Approximate Entropy as a Metric for Quantifying FMRI Changes Across Time

Kyle D. Steinke¹, David Frakes², Jose Rios³, Gabe Oland², Leslie C. Baxter¹

Functional MRI is of limited utility in longitudinal studies because it lacks a quantitative baseline. We tested whether a test of interest (e.g., a motor task) could be compared to a generally stable vision task to overcome this weakness. We used approximate entropy (ApEn) to characterize each time series signal. The motor task ApEn was then compared to the vision task ApEn to form a ratio, which remained stable after a caffeine intervention, as expected, but changed in surgical patients that had behavioral decline.