In juxtaposing the Indian and neuroscience conceptions of deep sleep, I have proceeded so far as if the Indian notion of dreamless sleep corresponds to NREM slow-wave sleep. But we can now see that this correspondence is too simplistic. The Indian conception of dreamless sleep suggests that we need a finer taxonomy of sleep states—a taxonomy that is not just physiological but also phenomenological, and that accommodates the ways that sleep may be culturally variable as well as flexible and trainable through meditative practices.
Consider that the fourth century C.E. author, Vyāsa, in his commentary on Patañjali’s Yoga Sūtras, distinguishes three types of sleep that are recalled upon awakening—peaceful sleep, disturbed sleep, and heavy sleep. According to the cosmology that informs Yoga, these three types of sleep result from whichever one of the three qualities or tendencies (guṇas) predominates in the psychophysical complex. Overall, the quality of dullness or the tendency to inactivity (tamas) dominates the mind in ordinary sleep. Sleep is heavy or stupefying when this quality is not modified by either of the two other qualities or tendencies. Sleep is disturbed and restless when the quality of excitation or tendency to activity (rajas) is present. And sleep is peaceful and refreshing when the quality of lightness or tendency to clarity (sattva) is present. When the Vedānta philosophers describe deep and dreamless sleep as blissful, it is deep sleep, with this quality of clarity, that they have in mind.
When sleep-lab participants are roused from NREM sleep, however, they sometimes report that they have been thinking while they were asleep, and often they describe going around in a repetitive loop of rumination. Although this kind of thinking probably occurs mainly in stage 2 NREM sleep, it is also reported during awakenings from deeper slow-wave sleep.
Owen Flanagan appeals to this finding to argue that there is no such thing as dreamless sleep and hence no sleep completely lacking in consciousness ( 2000). Contrary to the standard neuroscience view, Flanagan thinks we are always conscious while asleep because we are always dreaming. Dreaming, he proposes, is any conscious mental activity occurring during sleep, not just mental activity involving sensory imagery. If ruminative thinking occurring in NREM sleep counts as dreaming, and if this kind of mental activity can happen during slow-wave sleep, then all sleep stages involve dreaming and at least some degree of consciousness.
From the Indian perspective, however, we need to distinguish clearly between two things. One is whether there is such a thing as dreamless sleep; the other is whether we are conscious while we sleep. Yoga and Vedānta agree that consciousness is present while we sleep, but this is not because we are always dreaming, even if we define “dreaming” widely to mean any kind of thinking during sleep. On the contrary, what Yoga and Vedānta mean by “dreamless sleep,” as we have seen, is that sleep state in which there are no sensory or mental objects of awareness, that is, no images and no thoughts. Nevertheless, they maintain, there is awareness, so this state is a conscious state; it is a mode of consciousness without an object.
In the Yoga framework, reports of ruminative thinking upon awakening indicate a coarser or shallower sleep state—one closer to the surface of thinking consciousness—and a state with a strong quality of excitation or tendency toward movement of the mind.
Consider now the reasons that sleep scientist J. Allan Hobson gives for doubting the reliability of waking reports of ruminative thinking during slow-wave sleep:
Reports of antecedent mental activity elicited following awakenings from deep sleep are rendered unreliable by the brain fog through which they must pass […]. Even if the deeply sleeping brain were capable of the low-level ruminations sometimes implied by experimental reports, it is unlikely that they would survive the inertia of awakening. It may even be that the tumult of the awakening process triggers the chaotic and fragmentary mentation that is reported. And even when deep sleepers are sufficiently aroused to be interviewed, they may still generate huge slow waves in their EEGs, indicating that they are in a semistuporous state quite different from either sleeping or waking. Indeed, they may even hallucinate, become anxious, and confabulate as if they suffered from delirium. This is precisely what happens in the night terrors of children. (1999, pp. 142–143)
Clearly, this too is a far cry from the Indian conception of dreamless sleep. Neither reports of ruminative thinking nor waking hallucinatory confabulations correspond to the Yoga and Vedāntan descriptions of dreamless sleep as a peaceful or blissful state free of mental activity, from which one awakens feeling alert and refreshed. From the Yoga perspective, what Hobson describes are sleep states strongly marked by a quality of dullness combined with mental excitation upon awakening.
My point here is not at all that sleep science should refine its taxonomy using the Yoga framework. It is rather that ultimately we cannot map the Indian notion of dreamless sleep using already-established scientific categories, especially the physiologically-defined sleep stages, which, even from a scientific perspective, are now recognized as too crude to capture the moment-to-moment dynamics of electrical brain activity during sleep, let alone the experiences with which they may be correlated (Nir & Tononi 2009). Not only is the Indian notion phenomenological and metaphysical, rather than physiological, it is also embedded in a normative framework that understands sleep in contemplative terms. So, to bridge from sleep science and the neuroscience of consciousness to the Indian conception of dreamless sleep, we need to view sleep as a mode of being that is trainable through meditation.
From the Yoga perspective, entering a state of blissful dreamless sleep on a regular basis requires leading a calm and peaceful life guided by the fundamental value of nonviolence (ahiṃsā), practicing daily meditation, and treating going to sleep and waking up as themselves occasions for meditation—for watching the mind as it enters and emerges from sleep.
In addition, from a yogic perspective, we need to distinguish between ordinary dreamless sleep and lucid dreamless sleep. Ordinary dreamless sleep is the sleep of ignorance, in which awareness is described as being in total darkness. Lucid dreamless sleep is described as a state in which awareness is luminous and without an object (free of thoughts and images). Whereas lucid dreaming consists in knowing that you are dreaming, lucid dreamless sleep is said to consist in being able to witness the state of dreamless sleep and recall its phenomenal clarity upon waking up. Although the background metaphysics of Yoga, Vedānta, and Indo-Tibetan Buddhism differ in significant ways, they all describe lucid dreamless sleep as disclosing a basal level of pre-personal consciousness that lies deeper than the modes of awareness that characterize the ego-centred waking and dreaming states.[6]
At this point you may wonder whether we have strayed back into the realm of metaphysics. Does this conception of dreamless sleep really have any descriptive phenomenological content or is it simply a consequence of the Indian metaphysical views that identify the true self with pure consciousness (as in the case of Vedānta) or that maintain that there is no self but only an ownerless stream of consciousness that continues in dreamless sleep (as in Indo-Tibetan Buddhism)?
From a purely textual perspective, the metaphysical and the phenomenological are thoroughly intertwined in the Indian discussions. From a cognitive science perspective, however, we can ask whether the idea of inducing lucid dreamless sleep through certain types of meditation is experimentally testable, and, more generally, whether meditation is associated with altered sleep patterns or has measurable effects on sleep. Two neuroscience studies of sleep in relation to meditation are suggestive in this regard.
One recent study comes from the laboratories of Giulio Tononi and Richard Davidson (Ferrarelli et al. 2013). They examined slow-wave sleep in highly experienced Theravāda Buddhist and Tibetan Buddhist meditation practitioners. They found that the long-term meditators, compared to non-meditators, had significantly increased fast-frequency gamma activity, as recorded by high-density EEG, in a parietal-occipital region of the scalp during NREM sleep. In addition, the higher gamma activity was positively correlated with the length of meditation training. This finding is notable because gamma-frequency electrical brain activity is a well-known neural marker of conscious cognitive processes (Tononi & Koch 2008), including certain types of meditative states in long-term meditation practitioners (Lutz et al. 2004). Gamma activity has also been shown to distinguish lucid dreaming from non-lucid dreaming in REM sleep (Voss et al. 2009; see also Voss & Hobson this collection). During NREM sleep, however, gamma activity tends to decrease, so the higher gamma activity in the meditators could reflect a capacity to maintain some level of awareness. More generally, the study suggests that there may be distinct slow-wave sleep states associated with meditation practices.
Another older study examined long-term practitioners of TM (Transcendental Meditation) who reported what they called the subjective experience of “witnessing” during sleep (Mason et al. 1997). They described this experience as one of feeling a continuous and peaceful awareness without dreams while one sleeps and as resulting in one’s feeling refreshed upon awakening. The main finding was that the long-term meditation practitioners, compared to short-term practitioners and non-meditators, showed a unique EEG pattern during slow-wave sleep, one in which faster alpha and theta waves were superimposed on the slower delta waves. Although we cannot draw clear conclusions about what these distinctive physiological patterns mean, including whether they are due to TM practice or some other cause, the authors of the study interpret them as supporting the presence of a different kind of slow-wave sleep state in individuals who report witnessing of sleep.
These two studies reinforce the point that we cannot use already established categories from sleep science to map the Indian conception of dreamless sleep. This conception, besides being closely tied to a specific phenomenology, which in turn reflects a specific metaphysics, is embedded in a normative cultural framework that aims to bring about and promote certain kinds of contemplative sleep states. Instead of trying to fit these states into a physiological scheme derived from studying the way twentieth-century Americans and Europeans sleep in the sleep lab, we need to enlarge the conceptual framework of sleep science to include contemplative ways of training the sleeping mind. This project will require that sleep scientists, cognitive neuroscientists, cognitive anthropologists, and Western and Indian philosophers work together to map the sleeping mind. In short, we need a cross-cultural cognitive science and neurophenomenology (Lutz & Thompson 2003) of the wake–sleep cycle, one that draws on the combined expertise of Western and Asian theoretical traditions.
One benefit of such a cross-cultural cognitive science is that it could offer new data relevant to our guiding question about consciousness and dreamless sleep. Consider the following testable, neurophenomenological hypothesis: In highly-experienced practitioners of certain types of meditation, compared to individuals without this kind of experience, we should observe a stronger correlation between subjective reports of phenomenal qualities of sleep and various objective measures of brain activity. Specifically, if highly experienced meditators were able to provide reports upon awakening about qualities of their experience of the state they call dreamless sleep, and if cognitive neuroscientists were able to relate these reports to fine-grained features of sleep physiology and to familiar aspects of the neural correlates of consciousness, then we would have new evidence from experimental science that a certain type of dreamless sleep in certain individuals counts as a mode of phenomenal consciousness whose felt qualities can be made accessible to verbal report.[7]
This hypothesis also cast lights on our earlier discussion of sleep-state misperception. From a contemplative perspective, when little attention is given to sleep as an occasion for the practice of mindfulness, it is not surprising that sleep-state perception will be unreliable, even in ordinary individuals, let alone patients suffering from insomnia or other sleep disorders. In contrast, sleep-state perception may be more reliable when sleep is valued in a contemplative way and is treated as an opportunity for cultivating mindfulness. Whether these assumptions are correct is something that neurophenomenology should test.