DAIsy Developing AI ecosystems improving diagnosis and care of mental diseases

Goal

Today's health care system is becoming increasingly complex. More and more medical disciplines are involved in the diagnosis, treatment and follow-up of patients, and the number of available treatment options is rapidly increasing. However, the treatment available for mental illnesses in particular remains inadequate. Not only during the Corona pandemic, but also in the years before, mental illnesses have increased considerably. In Germany, depression is one of the most underestimated mental illnesses in terms of its effects. In Germany, a total of 8.2% of all adults develop a unipolar or persistent depressive disorder in the course of a year. About one in four women and one in eight men are affected by depression. With an estimated global lifetime prevalence of 16-20%, depression not only causes unbearable individual suffering (according to WHO estimates, >50% of all suicides occur against a background of major depression), but also a heavy social and economic burden. If one takes into account not only the direct diagnostic and treatment costs, but also secondary follow-up costs (for example. If one takes into account not only the direct costs of diagnosis and treatment, but also secondary costs (e.g. productivity losses due to inability to work or early retirement), the total annual costs of depression in Germany alone are estimated at at least 22 billion euros, with productivity losses accounting for the largest share. Therefore, the treatment of depression is not only important to reduce individual suffering, but also to avert economic damage. For this purpose, new, AI-based ways must be found to make treatment more efficient for those affected and to reduce the enormous health care expenditures. The aim of the DAIsy project is therefore to research novel, innovative therapy systems to improve diagnostic, interactive and individual approaches for patients suffering from a depressive illness. Two use cases are being pursued:

  1. Multimodal neurofeedback system: The aim of this outpatient therapy method is to record neuronal reactions to standardised test environments via a multimodal neurofeedback system and to present this to the patients as feedback. This is intended to activate self-regulating mechanisms of the brain to counteract dysfunctional neuronal activity patterns. Feature extraction from the recorded data is a highly demanding task that will be realised within the framework of this project with the development of innovative AI-based algorithms. In addition, the EEG-based neurofeedback system will be extended by an fNIRS to test whether the fusion of the sensor data of the two systems leads to an improved signal quality for neurofeedback. EEG-based approaches show good temporal resolution, while fNIRS-based approaches have their advantages in local resolution. A combination of the systems could lead to improved overall resolution.
  2. Virtual therapy assistant: The aim of this approach is to provide continuous support to patients in their daily lives through a virtual therapy assistant. In the form of a digital health application, this collects data on the patient's experience and behaviour, such as mood parameters, medication intake, the use of social media, physical activity, vital parameters or usage times of mobile devices. Through the application of AI-based methods, individual behavioural and mood patterns are to be recognised and situation-appropriate treatment measures derived. Through better and more precise knowledge of the patient's condition, symptom-based neurofeedback procedures can be optimally used and adjusted. In addition, the therapy assistance will provide various therapy elements - proven in classical psychotherapy - digitally processed and AI-assisted.

Both approaches will enable continuous monitoring of the mental state as well as an individualised therapy procedure based on this, in order to be able to recognise and treat a deterioration of the clinical picture at an early stage. Only timely intervention can minimise the consequences of a worsening course of the disease on the one hand and reduce therapy costs on the other.

In the German consortium of the Europe-wide ITEA project, the University Hospital Bonn, Materna Information & Communication SE, BEE Medic GmbH, Ascora GmbH and OFFIS are jointly facing this challenge.

Persons
Publications
Developing Advanced AI Ecosystems to Enhance Diagnosis and Care for Patients with Depression

Franziska Klein, Frerk Müller-Von Aschwege, Patrick Elfert, Julien Räker, Alexandra Philipsen, Niclas Braun, Benjamin Selaskowski, Annika Wiebe, Matthias Guth, Johannes Spallek, Sigrid Seuss, Benjamin Storey, Leo N. Geppert, Ingo Lück, Andreas Hein; Studies in Health Technology and Informatics; 0Oktober / 2023

From lab to life: challenges and perspectives of fNIRS for haemodynamic-based neurofeedback in real-world environments

Klein, Franziska and Kohl, Simon H. and Lührs, Michael and Mehler, David M. A. and Sorger, Bettina; Philosophical Transactions of the Royal Society B: Biological Sciences; 2024

Partners
Ascora GmbH
www.ascora.net
Materna Information & Commmunications SE
www.materna.de
Universitätsklinikum Bonn
www.ukbonn.de
BEE Medic GmbH
www.beemedic.de

Duration

Start: 01.11.2023
End: 31.10.2025

Source of funding