By now, it is commonly accepted in the scientific community that sleep is a necessary part of memory formation and consolidation. Over a century of clinical investigations has shown that sleep helps benefit the retention of memories and that sleep deprivation can negatively affect memory retention.
Though these connections between sleep and memory have been known for quite a while, scientists are still relatively in the dark about the specific neural mechanisms that underly the memory consolidation processes that occur during sleep. However, a recent study claims to have shed some light on the mystery of these neuronal memory mechanisms.
The study, published this past week in Nature, reports the observation of “ripples” of neuronal activity in the hippocampus during periods of sleep in human subjects. These ripples—oscillations of electrical activity in the brain—were associated with the “replay” of stimulus-specific neural activity elsewhere in the brain. The study found that the occurrence of these hippocampal oscillations and the replay of stimulus-specific activity during periods of sleep were heavily associated with more accurate memory recall in subsequent waking states. Thus, the observations imply that the hippocampal oscillations play an integral role in the consolidation of long-term memory during periods of sleep. The study marks the first direct physiological evidence of the connection between hippocampal oscillations and memory consolidation in human subjects.
Sleep & Memory
Initial theories of the role of sleep in memory consolidation had sleep playing more of a passive role. According to these theories, the state of lowered arousal during sleep allows that memory consolidations processes can go on unhindered. Sleep in itself, therefore, does not help with memory consolidation but makes it easier for the other parts of the brain to do so.
More recent theories of sleep and memory have sleep playing a much more active role in memory consolidation. Sleep, contrary to previous assumptions, does not just allow memory consolidation to go on unhindered but contributes actively to the consolidation of memories. Older research in this tradition focuses on the importance of periods of rapid eye movement (REM) sleep, a period of sleep in which a person exhibits heightened neurophysiological activity and quick eye movements. Now, however, the trend seems to be moving towards identifying the importance of slow brain wave activity and the associated neural mechanisms in memory consolidation.
It is well known that after initially encoding a memory, a person’s brain will “replay” the neural activity associated with that stimulus. This replay occurrence has been repeatedly observed in both humans and rodents, and it is well confirmed that the power of subsequent memory recall is related in some way to neural replay. In rodents, it has also been observed that replay activity during sleep is temporally coupled with characteristic hippocampal oscillations. Studies in rodents also suggest that hippocampal ripples are causally related to memory consolidation as interruption of sleep during periods of ripple activity contributes to decreased memory recall in later waking states.
Low temporal resolution of current fMRI imaging techniques prevents scientists from analyzing the temporal relationship between hippocampal oscillations and replay activity in human subjects. The only way to clearly measure these oscillations in humans is to implant a physical device in the brain that records electrical activity. Such devices are common in epilepsy patients, so for their study, the researchers pulled data from cranial implants in 12 epilepsy patients. First, they identified the neural profiles of stimulus-specific activity in the brain; neural activity that is correlated with a particular physical stimulus. Next, the researchers tracked the replay activity of the subjects during a mid-afternoon nap and related that information to oscillations in the hippocampus.
The researchers found that replay activity in the brain occurs regularly during periods of wake and rest, and the occurrence of replay activity alone, contrary to previous views, did not significantly predict later memory strength. However, they did find that replay activity triggered by hippocampal ripples occurred specifically for items that were later remembered only during periods of non-REM sleep. Thus, the observations suggest that hippocampal ripples during non-REM sleep provide a mechanistic explanation for the benefits of sleep on memory consolidation.
Additionally, the study found that only stimulus-specific activity in regions of the temporal lobe from relatively late poststimulus (500-1200ms) was related to subsequent memory while stimulus-specific activity from earlier (100-500ms) did not seem to matter much with respect to memory. This finding corroborates with previous research showing the response rate of memory selective neurons at ~550ms. This suggests that activity related to “deeper” semantic processing needs to reoccur during replay activity to reap the benefits of replay activity on memory. It is suggested that more “superficial” processing occurs in the ventral system, which seems to be related more toward perceptual processing. This finding accords with other studies that suggest “higher-order” representations, like linguistic representations, are more related to memory, as opposed to “lower-order” representations like perceptual or sensory representations.
Overall, the study provides direct physiological evidence for two major claims: (1) replay activity is related to hippocampal ripples and (2) ripple-triggered replay activity occurs specifically during non-REM sleep. Their observations also throw a wrench in the gears of previous views of the relation between sleep and memory consolidation. According to the study, replay activity alone is not significantly associated with remembering. Rather, replay activity triggered by hippocampal ripples during non-REM sleep seems to be the main locus of memory consolidation during sleep in humans. The asymmetry between replay activity in waking states and ripple-triggered replay activity during non-REM sleep suggests that the hippocampal ripples are the casually relevant variable that explains the benefits of sleep on memory consolidation.
To be sure, ripple-triggered replay activity is not the only mechanisms for memory formation. Humans can and frequently do consolidate memories while awake. However, this study gives an explanation for the particularly unique effects of sleep on memory consolidation. Despite being a thing commonly known and experienced by everyone, sleep is still somewhat of a mystery to us. Recent work such as this study brings us one step closer to explaining the phenomena of sleep, by showing the important function that sleep plays in memory consolidation.