Opposing the classical view, it soon became clear that ongoing activity carries information and is endowed with meaningful spatiotemporal structure, which
reflects previous learning and can bias the processing of stimuli (Engel et al., 2001 and Deco and Corbetta, 2011). The latter was first demonstrated by in vivo studies in cats combining microelectrode recordings with optical imaging (Arieli et al., 1996). These studies showed that low-frequency spatiotemporal fluctuations in ongoing activity could account for most of the trial-to-trial variability in sensory response amplitudes. Importantly, these fluctuations of ongoing activity were strongly synchronized across spatially distributed neuronal Doxorubicin order populations trans-isomer mw (Steriade et al., 1996a, Contreras and Steriade, 1997 and Destexhe et al., 1999), suggesting that processing of stimuli is biased not just by fluctuations in a local neuronal population but, actually, by the dynamics of coherently active networks. These coupling patterns in ongoing activity did not only involve low-frequency fluctuations in the delta-band (1–4 Hz) or below (Steriade et al., 1993, Contreras and Steriade, 1997 and Destexhe et al., 1999), but also faster frequencies in the theta- (5–8 Hz), alpha- (9–12 Hz), beta- (13–30 Hz), and gamma-frequency
range (>30 Hz) (Steriade et al., 1996a and Destexhe et al., 1999). Oscillations in these frequency bands are well known to be involved in a broad variety of cognitive processes (Singer, 1999, Fries, 2009, Engel and Fries, 2010 and Siegel et al., 2012). Oscillatory ongoing activity had also long been known from electroencephalography (EEG) studies of the human brain. However, the first demonstrations of spatially organized networks in ongoing activity were achieved using neuroimaging approaches such as fMRI (Biswal et al., 1995) Rutecarpine and positron-emission tomography (PET) (Raichle et al., 2001). These studies established
what became known as “resting state networks,” that is, networks of brain areas that show correlated fluctuations in the absence of a stimulus or task that the subject is engaged in (Fox and Raichle, 2007, Raichle, 2010, Deco and Corbetta, 2011 and Corbetta, 2012). In the past decade, a number of resting state networks have been extensively characterized using fMRI-based approaches. These include the default-mode and the dorsal attention network, as well as executive control, visual, auditory, and sensorimotor networks (Figure 1). Classically, the concept of resting state networks has been understood mainly in functional-anatomical terms, and it has been employed as a tool to map the structural organization and parcellation of brain systems (Yeo et al., 2011 and Buckner et al., 2013). As measured by fMRI, such networks show very slow (<0.