Engaging with digital tool to help detect early signs of anxiety, depression and borderline personality disorder in adolescents and young adults
| ISRCTN | ISRCTN55639530 |
|---|---|
| DOI | https://doi.org/10.1186/ISRCTN55639530 |
| Sponsor | University Medical Centre Maribor, Unit for pediatric and adolescent psychiatry |
| Funder | HORIZON EUROPE Framework Programme |
- Submission date
- 24/02/2026
- Registration date
- 09/03/2026
- Last edited
- 09/03/2026
- Recruitment status
- Recruiting
- Overall study status
- Ongoing
- Condition category
- Mental and Behavioural Disorders
Plain English summary of protocol
Background and study aims
Many children and teenagers experience mental health difficulties, but specialist services often have long waiting lists and limited staff, which can delay support. This study aims to test a new screening interview that can be used by a psychiatrist, a tablet-based chatbot, or a humanoid robot to see whether all three options can help identify mental health problems in young people in a clear, age-appropriate, and clinically useful way. The study also explores how young people feel about each type of interview and how mental health professionals rate the usefulness of the information they provide.
Who can participate?
Young people aged 12 to 20 years can take part. This includes those who are being seen by child and adolescent mental health services for anxiety, depression or borderline personality disorder, as well as young people without mental health diagnoses from schools and the community. All participants must speak Slovenian, be able to understand the questions and give informed consent or assent (with parent or guardian consent for those under 15).
What does the study involve?
Each participant attends a single study visit in a quiet room at University Medical Centre Maribor.Participants in both the clinical (with diagnoses of anxiety, depression and/or borderline personality disorder) and control (without mental health diagnoses) groups will be randomly allocated to one of two interview modes (chatbot or virtual agent). Randomisation will be performed at the individual level using a computer‑generated allocation list, with a 1:1 ratio between conditions. They sit in front of a tablet and complete a structured conversation led either by a text-based chatbot or a virtual agent. The session is video-recorded and sensors measure breathing, heart-related signals and the quality of the air they breathe out. Participants also complete short questionnaires about their mood, anxiety, personality traits, affinity to technology and how satisfied they were with the interview. A psychiatrist oversees the session, and at the end explains the true aims of the study and offers information on support services if needed.
What are the possible benefits and risks of participating?
Participants may benefit from reflecting on their feelings and from having contact with a mental health professional, but there is no direct medical benefit expected. The main risk is temporary emotional discomfort when answering questions about mood, anxiety or personal experiences, but participants can pause or stop at any time, and immediate psychiatric support is available if distress occurs. There are no invasive procedures: the study uses non-contact sensors and room air monitors, and all data are pseudonymised and handled under strict data protection rules.
Where is the study run from?
The study is run from the Child and Adolescent Psychiatry Unit, Clinic for Paediatrics, at University Medical Centre Maribor, Slovenia, in collaboration with the University of Maribor and international technology partners.
When is the study starting and how long is it expected to run for?
The broader CERTAIN project started in January 2025. This specific study is planned to start after ethics approval in late 2025 and is expected to run for around 6 months, or until about 120–240 participants have taken part.
Who is funding the study?
This research is funded by the European Commission's Horizon Europe research and innovation programme, project SMILE (Supporting Mental Health in Young People: Integrated Methodology for cLinical dEcisions and evidence-based interventions) , GA. 101080923 , and project CERTAIN (Certification for Ethical and Regulatory Transparency in Artificial Intelligence), GA. 101189650.
Who is the main contact?
Dr Izidor Mlakar, scientific supervisor, Faculty of Electrical Engineering and Computer Science, University of Maribor, HUMADEX Research group (email: izidor.mlakar@um.si)
Prof. Hojka Gregorič Kumperščak, clinical supervisor, child and adolescent psychiatrist at University Medical Centre Maribor (email: hojka.kumperscak@ukc-mb.si).
Contact information
Public, Scientific, Principal investigator
Koroška cesta 46
Maribor
2000
Slovenia
 ORCID ID |
0000-0002-4910-1879 |
|---|---|
| Phone | +386 2 220 7267 |
| izidor.mlakar@um.si |
Scientific, Principal investigator
Ljubljanska ulica 5
Maribor
2000
Slovenia
| ORCID ID |
0000-0001-8340-8674 |
|---|---|
| Phone | +386 (0)2 321 1000 |
| hojka.gregoric@guest.arnes.si |
Study information
| Primary study design | Observational |
|---|---|
| Observational study design | Cohort study |
| Participant information sheet | 49084 Participant Information Sheet - Slovenian.pdf |
| Scientific title | Multimodal pre-screening interviews with adolescents and young adults for developing an explainable artificial intelligence model and privacy-preserving synthetic data to assess risk of anxiety, depression and borderline personality disorder |
| Study acronym | SYN-MIND |
| Study objectives | This observational study is being conducted to understand whether structured, tablet-based pre-screening interviews with a chatbot or virtual agent can capture reliable digital signals of mental health problems in young people and support the safe development of explainable AI and synthetic data for psychiatry. Primary objective O1: To collect a high-quality, multimodal dataset of digital biomarkers (speech and language features, facial expressions, physiological signals and exhaled-air–related markers) from adolescents and young adults with and without anxiety, depression and borderline personality disorder, in order to enable the development and validation of explainable AI models for risk assessment and digital phenotyping. Secondary objectives O2: To compare interview acceptability (overall satisfaction, perceived communication quality and willingness to repeat) between chatbot‑led and virtual‑agent‑led pre-screening interviews. O3: To examine how interview acceptability differs by mental health status (anxiety, depression, borderline personality disorder, no diagnosis) and by personality traits and affinity to technology. O4: To explore associations between specific symptom dimensions (for example emotional dysregulation, identity disturbance, impulsivity, dissociation) and profiles of digital biomarkers. O5: To assess whether indoor air quality indicators, including volatile organic compounds related to exhaled breath, differ between diagnostic groups and are associated with mental health status. O6: To develop and evaluate methods for generating privacy-preserving synthetic data from the collected multimodal dataset while maintaining sufficient utility for model training and validation. |
| Ethics approval(s) |
Approved 13/10/2025, National Medical Ethics Committee of the Republic of Slovenia (Komisija za medicinsko etiko Republike Slovenije) (Štefanova ulica 5, Ljubljana, 1000, Slovenia; +386 1 478 60 01; gp.mz@gov.si), ref: 0120-466/2025-2711-3 |
| Health condition(s) or problem(s) studied | Early identification of anxiety, depression and borderline personality disorder in adolescents and young adults receiving outpatient and community mental health care |
| Intervention | Adolescents and young adults aged 12–20 years, with anxiety, depression, borderline personality disorder or no mental disorder (controls), will be recruited from child and adolescent mental health services and the community. Each participant will attend a single visit in a quiet clinical room and complete a structured pre‑screening interview on a tablet, led either by a chatbot or a virtual agent. The digital interview framework was designed under ISRCTN68006163 ( https://doi.org/10.1186/ISRCTN68006163) and described in the medRxiv preprint (https://www.medrxiv.org/content/10.1101/2025.10.13.25337952v1). The interview will be video‑recorded to capture speech, language and facial expressions. At the same time contactless sensors will passively record physiological signals (for example breathing and heart‑related signals) and devices will measure indoor air quality and exhaled‑air‑related markers during the session. Participants will, at the end, complete validated self‑report questionnaires on anxiety, depression, borderline personality features, personality traits, affinity to technology and satisfaction with the interview and communication, and the resulting multimodal dataset will be analysed to compare interview acceptability across modes and diagnostic groups, to explore associations between symptom dimensions, personality and digital/air‑quality biomarkers, and to train, validate and augment explainable AI models, including the development and evaluation of privacy‑preserving synthetic data. |
| Intervention type | Other |
| Primary outcome measure(s) |
|
| Key secondary outcome measure(s) |
|
| Completion date | 31/07/2026 |
Eligibility
| Participant type(s) | |
|---|---|
| Age group | Mixed |
| Lower age limit | 12 Years |
| Upper age limit | 20 Years |
| Sex | All |
| Target sample size at registration | 240 |
| Key inclusion criteria | Clinical group: 1. Adolescents and young adults aged 12–20 years at the time of consent/assent. 2. Currently in outpatient or inpatient care at University Medical Centre Maribor or in a collaborating child and adolescent psychiatry service. 2. Presence of symptoms consistent with at least one target condition: anxiety disorder, depressive disorder or borderline personality disorder, as determined by the treating clinician and/or clinical records. 3. Sufficient cognitive abilities to understand questions and provide meaningful answers. 4. Fluent in Slovenian. 5. Able and willing to complete the full study visit (interview, recordings and questionnaires). 6. Provision of informed consent by the participant (from age 15) and/or parent/legal guardian with assent from younger participants. Control group: 1. Adolescents and young adults aged 12–20 years at the time of consent/assent. 2. Recruited from schools, universities, community settings or paediatric services where they are being seen for non‑psychiatric reasons. 3. No history of formally diagnosed mental disorder and no clinically significant indication of mental disorder on screening questionnaires (borderline scores assessed case‑by‑case by the research psychiatrist). 4. Sufficient cognitive abilities to understand questions and provide meaningful answers. 5. Fluent in Slovenian. 6. Able and willing to complete the full study visit (interview, recordings and questionnaires). 7. Provision of informed consent by the participant (from age 15) and/or parent/legal guardian with assent from younger participants. |
| Key exclusion criteria | Clinical group: 1. Intellectual disability or developmental/behavioural disorder that prevents understanding or meaningful participation in the interview. 2. Acute, life‑threatening psychiatric condition or need for emergency intervention (for example severe agitation, aggression, acute psychosis). 3. Severe sensory or motor impairment (for example profound visual, hearing or speech impairment) that prevents effective interaction with the tablet‑based system and cannot be reasonably accommodated. 4. Severe unstable physical illness or acute/chronic pain that would interfere with participation or reliability of responses. 5. Acute intoxication with or withdrawal from psychoactive substances at the time of assessment. Control group 1. Any formally diagnosed mental disorder. 2. Clinically significant suspicion of mental disorder based on validated self‑report screening instruments, confirmed by the study psychiatrist. 3. Intellectual disability or developmental/behavioural disorder that prevents understanding or meaningful participation. 4. Acute psychiatric crisis or high current suicide risk requiring urgent care. 5. Severe sensory or motor impairment, severe unstable physical illness or acute intoxication/withdrawal, as above. |
| Date of first enrolment | 12/01/2026 |
| Date of final enrolment | 24/04/2026 |
Locations
Countries of recruitment
- Slovenia
Study participating centre
Maribor
2000
Slovenia
Results and Publications
| Individual participant data (IPD) Intention to share | No |
|---|---|
| IPD sharing plan summary | Not expected to be made available |
| IPD sharing plan |
Study outputs
| Output type | Details | Date created | Date added | Peer reviewed? | Patient-facing? |
|---|---|---|---|---|---|
| Other files | Consent form in Slovenian | 24/02/2026 | No | No | |
| Participant information sheet | in Slovenian | 24/02/2026 | No | Yes |
Additional files
- 49084 Consent Form - Slovenian.pdf
- Consent form in Slovenian
- 49084 Participant Information Sheet - Slovenian.pdf
- in Slovenian
Editorial Notes
24/02/2026: Trial's existence confirmed by National Medical Ethics Committee of the Republic of Slovenia (Komisija za medicinsko etiko Republike Slovenije).
