Aperiodic slope reflects glutamatergic tone in the human brain

Abstract

Excitatory and inhibitory neural processes are essential for every aspect of brain function, but current non-invasive neuroimaging methods to study these in the human brain are limited. Recent studies which separate oscillatory and aperiodic components of electrophysiological power spectra have highlighted a relationship between aperiodic activity and functional brain states. Studies in both animal models and humans suggest that the aperiodic slope of electrophysiological power spectra reflects the local balance of excitatory:inhibitory (E:I) synaptic transmission. Aperiodic slope varies across individuals, brain states, and clinical populations, which may reflect important differences in E:I balance. However, there is currently a lack of evidence linking aperiodic slope to other measures of excitation and inhibition in the human brain. Here, we show that flatter (less steep) aperiodic slopes from human electroencephalography (EEG) are associated with higher concentrations of the excitatory neurotransmitter glutamate measured with 7 tesla magnetic resonance spectroscopy (MRS) in the occipital lobe at rest. This suggests that individual differences in aperiodic neural activity reflect cortical glutamate concentrations, providing important insight for understanding changes in neural excitation across brain states and neuropsychiatric populations (e.g., schizophrenia) where glutamatergic function may differ. Our results support the use of aperiodic slope as a non-invasive marker for excitatory tone in the human brain.

Figure 1.

A) FOOOF model applied to eyes open resting data from an example participant. B) Difference between the aperiodic 1/f slope for eyes closed versus eyes open from the 6 occipital electrodes (O1, Oz, O2, PO3, POz, PO4; black box). C) Topoplot shows the difference in aperiodic slope between eyes open and eyes closed conditions. Red dots represent electrodes with a significantly greater slope with eyes closed (cluster corrected p < 0.05, one-tailed $t$-test). D) MRS VOI position: sagittal view with VOI in red. E) Segmentation of axial MRS VOI content into GM (orange), WM (yellow), and CSF (red). F) Example MR spectrum from an individual participant (black) plus LCModel fit (red), residual and baseline. The bottom rows show glutamate and GABA signals, as fitted by LCModel. G) Correlation between eyes open aperiodic 1/f slope from 6 occipital electrodes (O1, Oz, O2, PO3, POz, PO4; black box) and glutamate concentration. H) Topoplot showing the correlation between occipital glutamate concentration and eyes open aperiodic slope measured at each electrode. The interpolated color plot over the scalp shows the $r$ value. Points show the electrode locations, red indicates a significant correlation (cluster corrected p < 0.05), gray indicates electrodes that did not pass cluster correction (uncorrected p < 0.05).