After traumatic brain injury (TBI), some patients may recover completely, while others retain severe disabilities. Accurately evaluating prognosis is challenging in patients on life-sustaining thery. Though resting-state functional MRI (rs-fMRI) can assess neurological activity shortly after brain injury, it is unknown whether communication across brain regions at this early juncture predicts long-term recovery. Investigators from Mass General Brigham and collaborators in the U.S. and Europe analyzed data from three prospective cohorts comprising 97 patients who underwent rs-fMRI after injury, finding that early communication between three pairs of brain regions is associated with favorable six-month functional outcomes. Findings are published in PNAS .
"Using brain scans, we identified signature patterns of recovery after moderate or severe TBI," said lead author Sam Snider, MD, of the Division of Neurocritical Care and the Department of Neurology at Mass General Brigham. "These findings open new avenues for prognostic assessment in TBI, and emerging evidence suggests these patterns may be modifiable, raising the possibility of future therapeutic application."
Researchers investigated a special kind of brain activity known as "anticorrelated" brain activity, which is a hallmark of normal brain function. This means that when one brain region become active, a separate region deactivates.
The researchers analyzed brain scans (resting-state fMRI) from half of the participants with TBI, identifying two patterns in which different brain regions worked in opposite ways and a third in which regions worked together. People with any of the three patterns were more likely to have better outcomes, even after adjusting for factors like sedation and consciousness level. The researchers incorporated their findings into a model to predict outcomes after TBI. They tested the new method in the other half of the participants. This model did a better job of predicting recovery than older prediction models.
One pattern that strongly predicted positive outcomes involved parts of the salience network, a coordination "hub" for many brain networks, and the default mode network (DMN), which is active during rest. Together, these regions mediate conscious access to incoming information. Another pattern included regions involved in cognitive control and basic visual processing. Connectivity between the DMN and language network also helped predict outcomes.
Importantly, findings were consistent across patients with injuries of varying severity, treated in multiple hospitals with different MRI scanners in different countries. Future studies may explore the extent that these neurological connections are essential to recovery and their ability to guide prognostic decision-making after TBI.
Authorship: In addition to Snider, Mass General Brigham authors include Hui Shi, Calvin Howard, Alexandra J. Golby, Brian L. Edlow, and Michael D. Fox.
Additional authors include Yelena G. Bodien, Xiaoying Sun, Karl A. Zimmerman, Guido Bertolini, Sandra Magnoni, Vincent Dunet, Mauro Oddo, Neil SN Graham, Emma-Jane Mallas, Federico Moro, Pratik Mukherjee, Nancy R. Temkin, Sonia Jain, and David J. Sharp.
Disclosures: Fox has consulted for Magnus Medical, Soterix, Abbott, Boston Scientific, Tal Medical, and has received funds from Neuronetics and Nexstim. Fox has intellectual property on the use of brain connectivity imaging to analyze lesions and guide brain stimulation.
Funding: Snider was supported by grants from the NIH (1K23NS136767, U01EB034228) and the American Heart Association. Edlow was supported by grants from the NIH (R01NS128961; R01NS138258, DP2HD101400) and the Chen Institute MGH Research Scholar Award. Fox was supported by grants from the NIH (R01MH113929, R21MH126271, R21NS123813, R01NS127892, R01MH130666, UM1NS132358), the Kaye Family Research Endowment, the Ellison / Baszucki Family Foundation, the Manley Family, and Donna and Tom May.
Paper cited: Snider S et al. "Preservation of anticorrelated brain networks predicts recovery after traumatic brain injury" PNAS DOI: 10.1073/pnas.2518159122
