Temporal hierarchy of intrinsic neural timescales converges with spatial core-periphery organization.

TLDR: In this article, the autocorrelation window in short (ACW-50) and long ACW-0 windows were used to investigate intrinsic neural timescales during rest and task states.

Abstract: The human cortex exhibits intrinsic neural timescales that shape a temporal hierarchy. Whether this temporal hierarchy follows the spatial hierarchy of its topography, namely the core-periphery organization, remains an open issue. Using magnetoencephalography data, we investigate intrinsic neural timescales during rest and task states; we measure the autocorrelation window in short (ACW-50) and, introducing a novel variant, long (ACW-0) windows. We demonstrate longer ACW-50 and ACW-0 in networks located at the core compared to those at the periphery with rest and task states showing a high ACW correlation. Calculating rest-task differences, i.e., subtracting the shared core-periphery organization, reveals task-specific ACW changes in distinct networks. Finally, employing kernel density estimation, machine learning, and simulation, we demonstrate that ACW-0 exhibits better prediction in classifying a region’s time window as core or periphery. Overall, our findings provide fundamental insight into how the human cortex’s temporal hierarchy converges with its spatial core-periphery hierarchy. Golesorkhi et al. use a combination of magnetoencephalography data, machine learning and simulation to investigate intrinsic neural timescales during resting and task states. They provide insight into how the temporal hierarchy of the human cortex converges with its spatial core-periphery hierarchy.