Fast, efficient information transfer is essential for the brain to ensure survival. As recently shown in functional magnetic resonance imaging with high spatial resolution, turbulence appears to offer a fundamental way to facilitate energy and information transfer across spatiotemporal scales in brain dynamics. However, given that this imaging modality is comparably slow and not directly linked with neuronal activity, here we investigated the existence of turbulence in fast whole-brain neural dynamics measured with magnetoencephalography (MEG). The coarse spatial observations in MEG necessitated that we created and validated a empirical measure of turbulence. We found that the measure of edge-centric metastability perfectly detected turbulence in a ring of non-local coupled oscillators where the ground-truth was analytically known, even at a coarse spatial scale of observations. This allowed us to use this measure in the spatially coarse, empirical large-scale MEG data from 89 human participants. We demonstrated turbulence in fast neuronal dynamics and used this to quantify information transfer in the brain. The results demonstrate that the necessary efficiency of brain function is dependent on an underlying turbulent regime.