Subscribe to eTOC

In Older Adults, Less Frontal Volume and Slow-Wave Activity Add Up to Worse Memory


Aging is associated with grey matter atrophy, which is maximal in medial prefrontal cortex (left). This same brain region generates most of the slow wave oscillations (delta waves) observed in non-rapid eye movement (NREM) sleep in young adults. The degree of medial prefrontal grey matter atrophy predicts the magnitude of slow wave disruption (reduced slow wave activity) during NREM sleep in older adults (center). Reduced quality of NREM slow wave activity was directly associated with greater overnight forgetting of declarative memories (right) in older adults, even when adjusting for grey matter atrophy and age, suggesting that sleep may be an under-appreciated factor contributing to cognitive decline in later life.

Investigators conducted word association memory tests with older and young adults during the day and after a night's sleep, and found that the slow-wave activity of sleep, strongest in younger adults, was most correlated with better recall.

Brain atrophy, disrupted sleep, and memory loss are all part of the aging process, but are they related? New research strongly suggests that they are, and that loss of slow-wave activity during non-REM sleep is likely the key mediator between loss of brain volume on the one hand and reduced memory on the other.

The investigators performed memory tests on young and older cognitively normal adults while they were awake and after a night's sleep. During sleep, they measured slow-wave activity — defined as delta wave oscillations from 0.8 to 4.6 Hertz — using EEG recorded from multiple surface electrodes.

Slow-wave activity is the most important electrophysiologic predictor of sleep quality, and in recent years it has emerged as a central factor in the brain's ability to store episodic or biographic memories. Episodic memory refers to the “who, what, when, and where” of events we have experienced, versus procedural memories that aid us in performing rote tasks.

Slow-wave activity predominates during non-REM sleep, and experiments in young adults have shown that greater intensity of slow-wave activity correlates with better recall of memories learned the previous day, according to Bryce Mander, PhD, lead author on the online Jan. 27 Nature Neuroscience study and a post-doctoral fellow in psychology at the University of California at Berkeley.

A growing body of work has suggested that slow-wave activity that links the hippocampus to the medial prefrontal cortex is critically important in transferring episodic memories from the hippocampus, where they originate, to cortical areas for longer-term storage.

When forming a memory of, for instance, a reunion with a long-lost colleague, Dr. Mander said, “the theory is that initially declarative memory depends on the hippocampus to bind all the elements into one event: where you met, what you talked about, etc. — all of these sensory and cognitive elements are bound together in one whole.” The famous amnesia patient HM, for instance, lost the ability to form new memories because of damage to his hippocampus. But he retained older memories, because they were stored elsewhere.

“One hypothesis is that memories become transformed for more permanent storage,” a process mediated by the prefrontal cortex, Dr. Mander said. In this model, slow-wave activity during sleep allows a transfer of encoded memory from the hippocampus to the prefrontal cortex for dissemination throughout the brain.

If that is true, then could the loss of prefrontal volume explain age-related memory loss explicitly through reduction in slow-wave activity during sleep? While the idea is appealing, an alternative hypothesis is that while both cortical loss and memory loss occur in aging, they are independent processes. It was this that Dr. Mander set out to explore.


Dr. Mander first trained the cognitively normal older and younger adults in word association tests, in which a set of word-nonsense word pairs — for example, “bird jubu” — were presented together, each for five seconds. Their ability to recall the pair was tested either 10 minutes later, or the following day after a night of sleep in the sleep lab, during which slow-wave activity was assessed. The second recall test occurred while the subject underwent functional magnetic resonance imaging, to record activity in various brain regions during recall. Finally, a structural MRI was performed to assess atrophy.

The results showed that, as expected, older adults had about a 10 percent reduction in slow-wave activity compared with younger adults. They also had lower gray matter density across many brain areas, most significantly in the prefrontal cortex, where it was about 50 percent less than young adults. Two other sites that are also generators of slow-wave activity, the bilateral insula and the posterior cingulate, were also atrophied.

Dr. Mander's statistical analysis indicated that both age and atrophy predicted loss of slow-wave activity, but once atrophy was included in the statistical model, age no longer mattered: “Instead, the slow-wave activity decrease with age was significantly ‘mediated’ by reduction in medial prefrontal cortex gray matter,” meaning that atrophy, not age per se, was the determining variable in reduction in slow-wave activity.

Next, he looked at the effects of sleep on memory. Older adults did worse than younger ones on the initial recall experiment. Performance of both groups declined the next day, but the older subjects did considerably worse than the young ones, even after accounting for their poorer initial performance.

To understand which variable — age, prefrontal gray matter, or slow-wave activity — was most strongly associated with this differential loss of memory performance, Dr. Mander again turned to statistical analysis. While both age and gray matter were significantly associated with deficits in memory, he said, “neither significantly predicted the extent of overnight episodic memory retention when slow-wave activity was included in the model.” Whatever the contributions of age and gray matter loss, they were being mediated by slow-wave activity, rather than acting on memory independently of it.

Dr. Mander did several tests to rule out alternative explanations, including effects from loss of hippocampal volume, neurocognitive status, and circadian preferences of individual subjects (early birds versus nightowls). Finally, by conducting a similar test in a separate set of subjects, but separating the two tests by a waking — rather than a sleeping — period, he showed that a waking delay did little to prevent loss of recall in either older or younger subjects. Thus, the slow-wave activity of sleep, strongest in younger adults, was most correlated with better recall.

Dr. Mander was quick to point out that without a longitudinal study and the ability to manipulate slow-wave activity, it is too soon to say that slow-wave activity is directly responsible for better recall. “But what we can say is that there is strong evidence to implicate this pathway, which we think is supported by the literature as a whole, that brain structure is probably causing changes in sleep,” which is leading to reduced recall in aging brains. “The bottom line is, if you have more gray matter, it doesn't matter why, you will have higher slow-wave activity, and if you have higher slow-wave activity, it doesn't matter why, your memory is going to be better.”

The implications for cognitively normal adults are compelling: “It is possible that aggressively treating sleep disorders can have an impact on cognitive changes,” he said, although an interventional study to test that hypothesis still must be done. “The prefrontal cortex may be too disconnected from the hippocampus for slow-wave activity to have the effect it should have,” Dr. Mander said.


Antonio Culebras, MD, professor of neurology at SUNY Upstate Medical University in Albany, is more optimistic about the potential importance of sleep in dementia. “The major contribution of this article is to factor in sleep in dementia,” which is often discounted by neurologists and others. “We are beginning to get increasingly strong evidence that sleep is very important. If a person has dementia and sleeps well, likely the dementia is less expressive than if the patient has dementia and sleeps poorly.” But he is less certain that the prefrontal cortex is the main driver of sleep-related cognitive improvement, citing evidence showing an important role for the thalamus as well. “It is a very complex story,” he said.

Sara Mednick, PhD, assistant professor of psychology at the University of California, Riverside, suggested that future studies to address causality could build on preliminary work showing that transcranial magnetic stimulation can boost slow wave activity and improve memory in young adults (Massimini 2007).

“The question is whether that role changes as a person ages. Those are the studies that need to be done now,” she said. She also noted that exercise can improve sleep, and napping may be useful for making up sleep deficits and improving memory, as her own research has shown.


• Mander BA, Rao V, Walker MP, et al. Prefrontal atrophy, disrupted NREM slow waves and impaired hippocampal-dependent memory in aging. Nat Neurosci 2013; E-pub 2013 Jan 27.
    • Massimini M, Ferrarelli F, Tononi G, et al. Triggering sleep slow waves by transcranial magnetic stimulation. Proc Natl Acad Sci U S A. 2007;104(20):8496–8501.