Effects of age on resting-state cortical networks

Magnetoencephalography recordings of functional brain activity reveal large-scale cortical networks that are associated with cognition. The correlation of individualised networks with age and cognitive performance (age effects and cognitive performance effects respectively) was studied using a large cross-sectional healthy cohort (N=612, 18-88 years old) and accounting for a comprehensive set of confounds. Age effects were found in time-averaged functional networks in five canonical frequency bands (delta, theta, alpha, beta, gamma) that are consistent with the posterior-anterior shift with age observed in functional magnetic resonance imaging. Evidence from cognitive performance effects in time-averaged networks suggested the importance of maintaining alpha-band activity for cognitive health. A more detailed description of the functional activity was obtained by adopting an established machine learning approach (the Hidden Markov Model). Ten transient large-scale cortical networks with fast dynamics (~100 ms) were identified, which provided insight into age and cognitive performance effects that were not observed in the time-averaged analyses. The time spent in most networks increased with age, whereas the time spent in frontal networks decreased. The cognitive performance effects for the transient networks suggested that age effects in the frontal networks are compensatory. Thus, our study suggests both the maintenance of functional activity (lesser age effects) and the recruitment of compensatory functional activity can co-occur to produce good cognitive performance in older individuals. The time-averaged and transient functional networks have been made publicly available as a resource.