The AI research of our esteemed M4-HEALTH colleagues Rasmus Magnusson, Elin Nyman and Henrik Green enables more precise estimation of the time of death. The research has been carried out in collaboration with the Swedish National Board of Forensic Medicine, and is a valuable advance in forensic investigations.
The method analyses metabolites, small molecules in the blood that change in predictable ways as the organs and tissue begin to break down after death. Using nearly 4,900 blood samples from autopsies, researchers trained an AI model to estimate the time between death and autopsy.
The model can predict the post-mortem interval with an accuracy of around one day, even when as many as 13 days have passed since death. This is a significant improvement over established methods such as body temperature, rigor mortis and potassium levels in the eye, which become less reliable several days after death. The researchers are now working to develop even more accurate models that could eventually determine not only the date, but also the approximate time of day when a person died—providing an important new tool for forensic investigations.
Read a more detailed article on Linköping University website
