WP1 - Assess [Months: 1-48] WP Lead: REGIONH
- Automatic Assessment of Child-Parent Attachment.
- Electronic behavioural smartphone-based biomarker for depression and bipolar disorder and integration with Cognitive Behavioural Therapy.
- Automatic personalisation of mobile intelligent user interfaces to improve the effects of persuasive technologies for mental wellbeing.
- Intelligent techniques to support online mental health communities and recommend resources from 4 levels of stepped care.
Alberto Gonzalez Olmos, Host: UoG, Partners: NSPCC [Secondment 6 months]
The Manchester Child Attachment Story Task (MCAST) is the standard method used in middle childhood to assess attachment. This project will develop and evaluate the first automated technique for detecting and classifying attachment in realtime. This will be achieved using sensor-augmented toys and recorded storytelling and will be validated against MCAST. Alberto will investigate the sensing required to record doll and child movements and will then prototype sophisticated tracking systems to detect doll play and other related activity in real time. Alberto will develop novel algorithms that can do basic classifications of the 4 different categories of Attachment from mixed sensor inputs and will design a software application for the delivery of MCAST in an automated way. Alberto will undertake an intersectoral secondment at NSPCC comparing the results to a standard MCAST to assess effectiveness and accuracy.
2. Electronic behavioural smartphone-based biomarker for depression and bipolar disorder and integration with Cognitive Behavioural Therapy.
Sigurd Arne Melbye Host: RHP, Partners: DTU [Secondment 6 months]
This study is highly original and potentially ground breaking. It will introduce completely new methods for valid diagnosis and the subsequent treatment of mental disorders in young people. Sigurd will investigate key biomarkers in automatically generated smartphone data on social activity, physical activity and voice features to help discriminate young people with depression and bipolar disorder from healthy young people. This will be done in a partnership with young people with depression and bipolar disorder (subject to strict ethical guidelines), clinicians from Mental Health Services, Capital Region of Denmark and DTU. Sigurd to develop design guidelines for a smartphone-based cognitive behavioural therapy module, which will be developed in collaboration with Pegah Hafiz who will provide the technical build.
3. Automatic personalisation of mobile intelligent user interfaces to improve the effects of persuasive technologies for mental wellbeing.
Mohammed Khwaja, Host: TIA, Partners: UCD [Secondment 6 months]
This IRP will go beyond generic intervention designs by combining user modelling with personalization techniques to enable automatic adaptation of interfaces in a mobile persuasive system. Ultimately, the objective is to improve effectiveness of mobile technology-based prevention and treatment for young people. Mohammed will take advantage of already collected experimental data and will then build user models that represent the foundation for the personalization, e.g. for interaction structure, content, modality, and presentation of mobile persuasion. Mohammed will then undertake a second use-case on designing personalized persuasive interfaces for behavioural interventions. Mohammed will synthesis the findings of use-cases to identify the most effective personalization techniques.
4. Intelligent techniques to support online mental health communities and resource recommendations for 4 levels of stepped care.
Claudette Pretorius, Host: RO, Partners: UCD [Secondment 6 months]
Claudette will conduct a review of moderated online mental health resources for young people, focusing on evaluating service/moderation models. ESR4 will then undertake interviews and focus groups with RO moderators to gain experience of online systems that have been successfully deployed at a national scale. Claudette will investigate intelligent techniques for assessing mental health signals in online communities and develop a responsive online tool that will help young people and service moderators to identify difficulties (e.g. gender issues, exam stress, bullying, alcohol). This tool will combine self ratings and intelligent techniques to help identify difficulties, track shifts in mood and recommend appropriate next steps from four levels of stepped care: self-help (e.g. online resources); phone service (e.g. helplines); community/voluntary support; or a relevant medical service.