Researchers and clinicians at Haukeland University Hospital, the University of Bergen, and the University of Oslo have developed a new method to predict mood-states change in people with bipolar disorder using artificial intelligence. A patent application for the method has been submitted.
People with bipolar disorder experience extreme mood-state changes between depression and mania or hypomania. The disease is chronic, and the goal of treatment is to keep mood stable, which is not easy.
About one to two percent of the world’s population is diagnosed with bipolar disorder. The disease requires lifelong treatment, often consisting of both medication and cognitive therapy.
“Our goal is to support individuals in maintaining a balanced mood state, known as euthymia, consistently. However, this can be challenging. Treatment for people with bipolar disorder is often reactive, meaning they receive care after their mood has shifted into depression or hypomania.” says Ulysse Côté-Allard.
Côté-Allard is an associate professor at the Department of Technology Systems at the University of Oslo. He and other researchers and clinicians at Haukeland University Hospital, the University of Bergen, and the University of Oslo have developed a methodology that can predict when people with bipolar disorder will shift from one mood state to another before it happens.
Measuring Changes in Movement Patterns
At the “Science Impact” conference during “Oslo Innovation Week,” with the theme of AI and digitalisation in health and energy, Côté-Allard presented the results of a clinical study conducted on 49 people with bipolar disorder.
Ulysse Côté-Allard at the Science Impact 2024 conference. Photo: Elisabeth K. Andersen.
The 49 participants in the study wore a smartwatch, GENEActiv actigraph, that continuously monitored their movement patterns over an entire year. Eight of these patients experienced a confirmed transition from euthymia to either depression or mania.
In each case, the developed algorithm enabled the researchers to automatically determine the transition phase by observing changes in movement patterns. This was accomplished using machine learning and artificial intelligence.
“Individuals with bipolar disorder exhibit varying activity levels and movement patterns complexity depending on their mood states. Our goal was to identify the precise transition between mood-state to help prevent significant shifts into depression or mania. As when these shifts occur, hospitalization for treatment is often necessary, and returning to a stable phase can take time. This process is extremely disruptive for the person’s life and requires a lot of resources from the healthcare system.”says Côté-Allard.
The risk of suicide is 10 to 30 times higher for people with bipolar disorder, and suicide attempts often occur during a depressive phase or mixed phases combining depression and mania.
Funding Needed for Further Clinical Research
Petter Jakobsen, the lead author of the study, along with Côté-Allard and their collaborators, is now seeking funding to expand the clinical research in a new study.
Hao Wu is the project manager at Inven2, responsible for Côté-Allard, his research colleagues and their innovation project. As the project is a collaboration between researchers and clinicians at the University of Oslo and Haukeland University Hospital, the two technology transfer offices, Inven2 and VIS, share the project equally.
“This is an exciting project to work on. After our initial meeting with the researchers, we accelerated the DOFI evaluation process and, in collaboration with VIS, decided to move forward with a commercialization project. Steffen Boga at VIS is responsible for the commercialization aspect.,” says Wu.
From Inven2, Wu is joined by Irene Fjeldahl Johannesen, who is responsible for IPR and contributed to the IP identification and patentability assessment work. Anne Marie Bjørgo from Inven2 is also involved, overseeing the agreements between Inven2 and VIS.
Hao Wu is responsible for the project in Inven2. Photo: Inven2.
