If you want to make sure to remember something, sleep on it. Sleep research confirms this adage, and indeed, its chemical underpinnings may have surprising implications for Alzheimer’s disease treatment. Neurons communicate with one another to solidify newly acquired information into memories, and while acetylcholine (ACh) clearly is a crucial neurotransmitter in this process, the optimal timing and level of its activity have yet to be fully explored. In this week’s early online PNAS, Steffen Gais and Jan Born of the University of Lubeck, Germany, report that memory consolidation during deep slow-wave sleep (SWS) works best if cholinergic activity is at a minimum. At first blush, this may seem paradoxical, given that Alzheimer’s disease is marked by memory problems and a progressive cholinergic deficit.

During long-term memory consolidation, newly acquired memories stored in the hippocampus become reactivated and transmitted to neocortical networks. This dialogue requires that the inhibition of excitatory feedback synapses along the pathway be lifted, and a silencing of cholinergic activity is thought to facilitate this. Therefore, according to this model, high levels of ACh during SWS will thwart consolidation. In the present study, 29 healthy male volunteers learned a word-pair list (a declarative, hippocampus-dependent task) and a mirror tracing task, which is non-declarative and hippocampus-independent. After training, some men slept for three hours, during which they received the cholinesterase inhibitor physostigmine. Meanwhile, one control group remained awake on the same physostigmine infusion, and a second received a placebo after learning.

Following physostigmine administration, the researchers tested the subjects’ recall. They found that physostigmine diminished the consolidating effect of sleep on declarative memory, but not on non-declarative memory retention, compared to placebo. By contrast, the rise in ACh due to physostigmine administration did not impair memory retrieval in the wake group. In fact, these men appeared to do slightly better than placebo-treated ones in both the declarative and non-declarative memory performance tests. Overall cholinergic neurotransmission is known to dwindle in Alzheimer’s disease (AD), and cholinesterase inhibitors are commonly prescribed to patients (see Live Discussion). In light of their findings, the authors suggest that administration of these drugs before sleep be reevaluated.—Erene Mina

Erene Mina is a graduate student at the University of California, Irvine.


Q&A with Steffen Gais and Jan Born. Questions by Erene Mina.

Q: What exactly are the implications of your work for the use of the three cholinesterase inhibitor drugs donepezil, rivastigmine, and galantamine? Does the pharmacokinetics of these drugs allow for dosing them such that brain levels are low at night?
A: Our work has been done in healthy subjects and should not be used as a basis for treatment recommendations. However, our results suggest that daytime administration could be more effective than nighttime administration. The available drugs have different half-lives. Especially, those drugs usually given multiple times a day have pharmacokinetics which would allow this kind of administration.

Q: This is one small study; what other research is needed before physicians should reconsider the drug regimen they prescribe?
A: Acetylcholinesterase inhibitor administration in the morning should be compared to conventional treatment in a larger sample of AD patients. Especially, long-term memory (retention over more than 24 hours) should be tested.

Q: These drugs have been tested in thousands of patients; have you looked at the trial data to see if you can find indications of a memory deficit as predicted by your work?
A: Sleep-associated long-term memory formation has not been tested in AD. Generally, this aspect of memory formation has been totally ignored in patients with memory deficits.

Q: Could it be that the overall cognitive improvement observed in patients on these drugs masks a selective, sleep-related memory deficit?
A: Yes. According to known data, these drugs improve acquisition of new memories. Long-term consolidation of these memories, however, suffers from the high levels of ACh during sleep.

Q: Many AD patients have disturbed sleep cycles; they get up at night to "go to work," for example. Does this tie in with your findings?
A: It is known that high levels of ACh can interfere with sleep. Discontinuing treatment before sleep could improve sleep also in AD patients (as well as confusion during intermittent awakenings). However, this needs to be tested.

Q: Why did you decide to choose men only for your study?
A: Women are harder (and more expensive) to test in pharmacological trials (exclusion of pregnancy, intake of oral contraceptives, control of menstrual cycle, etc).

Q: How did the drowsiness, fatigue, motivation, and sleep quality of the subjects affect the results?
A: All these factors were controlled for and excluded as possible confounds.

Q: What did the subjects in the wake group do while the sleep group slept? Did this affect their performance?
A: The main comparison was between placebo and physostigmine. Because the subjects were doing the same things during placebo and physostigmine conditions (watching movies in the wake group), the conclusion regarding the effect of physostigmine does not depend on waking activity. In addition, other studies could show that staying awake for a short time (three hours) does not impair memory performance.

Q: Is there a "window of opportunity" for consolidation to occur? I mean, if you took the subjects off physostigmine and repeated the experiment, do you think you would see a difference (i.e., recovery)?
A: This may be the case. We believe that the process of sleep-related consolidation is strongest in the night after learning, but may continue, perhaps to a lesser degree, on subsequent nights. This is, in fact, the issue of a current study.

Q: Do your results merely suggest that cholinesterase inhibitors be administered when patients are awake? Or are they generally just not as useful as we think?
A: They are useful. Optimal levels of ACh are necessary for the acquisition of new memories. However, strengthening [consolidation] of newly learned memories seems to benefit from low levels of ACh during sleep. From these results, it might be beneficial to take the medication in the morning and not before sleep.


  1. Comment by Alireza Atri, Chantal Stern, and Michael Hasselmo

    There is an exciting link between the important recent article by Gais and Born on cholinergic mechanisms in memory and our own article in the current issue of Behavioral Neuroscience. In their study, Born and Gais showed that augmentation of cholinergic function during slow-wave sleep (SWS) by injection of physostigmine impaired the memory consolidation effect of SWS on learning of word-pairs in 29 young healthy male volunteers. This result is complementary to the findings of our recently published study (Atri et al., 2004).

    The major finding in our own study was that lowering cholinergic function in waking increases proactive interference—the interference of older memories with the learning and remembrance of new ones. Proactive interference influences common tasks such as remembering where we parked the car or where we left the keys. If one parks in the same lot every day, the memory of previous parking locations interferes when we try to encode and retrieve a new but similar parking place (this is called proactive interference).

    In our study, we transiently lowered cholinergic function in awake young healthy volunteers by injecting them with scopolamine. Scopolamine is a drug that transiently blocks acetylcholine (ACh) receptors and which has been commonly used for decades for a variety of purposes, including as a treatment for motion sickness. Using a word paired-associate memory paradigm, our results supported the a priori hypotheses that lowering cholinergic function, by scopolamine, should impair new learning of novel word pairs and increase proactive interference, but not impair memory for previously learned novel word pairs. The study involved 28 young healthy participants who were divided into three groups and either received scopolamine, glycopyrrolate (a more peripherally acting anticholinergic medication that has fewer effects in the brain) or no drug.

    The findings of Gais and Born are complementary with ours in that they both support predictions from theoretical models (Hasselmo, 1999). These models suggest that acetylcholine (ACh) is important to prevent proactive interference in the hippocampus during initial learning by suppressing retrieval of previously stored memories, to prevent them from interfering with new encoding, but that the release of this suppression is necessary to allow consolidation of new memories. Our study supports the idea that high brain ACh levels during wakefulness are important to acquire new memories and to reduce proactive interference, and that blocking the effects of ACh with scopolamine enhances proactive interference.

    The other side of this hypothesis is supported by the findings of Born and Gais: that during SWS, lower ACh levels are required for proper consolidation of newly acquired memories by allowing stronger excitatory feedback transmission to reactivate memories for consolidation within neocortical brain areas.

    It is interesting to speculate that these findings and model of cholinergic function may offer partial explanations for some of the memory and psychiatric symptoms (including delusions and hallucinations) that are found in many of the conditions known to be associated with lowered levels of ACh in the brain (such as Alzheimer's disease, Lewy body dementia, Parkinson's disease, and schizophrenia), and also for the partial improvement of these symptoms by the use of medications (acetylcholinesterase inhibitors) that increase brain ACh levels.

    Obviously, from the isolated standpoint of cognitive health, the chronic use of medications with anticholinergic effects by elderly individuals, especially those who are cognitively impaired, would be highly discouraged unless they are truly necessary. Use of anticholinergic medications, even in young healthy individuals, and particularly during waking hours, would be expected to interfere with acquisition and future recall of new and especially related memories. Until specific studies address these issues, these two studies may infer that a good strategy, at least in young healthy individuals, would be to take anticholinergic medications at bedtime, and procholinergic medications during the day. In the absence of additional data, to further stretch this inference to the timing of administration of medications for patients with Alzheimer's disease may well be a worthwhile strategy to consider in the interim.


    . Neuromodulation: acetylcholine and memory consolidation. Trends Cogn Sci. 1999 Sep;3(9):351-359. PubMed.

  2. I found your article interesting. I am not a reseacher, but my father has AD, and he tells me that Aricept makes him dream vividly throughout the night. I also worked in a sleep clinic for a number of years, and we were told by our supervisors that it was believed that REM sleep had something to do with memory consolidation. Therefore, I assumed that the increased REM activity Dad experiences had something to do with why drugs in the Aricept family slow memory loss. However, what you are describing is interesting as it is different from what I thought about sleep and memory. Anyway, I found some references that might be of interest to you.


    . Is Alzheimer's disease related to a deficit or malfunction of rapid eye movement (REM) sleep?. Med Hypotheses. 1993 Nov;41(5):435-9. PubMed.

    . Sleep and brain lesions: a critical review of the literature and additional new cases. Neurophysiol Clin. 2001 Dec;31(6):356-75. PubMed.

    . Dementia--the failure of hippocampal plasticity and dreams. Is there a preventative role for melatonin?. Med Hypotheses. 1987 Sep;24(1):59-68. PubMed.

Make a Comment

To make a comment you must login or register.


Other Citations

  1. Live Discussion

Further Reading


  1. . Blockade of central cholinergic receptors impairs new learning and increases proactive interference in a word paired-associate memory task. Behav Neurosci. 2004 Feb;118(1):223-36. PubMed.
  2. . Dissociable stages of human memory consolidation and reconsolidation. Nature. 2003 Oct 9;425(6958):616-20. PubMed.
  3. . Neuromodulation: acetylcholine and memory consolidation. Trends Cogn Sci. 1999 Sep;3(9):351-359. PubMed.
  4. . Transient memory impairment and hallucinations associated with tolterodine use. N Engl J Med. 2003 Dec 4;349(23):2274-5. PubMed.
  5. . Chronically mad as a hatter: Anticholinergics and Alzheimer's disease pathology. Ann Neurol. 2003 Aug;54(2):144-146. PubMed.

Primary Papers

  1. . Low acetylcholine during slow-wave sleep is critical for declarative memory consolidation. Proc Natl Acad Sci U S A. 2004 Feb 17;101(7):2140-4. PubMed.