Psychedelics, Meditation, and Self-Consciousness

Over the past 20 years, a large number of neuroimaging studies have investigated various styles of meditation, using electro-encephalography (EEG), functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) to measure changes in blood flow and electrical activity in the brain during meditation. This research program seeking to isolate the neural correlates of various meditation practices has come to be known as contemplative neuroscience. Interestingly, the wealth of data collected from these studies has begun to reveal that meditation practices from distinct traditions relying on similar mental techniques, also share some common neural correlates. A recent meta-analysis of 78 functional neuroimaging studies of meditation found dissociable patterns of activation and deactivation for four common styles of meditation: focused attention (FA), mantra recitation (MR), open monitoring (OM), and loving-kindness meditation (LK) (Fox et al., 2016; see Table 1).

FA was correlated with significant activation clusters in executive brain areas such as the premotor cortex and the dorsal anterior cingulate cortex, which may underlie top-down regulation of attention and monitoring of spontaneous thoughts. FA was also correlated with the deactivation of two important hubs of the so-called default-mode network (DMN) namely, the posterior cingulate cortex and inferior parietal lobule, involved in a plethora of introspection-related functions including self-reflection, mind-wandering, autobiographical memory recollection, mental time travel to the future and imagination more broadly. The main clusters of activation for MR were found in the motor control network (Broca's area, premotor cortex and supplementary motor cortex) and the putamen. The involvement of these regions in speech production and attention regulation is consistent with the practice of mantra recitation, which relies on the constant repetition of a phrase either in one's mind or out loud. MR was also associated with a significant deactivation cluster in the anterior insula, suggesting that the focus on a repeated phrase is linked to decreased awareness of bodily sensation.

By contrast, OM was found to be correlated with activation clusters in the insular cortex, involved in awareness of interoceptive signals, as well as in brain regions associated with the voluntary control of thought and action, such as the left inferior frontal gyrus, the pre-supplementary and supplementary motor areas, and the premotor cortex. OM was also associated with the deactivation of the right thalamus, a region involved in filtering out sensory stimuli (sensory gating), suggesting that the increased focus of awareness during OM is mediated by decreased sensory gating. Finally, significant activation clusters for LK were found in the right somatosensory cortices, the inferior parietal lobule and the right anterior insula, while no significant deactivation cluster was found. Although these four meditation styles are clearly dissociated by their neural correlates, Fox and colleagues found a few recurrent patterns of activity modulation, in particular in the insular cortex, an important multisensory area heavily involved in interoceptive awareness (Craig, 2002; Simmons et al., 2013).

It is worth noting that distinct styles of meditation were found to modulate the activity of the insula in different ways, namely activation for FA, OM, and LK, and deactivation for MR. Nonetheless, the involvement of the insula in all four styles of meditation points toward the central role of the attentional control of bodily awareness, and awareness of breathing in particular, during various contemplative practices. Other convergent patterns of activity were found to a lesser extent in regions involved in the regulation of attention such as the premotor cortex, supplementary motor area and dorsal cingulate cortex (Fox et al., 2016).

Importantly, and albeit not emphasized by Fox et al. (2016), attenuation of either activity or functional connectivity in the default mode network (DMN), a large-scale intrinsic network which is highly active at rest but less active during goal-directed tasks (Raichle et al., 2001), was shown for MR (Berkovich-Ohana et al., 2015), and in many studies of mindfulness meditation (combining FA and OM in different degrees). These studies reported a decrease of activity in key nodes of the DMN, particularly in the medial prefrontal cortex (Farb et al., 2007, 2013; Brewer et al., 2011; Scheibner et al., 2017) and in the posterior cingulate cortex (Brewer et al., 2011; Ives-Deliperi et al., 2011; Pagnoni, 2012; Brewer and Garrison, 2014; Lutz et al., 2016; Scheibner et al., 2017), compared to meditation-naïve controls or within-group resting state (reviewed by Tang et al., 2015). While DMN deactivation is not specific to mindfulness meditation, it has been found to be more pronounced during meditation than during other cognitive tasks (Garrison et al., 2015). Moreover, several studies reported altered connectivity during different types of meditation most consistently in association with the DMN (Tang et al., 2015). Specifically, within-DMN connectivity was found to be reduced during LK (Garrison et al., 2014).

Recent years have seen a renaissance of scientific research on psychedelic drugs, using modern neuroimaging techniques to gather insight into the neural correlates of their vivid subjective effects (see dos Santos et al., 2016 for a review). Early neuroimaging studies using PET and single-photon emission computed tomography (SPECT) after administration of psilocybin and mescaline found excitatory effects on frontal cortical areas, medial temporal lobes and the amygdala (Hermle et al., 1992; Vollenweider et al., 1997, 1998; Gouzoulis-Mayfrank et al., 1999). By contrast, recent fMRI of psilocybin and ayahuasca found significant reductions in activity across many brain areas, including frontal and temporal cortical regions, as well as hubs of the DMN (Carhart-Harris et al., 2012; Palhano-Fontes et al., 2015; see Table 2). This apparent discrepancy could be due to the much greater timescales used by PET/SPECT compared to fMRI (Carhart-Harris et al., 2012).

Recent fMRI studies also revealed alterations of resting-state functional connectivity in key nodes of the DMN after administration of psilocybin (Carhart-Harris et al., 2012, 2013; Tagliazucchi et al., 2014; Lebedev et al., 2015), ayahuasca (Palhano-Fontes et al., 2015) and LSD (Carhart-Harris et al., 2016b; Tagliazucchi et al., 2016; see Table 2). Researchers found increased integration between cortical regions under ayahuasca (de Araujo et al., 2012), psilocybin (Carhart-Harris et al., 2013; Roseman et al., 2014), and LSD (Carhart-Harris et al., 2016b; Tagliazucchi et al., 2016; Müller et al., 2017). Finally, psilocybin and LSD were found to produce an enhanced repertoire of dynamical brain states (Tagliazucchi et al., 2014; Atasoy et al., 2017) and increased spontaneous MEG signal diversity (Schartner et al., 2017).

Psychedelic drugs are known to produce short-term, dramatic effects on self-consciousness, especially at higher doses. This phenomenon is known as drug-induced ego dissolution (DIED); it is described a loss of one's sense of self and self-world boundaries, together with a concomitant oceanic feeling of “oneness” or “unity” (Letheby and Gerrans, 2017; Millière, 2017; Nour and Carhart-Harris, 2017). Such effects on self-consciousness had already been reported in early studies with mescaline (Mayer-Gross and Stein, 1926; Beringer, 1927; Guttmann and Maclay, 1937), LSD (Anderson and Rawnsley, 1954; Savage, 1955; Von Mering et al., 1955; Bercel et al., 1956; Klee, 1963; Sedman and Kenna, 1964; Pahnke and Richards, 1966) and psilocybin (Rümmele and Gnirss, 1961; Pahnke, 1966). Recent studies have sought to investigate the neural correlates of DIED, using correlations between questionnaire items and neuroimaging data. The main correlates of DIED, summarized in Table 3, include increased functional connectivity between the DMN and a task-positive network (Carhart-Harris et al., 2013), decreased integrity of the DMN (i.e., decreased within network correlation of timecourses) (Carhart-Harris et al., 2016b) and salience network (Lebedev et al., 2015), decreased connectivity between the parahippocampal and retrosplenial cortices (Carhart-Harris et al., 2016b), increased entropic brain activity (Lebedev et al., 2016) and spontaneous MEG signal diversity (Schartner et al., 2017), and decreased mean energy and fluctuations of low frequency connectome harmonics (Atasoy et al., 2017).

The familiar experience of thinking about oneself is perhaps the most salient form of self-consciousness. In philosophy of mind, this is also known as having de se thoughts, namely thoughts that involve the first-person concept and are naturally expressed using the first-person pronoun (García-Carpintero, 2015). De se thoughts themselves come in different flavors, which are more or less egocentric. Thus, one may explicitly reflect on one's personality traits or one's life trajectory, both of which are important elements of an individual's identity. This category of de se thought broadly pertains to the entertainment of core self-related beliefs, and is often linked to the notion of narrative selfhood—the stories we tell ourselves about the kind of person we are or want to be. Admittedly, this kind of de se thought only occurs sporadically in the waking state, because we are not constantly reflecting on our personal identity. However, de se thoughts also include more mundane and pervasive instances of mind-wandering, such as wondering what one will have for dinner. Such thoughts also link back to the self, insofar as they engage the first-person concept, even if they are not directly related to fundamental beliefs about one's identity. More generally, this family of self-referential cognitive content includes not only de se thoughts about the present moment, but also autobiographical memory retrieval and self-centered mental time travel to the future, both of which also crucially involve the self. Together, these self-referential mental episodes constitute what may be called narrative self-consciousness, namely the complex sequences of self-centered thoughts, memories and imaginings that weave the narrative of our daily lives and shape our core self-related beliefs (Damasio, 1999; Gallagher, 2000; Schechtman, 2011).

There are at least two ways in which narrative self-consciousness may be disrupted. First, the rate of occurrence of self-referential thought and mental time travel may be dramatically reduced, or altogether suppressed, during a certain time interval. There is convincing evidence from experience sampling studies that mind-wandering and mental time travel is more ubiquitous in the waking state than we might think, due to the fact that we are often unaware of such episodes (Smallwood and Schooler, 2006; Killingsworth and Gilbert, 2010; Baird et al., 2011). Moreover, daily mind-wandering episodes are predominantly future-focused, and frequently involve the planning and anticipation of personal goals, known as autobiographical planning (Baird et al., 2011; D'Argembeau, 2016). Thus, it is not exaggerated to say that a large part of our lives is spent entertaining self-involving thoughts, insofar as this includes spontaneous de se thoughts about the future. Most contemplative practices, especially in the Buddhist tradition, explicitly aim at increasing meta-awareness of mind-wandering, namely monitoring and taking explicit note of spontaneous thoughts, in order to disengage from them and re-focus attention back onto a particular object (such as the breath or a mantra) or on a wider awareness of the present moment (Hasenkamp, 2018). Once attention has been stabilized and the mind “quieted,” meditators can undergo prolonged conscious episodes entirely lacking in self-referential thoughts3.

This is consistent with the neurophysiological basis of mindfulness meditation reviewed in the previous section. While key nodes of the DMN such as the mPFC and the PCC are especially active during mind-wandering (Christoff et al., 2009; Fox et al., 2015), a number of studies have found that these regions are deactivated during mindfulness meditation—which is consistent with the practice's focus on attentional control of mind-wandering episodes (Farb et al., 2007; Brewer et al., 2011; Lutz et al., 2016; Scheibner et al., 2017). Importantly, the decrease in DMN activity is more significant during meditation than during other cognitive tasks (Garrison et al., 2015), which speaks to the crucial issue of specificity and may be characteristic of the suppression of mind-wandering that can be achieved by trained meditators. It is also worth noting that experiences of flow (Csikszentmihalyi, 1975, 1990), namely states of intense focus in which self-referential processing is inhibited (such as intense practice in expert athletes or jazz improvisation in expert musicians), are also associated with decreased activity in the mPFC (Ulrich et al., 2014). Interestingly, psychedelic drugs have been shown to decrease activity in the mPFC and PCC as well (Carhart-Harris et al., 2012; Palhano-Fontes et al., 2015). These nodes of the DMN have been shown to be more active during various types of self-related stimuli (Northoff et al., 2006; Sugiura et al., 2008; Qin and Northoff, 2011; Tacikowski et al., 2017), self-reflection (Jenkins and Mitchell, 2011; D'Argembeau, 2018) and autobiographical memory retrieval (Cabeza and St Jacques, 2007). A recent study using dynamic causal modeling suggested that self-referential processes are driven by PCC activity and modulated by the regulatory influences of the mPFC (Davey et al., 2016).

Aside from changes in cerebral blood flow in these regions, psilocybin, ayahuasca and LSD have also been consistently linked to decreased functional integrity of the DMN (Carhart-Harris et al., 2012, 2016b; Palhano-Fontes et al., 2015). Moreover, DMN disintegration was correlated with reports of ego dissolution (Carhart-Harris et al., 2016b) and decreased mental time travel to the past (Speth et al., 2016). These findings are intriguing, because strong DMN connectivity at rest is associated with increased tendency for mental time travel, in particular spontaneous thoughts about the future (Schacter et al., 2007; Godwin et al., 2017; Karapanagiotidis et al., 2017; Wang et al., 2017). This suggests that reports of drug-induced ego dissolution may be related to the experience of decreased self-referential thought and mental time travel, which is also fundamental to the practice of meditation.

While the temporary cessation of self-referential thoughts is one way in which narrative self-consciousness may be altered, it may also be disrupted by a total loss of access to autobiographical memories and self-related beliefs. These two cases are not unrelated, given that the unavailability of personal memories and beliefs precludes the possibility of at least some forms of de se thought and mental time travel. However, the experience of losing access to these memories and beliefs might differ from the mere cessation of de se thought from a phenomenological point of view. Indeed, anecdotal evidence from narrative self-reports suggests that drug-induced ego dissolution may be related in some cases to reversible retrograde amnesia, specifically regarding abstract information about oneself (i.e., semantic autobiographical memory). For example, one subject responding to an online questionnaire on ego dissolution reported “forgetting that I was a male, a human, a being on Earth—all gone, just infinite sensations and visions,” while another stated “I no longer felt human. I didn't remember what a human was” (R.M., unpublished data from an online survey on drug-induced ego dissolution completed by experienced users of psychedelic drugs)4.

Interestingly, decreased functional connectivity within the DMN is associated with age-related memory deficits (Damoiseaux et al., 2008; Ward et al., 2015) and Alzheimer's disease (Greicius et al., 2004; Dennis and Thompson, 2014; Dillen et al., 2017). Given the link between DMN disintegration (decreased within-network connectivity) and ego dissolution, it is intriguing to speculate that the temporary loss of access to semantic autobiographical information that can occur in the psychedelic state may be mediated by pronounced reductions in DMN integrity. Interestingly, post-hoc mediation analyses from Dillen and colleagues revealed that the retrosplenial cortex enables communication between the hippocampus and DMN regions in healthy controls, but not in individuals reporting cognitive decline or diagnosed with prodromal Alzheimer's disease (Dillen et al., 2017). Ratings of ego dissolution under LSD have been shown to correlate with decreased resting-state functional connectivity between the retrosplenial cortex and the parahippocampus (Carhart-Harris et al., 2016b). If the retrosplenial cortex acts as a gateway between the hippocampal formation and specific DMN regions for memory retrieval, the correlation between this change in connectivity and ego dissolution may be related to a loss of access to autobiographical memories. It is also interesting to note that damage to the retrosplenial cortex has been linked with retrograde amnesia (Valenstein et al., 1987; Takayama et al., 1991; Gainotti et al., 1998; Oka et al., 2003), and damage to the hippocampal formation has been associated with both retrograde amnesia (Scoville and Milner, 1957; Squire, 1992; Steinvorth et al., 2005) and loss of general world knowledge (Rempel-Clower et al., 1996; Bayley et al., 2006; Gregory et al., 2014).

In summary, narrative aspects of self-consciousness can be radically altered during a specific conscious episode in two ways: through a temporary cessation of self-referential thought and mental time travel, or more dramatically through a temporary loss of access to semantic autobiographical information, resulting in a complete breakdown of one's personal identity. While the former can seemingly be induced both by contemplative practices and by psychedelic drugs, the latter appears to be more specific to psychedelics, and may be related to more pronounced reductions in DMN integrity—which does not appear to be such a reliable feature of meditation.

In the first case, the retrieval of self-related information is effectively reduced, via attentional control in the case of meditation, and perhaps via an involuntary attentional shift in the case of psychedelic states. In the second case, even the dispositional ability to retrieve such information is temporarily impaired, which amounts to a form of reversible amnesia. It is important to note that these alterations of narrative aspects of self-consciousness come in degrees. At low doses of psychedelics or in novice meditators, self-referential spontaneous thoughts and mental time travel are unlikely to disappear entirely for an extended period of time. By contrast, states of deep absorption achieved by advanced meditators or with higher doses of psychedelics are more likely to involve a complete cessation of these thoughts. Similarly, the disruption of access to self-related information may be partial or total in different cases. For example, there is a difference between the inability to remember what one has done the day before, the inability to remember one's name, and the inability to remember anything at all about oneself, including that one belongs to the human species (which has been reported, post-hoc, with certain doses of some psychedelics). Thus, narrative aspects of self-consciousness may be temporarily disrupted in different ways and to varying degrees by meditation and psychedelics and it is not yet clear whether these different degrees or grades lie on a linear or nonlinear scale.

While self-related thoughts are a paradigmatic example of self-consciousness, it is widely agreed that self-consciousness is not strictly limited to the cognitive domain. In particular, a number of authors have stressed the need to distinguish between the “narrative self,” congruent with narrative aspects of self-consciousness outlined in the previous section, and the “minimal” or “embodied” self (Damasio, 1999; Gallagher, 2000; Legrand and Ruby, 2009; Christoff et al., 2011; Musholt, 2013; Zahavi, 2014). For example, Gallagher defines the minimal self as “a consciousness of oneself as an immediate subject of experience, unextended in time” (Gallagher, 2000, p. 15), by opposition with the temporal thickness of the narrative self-woven by autobiographical memories and self-projection to the future. Moreover, many authors have insisted on the idea that the minimal self is crucially linked to embodiment and agency, equating this basic form of self-consciousness with an awareness of oneself as an embodied agent (Legrand, 2006; Bayne and Pacherie, 2007; Christoff et al., 2011; Seth, 2013). While the distinction between high-level/narrative and minimal/embodied selfhood is helpful as a first pass to clarify the umbrella notion of self-consciousness, it remains somewhat ambiguous and potentially simplistic as such. As we have seen, the “narrative self” is better construed as a family of distinct self-referential processes, which may or may not involve mental time travel, be spontaneous, or recruit abstract semantic information. Likewise, the “embodied” or “minimal” self may be construed as a complex set of somatosensory and agentive aspects of self-consciousness which can come apart in special cases (Blanke and Metzinger, 2009).

At least three constructs that have been related to a basic form of self-consciousness rooted in multisensory processing may be distinguished, namely: (a) the sense of body ownership, namely the alleged sense of “mineness” that one experiences with respect to one's own body or individual limbs; (b) bodily awareness in general, namely the awareness of any bodily sensation, either internal (interoception and proprioception) or external (tactile); and (c) spatial self-location, namely the experience of being located somewhere in space with respect to one's perceived environment.

Non-dual awareness meditation (NDA) refers to a family of practices which can be found in several Eastern contemplative traditions, including Dzoghen and Mahāmudrā within Tibetan Buddhism, and Advaita Vedanta and Kashimiri Shaivism within Hinduism (Josipovic, 2010; Dunne, 2011). NDA meditation rests on three core assumptions: (a) ordinary experience is “dual” or dichotomous, insofar as it is structured around a subject-pole and an object-pole; (b) this subject-object dichotomy is illusory, because conscious awareness as such is not fundamentally dual; (c) it is possible, through the practice of NDA meditation, to dispel this illusion and directly experience conscious awareness as non-dual. All of these assumptions are worth discussing. The first assumption, in particular, requires clarification. It is rather uncontroversial that all conscious mental states have a subject of experience, insofar as “an experience is impossible without an experience” (Frege, 1956, p. 299, translation modified). This is merely a metaphysical requirement of conscious experience8. However, assumption (a) goes further in claiming that the phenomenal character of conscious experience is itself structured by a subject-object dichotomy.

There are at least two ways to understand this claim. As we have seen in the previous section, a number of components of ordinary experience can be related to a form of self-consciousness—including self-related thought, body ownership, bodily awareness and spatial awareness. One can claim that conscious experience normally involves a background awareness of oneself which is reducible to one or several of these components (Damasio, 1999; Blanke and Metzinger, 2009; Bermúdez, 2011; Seth, 2013). According to this reductionist interpretation, ordinary consciousness is structured by a subject/object dichotomy insofar as we are normally aware of ourselves via thought, perception and bodily sensations in addition to being aware of external objects. A second interpretation holds that there is a form of sui generis self-awareness in experience which is irreducible to the cognitive, bodily and spatial features of experience: “being presented with something necessarily involves being pre-reflectively and pre-conceptually aware of being the subject to whom something is presented” (Nida-Rümelin, 2017, p. 66; see also Strawson, 1999; Kriegel, 2009; Zahavi, 2014; Guillot, 2017). According to this second interpretation, the seemingly dichotomous nature of experience does not rest on a specific kind of self-representing content, but on the very nature of conscious representation in general, which is structured by an implicit distinction between the represented objects and the subject to whom those objects are presented. In similar fashion, Evan Thompson has argued that ordinary experience is infused with a sense of mineness such that every thought, emotion, perception or sensation is experienced as one's own, and has also suggested that this feature can disappear during meditation (Thompson, 2014b, p. 362).

On both of these interpretations, conscious experience is structured by a subject-object dichotomy insofar as it involves an awareness of oneself in addition to the awareness of external objects. Accordingly, non-dual awareness states can be construed as conscious states which lack the background self-awareness normally present in experience. However, a core assumption of non-dual awareness meditation practices is that the subject-object dichotomy supposedly found in ordinary experience is illusory, in line with the so-called “no self” doctrine of Buddhism (anātman in Sanskrit or anattā in Pali). Thus, non-dual awareness practice is supposed to reveal that the putative phenomenological distinction between oneself and one's experience of the external world is ultimately an illusion9.

While this general idea appears to be consistent with recent proposals regarding the notion of non-dual awareness, it is not always clear which of the two interpretations outlined above is favored. For example, Wolfgang Fasching has argued that in normal experience, subjects are aware of their body and location in addition to objects of the external world, which suggests that he favors the reductionist interpretation of the subject/object dichotomy. In the same vein, he suggests that self-consciousness is rooted in the identification of oneself with “certain configurations of experienced contents as opposed to others” (Fasching, 2008, p. 476), which is also consistent with the first interpretation. However, he goes on to argue that “in perception I am necessarily co-conscious of myself” (Fasching, 2008, p. 472), because the subject/object polarity is built in all conscious representational states. He further claims that some meditation practices can reveal that “I am not something ‘inner' as distinct from external objects” and that “there is no ‘I' to which things are given, there is just the event of givenness” (Fasching, 2008, p. 478). In this context, meditation is conceived as a way of becoming aware of consciousness as such, without the illusory distinction between the subject and the objects of experience. Fasching cites a central text of the Vedic tradition, the Māṇḍūya Upaniṣad, which describes the state of non-dual awareness reached through meditation as one in which the subject is “not conscious of the internal world, nor conscious of the external world” (Deutsch, 1969, p. 62). Similarly drawing on the Vedic tradition, Miri Albahari suggests that non-dual awareness meditation involves “a direct realization that consciousness… is ownerless” (Albahari, 2010, p. 104).

These descriptions seem consistent with the second interpretation above: in ordinary conscious states, subjects have a sense that all of their experiences are theirs, given to them, or owned by them; in turn, this introduces an artificial distinction between oneself as the subject of experience and the experiences themselves; finally, trained meditators can dispel this illusion by becoming aware of consciousness itself as a non-dual process. Although we cannot discuss this proposal at length within the scope of this article, it is worth underlining that it rests on a controversial picture of ordinary experience. Indeed, the idea that consciousness is normally infused with a special sense of phenomenal mineness or awareness of oneself as the owner of one's experiences is far from obvious and has been met with a number of objections (see Howell and Thompson, 2017; O'Conaill, 2017; Chadha, 2018; McClelland, forthcoming; Wu, forthcoming).

By contrast, the first interpretation of non-duality outlined above seems consistent with the evidence presented in the previous section: the inhibition of self-related thoughts, body ownership, bodily awareness and self-location should entail a blurring or dissolution of the boundary between self and world, and the associated “unitive experience”—identified by Walter Stace as the core feature of so-called “mystical-type experiences” (Stace, 1960). Although narrative and multisensory aspects of self-consciousness are not illusory as components of ordinary experience, their transient cessation in meditation or drug-induced states may be interpreted as revealing that their association with a more substantial notion of selfhood is fallacious (see also Metzinger, 2003; MacKenzie, 2016; Letheby and Gerrans, 2017). It is worth noting, however, that classic accounts of non-dual awareness in the Buddhist and Hindu traditions emphasize its reliance on the loss of narrative aspects of self-consciousness more than the loss of multisensory aspects. Non-dual Mahāmudrā practice, for example, instructs students to “drop thoughts of past, present and future and release the mind into its natural state of clear, non-conceptual awareness” (Dunne, 2011, p. 81). Consequently, it is possible that global states of consciousness reached through NDA meditation may preserve some awareness of bodily sensations10.

In summary, we have suggested that the notion of NDA states can be understood in two different ways. According to the first interpretation, they are conscious states in which both narrative and perhaps multisensory aspects of self-consciousness are radically disrupted, such the distinction between oneself and the external world no longer has an experiential basis. According to the second interpretation, ordinary conscious states involve a minimal form of self-awareness which is not tied to any specific content, but to a special sense of “self-givenness” or “mineness” built in normal experience; in NDA states, this feature supposedly goes missing, such that subjects become aware of consciousness itself as an ownerless process. While the second interpretation rests on a controversial assumption about normal experience, the first is consistent with the evidence presented so far regarding the phenomenology of alleged “selfless” states in meditation and drug-induced states. 0011

Although there have been few neuroimaging studies of NDA meditation, the available evidence is intriguing. In the resting state, activity in the default-mode network has been found to be negatively correlated with activity in a set of regions commonly recruited in attention-demanding tasks, in particular the sensory (Golland et al., 2007), dorsal attention network and fronto-parietal control network (Greicius et al., 2003; Fox et al., 2005; Uddin et al., 2009; Chai et al., 2012; Carbonell et al., 2014). A similar pattern of anticorrelation has been reported during mindfulness meditation (Brewer et al., 2011; Kilpatrick et al., 2011; Hasenkamp et al., 2012), and there is some preliminary evidence of increased anti-correlation between the DMN and a task-positive network during focused attention meditation compared to the resting state (Josipovic et al., 2012). However, Josipovic and colleagues found that this anticorrelation was significantly decreased during NDA meditation compared to both focused awareness meditation and the resting state (Josipovic et al., 2012; Josipovic, 2014). It is worth noting that a similar pattern of increased correlation between the DMN and the habitually anti-correlated networks has been observed during various other forms of meditation, such as mantra recitation (Berkovich-Ohana et al., 2015) as well as choiceless-awareness, loving-kindness and concentration in experienced meditators compared to novice meditators (Brewer et al., 2011). Thus, it is questionable that this change of connectivity is unique to NDA meditation. Interestingly, a similar decrease of the anticorrelation between the DMN and a task-positive network has been observed after administration of psilocybin, and was hypothesized to correlate with decreased separatedness between internally and externally focused states (Carhart-Harris et al., 2013). Although this hypothesis is speculative, it is possible that conscious states induced through NDA meditation have a lot in common with certain states of “ego dissolution” induced by psychedelics from a phenomenological point of view. Before discussing this potential convergence, we will assess how NDA may relate to so-called “pure consciousness.”

The notion of pure consciousness seems to originate in Stace (1960, p. 86), who applies it to mystical experiences. A purely conscious state is characterized by “an ‘emptying out' by a subject of all experiential content and phenomenological qualities, including concepts, thoughts, sense perception, and sensuous images” (Gellman, 2017). States which have been qualified as purely conscious include experiences reached through certain forms of meditation (Nash and Newberg, 2013) and mystical experiences (Stace, 1960; Shear, 1994).

There has been considerable debate about whether purely conscious states are even possible. Some have argued that they are impossible because one cannot be conscious without being conscious of anything at all (Bayne, 2007, p. 16) or because one cannot have episodic memory of a state devoid of experiential content (Bagger, 1999, p. 102; Gennaro, 2008). Others have argued that pure consciousness is not ruled out by logical or phenomenological considerations (Woodhouse, 1990; Shear, 1994), and that we might even find empirical evidence of pure consciousness in contemplative practices (Koch, 2004; Baars, 2013; Metzinger, 2013; Nash and Newberg, 2013) or dreamless sleep (Windt et al., 2016). Note that this apparent disagreement could be a mere verbal dispute, because there is an important ambiguity in the notion of a conscious mental state devoid of any content. Indeed, “content” might refer here to representational or phenomenal content.

Although controversial, the hypothesis that some conscious mental states lack representational content has been defended on the ground that states such as moods, pain or orgasm which do not seem to represent anything (Rey, 1998; Kind, 2007; Aydede, 2009). This is in fact the basis of a prominent objection to representationalism, the thesis that all phenomenally conscious states have representational content, or in its stronger version that the phenomenal character of all conscious mental states is wholly constituted by their representational content. But representational contentlessness is presumably not what proponents of pure consciousness have in mind, as they suggest that purely conscious states lack “all determinate phenomenological contents…whatsoever” (Shear, 1994, p. 320). Thus, the strict definition of a purely conscious state would be a conscious mental state lacking phenomenal character. However, this definition cannot be taken literally if we are talking about phenomenal consciousness: by definition, a phenomenally conscious mental state is a mental state such that there is something it like for a creature to be in it (Nagel, 1974), and if there is something it is like to be in a mental state then the mental state has phenomenal character. Therefore, the notion of a (phenomenally) conscious mental state literally lacking phenomenal character is absurd (Gennaro, 2008; Strawson, 2013): the experience of absence is not equivalent to the absence of experience.

There is, however, a more plausible, non-literal definition of pure consciousness as a conscious state lacking ordinary phenomenal content (Gennaro, 2008; Strawson, 2013). The question, then, is what counts as an ordinary phenomenal content, and how phenomenally “bare” or “sparse” purely conscious states can get. Presumably, if there are phenomenal properties pertaining to self-consciousness in ordinary experience (i.e., the narrative and somatosensory aspects of self-consciousness previously discussed), then such ordinary phenomenal properties should be missing in purely conscious states, among other ordinary phenomenal properties. Defined in this way, pure consciousness appears to be at least conceivable, and constitutes a plausible candidate for a wholly “selfless” state, lacking self-referential thoughts, bodily ownership, body awareness and self-location. The question is now whether purely conscious states can actually occur, specifically during meditation or after psychedelic intake, and whether they differ from states of alleged “non-dual awareness.”

In the Hindu and Buddhist traditions, the meditative practice called Samadhi aims at inducing a state of deep absorption (Shankman, 2008). The practice of Samadhi includes four stages or jhanas, culminating in a “formless” or “objectless” state. It relies on successive shifts of attention, first from the breath to somatosensory pleasure, then to mere positive affect, until all intentional objects have been stripped away and only “one-pointed” awareness remains. In the final stage of Samadhi practice (i.e., the last jhana), “there is no directing of attention, no representation held in focal attention, no pleasure and no aversion, affective contentment has dropped away, and all that remains is an alert, clear, one-pointed, equanimous awareness.” (Yamashiro, 2015, p. 7).

The aim of this practice is frequently described as reaching a state of pure consciousness which involves a complete loss of self-consciousness: “attaining Samadhi is to reach the silent state of pure consciousness where there is no phenomenological content and a loss of any sense of individual self or duality” (Wahbeh et al., 2018, p. 6)12. Miri Albahari has suggested that the state of pure consciousness allegedly reached through Samadhi meditation could provide evidence for the existence of what she calls “witness-consciousness,” which she defines as the neutral common denominator between all conscious experience (Albahari, 2009). She argues that witness-consciousness has its own intrinsic phenomenal character, which explains why purely conscious states in Samadhi are not devoid of phenomenal content, and thus not logically inconsistent. However, this hypothesis seems unduly inflationary: rather than postulating that a special “objectless” phenomenology is ubiquitous in consciousness and is the only thing remaining in purely conscious states, one may seek to describe such states as a form of deep absorption associated with extreme sensory gating which does not have much in common with ordinary experience.

Unfortunately, the neurophysiological evidence regarding Samadhi practice is still extremely sparse (Yamashiro, 2015). A single-participant study found decreased activity in Brodmann areas 5 and 7, which may be associated with the representation of the body's orientation in three-dimensional space (Hagerty et al., 2013). This reflected the participant's experience of losing body boundaries. Nash and Newberg (2013) also suggest that activity of the posterior parietal lobule is modulated by Samadhi, and “might be critical for distinguishing between the self and the external world” (Nash and Newberg, 2013, p. 8). Another contemplative tradition which aims at inducing pure consciousness is transcendental meditation (TM), whose goal is the “loss of boundaries of time, space, and body sense that defines the framework for typical waking experience,” but neurophysiological data on TM is also very limited (Travis and Pearson, 2000; Wahbeh et al., 2018).

It is unclear to what extent the notions of non-dual awareness and pure consciousness should be distinguished, both in terms of their conceptualization in the relevant literature on contemplative traditions, and in terms of the phenomenology and neurophysiology of the associated states induced by meditation. According to Josipovic, NDA meditation is not content-driven like focused attention and open monitoring, because it seeks awareness of the background of experience rather than any specific bodily or perceptual content (Josipovic, 2014). Moreover, he argues that NDA meditation also differs from “objectless” meditation “in which the mind is emptied of content and held in an empty state through the force of concentration” (Josipovic, 2014, p. 12)13. Indeed, Josipovic contends that pure consciousness practice like Samadhi aims at actively “eliminating” each pole of the subject/object dichotomy, while NDA meditation involves the recognition of a background non-dual awareness that “precedes conceptualization and intentionality” (Josipovic, 2014, p. 12). Likewise, John Dunne argues that non-dual awareness differs from objectless meditation techniques such as Samadhi insofar as it should not involve any cognitive effort (Dunne, 2011). However, these conceptual distinctions are not straightforward, and there is not yet enough data on the phenomenology and neurobiology of alleged states of “pure consciousness” and “non-dual awareness” to determine whether these are valid and distinct constructs.

As we have suggested, one possible construal of non-dual awareness is in terms of the inhibition of both narrative and multisensory aspects of self-consciousness, leading to a temporary loss of experiential boundary between self and world, or between endogenous and exogenous stimuli. In turn, purely conscious states could be defined as states of extreme absorption involving high sensory gating, whose sparse phenomenal content has little overlap with the rich phenomenology of ordinary wakeful experience. If these tentative definitions are sound, states of pure consciousness should also be states of non-dual awareness, because they should lack ordinary phenomenal content including the narrative and multisensory components of self-consciousness. By contrast, non-dual awareness might not entail pure consciousness, because states of total self-loss (lacking narrative and multisensory self-consciousness) need not have very sparse phenomenal content.

This raises the question of whether psychedelic drugs can induce virtually “contentless” states similar to the descriptions of pure consciousness from mystical and contemplative traditions. While most reports of drug-induced ego dissolution involve a rich sensory phenomenology, including vivid hallucinations, this is not always the case. In particular, there is anecdotal evidence that certain compounds such as 5-MeO-DMT may induce states of radical absorption reminiscent of Samadhi practice14. Indeed, users of 5-MeO-DMT frequently describe an experience of “emptiness,” “nothingness” or “void” which is associated with a cessation of thoughts, extreme sensory deprivation and a complete loss of self-consciousness: “my brain was not conveying anything meaningful from what my senses were receiving; I was completely unaware of my body, experiencing profound stillness” (report #99920 from Erowid.org); “I felt that there was nothing to me and there was nothing around me” (#11701); “the reality around me disintegrated into nothing. I fell into a void [that] I can't even describe” (#23487); “I wasn't anything anymore. I had been broken down into nothingness, into oblivion” (#87426); “my thoughts ceased to exist, and my senses shut off completely. I could not hear, see, smell, taste or feel anything” (#37301). This suggests that states of drug-induced ego dissolution may vary not only according to the extent to which narrative and multisensory aspects of self-consciousness are inhibited, but also according to the richness or sparsity of their phenomenology. While some forms of self-loss induced by psychedelics may involve a rich sensory phenomenology, others are more similar to descriptions of “pure consciousness,” and may be almost devoid of sensory content.

Accumulating evidence shows that meditation's state effects linger into daily life, to become long-term, trait alterations (Cahn and Polich, 2006). These trait effects include alterations in resting state function or connectivity, as well as structural changes, compared to control groups (Fox et al., 2014; reviewed in Tang et al., 2015). A full review of all the accumulated evidence is beyond the scope of this paper. Instead, here we focus on trait-effects of alterations in self-consciousness, either from the neurophysiological (underlying mechanisms, specifically regarding the DMN) or first-person perspective.

Several studies have provided evidence that meditation practitioners, compared to controls, exhibit reduced resting state DMN activity and connectivity, either using electrophysiology (Berkovich-Ohana et al., 2012, 2013) or fMRI (Hasenkamp and Barsalou, 2012; Berkovich-Ohana et al., 2016). Specifically, resting state functional connectivity analyses showed that following 8-weeks of mindfulness-training, participants demonstrated significantly increased functional connectivity of the anterior DMN region with the auditory/salience network (Kilpatrick et al., 2011). Similarly, long-term mindfulness practitioners showed compared to controls increased functional connectivity between DMN and visual regions (Berkovich-Ohana et al., 2016). In other studies, proficient mindfulness meditators showed increased functional connectivity between the DMN and fronto-parietal control network compared to novices (Brewer et al., 2011; Hasenkamp et al., 2012; Taylor et al., 2013), although there is some evidence of increased anti-correlation between the DMN and a task-positive network during focused attention meditation compared to the resting state (Josipovic et al., 2012), or the DMN and visual regions following MBSR (Kilpatrick et al., 2011). While fMRI studies largely show that mindfulness meditation practice is associated with reduced DMN activity, the related structural effects in the DMN are less clear (recently reviewed by Fox et al., 2014). Specifically, several neuroanatomical studies reported changes in PCC gray matter thickness: while a few studies indicate reduction in meditators relative to controls (Kang et al., 2013), another study failed to find group differences (Grant et al., 2013), and yet another reported gray matter increases following a short 8-weeks mindfulness meditation intervention (Hölzel et al., 2011). Initial evidence suggests that such changes in DMN functioning are indeed related to a reduced tendency to engage in self-referential processing. One earlier study (Farb et al., 2007) showed that after an 8-week mindfulness intervention, participants were able to disengage from a “narrative self-focus” in the trait judgment task (typically employed to investigate self-referential processes, cf. Northoff et al., 2006), as evidenced by more pronounced reductions in the DMN compared to the control group. Similarly, another study found reduced bias in neural responses to the self vs. an other's face for long-term meditators compared to controls (Trautwein et al., 2016).

Less is known about whether meditation practice affects multisensory aspects of self-consciousness in a trait-like manner. Studies investigating neural and phenomenological state effects of meditation suggest that long-term meditators are capable of flexibly modulating and reducing bodily self-awareness (Ataria et al., 2015; Dor-Ziderman et al., 2016). However, controlled studies have not yet systematically assessed whether this is an outcome of long-term practice (some effects also seem to occur in novice practitioners, cf. Dambrun, 2016) and to which degree such flexibility affects daily life functioning.

There is no evidence that the experience of drug-induced ego dissolution may have long-term effects on narrative and multisensory aspects of self-consciousness, such as a reduction of self-related thoughts. This question may be answered by future longitudinal studies.

While the long-term consequences of ego dissolution on self-consciousness are uncertain, there is preliminary evidence that drug-induced alterations of self-consciousness may mediate therapeutic outcomes. The current model of psychedelic-assisted therapy originated in the 50's and 60's (Pahnke et al., 1971; Dyck, 2006). In this model, subjects receive a high dose of a psychedelic drug in a supportive environment, with the aim of experiencing a mystical-type or “peak” experience (Maslow, 1959; Stace, 1960). This experience is conceived as a “unitive” state involving the loss of self-world boundaries, as described in sections Alterations of self-consciousness induced by meditation and psychedelics and Pure consciousness, non-dual awareness and total selflessness. The therapeutic model suggests that this experience mediates long-term outcomes (Pahnke et al., 1971; Grof et al., 2008; Majić et al., 2015; Richards, 2015). In recent years, many controlled studies have shown that that the magnitude of the mystical-type experience predicts positive psychological outcomes for depression, addiction, palliative care, and general well-being (O'Reilly and Funk, 1964; Klavetter and Mogar, 1967; Pahnke et al., 1971; Kurland and Grof, 1972; Richards et al., 1977; Griffiths et al., 2006, 2016; MacLean et al., 2011; Garcia-Romeu et al., 2014; Bogenschutz et al., 2015; Ross et al., 2016; Johnson et al., 2017; Roseman et al., 2018). Furthermore, a recent analysis of the psychedelic experience in uncontrolled environments suggests that mystical-type experiences predict changes in well-being (Carhart-Harris et al., 2018). It is also suggested that mystical-type experiences are linked to the positive emotional outcomes of meditation (Russ and Elliott, 2017).

Two variables appear to predict the occurrence of mystical-type experience in both meditation and psychedelics: trait absorption and surrender state. Absorption (Tellegen and Atkinson, 1974) is a trait characterized by the disposition to have episodes of “total” attention to one's representational resources. Subjects with higher absorption tend to have stronger mystical-type experiences both under psychedelics (Studerus et al., 2012) and in meditation (Russ and Elliott, 2017). The second variable is the pre-experience state characterized by the disposition to “let-go” or “surrender” to whatever experience comes, sometimes called “surrender state” (Richards, 2015). This state is not only related to the subject's personality but also to interactions with the environment (e.g., trust toward the therapist). Higher ratings of willingness to surrender are associated with stronger mystical-type experience in both psychedelic experiences (Carhart-Harris et al., 2018) and meditation (Russ and Elliott, 2017).

Do experiences of “self-loss” induced by meditation or psychedelics have a relevance for therapeutic outcomes or general well-being? There is preliminary evidence that the therapeutic effects of psychedelics for treatment-resistant depression is mediated by meaningful “breakthrough” experiences, which may include ego dissolution (Roseman et al., 2018). By contrast, long-term effects of contemplative practices on well-being do not appear to be necessarily mediated by intense experiences, but rather by training of different cognitive mechanisms, such as attentional control mediating meta-awareness of mind-wandering. Interestingly, one suggested mechanism is a (gradual) shift in the perspective on the self-described as “detachment from identification with a static sense of self” (Hölzel et al., 2011). In the Buddhist tradition, this process is thought to be a crucial factor for the attainment of stable well-being (Olendzki, 2006; Shiah, 2016). Some psychometric research supports this link (Dambrun, 2017). Moreover, a related construct called “decentering” has been found to mediate treatment effects of mindfulness based interventions (reviewed in Bernstein et al., 2015). Anecdotal reports and traditional sources do also claim that peak meditative experiences involving a loss of the sense of self can have lasting effects on well-being (Austin, 2009); however, perhaps due to the difficulty to induce such experiences in the lab, they have not been systematically investigated. In a first qualitative survey on experiences associated with long-term meditative practice, changes in the sense of self, including narrative and bodily levels were frequently reported (more than 75%) (Lindahl et al., 2017). Of note, the affective response to these experiences was not always positive, but ranged “from neutral curiosity, to bliss and joy, to fear and terror” (p. 20). Interestingly, ayahuasca administration was also found to produce after 24 h a significant reduction in judgmental processing of experiences measured by the Five Facets Mindfulness Questionnaire (FFMQ), as well as a significant increase in decentering ability measured by the Experiences Questionnaire (EQ) (Soler et al., 2016). Thus, it is possible that the medium- to long-term effects of meditation and psychedelic experiences involve a reduced tendency to be personally and emotionally engaged in one's thoughts and feelings, such that “the contents of consciousness are less filtered through considerations of self-relevance than is usual” (Letheby and Gerrans, 2017, p. 7).

A further question worth investigating in future research regards the specific kind of “self-loss,” if any, which may mediate long-term therapeutic outcome and increased well-being. For example, are such effects more likely to be mediated by the loss of narrative or multisensory aspects of self-consciousness? Given the potential link between mind-wandering and unhappiness (Killingsworth and Gilbert, 2010), it is intriguing to speculate that increased control of spontaneous thoughts might mediate increased well-being in experienced meditators, which would indicate that the disruption of narrative aspects of self-consciousness may have a positive effect, to some extent. At this stage, however, this speculative hypothesis remains unsubstantiated by significant evidence.

Besides individual well-being, one potential outcome of meditative practice is an increase in empathy and compassion, which are regarded as antecedents of prosocial behavior. There is indeed evidence that meditation fosters these aspects of prosociality (Condon et al., 2013; Ashar et al., 2016; see Kreplin et al., 2018 for a recent meta-analysis; Leiberg et al., 2011; Weng et al., 2013). Importantly, it has been hypothesized that these effects rely on a change in the sense of self. For example, Dambrun and Ricard (2011) suggested that meditation can support a shift from self-centered to selfless functioning, characterized by “a weak distinction between self and others, and self and the environment as a whole,” which in turn “is closely related to characteristics such as altruism, kindness, respect, empathy, compassion and the search for harmony” (Dambrun and Ricard, 2011, p. 140). Providing evidence for one part of this relationship, recent studies demonstrated that manipulation of bodily levels of self-awareness can affect empathy and social cognition (Maister et al., 2013, 2015). However, the link to meditation is still largely hypothetical, although one study found that reduced self-related processing in long-term meditators was correlated with increased trait levels of self-reported compassion (Trautwein et al., 2016). Intriguingly, the administration of a high dose of psilocybin 1 or 2 months after a program of meditation practice was found to occasion enduring trait-level increases in prosocial behavior (Griffiths et al., 2018). This finding raises the question of whether psychedelic intake may potentiate in certain cases the putative long-term effects of meditation on prosociality. In summary, it is possible that there is a link between “selflessness” as a conscious episode lacking self-consciousness and trait increase in prosocial behavior, but these two constructs should not be conflated. Future research could determine whether such an association exists for meditation training and perhaps psychedelic use, and get clearer on the mechanisms involved.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer MC and handling Editor declared their shared affiliation.