Assessing blood oxygen level–dependent signal variability as a biomarker of brain injury in sport-related concussion

Authors

Evan D. Anderson, University of Illinois Urbana-Champaign, Wright-Patterson AFB
Tanveer Talukdar, University of Illinois Urbana-Champaign
Grace Goodwin, University of Illinois Urbana-Champaign, University of Nevada
Valentina Di Pietro, University of Illinois Urbana-Champaign, University of Birmingham, University Hospitals Birmingham NHS Foundation Trust
Kamal M. Yakoub, University of Birmingham, University Hospitals Birmingham NHS Foundation Trust
Christopher E. Zwilling, University of Illinois Urbana-Champaign
David Davies, University of Birmingham, University Hospitals Birmingham NHS Foundation Trust
Antonio Belli, University of Birmingham, University Hospitals Birmingham NHS Foundation Trust
Aron K. Barbey, University of Illinois Urbana-Champaign, University of Nebraska-Lincoln

Date of this Version

8-16-2023

Document Type

Article

Citation

BRAIN COMMUNICATIONS 2023. https://doi.org/10.1093/braincomms/fcad215

Comments

Open access.

Abstract

Mild traumatic brain injury is a complex neurological disorder of significant concern among athletes who play contact sports. Athletes who sustain sport-related concussion typically undergo physical examination and neurocognitive evaluation to determine injury severity and return-to-play status. However, traumatic disruption to neurometabolic processes can occur with minimal detectable anatomic pathology or neurocognitive alteration, increasing the risk that athletes may be cleared for return-to-play during a vulnerable period and receive a repetitive injury. This underscores the need for sensitive functional neuroimaging methods to detect altered cerebral physiology in concussed athletes. The present study compared the efficacy of Immediate Post-concussion Assessment and Cognitive Testing composite scores and whole-brain measures of blood oxygen level–dependent signal variability for classifying concussion status and predicting concussion symptomatology in healthy, concussed and repetitively concussed athletes, assessing blood oxygen level–dependent signal variability as a potential diagnostic tool for characterizing functional alterations to cerebral physiology and assisting in the detection of sport-related concussion. We observed significant differences in regional blood oxygen level– dependent signal variability measures for concussed athletes but did not observe significant differences in Immediate Post-concussion Assessment and Cognitive Testing scores of concussed athletes. We further demonstrate that incorporating measures of functional brain alteration alongside Immediate Post-concussion Assessment and Cognitive Testing scores enhances the sensitivity and specificity of supervised random forest machine learning methods when classifying and predicting concussion status and post-concussion symptoms, suggesting that alterations to cerebrovascular status characterize unique variance that may aid in the detection of sport-related concussion and repetitive mild traumatic brain injury. These results indicate that altered blood oxygen level–dependent variability holds promise as a novel neurobiological marker for detecting alterations in cerebral perfusion and neuronal functioning in sport-related concussion, motivating future research to establish and validate clinical assessment protocols that can incorporate advanced neuroimaging methods to characterize altered cerebral physiology following mild traumatic brain injury.