Hemisphere- and condition-specific alpha oscillations support semantic and spatial cognition in aging.

Neuroimaging studies have shown age-related alterations in brain structure and function supporting semantic knowledge, although the significance of these is not fully understood. Herein, we report novel temporal, spectral, and spatial information on age-related changes from the largest dynamic functional mapping study of semantic processing. Participants (N = 130, age range 21-87 years, M = 51.05, SD = 14.73, 68 females) performed a semantic judgement task during magnetoencephalography (MEG), and significant task-related oscillatory responses were projected into anatomical space using a beamformer. Voxel-wise linear mixed-effects models were performed to assess semantic-related neural oscillations, irrespective of and influenced by age. Mediation analyses were performed to assess if local oscillations mediated the relationship between age and reaction time. Whole-brain analyses revealed stronger left-lateralized alpha/beta oscillations in frontotemporal cortices during semantic trials and stronger right-lateralized alpha/beta responses in temporoparietal regions during length trials (all ps <.001). Older adults showed stronger left temporoparietal alpha and left frontal beta during semantic processing and stronger alpha in the right temporal cortex during the length condition (all ps <.001). Alpha oscillations further mediated the relationship between age and reaction time in a hemisphere- and condition-specific manner, whereby right temporal activity mediated the age-behavior relationship in the length but not semantic condition (Z = 2.01, p =.022), while left temporoparietal activity significantly mediated this relationship in the semantic but not length condition (Z =  - 2.41, p =.008). Altogether, our findings demonstrate accentuated oscillations in aging which are hemisphere- and condition-specific and support compensatory processing to aid in maintaining adequate behavioral performance, lending clear support to leading neuroscientific models of aging.