The new tool fueled with AI could increase the traumatic investigations of brain lesions in forensics and law enforcement

A team of researchers at the University of Oxford, in collaboration with Thames Valley Police, the National Crime Agency, John Radcliffe Hospital, Lurtis Ltd. and Cardiff University, has developed an advanced tool based on Physics to help foreign forensic investigation of traumatic brain lesions (TBI).

The study, published in Communications engineeringIt introduces an informed automatic learning framework for the mechanics to help the police and forensic teams accurately predict the TBI results based on the documented assault scenarios.

TBI is a critical public health problem, with severe and long -term neurological consequences. In the medical -legal investigations, the determination if an impact could have caused a reported injury is crucial for legal proceduresHowever, there is currently no standardized, quantifiable approach to do so.

The new study demonstrates how automatic learning tools informed by mechanical simulations could provide evidence -based injury predictions to improve the accuracy and consistency of TBI investigations.

The main researcher Antoine Jérusalem, professor of mechanical engineering in the engineering department of the University of Oxford, said: “This research represents a significant step before in forensic biomechanics. An unprecedented tool to evaluate TBI objectively. “

The framework of the study, trained on real anonymous police reports, and forensic data, has obtained a remarkable prediction precision for TBI injuries:

• 94% Precision for skull fractures
• 79% precision for losing consciousness
• 79% Precision for intracranial hemorrhage (bleeding in the skull)

In each case, the model showed a high specificity and a high sensitivity (a low rate of positive and false negative results).

The frame uses a mechanical model of general head and neck calculation, designed to simulate how different types of impacts – such as fists, palms or strikes on a flat surface – affects various regions. This provides a basic prediction that an impact is likely to cause tissue deformation or stress.

However, it does not predict any risk of injury on their own. This is achieved by a upper layer that incorporates this information with any additional relevant metadata, such as the age and height of the victim, before providing a prediction for a given injury.

The researchers trained the general frame on 53 anonymous real reports on the assault cases. Each report included information on a number of factors that could affect the severity of the blow (for example, age, sex, building the victim/offender). This has led to a model capable of integrating mechanical biophysical data with medical -legal details to predict the probability of different lesions.

When the researchers evaluated which factors had the greatest influence on the predictive value for each type of injury, the results were remarkable in accordance with the medical results.

For example, when predicing the skull fracture, the most important factor was the largest amount of stress experienced by scalp and skull during an impact. Similarly, the most powerful predictor of losing consciousness has been the stress values ​​for the brain trunk.

The research team insists that the model is not intended to replace the involvement of human and clinical forensic experts in investigating assault cases. Rather, the intention is to provide an objective estimate of probability that a documented attack is the true cause of a reported injury.

The model could also be used as a tool to identify high risk situations, to improve risk assessments and to develop preventive strategies to reduce the occurrence and severity of head injuries.

Professor Jérusalem added: “Our frame will never be able to identify, no doubt, the culprit that caused an injury. All it can do is tell you if the information provided is correlated with a certain result. Since the quality of production depends on the quality of the information entered in Model, having detailed statements of witnesses is still crucial.

DNA Sonya Baylis, Senior Manager at the National Crime Agency that has supported this research project, said: “Understanding brain injuries using innovative technology to support an investigation of the police, which is based on limited information, will increase the necessary interpretation from a medical perspective until the criminal prosecution supports. “

Dr. Michael Jones, a researcher at Cardiff University, and forensic consultant, said: “A” Achilles heel “is evaluation whether a witness or deduced mechanism often matches the observed lesions.

“With the application of automatic learning, each additional case contributes to the general understanding of the association between the mechanism, primary damagepathophysiology and result. “

The study was conducted by an interdisciplinary team of engineers, forensic specialists and medical professionals at the University of Oxford, Thames Valley Police, National Crime Agency, Cardiff University, Ltd., John Radcliffe Hospital and other partner institutions.