, who examined changes in cortical responsiveness across isolated periods of SWS, these studies examined effects of sleep as a whole comprising the repeating sequence of SWS and REM sleep. Thus, they basically do not exclude the possibility that REM sleep contributes to the net downscaling effect observed after sleep. Likewise, Chauvette et al. cannot exclude such a possibility, because they did not manipulate REM sleep.

Fortunately, also in this issue of Neuron, Grosmark et al. (2012) provide data suggesting such a contribution of REM sleep to processes of downscaling. Across triplets of NonREM-REM-NonREM sleep, they revealed a significant decrease in firing rates of rat hippocampal pyramidal cells and interneurons, consistent Selleck Alectinib with the occurrence of downscaling across

sleep ( Figure 1). However, analyzing the firing dynamics within each NonREM and REM sleep period in detail revealed a substantial decrease in firing rates only during REM sleep; NonREM sleep periods instead were associated with an increasing firing rate. As the REM-associated decrease in firing rate outreached the firing increase during NonREM sleep, a net decrease in firing resulted across the whole sleep period. Interestingly, BMN 673 chemical structure the mean decrease in firing rate from one to the next NonREM sleep period was significantly correlated to EEG theta power during the interleaving REM sleep period, suggesting that theta is involved in the downscaling process. Reductions in firing rates do not necessarily reflect synaptic downscaling. Also, because hippocampal sleep differs from neocortical sleep,

it remains and unclear whether similar firing relationships occur in cortical neurons. Nevertheless, these data open a new perspective on how sleep could contribute to synaptic homeostasis by suggesting a possible involvement of REM sleep in downscaling. Rather than SWS alone, the sequence of SWS and REM sleep periods might be important (Giuditta et al., 1995). In combination, the findings by Chauvette et al. (2012) and Grosmark et al. (2012) do not question the concept of global synaptic downscaling during sleep but instead suggest that processes during REM sleep should be taken into consideration. Beyond this, Grosmark et al.’s findings offer an interesting link between global processes of downscaling and the consolidation of specific memories in local networks, because they analyze firing occurring in the presence of hippocampal ripples, which regularly accompany the neuronal replay of newly encoded memory representations from the prior waking period (O’Neill et al., 2010). Ripple-associated replay during SWS has been considered the key mechanism launching the consolidation of newly acquired episodic memories (Diekelmann and Born, 2010). Grosmark et al. report that during ripples, cells fire more synchronously, and this firing paradoxically increases across NonREM-REM-NonREM sleep triplets.