Recent advances in language and multimodal foundation models have created new opportunities to understand psychological states and to support mental health and psychotherapeutic interactions. However, current research remains predominantly text-centric, English-focused, and grounded in Western contexts, leaving multimodal signals, low-resource languages, and culturally diverse expressions of mental health insufficiently explored.
The First Workshop on Multimodal, Multilingual, and Multicultural Mental Health and Psychotherapy (MultiPsyche 2026) provides an interdisciplinary forum for addressing these limitations. We welcome empirical, methodological, resource, evaluation, survey, and position papers from natural language processing, multimodal learning, speech and vision, human-computer interaction, clinical psychology, psychiatry, and related fields. Submissions may address one or more of the workshop’s three core dimensions. Broader work on mental health and psychotherapy is also welcome.
We welcome submissions on topics including, but not limited to:
Multimodal Mental Health and Psychotherapy
Mental State Assessment from Text, Speech, Vision, and Physiological Signals
Cross-Modal Representation, Alignment, and Grounding
Temporal Modeling of Multimodal Interactions
Learning with Missing, Noisy, or Weakly Aligned Modalities
Multimodal Foundation Models, Generative Systems, and Embodied Interactions
Datasets, Benchmarks, and Evaluation for Multimodal Research
Multilingual Mental Health and Psychotherapy
Mental Health Analysis and Therapeutic Communication across Languages
Cross-Lingual Transfer, Code-Switching, and Low-Resource Adaptation
Translation and Language Mediation in Mental Health Settings
Multilingual Foundation Models for Mental Health Applications
Cross-Lingual Evaluation and Validity of Mental Health Measures
Multicultural Mental Health and Psychotherapy
Cultural Variation in Mental Health Expression, Stigma, Coping, and Help-Seeking
Culturally Grounded Assessment, Measurement, and Therapeutic Interaction
Cultural Bias, Fairness, and Pluralistic Alignment
Culturally Diverse Datasets, Annotation, and Benchmarks
General Topics in Mental Health and Psychotherapy
Mental Health Assessment, Risk Identification, and Longitudinal Modeling
Supportive Dialogue, Intervention Design, and Personalization
Foundation Models, Retrieval-Augmented Systems, and Agents
Data Collection, Annotation, Benchmarking, and Synthetic Data
Model Evaluation, Robustness, Explainability, and Uncertainty
Safety, Fairness, Privacy, Ethics, and Governance
Human-AI Collaboration, Clinical Validation, and Deployment
Archival submissions must describe original and unpublished work. We welcome long papers of up to eight pages and short papers of up to four pages of content, with unlimited pages for references. Accepted papers will be given one additional page in the camera-ready version.
Papers may be submitted directly to MultiPsyche or committed through ACL Rolling Review. Direct submissions must follow the official ACL style template, be fully anonymized, and undergo two-way anonymized peer review. ARR commitments will be assessed based on their existing reviews and meta-review, their relevance to the workshop, and program capacity.
Accepted archival papers will be published in the official workshop proceedings and included in the ACL Anthology, subject to the completion of all camera-ready, registration, and presentation requirements. All submissions must comply with applicable ACL policies, including those on originality, anonymity, publication ethics, and responsible research. A Limitations section is required, and an Ethical Considerations section is encouraged where relevant.
The non-archival track welcomes work in progress, preliminary findings, early-stage interdisciplinary research, work currently under review, and work already accepted or published elsewhere. It also welcomes position papers and emerging ideas that would benefit from community discussion. Non-archival submissions must be submitted directly through the workshop submission site; ARR commitments are not accepted for this track.
Unpublished submissions should take the form of an anonymized extended abstract of two to four pages of content, with unlimited pages for references. For work already accepted or published elsewhere, authors may submit either the original paper or a two-to-four-page summary and should clearly disclose its publication status.
Non-archival submissions will be evaluated for their relevance, clarity, scientific value, and potential to stimulate discussion. Accepted contributions will be presented at the workshop but will not be included in the workshop proceedings or the ACL Anthology. Authors of work under review elsewhere are responsible for ensuring compliance with the other venue’s submission policies.