The history of research that sleep contributes to the consolidation of memory began in 1885.1 In 1914,2 and again in 1924,3 other similar studies confirmed that nonsense syllables learned just before going to bed were more likely to be remembered the following day after a period of sleep than when a similar time period of wakefulness occurred between learning and recall.

The discovery of rapid eye movement (REM) sleep in 1953 made it clear that sleep was not a homogeneous state of passive rest for the brain. In fact, the low-voltage, mixed-frequency electroencephalographic (EEG) brainwaves of REM sleep were seen to be not all that different from those of waking brain activity. Sleep and memory research remained dormant for some time, but by the late 1980s and early 1990s, interest in the field of sleep and memory was reignited. Studies involving trampolining,4 intensive learning of a foreign language,5 and Morse code learning6 all demonstrated increases in REM sleep following successful learning.

Get breaking research findings delivered straight to your e-mail inbox. Subscribe to Sleep Report.


Specific EEG waveforms during REM sleep may be important in the process of memory consolidation. Sawtooth waves in the theta frequency seen during REM sleep may occur right before a burst of rapid eye movements.7 What is incredibly thought provoking is that during prior wakefulness, the process of temporarily storing a memory trace in the hippocampus also occurs at this theta frequency.8 Not simply the frequency but also the actual pattern of hippocampal neuronal firing seen during waking learning may later be witnessed during subsequent REM sleep.9 This process has been witnessed in rats running wheels and learning mazes as well as in zebra finches learning new songs.10

Although there is much evidence supporting the idea that memories are consolidated during REM sleep, forgetting or pruning of unnecessary memories may also occur during this time.11,12 The hippocampus deep in the temporal lobe allows the formation of new associations and episodic memories by blending such disparate information as the who, what, where, and when of a memory trace. Networks of hippocampal neurons show reactivation of the heterogeneous firing pattern of the experience initially formed during waking learning.13,14

Through a novel method to detect the temporal sequence in the firing of individual hippocampal neurons,15 Poe et al have demonstrated that when neurons encoding novel memories are active on the peak of hippocampal theta waves,14 long-term potentiation,16,17 or strengthening of the connection occurs, while when neurons are active on the trough of the waveform, the result is weakening of the connection, or long-term depression.18,19 So it is the timing of firing of hippocampal neurons during either the peak or the trough of a theta wave that may make the difference between strengthening and weakening of memories in the hippocampus.20 These studies provide clues as to how our life experiences may slowly be transformed into semantic memory. We all know the layout of our home town, but which one of us remembers each specific experience that, when combined, allowed us to learn this information? It seems that through time, the less important aspects of a memory are pruned while the core of what we really need to know is consolidated.

The sequence of sleep stages within an ultradian sleep cycle is predictable. It seems unlikely that this order is accidental. Apparently, at least some degree of processing of the memory trace likely occurs prior to REM sleep.21 The sleep spindles of stage II sleep and the slower higher-amplitude brainwaves of slow wave sleep are also clearly important. The sleep spindles seen in stage II sleep trigger an influx of calcium ions into cortical neurons.22,23 This calcium influx may be involved with gene expression in cortical neurons, which can then lead to synaptic sprouting with other neurons.24,25 In motor procedural tasks, an increase in the total number of sleep spindles may be seen for the groups that do well on post-training retesting.26 This may be especially true for the sleep spindles obtained in the last quarter of the night.27 The oscillatory 12-14 Hz buzzing of sleep spindles may be priming cortical neurons for synaptic plasticity.16,17

Timothy J. Walter, MD

As the 1-2 Hz large-amplitude slow EEG pattern of slow wave sleep (SWS) is very similar to that of the most common brainwave pattern seen in coma, some authors have described SWS as an “abject annihilation of consciousness.”28 The key in understanding SWS is that it is not at all what it initially seems to be.29 Within each wave of SWS is the firing of many fast spiking brainwaves. The reason that they appear to be slow is that the whole cortex at once is firing together in synchrony. The predominantly frontal lobe synchronous firing during SWS is involved in the acquisition of new memories.30-32 Not surprisingly, there is a link in timing of these slow waves and firing patterns of cells in the hippocampus.33


Which sleep stage is most important in the consolidation of memory? Usually, when there are multiple sides to a story, the truth typically resides somewhere in the middle. The body of research as a whole is pointing toward the conclusion that REM sleep, the sleep spindles of stage II sleep, and SWS are all important in memory, for different but complementary reasons.

The idea that the sequence of the sleep stages is what is important was first documented in 1977 by Giuditta.34 He believed that newly acquired memory traces first underwent a processing step during slow wave sleep that allowed their further processing during REM sleep. His initial theory in 1977 was that slow wave sleep was necessary for clearing unnecessary memories and experiences, and that REM sleep must be more important in consolidating new memories. It seemed to make sense to erase less important memories first, before implementing the step that keeps the good ones. The sequence of slow wave sleep invariably preceding REM sleep present throughout the cycles of the sleep of all mammals certainly seemed to support this logic. Additionally, human fetal sleep is largely REM sleep35 and the percentage of REM sleep decreases as we age.36 The idea is that REM sleep must be high initially as the animal is learning from scratch and that there is little need for slow wave sleep initially if there is nothing yet to forget.37

Since the publication of his initial theory in 1977, Giuditta has refined his theory to some degree38 to account for the replay of firing of memory sequences during slow wave sleep, but he certainly deserves the credit for being the first to realize that it is the sequence of the sleep stages that is paramount.

By comparing functional imaging studies during sleep stages following learning with various neuropsychological tests, Walker has done much to elucidate the various neurological contributions during the different sleep stages on different types of memory. Poe et al has demonstrated that the brain wave frequencies of REM sleep allow the formation or pruning of neuronal synaptic connections.20 Born et al has demonstrated that transcranial electrical stimulation (TES) provided to the frontal lobes during slow wave sleep seems to enhance learning,39 and Tononi has used transcranial magnetic stimulation (TMS) to trigger the slow waves of slow wave sleep.40


Is there a way to enhance the learning process to improve upon our ability to learn? Clearly, if learning is impaired by something as simple as sleep deprivation, then this should be addressed first. Although few scholars are currently listening to tapes of recorded knowledge while sleeping or sleeping with books under their pillows, one study has demonstrated enhanced learning when auditory clicks were presented while learning and then during bursts of rapid eye movements in REM sleep.41 Neuronal excitation may be enhanced when auditory clicks are presented during the sleep spindles of stage II sleep during the ascent from slow wave sleep to REM sleep.42 Born used TES through the scalp over 30-minute periods to modulate excitability in the frontal lobes during slow wave sleep. This study claims that the direct cortical stimulation significantly improved memory in regard to memorizing word pairs. When the same stimulation was applied during wakefulness, the stimulation did not improve the memory for the word pairs. The conclusion was that the effects of transcranial direct cortical stimulation enhanced the generation of slow waves and thus facilitated the processes of neuronal plasticity.39


Theoretically, a device could be devised that could enhance normal learning. This device would presumably provide a stimulus during the learning of conscious awareness, and then again during the occurrence of certain neurophysiologic waveforms of sleep.

TMS stimulation could be provided both during waking learning and during sleep. TMS could be provided either over the temporal lobes superficial to the hippocampus, or over the appropriate cortical areas corresponding to the learning at hand. For instance, a task involving dexterity of the right hand may warrant TMS stimulation over the precentral gyrus of the left hemisphere corresponding with motor control of the right hand. During sleep, TMS stimulation could occur during the slow waves of slow wave sleep, possibly during the predicted onset of incline in slope of the frontal slow waves. Alternatively, TMS could also be provided during the sleep spindles of stage II sleep, especially for the memory processing of motor procedural tasks. As a simple finger tapping task demonstrated that the stage II sleep in the last quarter of the sleep night may be especially important,43 it may be important to provide TMS during sleep spindles at this time. TMS could also be given during the sawtooth waves of REM sleep that typically occur just prior to the phasic bursts of rapid eye movements, or during the actual phasic eye movements of REM sleep. Poe’s research may indicate that the timing of stimulation on the peak versus the trough of REM sleep EEG waveforms in the theta frequency, presumably sawtooth waves, may make the difference between consolidation and pruning.

As TMS may be expensive, potentially epileptogenic, and of limited availability, other tools may potentially be used to enhance memory consolidation. Auditory cues given both during waking learning and during these specific neurophysiologic waveforms of sleep may also conceivably enhance learning. Short auditory snippets, similar to various cell phone ring tones, or short clips of well-known songs, each unique and each lasting several seconds in duration, could be played while learning a specific task, procedure, fact, or relationship to be remembered by the subject. A different auditory snippet could be used for each specific thing to be learned, but the same auditory snippet would be used if the same information is rehearsed more than once, so that a specific auditory snippet is specific to each task, procedure, fact, relationship, or other such thing to be remembered. The auditory snippets could also be given first during learning and then during sleep, specifically during the sawtooth waves of REM sleep, the actual phasic eye movements of REM sleep, during the sleep spindles of stage II sleep especially for the stage II sleep seen following slow wave sleep just prior to the onset of REM sleep, or during the incline of slope of the slow waves of slow wave sleep. Such timing may theoretically be important in order to correctly orchestrate the process of remembering the important information while actively erasing any erroneous or unnecessary associations.

It is also conceivable that visual, somatosensory, or olfactory stimulation could be provided during waking learning and later during sleep in a similar fashion as that described for auditory snippets. Born has demonstrated that the olfactory scent of a rose provided during learning and then again during SWS enhanced learning as retested the following day.44 In a similar way, various combinations of sensory modalities or sensory modalities together with TMS may also prove to be promising.

Is the enhancement of the normal learning process really possible? Only time will tell. It is not clear if we will all soon be wearing electrode grids or hearing auditory snippets while we learn and while we sleep so that we can all learn at a superhuman rate. In the meantime, watching how the research pans out will sure be interesting.

Timothy J. Walter, MD, is the author of REM Illumination Memory Consolidation (available at, and is currently the co-medical director of Capitol Sleep Medicine Sleep Diagnostic Center at Grove City, Ohio. He can be reached at [email protected].


  1. Ebbinghaus H. Uber das Gedachtnis.Untersuchungen zur Experimentellen Psychologie. Leipzig: Duncker und Humblot; 1885.
  2. Heine R. Uber der Wiedererkennen und ruckwirkende Hemmung. Zeitschrift fur Psychologie. 1914;68:161-236.
  3. Jenkins JG, Dallenbach KM. Obliviscence during sleep and waking. Am J Psychol. 1924;35:605-612.
  4. Buchegger J, Meier-Koll A. Motor learning and ultradian sleep cycle: an electroencephalographic study of trampoliners. Percept Mot Skills. 1988;6:635-645.
  5. De Koninck J, Lorrain D, Christ G, et al. Intensive language learning and increases in rapid eye movement sleep: evidence of a performance factor. Int J Psychophysiol. 1989;8:43-47.
  6. Mandai O, Guerrien A, Sockeel P, et al. REM sleep modifications following a Morse code learning session in humans. Physiol Behav. 1989;46:639-642.
  7. Walter T. Cycling through the brainwave patterns of sleep. In: REM Illumination Memory Consolidation. Lotus Magnus; 2007:34-35.
  8. Hirase H, Leinekugel X, Czurko A, et al. Firing rates of hippocampal neurons are preserved during subsequent sleep episodes and modified by novel awake experience. Proc Natl Acad Sci USA. 2001;98:9386-9390.
  9. Pavlides C, Winson J. Influences of hippocampal place cell firing in the awake state on the activity of these cells during subsequent sleep episodes. J Neurosci. 1989;9:2907-2918.
  10. Dave AS, Margoliash D. Song replay during sleep and computational rules for sensorimotor vocal learning. Science. 2000;290:812-816.
  11. Crick F, Mitchison G. The function of dream sleep. Nature. 1983;304:111-114.
  12. Marr D. Simple memory: a theory for archicortex. Philos Trans R Soc Lond B Biol Sci. 1971(841);262:23-81
  13. Louie K, Wilson MA. Temporally structured replay of awake hippocampal ensemble activity during rapid eye movement sleep. Neuron. 2001;29(1):145-56.
  14. Jablonski P, Poe GR, Zochowski M. Structural network heterogeneities and network dynamics: a possible dynamical mechanism for hippocampal memory reactivation. Phys Rev E Stat Nonlin Soft Matter Phys. 2007;75(1 Pt 1):011912. Epub 2007 Jan 16.
  15. Waddell J, Dzakpasu R, Booth V, et al. Causal entropies—a measure for determining changes in the temporal organization of neural systems. J Neurosci Methods. 2007;162(1-2):320-32. Epub 2006 Dec 22.
  16. Steriade M, Amzica F. Coalescence of sleep rhythms and their chronology in corticothalamic networks. Sleep Research Online. 1998;1(1):1-10.
  17. Steriade M. Coherent oscillations and short term plasticity in corticothalamic networks. Trends Neurosci. 1999;8:337-345.
  18. Huerta PT, Lisman JE. Bidirectional synaptic plasticity induced by a single burst during cholinergic theta oscillation in CA1 in vitro. Neuron. 1995;15:1053-1063.
  19. Pavlides C, Greenstein YJ, Grudman M, et al. Long-term potentiation in the dentate gyrus is induced preferentially on the positive phase of theta-rhythm. Brain Res. 1988;439:383-387.
  20. Booth V, Poe GR. Input source and strength influences overall firing phase of model hippocampal CA1 pyramidal cells during theta: relevance to REM sleep reactivation and memory consolidation. Hippocampus. 2006;16(2):161-73.
  21. Walter T. The sequential hypothesis. In: REM Illumination Memory Consolidation. Lotus Magnus; 2007:173.
  22. Sejnowski TJ, Destexhe A. Why do we sleep? Brain Res. 2000;886:208-223.
  23. Ghosh A, Greenburg ME. Calcium signaling in neurons: molecular mechanisms and cellular consequences. Science. 1995;268:239-247.
  24. Abel T, Nguyen PV, Barad M, et al. Genetic demonstration of a role for PKA in the late phase of LTP and in hipppocampus-based long-term memory. Cell. 1997;88:615-626.
  25. Li W, Llopis J, Whitney M, et al. Cell-permanent caged InsP3 ester shows that Ca++ spike frequency can optimize gene expression. Nature. 1998;392:936-941.
  26. Fogel S, Jacob J, Smith C. Increased sleep spindle activity following simple motor procedural learning in humans. Actas de Fisiologia. 2001;7:123.
  27. Walker M, Brakefield T, Morgan A, et al. Practice with sleep makes perfect: sleep dependent motor skill learning. Neuron. 2002;35:205-211.
  28. Eccles JC. Chairman’s opening remarks. In: Wolstenholme GEW, O’Connor M, eds. The Nature of Sleep. London: Churchill; 1961:1-3.
  29. Walter T. Slow wave sleep may deliver the memory upload. In: REM Illumination Memory Consolidation. Lotus Magnus; 2007:165-170.
  30. Marshall L, Helgadottir H, Molle M, Born J. Boosting slow oscillations during sleep potentiates memory. Nature. 2006;444(7119):610-3.
  31. Kudrimoti HS, Barnes CA, McNaughton BL. Reactivation of hippocampal cell assemblies: effects of behavioral state, experience, and EEG dynamics. J Neurosci. 1999;19:4090-4101.
  32. Nadasdy Z, Hirase H, Czurko A, et al. Replay and time compression of recurring spike sequences in the hippocampus. J Neurosci. 1999;19:9497-9507.
  33. Molle M, Yeshenko O, Marshall L, et al. Hippocampal sharp wave-ripples linked to slow oscillations in rat slow-wave sleep. J Neurophysiol. 2006;96(1):62-70. Epub 2006 Apr 12.
  34. Giuditta A. The biochemistry of sleep. In: Davidson AN, ed. Biochemical Correlates of Brain Structure and Function. New York: Academic Press; 1977:293-337.
  35. Roffwarg HP, Muzio JN, Dement W.C. Ontogenetic development of the human sleep-dream cycle. Science. 1966;152:604-619.
  36. Anders T, Sadeh A, Appareddy V. Normal sleep in neonates and children. In: Ferber R, Kryger MH, eds. Principles and Practice of Sleep Medicine in the Child. Philadelphia: Saunders; 1995:7-18.
  37. Giuditta A, et al. The neurochemical study of sleep. In: Lajtha A, ed. Handbook of Neurochemistry. Vol 8. 2nd ed. New York: Plenum Press; 1984:443-476.
  38. Giuditta A, et al. The role of sleep in memory processing: the sequential hypothesis. In: Maquet P, Smith C, Stickgold R, eds. Sleep and Brain Plasticity. New York: Oxford University Press; 2003:157-178.
  39. Marshall L, Molle M, Hallschmid M, Born J. Transcranial direct current stimulation during sleep improves declarative memory. J Neurosci. 2004;24:9985-9992.
  40. Massimini M, Ferrarelli F, Esser SK, et al. Triggering sleep slow waves by transcranial magnetic stimulation. Proc Natl Acad Sci USA. 2007;104(20):8496-501. Epub 2007 May 4.
  41. Smith C, Weeden K. Post training REMs coincident auditory stimulation enhances memory in humans. Psychiatric Journal of the University of Ottawa. 1990;15:85-90.
  42. Church MW. Johnson LC. Human EEG and cardiovascular responses to spindle synchronous and spindle asynchronous clicks during stage 2 sleep. Sleep Research. 1977;6:24.
  43. Walker M, Brakefield T, Morgan A, et al. Practice with sleep makes perfect: sleep dependent motor skill learning. Neuron. 2002;35:205-211.
  44. Rasch B, Büchel C, Gais S, Born J. Odor cues during slow-wave sleep prompt declarative memory consolidation. Science. 2007;315(5817):1426-9.