A machine learning-based predictive framework for head impact injuries

Traumatic brain injury (TBI) remains a major challenge in both healthcare and engineering, where accurate prediction and prevention strategies are critical to mitigating its impact. Modern technologies, particularly artificial intelligence (AI), have rapidly advanced, offering transformative solutions in this domain. Machine learning (ML), a prominent subset of AI, has proven effective in addressing challenges by discovering intricate relationships within datasets. This research leverages artificial neural networks (ANNs) to predict intracranial pressures, a critical parameter linked to TBI, using kinematic linear accelerations as inputs. The proposed model incorporates three kinematic inputs (ax, ay, az) and time as inputs to the ANN. The developed ANN model successfully predicted intracranial pressure (ICP) values for five distinct brain regions — as frontal, parietal, left temporal, right temporal and occipital — achieving a correlation coefficient (R = 0.98) and root mean square error (RMSE = 0.008 MPa). Both feedforward and time series ANNs were trained on 80 samples with 177 time-steps data each. This study demonstrates the potential of artificial neural networks (ANNs) for accurate and computationally efficient ICP prediction. The findings further underscore the importance of integrating artificial intelligence with biomechanics to advance traumatic brain injury (TBI) research and support its translation into clinical practice.