Long-Range Input to Cortical Microcircuits Shapes EEG-BOLD Correlation

Although generated by different mechanisms, electroen-cephalography (EEG) rhythms and blood-oxygen-level-dependent (BOLD) activity have been shown to be correlated. The level of correlation varies between EEG frequency bands, brain regions and experimental paradigms, but the underpinning mechanisms of this correlation remain poorly understood. Here we create a mathematical, data-informed model of a cortical microcircuit that encompasses all major neuron types across cortical layers, and use it to generate EEG and BOLD under various external input conditions. The model exhibits noise-driven fluctuations mimicking EEG rhythms, with external inputs modulating EEG spectral characteristics. Consistent with experimental results, we observe negative alpha-BOLD correlation and positive gamma-BOLD correlation across different input configurations. Temporal variability of the input is found to increase EEG-BOLD correlation and to improve the correspondence with experimental results. This study provides a mathematical framework to theoretically study the correlation of EEG and BOLD features in a comprehensive way.