. Together Yet Apart: Multimodal Representation Learning for Personalised Visual Art Recommendation. In Proceedings of the 31st ACM Conference on User Modeling, Adaptation and Personalization (Limassol, Cyprus) (UMAP ’23). Association fo...

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.