Smartphone Technology Supporting Mental Health Care (multiple projects)

The master projects outlined below are a part of the research project INtroducing personalized TReatment Of Mental health problems using Adaptive Technology (INTROMAT) financed by the Research Council of Norway, 2016-2022, as one of three IKTPLUSS lighthouse projects, see (in Norwegian):

Announcement Research Council of Norway

The main objective of the project is to increase access to mental health services for common mental health problems by developing technology which can guide patients with limited or no therapist support through a treatment process adapted to each user.

The specific goal of these Master projects is to contribute to developing a patient monitoring and support system that consists of collecting data from the patient for modelling behaviour and emotional state, to be used for prediction and adaptation of the treatment. The work would be building on the state-of-the-art knowledge in sensors, apps, behavioral models and machine learning.

Mental health problems are a major public health concern worldwide. New technologies represent huge opportunities for innovations in the field of mental health. The vision of INTROMAT is to improve public mental health with innovative ICT. The main objective for INTROMAT is to increase access to and quality of mental health services by developing a scalable and flexible infrastructure with engaging and adaptive interfaces that can be integrated with existing systems. The system will allow use across individuals, mental health problems and clinical settings, with a documented impact on health.

INTROMAT will progress state- of-the-art within machine learning, mobile technology, multimedia technology, system modelling, interaction design and psychology. The results of INTROMAT are expected to increase access to evidence-based mental health services, increase digital collaboration between the health services, and increase uptake and use of treatment technologies as pure self-help and in the primary and secondary health services.

The project is managed by Haukeland University Hospital, Bergen. As one of the partners in the project, the ROBIN group at UiO would be contributing with developing patient monitoring and support systems running as smartphone apps. That is, developing systems consisting of collecting data from the patient for modelling behaviour including emotional state to be applied for prediction of these in the future to reduce mental health challenges. The developed technology will provide a support system for patients and clinicians by recommending appropriate possible actions and treatments.


A number of robotics master projects are available related to the INTROMAT project:

Collecting and Classifying Sensor Data for Emotional State Prediction

Today´s smartphones contain a variety of built in sensors. We would in this master project like to investigate whether machine learning techniques can be applied for classifying and predicting change in the mood and emotions of it´s user (mood is a more long term state than often rapidly changing emotions). That is, if the person is happy, angry, stressed, sad, or in some other state. This is useful for miscellaneous applications including mental health treatment.

The project should build on earlier undertaken work, including the following master thesis, Paper Phonesense 2011 and Microsoft Research

The tasks of the project:

·      Get an overview of mood and emotion classification using smartphone sensors including algorithms being applied.

·      Collect data for different emotions for a number of different users by developing an app (or revising an existing app).

·      Compare various machine learning algorithms for different set of sensors to see which are most effective for classifying emotions.

·      Write master thesis.

Qualifications: Machine learning (e.g. INF3490/INF4490), programming, app development (e.g. INF5261)


Exploration of Mood Prediction using Speech

How people speak can be an important indicator about their psychological state. Audio signal analysis of speech is one tool that can be used to explore this. In addition it would be interesting to study how data from other sensors can help to improve the predictions based on the audio content. Thus, in this master project, we are interested in researching how the changes in the voice in combination with sensor data can be used to predict the mood of a person.

Goal: A system that uses audio signals from speech in combination with sensor data to make predictions about the mood of a person.

The tasks of the project:

·      Get an overview of earlier work on speech analysis and other sensor data with regards to mood prediction.

·      Compare various machine learning algorithms for different set of sensors data to see which are most effective for mood prediction.

·      Write master thesis.

Qualifications: Machine learning (e.g. INF3490/INF4490), mathematical skills, programming skills, app development (e.g. INF5261)  


Classifying Collected Sensor Data for Bipolar Disorder Prediction

Bipolar disorder is a brain disorder that causes unusual (large) shifts in mood, energy, activity levels. A goal of this project is to try to predict when a negative condition is in progress to be able to initiate treatment early. We collaborate with clinical partners collecting data from bipolar patients. These include speech data and various data about the activation of the patient including using the empatica E4 watch.

The goal of this master project is to apply state-of-the-art machine learning techniques to develop effective mechanism for predicting upcoming negative periods as early as possible. Some relevant earlier work has been published by Faurholt-Jebsen.

The tasks of the project:

·      Get an overview of earlier work on classifying sensor data with regards to bipolar disorder prediction.

·      Compare various machine learning algorithms for different set of sensors data to see which are most effective for bipolar disorder prediction.

·      Write master thesis.

Qualifications: Machine learning (e.g. INF3490/INF4490), signal processing (e.g. INF4470), programming


Machine Learning for Modelling and Predicting Human State

Design a set of models using machine learning for representing human behavior including emotional states using self-reporting and features extracted from sensors from either a single person or a population of persons. Models with a flexibility in abstraction level and time scale should be implemented and compared.

The tasks of the project:

·      Get an overview of earlier work on human behavior modelling.

·      Compare various methods for human behavior modelling using machine learning algorithms.

·      Write master thesis.

Qualifications: Machine learning (e.g. INF3490/INF4490), programming


Sensor Data Analysis of Mental Conditions Datasets such as Social Anxiety Disorder and Attention-Deficit/Hyperactivity Disorder (ADHD)

Analyze wearable sensor datasets related to social anxiety disorder and/or attention deficit/hyperactivity disorder to find meaningful associations between mental states. Develop machine learning models to automatically predict different mental states based on the sensor data.

The tasks of the project:

  • Perform data analysis tasks on the datasets (exploratory data analysis, visualization, etc.).
  • Analyze the correlation between sensor data and different mental states.
  • Apply machine learning algorithms to automatically predict mental states from sensor data.
  • Write master thesis.

Qualifications: Machine learning (e.g. INF3490/INF4490), programming


Benefits of taking master project related to an externally funded research project:

  • close collaboration with doctoral students (PhD) and postdoctoral researchers working on related topics
  • strong focus on progressing international state-of-the-art research and publication in leading international journals and conferences (beneficial for later application to PhD or researcher positions)
  • funding available for publishing results available at international conferences

Competences relevant for jobs in industry: programming of embedded systems, system modeling, prototyping of sensor/mechanical systems, smartphone app development, experience from an upcoming application area (smartphone as a part of the health care, mental health treatment, …)


Contact for further information: Enrique Garcia Ceja and Jim Tørresen

Emneord: mental health, machines learning, sensor data analysis, prediction
Publisert 26. sep. 2016 14:57 - Sist endret 11. sep. 2017 17:06

Omfang (studiepoeng)