Linking age changes in human cortical microcircuits to impaired brain function and EEG biomarkers

Abstract

Human brain aging involves a variety of cellular and synaptic changes, but how these changes affect brain function and signals remains poorly understood due to experimental limitations in humans, meriting the use of detailed computational models. We identified key human cellular and synaptic changes occurring with age from previous studies, including a loss of inhibitory cells, NMDA receptors, and spines. We integrated these changes into our detailed human cortical microcircuit models and simulated activity in middle-age ([~]50 yrs) and older ([~]70 yrs) microcircuits, and linked the altered mechanisms to reduced spike rates and impaired signal detection. We then simulated EEG potentials arising from the microcircuit activity and found that the emergent power spectral changes due to these aging cellular mechanisms reproduced most of the resting-state EEG biomarkers seen in human aging, including reduced aperiodic offset, exponent, and periodic peak center frequency. Using machine learning, we demonstrated that the changes to the cellular and synaptic aging mechanisms can be estimated accurately from the simulated EEG aging biomarkers. Our results link cellular and synaptic mechanisms of aging with impaired cortical function and physiological biomarkers in clinically-relevant brain signals.