Brain topology and cognitive outcomes after cardiac arrest: A graph theoretical analysis of fMRI data

Introduction

Half of all cardiac arrest patients experience long-term cognitive impairment. Identifying patients at risk is challenging and underlying mechanisms are incompletely understood. Since postanoxic encephalopathy is diffuse, measures of global network organization might contribute to identifying these patients at risk. We studied MRI-based whole-brain and subnetwork topology in relation to short- and long-term cognitive outcomes after cardiac arrest.

Methods

We performed a multicenter prospective cohort study in cardiac arrest survivors. Patients underwent resting-state functional MRI during hospitalization (within one month after cardiac arrest). We analyzed 264 regions of interest using Power's atlas, extracting mean timeseries and calculating pairwise connectivity with Pearson's correlation. Overall functional connectivity, global efficiency, clustering coefficient, and modularity were calculated for the whole brain and five subnetworks. Cognitive function was measured using the Montreal Cognitive Assessment (MoCA) during hospitalization, at three, and twelve months, and using neuropsychological examination at twelve months. Mixed effects models were used to examine relationships between topology measures and cognitive outcomes. We corrected for multiple testing.

Results

We included 80 patients (age 60 ± 11 years, 70 (90%) male). Our analyses showed consistent relations between various topology measures of the sensory/somatomotor network (SSN) and MoCA score during hospitalization and memory, attention, and executive functioning at twelve months follow up. After correction for multiple testing, we found no statistically significant relations between whole-brain and subnetwork graph measures and cognitive outcomes.

Conclusion

Early whole brain functional topology was not related to short- or long-term cognitive outcome after cardiac arrest in this analysis. Potential relations between SSN topology and cognitive outcome point towards spatial heterogeneity of postanoxic encephalopathy. Possible predictive values of SSN structure or function need further investigation.