Further support for a role of low-frequency oscillations derives from a macaque resting-state study that showed cross-correlations between LFP power at one cortical site (frontal, parietal, or visual cortex) and simultaneously acquired BOLD signals at distant sites (Schölvinck et al., 2010). Although gamma-frequency contributions were emphasized, theta- and alpha-frequency oscillations at times showed the strongest correlation with BOLD signals, consistent with our study. Because different functional networks can recruit distinct
frequency bands (Siegel et al., 2012), the particular low frequencies of neural oscillations that predominantly contribute to BOLD connectivity across the brain may be network dependent. It has been suggested that the biophysical properties of neural circuits determine the frequencies of network interactions (Siegel et al., 2012; Wang, 2010). For example, conduction selleck screening library delays
between distant network nodes may be one important factor contributing to the frequency range of cortical network interactions I BET151 (Kopell et al., 2000; von Stein and Sarnthein, 2000). Long conduction delays between distant brain regions may limit the frequency of large-scale network oscillations to a low-frequency band, accounting for the low-frequency oscillations observed during our multisite recordings. Evidence suggests that low-frequency oscillations (e.g., theta and alpha) can be generated locally in thalamic nuclei (Hughes and Crunelli, Carnitine palmitoyltransferase II 2005; Lörincz et al., 2008) or the deep layers of high-order visual cortex (Bollimunta et al., 2008, 2011; Lopes da Silva, 1991; Lopes da Silva and Storm Van Leeuwen, 1977) and propagated to other network nodes. Because previous studies of the neural basis of BOLD connectivity (He et al., 2008; Nir et al., 2008) focused mainly
on the primary sensory cortices (which reportedly have different oscillation-generating mechanisms; Bollimunta et al., 2008; Mo et al., 2011), rather than on these generators, the low-frequency contributions of oscillations to the BOLD signal may have been more difficult to detect. Different oscillatory frequency bands may also be associated with different functional properties. It has been suggested that different frequencies reflect different directions of cortical information transmission (Buffalo et al., 2011; Buschman and Miller, 2007; von Stein et al., 2000), specifically, gamma-band coherence for feedforward processing, and lower-frequency coherence for feedback processing. In our study (similar to other resting-state studies), the absence of visual stimulation in a completely dark room possibly reduced gamma activity in bottom-up processing and relatively increased the contribution from lower-frequency oscillations. However, an alternative interpretation of the finding of prominent alpha oscillations in deep cortical layers (Buffalo et al.