Abstract
Sports-related concussion, classified clinically as mild Traumatic Brain Injury (mTBI), is one of the most common yet least objectively diagnosed injuries in competitive athletics. An estimated 1.6 to 3.8 million sports concussions occur annually in the United States, and the Lancet Neurology Commission on TBI identifies mTBI as the dominant clinical form of brain injury, accounting for over 90% of cases, yet one for which evidence to guide diagnosis and management remains scarce. Artificial Intelligence (AI) and neuroimaging offer promising tools to address this gap; however, it remains unclear whether three decades of computational research have prioritised the high-incidence, 'invisible' mTBI or focused on more imaging-visible moderate-to-severe cases.
This study systematically examines this prioritisation over three decades. This study presents a bibliometric analysis of 1,950 publications retrieved from the Web of Science (1996 2026). Using VOSviewer for publication trend, keyword co-occurrence, and collaboration network analyses, we map the intellectual structure and thematic evolution of AI and neuroimaging research for TBI and concussion. Our analysis reveals a fourfold growth in annual research output over the past decade (63 publications in 2016 to 257 in 2025), with the United States leading (979 publications), followed by China (270) and the United Kingdom (169). Dominant themes include traumatic brain injury, machine learning, biomarkers, concussion, and deep learning. Critically, sports-specific concussion terminology occurs 4.4 times less frequently than severe-TBI clinical markers in the keyword landscape, despite its larger global burden.
These findings provide a roadmap for the computational sports science community to target the most underserved challenge in brain injury diagnosis: objective, imaging-based tools to support return-to-play decisions in athletic populations, particularly for mTBI. The analysis also highlights emerging AI techniques, such as explainable AI and federated learning, as promising avenues for future research.