Neuro-oscillations in memory consolidation and forgotten parts of the brain: Commentary on Weiner et al., 2023

The precise synchronization of neuro-oscillatory patterns during NREM sleep, called grapho-elements, has been implicated in memory consolidation in several studies. More precisely, coupling of spindles to the up state of slow oscillations within cortical regions and coupling of hippocampal ripples to cortical slow waves and spindles is thought to be an especially important electrophysiologic basis of this process (Klinzing et al., 2019). What exactly is happening on a cellular level during grapho-elements is still up to debate. Several hypotheses have been formulated as to what changes do synapses and neurons undergo to consolidate newly acquired information for later retrieval (Klinzing et al., 2019; Tononi & Cirelli, 2006). In this issue, Weiner et al. (2023) demonstrate that subjects whose spindles occur closer to the peak of the up state show better memory performances, thus furthering our understanding of the electrophysiologic architecture required for proper memory consolidation. Interestingly, they demonstrate that within subjects, the preferred phase of the slow wave the spindles lock to stays the same before and after a memory task, thus indicating that this feature of spindle-slow wave-locking might be a fixed trait in each individual. However, another study compared the coupling of younger and older adult humans and found a shift from the peak towards the ascending part of the up state with increasing age (Helfrich et al., 2018). Thus, while spindle-slow wave-locking could be independent of factors such as the amount of novel information to be stored, it seems to undergo shifts over time and the finding of Weiner et al. could point towards some brains retaining a more ‘youthful’ architecture as they age. Open questions in this regard concern the thalamo-cortical anatomy and channel configuration of neurons. While brains are known to undergo atrophy with age, a simple reduction in cell number should not necessarily cause a change in the architecture of neuro-oscillations. It is known however that the brain atrophies unevenly with age (Raz & Rodrigue, 2006); thus, a relative change in the connectivity of the anatomical triangle consisting of cortex, thalamic nuclei and reticular thalamic nucleus is insinuated in the current data. Furthermore, age-related changes in the channel configuration of single neurons (Petralia et al., 2014) could play a role as well. However, both mechanisms have not been investigated in relation to electrophysiologic architecture during sleep. They additionally found that a stronger locking of spindles to slow waves, that is, a steeper rise and fall of spindle amplitude during the up state, is associated with memory performance, indicating that oscillatory synchrony of a greater set of neurons during a sleep spindle is conducive to memory consolidation. Here again, what happens precisely to neurons and synapses during spindles is not 100% clear. In vitro data point towards induction of long-term potentiation (LTP) (Rosanova & Ulrich, 2005), and in vivo data demonstrate replay of cortical firing patterns right before sleep spindles (Peyrache et al., 2009). However, it is not clear yet how and if these two observations are linked to the formation of memory engrams. The precise coordination of neuro-oscillations across brain regions might help to synchronize replay and the induction of LTP across different parts of the brain to become more connected and thus form a memory engram. Lastly, upon researching literature for this comment, the heavy focus on the prefrontal cortex as the anatomical substrate for memory consolidation during sleep was very striking. Early theories of memory consolidation stated information were stored in the hippocampus temporarily and directly, while it was processed and classified in the neocortex, without stating that it was the prefrontal cortex exclusively. As the field advanced, it was proposed that early in the memory consolidation process the hippocampus ‘binds’ synapses distributed across the neocortex, which form an ‘engram’, with the different sensory information encoded in the respective parts of the cortex (Klinzing et al., 2019). During sleep, the engram then becomes slowly hippocampus independent. In both models, the sensory and associative cortices are said to participate in the storage of memories as well, yet the prefrontal cortex has received most of the attention, at least in the electrophysiological field. However, fMRI data point toward a participation of the precuneus in the retrieval of word pair memory (Krause et al., 1999), and two-photon data demonstrated the involvement of the sensory cortex in memory retrieval (Xie et al., 2014). Descriptive studies show that these parts of the brain are electrophysiologically coupled to the hippocampus during sleep as well (Jiang et al., 2019). It stands to reason that grapho elements occurring in sensory and associative cortices should be more investigated in their role of memory consolidation as well, as they are necessarily involved in acquiring information that will later be translated to the declarative parts of memory. Open access funding provided by Universitat Bern. The author declares no conflicts of interest. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/ejn.16056.