Previous studies on the effect of aging and AD on power and FC
Alpha FC has been shown to be more reproducible than other frequencies (Duan et al., 2021), allowing high subject identifiability in resting state MEG, especially in the visual subnetwork (Sareen et al. , 2021). Consequently, the reduction in alpha FC with aging has been demonstrated using different recording setups (MEG, EEG) and techniques. For example, a recent MEG study showed that FC in Upper Alpha (UA) band (10 – 14 Hz) falls with age (Pathak et al. , 2022), owing to the reduction in individual peak alpha frequency (IAPF) with age (18 – 88 yrs), even though the FC (PLV and Phase lag index (PLI)) at the peak alpha frequency remained invariant. This matches with the falling trend of alpha (8 – 12 Hz) FC with age in our data (Figure 6a). Another study in eyes-closed resting state EEG (Scally et al. , 2018) reported reduced alpha power and FC in the conventional UA band (10 – 12 Hz) in older adults compared to young adults, although power and FC at IAPF were similar in the two age groups. Note that our results are not inconsistent with previous studies that have shown reduction in alpha power (Scally et al. , 2018) and IAPF with aging (Pathak et al., 2022), since these studies used a much wider age distribution than ours. We have previously shown that alpha power was indeed significantly less in middle-aged subjects (50-64 years) compared to young subjects between 20-49 years (Murty et al. , 2020). Further, since we computed the PSDs over 500 ms of data, the resulting frequency resolution of 2 Hz was not enough to study changes in IAPF. Another study in resting state EEG (Moezzi et al. , 2019) showed reduction in alpha FC (measured using imCOH) within old adults compared to the young, along with gamma FC, especially within the connections involving the occipital electrodes.
With AD progression, Dai and colleagues showed prominent disruption in FC in long-range connections (Dai et al. , 2015). These long range connections across the brain have been proposed to play an important role in aiding human cognitive function (Bullmore & Sporns, 2012). In a MEG study, Berendse and colleagues showed reduced fronto-parietal FC in 2 to 22 Hz range in AD patients over controls (Berendse et al. , 2000). FC measures like phase coherence and amplitude correlation have been shown to provide insights into large-scale neuronal interactions, which degrade in neuropsychiatric disorders (Siegel et al. , 2012).
In addition, FC of low-frequency oscillations such as delta and theta have also been shown to get altered with AD/ MCI. For example, a recent study showed disrupted event-related EEG FC over intra-hemispheric and midline theta band in AD subjects compared to MCI and healthy controls (Fide et al. , 2022). Another study in resting EEG reported reduced delta power in MCI patients (Liddell et al. , 2007) along with ERP components such as N300 which is related to working memory. Penttila and colleagues (1985) reported reduced theta power in mid-temporal area and reduced alpha in the occipital area in subjects with moderate AD over healthy controls. In addition, the study highlighted the slowing of alpha rhythm, i.e., IAPF reduction in AD patients. Olichney and colleagues (2002) showed that P600, an ERP measure of episodic memory encoding, decreased in MCI patients.