The ISBGroup Blog

Two days after STRATIF-AI was featured at the BME@LIU-day in Linköping, Sweden, it is featured again in the ISMH conference in Technirghiol, Rumania. In Linköping, we had a shared booth with SUND (the spin-off company that converts the digital twins to eHealth products), and produced a roll-up and a poster. This roll-up has now been transported to Rumania, together with Gunnar Cedersund, to be displayed at a 2h session at the ISMH conference. This happens today May 23, at 17-19, CET, and can be viewed online.

ISMH stands for the International Society of Medical Hydrology and Climatology. In other words, the conference deals with water-based treatments, such as spas (German: kurort), mineral waters, mud baths, as well as temperature-based treatments, such as saunas, ice-baths, etc. These are old traditional treatments, that dates back to ancient Greek, at the least, but there is also a growing scientific interest, with 10 times more papers published per year today, compared to 20 years ago. Still, the ISMH is old, by scientific standards, and the conference celebrated its 125th birthday yesterday. In the picture below, you see Prof, MD Gelu Onose, one of our main collaborators in STRATIF-AI, and also one of the the main organizer of the event, who cut the first piece of the cake, during yesterday’s celebration.

While the official motto of the conference is “Evidence-based balneology”, while there are lots of medical doctors and professors here, and while balneology is an established speciality in medicine in quite a few countries in Europe, it is still a field that lies on the fringes of medicine. For instance, in Sweden, balneology is not an established discipline of medicine, and MDs cannot become specialized in this field, as they can in e.g. cardiology, or hepatology. Therefore, balneology is to some extent similar to other alternative treatments, such as yoga, regular exercise and gym-based personal trainers, health coaches, etc. They exist in society, they work with health, there is some scientific support for the health benefits of these activities, but they are to a large extent not integrated in conventional healthcare. This is where the activities and visions of STRATIF-AI can contribute. Our ambition is to go from a doctor-centric fragmented system, to an integrated patient-centric eco-system system (see picture below). In this eco-system, e.g. personal trainers can take over where nurses leave a patient, following a health conversation (as in our clinical study 2 in STRATIF-AI). And in this eco-system, one can add new actors, such as spas and kurorts, as long as there is evidence that their methods work, and as long as the underlying mechanisms can be added to the digital twin’s models and visualizations.

The 2h session we will have is structured as follows: First, Gunnar Cedersund, the coordinator of STRATIF-AI gives an overview of STRATIF-AI, and of the digital twins, which converts knowledge and data to computational models. Thereafter Prof Gelu Onose from SCUBA (in the picture above), and his colleagues Constantin Munteanu and Cristina Popescu (both MDs), give an overview of the scientific basis for prevention and treatment of stroke using different modalities, including balneology. Thereafter, we look at the data integration aspects: Jesper Fellenius from Z2 gives an overview of the personal data vault, Cati Martinez from University of Murcia talks about their work with semantic harmonization and inter-operability, and SIMAVI talk about their work on Electronic Healthcare Records, and app developments, e.g. for rehabilitation of stroke. The session is concluded with a round-table discussion. All things can be followed online, at this link. The general link to the conference is available here.

Tomorrow, May 21, 2024, we are arranging a new edition of the BME@LIU-day. BME stands for biomedical engineering, i.e. the interplay between biomedicine (from basic science to clinical practice) and engineering. In practice, BME, entails fields such as modelling and AI, digital twins, wearable sensors, new biomaterials, eHealth, bioinformatics, imaging, bio-optics, etc etc. BME is therefore also one of the core areas of LIU, and it is – in fact – one of the reasons why Linköping university (LIU) was founded, in 1975: to create a bridge between the needs for new engineers at SAAB, and the need for new biomedical research and personnel at the hospital. Around this need, LIU was founded, and at the heart of this dream of a bridge between these two worlds was the Department of Biomedical Engineering (IMT).

IMT eventually attracted Gunnar Cedersund, who started there gradually from around 2011, and then gradually moved his research group over to this department. IMT has since then also expanded its mission to support biomedical engineering beyond its own department, i.e. at the university level. And the arrangement of this BME@LIU-day is a central part of this new broader mission.

In practice, the day will feature a lot of activities at LiU, and also a lot of our own activities. For instance: at 10-10.30, postdoc William Lövfors and MD/Ph.D. student Valentin Kindesjö will introduce the latest developments of our digital twins, using a so-called double presentation where a clinician and an engineer each presents from their specific perspective. In the same session on modelling and AI, at 10.50-11.10, we also see presentations by our newly transitioned postdoc Nicolas Sundqvist, who will present a new story, where we quantify the importance of different cell types for the neurovascular coupling. Finally, also in that session, at 11.20-11.40, we have one of our newest postdocs, Dirk de Weerd, who presents a work he has done together with our incoming postdoc Rasmus Magnusson, on the usage of Variational AutoEncoders, to identify a compressed space for gene expression variations, and apply this to disease modelling. A final oral presentation, not by us, but by one of our closest collaborators, is that 10.50-11.10 by Jesper Fellenius, who will present his perspectives from 20 years of app developments for commercial companies (Ericsson, Swedbank etc) and his 75 MSEK development of a personal data vault. Link to the modelling presentation is here and link to Jesper’s session is here.

Finally, apart from oral presentations, we will also be featured in various places in the poster session combined with the arena. There will be 6 student-based posters, who present 6 large B.Sc. projects, on various modelling tasks within the digital twin development: i) on blood pressure and exercise, ii) on exercise and VO2max, iii) on dysfunctional fat storage in large adipocytes and liver fat, iv) on the neurovascular coupling, v) on drugs and cognitive function, and vi) on alcohol metabolism. Apart from that, there will be many other research-based posters from our group. Finally, we will also – for the first time – be featured not only by our original spin-off company SUND sound medical decisions, but by the newly created daughter company with Z2 and Johan Holmsäter – with the working title Lev Skönare.

All these companies and research projects will be featured with oral presentations, posters, or exhibitions, where you also can test to create your own digital twin, if you want.

Welcome!

Gunnar Cedersund, chair of the program committee

Last week, I had the honor of giving the concluding keynote lecture at the event “Data-driven mechanistic modelling in life sciences”. This follows a trend of being invited to give more and more keynote and plenary lectures at events, for which I am very grateful. Such longer lectures also give me the chance to expand a little bit more on my point-of-view. The focus of this particular workshop is also something I am very keen to promote, since I think that this particular overlap (mechanistic and data-driven modelling) is under-represented in many communities and conferences.

The centrality of this overlap is actually seen even in our group logo (Fig A). a) The fact that it has an open non-black box, represents the fact that we do mechanistic modelling. b) The data-driven aspect of our models is represent by the purple core in the middle, which represents the fact that we always look for core predictions. Core predictions are predictions that are well-determined from the current prior knowledge and data, even when taking all uncertainties in data and prior knowledge into account.

While I personally think that this is the way to work, and while we have a very well-established workflow for how to develop models in this fashion, mechanistic modelling and data-driven modelling are unfortunately often done in two disjoint communities, with too little overlap (Fig B). Mechanistic modelling often results in mere simulation-based results, which have not been validated using independent data, i.e. data that has not been used to train the model. This is often the case for e.g. PDE and agent-based models, but also common in e.g. theoretical ecology, theoretical biology, etc. It was therefore encouraging too see that one of the presentations at the workshop (by Joshua Bull) looked at spatial models, and on how to quantify the comparison between simulations and data also for spatial models. Data-driven modelling is too often interpreted to mean only machine-learning, narrow AI, and other black-box modelling techniques. While these are big and very hyped communities and approaches at the moment, they are not the only techniques that can be used to do data-driven modelling. In other words, while these black-box models include important techniques, which are useful if one has standardized large-scale data, they also have critical short-comings. Black-box models e.g. have big problems incorporating the type of data that is present in most biological papers, including the prior that is knowledge available. For these reasons, explainability and trustworthiness are challenges. I therefore think that hybrid modelling is the way forward (see e.g. this review, and this example). At the conference, there was also an excellent opening keynote of day 2, by Alvaro Köhn-Luque, which showed some additional and interesting examples of hybrid models.