Neural Mechanisms and Psychology of Psychedelic Ego Dissolution

1. Thalamic Signal Filtration

The early understanding of psychedelic mechanisms—as described in filtration theory—suggests that consciousness is winnowed by selection processes that exclude subconscious material (Marshall, 2005; Swanson, 2018). The thalamus is thought to serve such a function by filtering sensory information reaching the cortex. The thalamus is an information filter and relay hub located in the center of the brain, with projections in all directions, responsible for conveying signals between the brain and body (Torrico and Munakomi, 2019). The thalamus is also involved in the regulation of consciousness level, including sleep, wakefulness, alertness, and sensorimotor arousal (Torrico and Munakomi, 2019). It shares connectivity profiles with limbic structures (Steriade and Llinas, 1988; Stein et al., 2000; Voets et al., 2015) and is implicated in medial temporal lobe epilepsy (Barron et al., 2014), which produces psychedelic-like sensory illusions and déjà vu experiences. Disruptions to the thalamus—within the cortico-striato-thalamo-cortical loop—speak to the role of the thalamus in schizophrenia (Behrendt and Young, 2004; Halberstadt, 2015), sometimes referred to as a cognitive dysmetria. Certain portions of the thalamus (e.g., the pulvinar) have been associated with setting the precision of cortical message passing, of the sort implicated in selective attention and feature ground separation in the visual system (Kanai et al., 2015).

Commonalities—between the positive symptoms of schizophrenia and hallucinations produced under psychedelics—suggest that the cortico-striato-thalamo-cortical circuitry may be involved in hallucinations and mediate increased bottom-up connectivity from the thalamus to the cortex (Geyer and Vollenweider, 2008; Tylš et al., 2016). Examinations of functional connectivity (FC), under LSD, confirm an increase in correlations between the thalamus and cortical hubs (Tagliazucchi et al., 2016; Müller et al., 2018). FC analysis under LSD also found increases in correlated activity between the thalamus and insula and fusiform gyrus, thought to be related to the perceptual effects of psychedelics (Müller et al., 2017). Functional connectivity refers to correlations between neuronal activity in different systems that is mediated by effective connectivity, namely the directed influence that one neuronal system exerts over another.

Dynamic causal modeling (DCM) for resting-state fMRI (Friston et al., 2014; Razi et al., 2015; Razi et al., 2017) quantified the effective connectivity of the thalamus connectivity under LSD. The results showed increased effective connectivity from the thalamus to the posterior cingulate cortex (PCC) and decreased top-down connectivity from the PCC to the thalamus (Preller et al., 2019). These findings support the general notion of psychedelic-enabled increase in bottom-up effects on the cortex (Vollenweider and Preller, 2020). This is consistent with a relaxation of prior precision in high levels of cortical hierarchies, to which the thalamus projects. In short, the thalamus appears to play a key role in enabling ascending or bottom-up connectivity under psychedelics by opening a “thalamic filter” (Fig. 4). Further review of thalamic connectivity changes under psychedelics is featured in Unifying Neuroimaging Evidence of Psychedelic Mechanisms.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4

Opening the thalamic filter under psychedelics. Flatheads represent top-down inhibition of bottom-up signals, and arrowheads represent uninhibited signals. Reduced top-down inhibition from the cortex enables increased bottom-up connectivity to the cortex.

1. Medial Temporal Lobe

The MTL regions involve limbic structures that are associated with emotion and memory, which may relate to the therapeutic effects of psychedelics (Carhart-Harris, 2007; Ritchey et al., 2019). Cortical midline structures such as the PCC, anterior cingulate cortex (ACC), and medial prefrontal cortex (mPFC) are connected to MTL structures, including the parahippocampus (PHC), retrosplenial cortex (RSC), and amygdala (AMG) (Milad and Quirk, 2002; Etkin et al., 2011; Andrews-Hanna et al., 2014). Like the thalamus, increased signaling from the MTL following psychedelics is thought to affect the influence of MLT on higher associative areas and subsequently conscious awareness (Carhart-Harris et al., 2014a). Abnormal MTL activity can produce dreamy states and a sense of depersonalization, resembling psychedelic subjective effects (Halgren et al., 1978; Lemche et al., 2016). This suggests that psychedelic-induced alterations of the MTL may produce similar effects. However, the MTL’s importance in memory and emotion point to its connectivity with other cerebellar and cortical regions that may play a role in acute and lasting therapeutic effects. Moreover, decoupling between cortical regions and the MTL has been suggested to play a role in ego dissolution (Lebedev et al., 2015). Understanding connectivity changes of the MTL under psychedelics may help discern its contribution to subjective and therapeutic effects. Partitioning of the MTL into its constituent systems, including the AMG, PHC and RSC, provides a more granular view of their role under psychedelics.

2. Amygdala

The AMG is involved in perception and emotional processing (Phelps and LeDoux, 2005; Kraehenmann et al., 2015a) and identifying emotionally salient stimuli. Activity of the AMG can be summarized as eliciting experiences of heightened emotional arousal mediated by top-down connectivity. The AMG comprises substructures with distinct functions and connectivity to various brain regions. Although few psychedelic studies report the change to substructure connectivity of the AMG—likely due to the challenges of accurately locating and aligning group connectivity changes—investigating the differential roles of amygdala substructures under psychedelics for their role in emotion and therapeutic outcomes is crucial to illustrate the neural mechanisms of psychedelics accurately. For example, the basolateral amygdala is the largest substructure and projects to layer V pyramidal neurons (Marek, 2018). An increase in basolateral AMG pyramidal and stellate neurons is implicated in chronic stress (Lu et al., 2021), and its projections to the cortex are cited in preclinical research of addiction (Roura-Martínez et al., 2020). Its connectivity with the prefrontal cortex is also important to extinction learning (Quirk and Mueller, 2008; Herry et al., 2010; Milad and Quirk, 2012; Baldi et al., 2021), and dysfunction of this connectivity marks fear disorders such as PTSD (Rauch et al., 2006; Shin et al., 2006). Under psilocybin, the basolateral AMG demonstrates an elevated expression level of c-Fos (a marker of neural activity) (Davoudian et al., 2022) and is implicated in mice under nonclassic psychedelics such as 3,4-methylenedioxymethamphetamine (Glavonic et al., 2022). However, destruction of the basolateral AMG had a negligible effect on prefrontal 5-HT cortical excitation in preclinical research. The negligible effect suggests that the AMG is not a bottom-up pathway to prefrontal serotonergic excitation in the rat brain (Marek et al., 2001). Future psychedelic work may also wish to investigate the differential roles of amygdala substructures under psychedelics for their role in emotion and therapeutic outcomes.

Whole AMG analyses indicate top-down AMG connectivity may play a role in mediating visual salience. This emotional processing circuitry can be sensitive to the effects of psychedelics. The AMG receives top-down projections from midline cortical regions, like the ACC with indirect connections to the mPFC (Etkin et al., 2011; Robinson et al., 2013; Kraehenmann et al., 2015b). For example, top-down connectivity between the ventral ACC and AMG influences emotional regulation and reappraisal (Etkin et al., 2011). AMG circuitry is thought to respond to threats to both beliefs and physical safety (Kaplan et al., 2016). This implicates the AMG in ego resistance and defenses that may require resolution to produce ego dissolution experiences. However, the AMG’s association with anxiety and fear should not overshadow its more general role in the regulation of emotions (Price et al., 2009) and detection of emotional salience (Adolphs, 2010; Santos et al., 2011). AMG activation does not determine the emotional valence of subjective experiences. For example, increased AMG activation was found after treating (treatment-resistant) depression with psilocybin (Roseman et al., 2018a). This effect was attributed to acceptance and emotional reconnection reported by patients (Watts et al., 2017; Roseman et al., 2018a). Conversely, another psilocybin study associated positive mood change in healthy subjects, with decreased AMG response measured acutely (Kraehenmann et al., 2015a). This suggests that AMG deactivation was associated with a reduction in negative emotion and aligns with other psilocybin investigations showing a reduction of cerebral blood flow in the AMG that correlates with a reduction of depressive symptoms (Carhart-Harris et al., 2017).

Such contradictory associations between AMG activation and behavioral outcomes are evident throughout imaging literature. The reduced structural volume of the AMG in meditators is thought to represent decreased negative affect (Gotink et al., 2018), whereas the same finding in borderline personality disorder is believed to indicate decreased emotional regulation (Weniger et al., 2009). These findings suggest that the structural volume and activation of the AMG in isolation do not disambiguate the quality of psychologic experience. More accurate associations between the AMG and behavior require attention to the subject’s experience during measurement and the time when AMG activity is assessed: the AMG appears to deactivate under acute psychedelic effects in response to negative emotional stimuli in participants in supportive conditions. Increased activation of the AMG, however, appears in the period following clearance of the drug.

Understanding top-down connectivity may help disambiguate the changes in AMG relating to psychologic function and therapeutic outcomes. Top-down connectivity generally has an inhibitory influence on bottom-up projections. Disruption caused by psychedelics to top-down connectivity is seen clearly in changes in the visual-limbic-prefrontal network involved in detecting visual threat cues. This circuitry is thought to evince increased top-down influences in pathologic conditions (Disner et al., 2011). Under acute effects of psilocybin, reduced top-down threat processing in healthy adults is demonstrated by decreased effective connectivity from the AMG to the primary visual cortex (Kraehenmann et al., 2015b). Cortical connectivity to the AMG also mediates emotional threat processing, and LSD reduces AMG responses to fearful faces in healthy adults under acute effects (Mueller et al., 2017).

As previously mentioned, an opposite pattern of increased AMG activity is seen using similar task conditions 1 day after psilocybin in the (right) AMG (Roseman et al., 2018a), whereas individuals with borderline personality disorder—which is characterized by over-reactive responses to threat—show hyper-AMG activation to emotional face tasks (Donegan et al., 2003). This signifies the important role of top-down connectivity when interpreting AMG activity. Follow-up research measuring psilocybin effects 1 day after intake suggests that AMG activity increases are accompanied by decreased connectivity between the prefrontal cortex and the AMG (Mertens et al., 2020). Although this research did not determine the direction of influence, the authors propose connectivity involving the AMG may be an important mechanism underlying previously reported therapeutic outcomes of emotional reconnection (Watts et al., 2017).

AMG and affect responses to negative faces have also been measured up to 1 month postpsilocybin in healthy participants. This research shows sustained reduction of AMG responses to negative faces lasting up to 1 month. Moreover, this research indicated that increased positive responses to emotionally conflicting (resistance arousing) stimuli, lasting up to 1 week, were associated with cortical projections to the dorsolateral prefrontal and medial orbitofrontal cortex (Barrett et al., 2020a). Positive responses to emotionally conflicting (resistance arousing) stimuli—involving the cortex and changes to the AMG—are evidence that psychedelics can have enduring effects on neuroplasticity, lasting well beyond the half-life of psilocybin. These results also speak to the importance of emotional regulation of the AMG by the cortex in long-lasting therapeutic outcomes. Reduced acute activation of the AMG by negative stimuli, under psychedelics, may contribute to reduced ego resistance and positively felt ego dissolution.

3. Parahippocampus and Hippocampus

The PHC is an associative hub of the MTL involved in memory retrieval, associative memory, contextual processing and detection of familiarity (Aminoff et al., 2013). The hippocampus (HC) is involved in decision making, learning, and memory consolidation and can be further divided into the dentate gyrus, the hippocampus proper, and the subiculum (Anand and Dhikav, 2012; Fogwe et al., 2021). The PHC is connected with the HC and may mediate HC connections with the associative cortex (Ward et al., 2014; Millière, 2017). This connectivity is relevant in limbic-cortical dysregulation impacting mood regulation (Zeng et al., 2012), and increased FC between the HC and prefrontal cortex (e.g., mPFC) is seen in depression (Kaiser et al., 2015). Decreased resting-state FC between the PHC and prefrontal cortex was identified under psilocybin and was predictive of treatment response in a study of psilocybin for treatment-resistant depression (Carhart-Harris et al., 2017). Moreover, these connectivity changes also predicted ego dissolution (Carhart-Harris et al., 2017). Dysconnectivity of the PHC is also implicated in medial temporal lobe epilepsy and déjà vu experiences (Illman et al., 2012), suggesting that, under psychedelics, PHC activity in the detection of familiarity may contribute to disruption of the narrative self (discussed later). Following this notion, information theoretical analysis of fMRI brain connectivity identified that reduced connectivity (as measured by lower diversity coefficient) between the anterior PHC and cortical regions increased the likelihood of experiencing ego dissolution under psilocybin (Lebedev et al., 2015). Speculatively, reduced connectivity may sensitize neuronal message-passing to the effects of 5-HT2AR activation.

Music also influences PHC connectivity with cortical regions under psychedelics and may relate to PHC function in associative memory and contextual processing. A study using DCM for fMRI data reported increased connectivity of the PHC to visual regions and attenuated top-down connectivity under LSD (Kaelen et al., 2016). This suggests that psychedelic-induced dissolution of hierarchical processing may decouple the PHC and the visual system (Kaelen et al., 2016). It also suggests free interactions lower in the cortical hierarchy, labeled anarchy (Carhart-Harris et al., 2014b). Moreover, hierarchical decoupling identified by reduced hippocampal glutamate levels under psilocybin was associated with positively felt ego dissolution (Mason et al., 2020). Although it remains uncertain whether reduced glutamate levels are a cause or consequence of ego dissolution, some authors propose that the decoupling of hippocampal connectivity may inhibit semantic autobiographical memory from reaching the cortex (Millière et al., 2018).

Oscillatory changes are also seen in psychedelic states. Cortical oscillations are rhythmic electrical activity or waves that are produced by interactions among neurons. Synchronous patterns of neuronal firing are commonly associated with states of arousal and sleep (Poulet and Crochet, 2019). EEG recordings under psilocybin show decreased PHC oscillations—also observed in coherence between the ACC and PCC (Kometer et al., 2015). These findings suggest that changes to cortical FC and the PHC contribute to the psychedelic experience. The PHC also contributes to spatial navigation functions of the RSC (Epstein, 2008; Vann et al., 2009).

4. Retrosplenial Cortex

The RSC supports a sense of head orientation and cognitive capacities such as visualization, autobiographical memory, future-oriented thinking, retrieval of memory, and navigation. The RSC’s role in navigation enables awareness of self-location, relative to spatial context, and the translation of spatial context to cognitive maps, which are independent of self-reference (Vann et al., 2009). The role of the RSC in generating reference-independent self-orientation has also been suggested to extend to representations of permanence, both of environmental features (Auger et al., 2012) and representations of permanence more generally (Auger and Maguire, 2018; Kim and Maguire, 2018). These representations may underwrite a sense of self, time, and space that are affected by ego dissolution. A study of LSD, integrating fMRI, MEG, and arterial spin labeling–MRI, showed decreased communication between the RSC and PHC, which correlated with subjective measures of ego dissolution and the maintenance of a sense of self (Carhart-Harris et al., 2016b). Desynchronization of δ-band oscillations, measured between the RSC, PHC, and lateral orbitofrontal areas, was also seen under psilocybin: this desynchronization was associated with spiritual experience and insight, which may be considered aspects of ego dissolution (Kometer et al., 2015). It was conjectured that the implicit effects on connectivity may facilitate insight-enhanced reprocessing of autobiographical memories (Kometer et al., 2015). Conversely, memory retrieval tasks under psilocybin do not support this supposition (Carhart-Harris et al., 2012a). However, it is worth noting that the memory retrieval studies included only a few subjects (n = 10) and did find lasting therapeutic effects of psilocybin (Carhart-Harris et al., 2012a). Despite the absence of clear evidence for the functional correlates of psychedelic effects on the RSC, its role in spatial orientation and representation of permanence suggests a role in ego dissolution.

5. Subcortical Summary

Evidence relating hierarchical connectivity changes between subcortical limbic regions and the cortex under psychedelics is consistent across studies and may relate to ego dissolution and therapeutic effects. However, further investigations are needed to identify the specific effects on directed (effective) connectivity and their functional role in sentience and ego dissolution. The effects on thalamic connectivity appear to be an increase in bottom-up connectivity, which may facilitate hallucinations/hallucinosis through reduced sensory filtration. The role of the thalamus in ego dissolution and the cognitive aspects of psychedelic experiences is uncertain and may be indirect. The connectivity of the AMG with cortical and visual regions speaks to an effect of psychedelics on emotional regulation. This connectivity also appears to be rendered increasingly plastic by psychedelics. In positively felt ego dissolution, connectivity between the AMG and cortical regions may facilitate reduced threat responses (defenses) and ameliorate ego resistance required for therapeutic outcomes. The effect of psychedelics on associative processes in the PHC may have a role in ego dissolution and mediate therapeutic effects through routing of neuronal signals between limbic structures and the cortex. Psychedelics may also decouple the PHC from regions lower in the hierarchy due to reduced top-down inhibition from higher levels of the cortex. The ensuing flexibility, or anarchy, may also account for increased context sensitivity experienced by subjects under psychedelics. A role of the RSC in the dissolution of ego permanence is conjectured. The RSC is related to sense of permanence and may mediate some of the therapeutic effects of psychedelics.

Changes in subcortical connectivity, with higher cortical areas, indicate the importance of hierarchical message-passing under psychedelics. Cortical regions support high-level processes, including self-related thinking, salience, and attention, and are responsible for top-down connectivity, which influences bottom-up projections from subcortical regions. The major cortical networks—that feature in the psychedelic imaging literature—are considered next.

1. The Default Mode Network

The default mode network (DMN) is a brain network that shows increased endogenous fluctuations when an individual is awake but not engaged in a task requiring attention (Buckner et al., 2008). It was originally identified using PET and fMRI, which showed a decrease in activity from baseline—in certain brain regions—during goal-directed behavior. This decreases regional activity in the absence of attentional- and stimulus-bound processing (Raichle et al., 2001). Advances in our understanding of the role of the DMN rest on its connectivity to other parts of the association cortex (Hagmann et al., 2008; Bonnelle et al., 2012). DMN connectivity is believed to emerge over the course of human evolution and at a developmental timescale over the lifespan (Van Essen and Dierker, 2007; Supekar et al., 2010).

The free energy principle, and associated Bayesian brain hypotheses, rests on the notion that the brain is trying to fit or invert an internal model that generates the sensory consequences of hidden or latent states in the world. On this view, the brain embodies a hierarchical generative model of how sensations are generated. In terms of evolution, hierarchical brain organization is thought to reflect recently evolved high-level associative functions that have a predictive, supervisory and directive role over subordinate primal regions (Holmes and Nolte, 2019). The notion of hierarchy suggests an increased complexity of representations from sensory to associative regions of the brain (Felleman and Van Essen, 1991). Furthermore, complex sensory events involving longer sequences of representations are processed in superordinate levels of the hierarchy capable of generating ordinal sequences and narratives (Kiebel et al., 2008). In predictive coding formulations of the Bayesian brain, the highest levels of the hierarchy are thought to generate top-down predictions of sensory information. The DMN is believed to be situated at the top of the cortical hierarchy and may control bottom-up (prediction error) signals from lower levels of the hierarchy (e.g., the MTL) (Andrews-Hanna et al., 2010; Menon, 2011). The cycle of updating top-down predictions based on the flow of bottom-up prediction errors is referred to as hierarchical predictive coding (Rao and Ballard, 1999; Kanai et al., 2015; Friston, 2018).

2. Default Mode Network Function and Priors (Beliefs)

The function of the DMN centers on adaptive and internally oriented mental processes that allow humans to be aware of environmental threats, such as predators, in the absence of goal-directed attention (Raichle et al., 2001). The DMN’s role in directing attention inward—that manifests as self-oriented thoughts—also suggests its relationship to personal identity and the construct of narrative self, which is closely related to the Freudian concept of ego (Buckner et al., 2008; Carhart-Harris and Friston, 2010; Lebedev et al., 2015). In Freud’s theory, the ego regulates behavior by suppressing the free energy of underlying subordinate structures, such as the limbic and paralimbic systems (Cieri and Esposito, 2019). The DMN similarly attenuates free energy and regulates behavior by controlling the selection of information channels.

The DMN is rich in 5-HT2AR (Beliveau et al., 2017). Mounting evidence suggests that psychedelics decrease DMN connectivity; see Table 1, which provides a summary of connectivity changes under various psychoactive compounds (Ruban and Kolodziej, 2018). Acute decrease in DMN connectivity is observed under both psilocybin and LSD (Carhart-Harris et al., 2012b; Müller et al., 2018). Decreased DMN connectivity likely relates to ego dissolution and may involve the dissolution of self-related priors (beliefs) sustained by the DMN. These priors may support a narrative sense of self that disintegrates during ego dissolution. Self-related priors in the DMN are also thought to control the top-down inhibition of prediction errors from subordinate hierarchical levels (Dillon and Pizzagalli, 2007; Carhart-Harris and Friston, 2010; Carhart-Harris et al., 2014a). If top-down inhibitory mechanisms are disengaged by psychedelics, then DMN regions may become more receptive to bottom-up influences (e.g., from the MTL) that enable revision and updating of beliefs (Carhart-Harris and Friston, 2019).

3. Default Mode Network and Priors in Mental Health

Maladaptive priors are surmised to be part of ego resistance. Disinhibition of ascending prediction errors by psychedelics may reduce ego resistance and play a permissive and foundational role in the updating of priors. Reported psychedelic therapeutic outcomes and increased communication between the subcortex and cortex is reflected in their role in the cognitive control of emotion (Ochsner and Gross, 2005). The interaction between cognition and emotion may help disarm maladaptive ego defenses. In the context of psychedelic therapy, supportive preparation and prior application of behavioral therapies may also disarm maladaptive ego defenses and help guide positively felt ego dissolution and facilitate cognitive-emotional engagement (Carhart-Harris et al., 2018a), emotional reconnection (Watts et al., 2017), and a sense of being attuned with one’s emotions (Roseman et al., 2018a).

The importance of DMN connectivity in mental well-being is evidenced in psychopathology research. For example, connectivity variability between the mPFC and PCC is greater in depressed individuals than in healthy controls (Wise et al., 2017), and increased connectivity of the DMN is associated with ruminative thought, low mood, and depression (Speth et al., 2016). Furthermore, reduced task-related mPFC suppression and increased functional connectivity of the DMN has been identified in schizophrenia (Whitfield-Gabrieli et al., 2009; Whitfield-Gabrieli and Ford, 2012), and decreased anterior and increased posterior DMN connectivity is reported in addiction (Zhang and Volkow, 2019).

These reports speak to the importance of DMN connectivity for psychologic well-being. For example, decreases in DMN functional connectivity is related to reduced references to the past by subjects post-LSD in task-based studies, suggesting that the influence of psychedelics on DMN function may relate to rumination (Speth et al., 2016). This could mean that psychedelics—similar to meditation—cultivate a more present-centered focus (Brewer et al., 2011). Moreover, this present-centered focus may represent reduced defense mechanisms.

Defense mechanisms associated with psychologic threats may have evolutionary origins in DMN monitoring for physical threats. To minimize surprise, the DMN remains aware of the environment and can mentalize temporally extended sequences of events (Andrews-Hanna et al., 2014). The DMN also responds to nonphysical forms of threat. The DMN displayed activity increases in subjects whose political beliefs have been challenged (Kaplan et al., 2016). These findings suggest that DMN mentalizing functions predict the consequences of threats to beliefs like physical threats. Defense mechanisms may have evolved to mitigate psychosocial hazards and minimize surprise. However, defense mechanisms also mark instability and suggest the difficulty of the ego to effectively estimate the value of immediate relative to temporally delayed action or gratification (i.e., regulate behavior) in a dynamic environment (Dayan and Daw, 2008; Dohmatob et al., 2020). At their extreme, defense mechanisms may form maladaptive beliefs that underwrite psychopathologies.

Investigation of the brain regions that form the DMN and their relationship with self-related processes may help discern mechanisms of action of psychedelics that mediate ego dissolution and therapeutic outcomes.

4. Posterior Cingulate Cortex

Within the DMN, a cardinal structure associated with the self is the PCC. The PCC absorbs around 20% more metabolic energy than most other brain regions (Raichle et al., 2001; Buckner et al., 2008). Its prominent role may be reflected in its structurally fortified location and access to ample blood supply (Raichle et al., 2001). The PCC remains active, automatically and continuously alert to surroundings (Raichle et al., 2001), and there is robust evidence of PCC involvement in processes of self-reflection (Vogt and Laureys, 2005) and a sense of self (Sampedro et al., 2017). This suggests that its activity underwrites a narrative sense of self (the ego) (Carhart-Harris and Friston, 2010; Carhart-Harris et al., 2012b). The PCC is also a hub that routes and regulates neuronal message-passing throughout the brain and has been shown to be abnormal in schizophrenic patients (Liang et al., 2020).

The importance of the PCC in the preservation of ego was suggested by an fMRI study of regional decreases of cerebral blood flow in the PCC (by up to 20%) that were correlated with ego dissolution induced by psilocybin (Carhart-Harris et al., 2012b). The relationship between the PCC and ego dissolution was further characterized by effective connectivity analysis of MEG data using DCM. DCM analysis demonstrated that psychedelics desynchronized cortical oscillatory rhythms of the PCC by decreasing α power (Muthukumaraswamy et al., 2013; Carhart-Harris et al., 2016b). This finding suggested that the α (8–13 Hz) frequency band of the PCC may be a correlate of ego integrity (Carhart-Harris et al., 2014)). This observation was reinforced by later findings, from fMRI under LSD, which also detected reduced α power in the PCC in association with ego dissolution (Carhart-Harris et al., 2016b).

PCC functionality is differentiated by its ventral and dorsal segments. The ventral PCC shares connections with the DMN during internally focused states and is thought to orient attention toward the self, whereas the dorsal PCC—a node within multiple resting-state networks—is involved in the dynamic coordination of attentional focus between internal and external thoughts by altering the metastability of resting-state networks (Leech and Sharp, 2014; Preller et al., 2019). The connectivity of the dorsal and ventral PCC with other brain regions—in conjunction with evidence of its alteration by psychedelics—suggests that it may be important to track these pathways to understand the mechanisms of psychedelic effects. Evidence of connectivity between the PCC and another cardinal node of the DMN, the mPFC, has also been highlighted in accounts of ego dissolution.

5. Connectivity to the Medial Prefrontal Cortex

The mPFC is implicated in executive cognitive functions, involving memory and decision making (Euston et al., 2012), and contextual associations between memory and responses by relating times, places, and events to adaptive emotional and physical responses (Euston et al., 2012). The mPFC integrates bottom-up internal and external information, conveyed by connectivity to the PCC, which may play a self-reflective role in decision making (Gusnard et al., 2001; Northoff et al., 2006; Brewer et al., 2011). Relatedly, functional connectivity between the PCC and mPFC is identified in postdecision dissonance (Tompson et al., 2016). The mPFC also has a role in selecting which bottom-up signals reach the DMN, which points to its significance in emotional processing (Gusnard et al., 2001; Ochsner and Gross, 2005). Under LSD, evidence of mPFC emotional mediation is demonstrated by reduced reactivity of right mPFC and left AMG functional connectivity in response to fearful faces (Mueller et al., 2017). Neuroimaging studies support the role of the mPFC in the suppression of limbic signals (e.g., in posttraumatic stress disorder) (Davidson et al., 2000; Hariri et al., 2000; Milad and Quirk, 2002; Phillips et al., 2003; Shin et al., 2004; Etkin et al., 2006; Hopper et al., 2007; Carhart-Harris and Friston, 2010). In short, changes in the mPFC facilitation of bottom-up signals, under psychedelics, may be important in ego dissolution and therapeutic outcomes.

Mindfulness research has identified reduced mPFC and PCC connectivity in meditators following training compared with control subjects (Brewer et al., 2011; Simon and Engström, 2015). Similarities between psychedelic experiences and meditation practice provides a reference for psychedelic-induced connectivity changes and therapeutic outcomes, and moreover suggest a potential synergy. Smigielski et al. (2019b) report a study in which experienced meditators practiced an open-awareness style of meditation under a single dose of psilocybin during a 5-day mindfulness retreat (Smigielski et al., 2019b). These meditators then underwent fMRI imaging 1 day after the retreat. Compared with controls, meditators exhibited reduced mPFC-PCC connectivity that correlated with ego dissolution. Moreover, lasting therapeutic outcomes were evident 4 months after administration. These results are significant because, despite being trained with an average 5000+ hours, the treatment group showed lasting psychologic gains beyond the benefits of extensive meditation alone (Smigielski et al., 2019a,b). Interestingly, during the follow-up fMRI, meditators in the psilocybin group were able to better modulate experiences of self-transcendence through their meditation associated with the mPFC-PCC decoupling (Smigielski et al., 2019b), suggesting psychedelic-induced neural plasticity in these projections. This study also noted reduced connectivity between the mPFC and angular gyrus, another node of the DMN, that was associated with a loosening of self-reference (Smigielski et al., 2019b). Increased self-transcendence has also been associated with cortical thinning of the PCC in regular ayahuasca users (Bouso et al., 2015). These research findings again highlight that psychedelics may facilitate lasting changes—in this case, self-transcendence through mPFC-PCC connectivity under meditation—that may also facilitate longer-term psychologic well-being.

An earlier fMRI psilocybin study found decreased functional connectivity between the PCC and mPFC. The link was interpreted as a reduced PCC influence on the mPFC and was related to ego dissolution (Carhart-Harris et al., 2012b). However, this assertion was not supported by an effective connectivity analysis of the directed influence. Another study using ayahuasca failed to find significant decoupling between PCC and mPFC using fMRI. Instead, and similar to the research of mindfulness meditators (Brewer et al., 2011), the PCC and mPFC showed reduced self-connectivity (Palhano-Fontes et al., 2015). The deactivation of mPFC has been related more generally to the subjective effects of psilocybin (Carhart-Harris et al., 2012b; Ruban and Kolodziej, 2018). In contrast, increased mPFC activation—which shares association with hyperfrontality—identified by higher levels of glutamate was associated with anxious ego dissolution (Mason et al., 2020). Decreased connectivity within or between the mPFC and PCC may indicate self-related attention, contextual judgement, emotional regulation, and suppression of bottom-up signals as possible mechanisms of ego dissolution. This evidence points to the importance of the midline regions of the DMN under psychedelics. However, further specification of the roles of these connections is required before any more definitive conclusions can be drawn.

6. Salience Network

The salience network (SN) serves important roles in sentience and conscious awareness. It detects and evaluates salient events (Menon and Uddin, 2010), monitors the environmental features relevant to goal-directed thinking (Seeley et al., 2007), and acts as the switching mechanism, coordinating attention between internal and external stimuli (Liang et al., 2015; Seeley et al., 2007). Moreover, the SN contributes to self-awareness (Menon, 2015). The SN is usually considered to be composed of the dorsal anterior cingulate cortex (dACC) and anterior insula (AI). Both are consistently coactivated across cognitive tasks (Swick et al., 2011), although the dACC is more involved in response selection and conflict monitoring (Menon, 2011; Ide et al., 2013). The AI receives more multimodal sensory input (Vogt and Pandya, 1987; Averbeck and Seo, 2008), detects behaviorally relevant stimuli (Menon, 2015), and coordinates the dynamic interactions of networks involved in external-oriented attention and internally self-oriented mental processes (Sridharan et al., 2008; Menon and Uddin, 2010). The SN is also associated with reasoning under uncertainty (Singer et al., 2009; Donoso et al., 2014; Lebedev et al., 2015), perception of time (Craig, 2009), and self-agency (Farrer and Frith, 2002). Many of these associations share a similarity to ego dissolution effects under psychedelics. Psychedelic-induced alterations of self-agency, the perception of time (Shebloski and Broadway, 2016), and reasoning under uncertainty indicate a possible role of the SN in the awareness of psychedelic experiences.

7. Anterior Cingulate Cortex

The ACC is a cortical midline hub region of the SN involved in cognitive control and is frequently highlighted in connection with psychedelic-induced altered states. It is highly connected to hub regions of the DMN and plays a role in self-referential processing (Northoff et al., 2006). It is also situated between emotional and cognitive domains, suggesting that it may bridge and mediate emotional responses (Liotti et al., 2000; Etkin et al., 2011; Stevens et al., 2011). Similar to the PCC, the ACC is a large region with dorsal and ventral subdivisions.

The dACC is connected to regions implicated in cognition, such as the (lateral) prefrontal cortex, parietal cortex, and premotor and supplementary motor areas (Asemi et al., 2015; Heilbronner and Hayden, 2016). dACC connectivity may be expressed in social pain. A task-based fMRI study exposed controls and participants under psilocybin to social exclusion cues. Under psilocybin, negative responses to social exclusion were reduced with decreased dACC connectivity (Preller et al., 2015). The dACC is a large cortical region with multiple functions, including performance monitoring (Ham et al., 2014). This may suggest that reduced dACC activity represents abnormal self-monitoring in response to social cues under psilocybin. Changes in AMG responsiveness under psychedelics suggests that changes to top-down mechanisms may indeed mediate emotional responses (Preller et al., 2020), with the ACC implicated (Ghashghaei et al., 2007; Fusar-Poli et al., 2010). Moreover, the rostral ACC (rACC) subarea of the ACC is related to emotional expression (Etkin et al., 2006) and—under psilocybin—has been related to subjective scores on emotional subcategories of the 5D-ASC (more on this below) (Lewis et al., 2020). Magnetic resonance spectroscopy, after ingestion of ayahuasca, showed increased connectivity between the ACC and MTL regions associated with enhanced self-compassion (Sampedro et al., 2017). Increased connectivity between the ACC and PCC was also detected. This study suggested that the ACC mediated the interaction between cognition and emotion that contributes to psychedelic afterglow effects and the enhanced mindfulness capacities measured in study participants. Mindfulness capacities cultivate a more present-centered default mode (Brewer et al., 2011), develop increased awareness of mental contents (Schooler et al., 2011; Dahl et al., 2015), and reduce maladaptive DMN activity associated with various forms of psychopathology (Soler et al., 2014; Sampedro et al., 2017; Ramirez Barrantes et al., 2019). Mindfulness functions may also help reduce ego resistance, and their increase following psychedelics again suggests evidence of neuroplasticity.

Despite its large size and differing functions, ACC subdivisions are not always reported in psychedelic literature (Stevens et al., 2011). This may contribute to conflicting findings. For example, past PET imaging evidence describes increased ACC and MTL activity under psilocybin (Vollenweider et al., 1997). However, contrasting evidence of reduced ACC neural oscillations in a network involving the PHC, PCC, and ACC, measured by EEG, have also been reported under psilocybin (Kometer et al., 2015), and further evidence suggests increased synchronization and entropy between the ACC and hippocampus under psilocybin (Tagliazucchi et al., 2014). The lower spatial resolution of PET and EEG may explain these contradictory findings. Spatial specificity of ACC changes under psychedelics may help identify associations between its connectivity, function, and phenomenology under psychedelics. For example, the ventral stream of ACC connects to affective brain regions, including the AMG (Etkin et al., 2011; Andrews-Hanna et al., 2014), that could be investigated in relation to therapeutic outcomes under psychedelics.

The structural size of the ACC has also been related to psychedelics. The subjective effects of a standardized psychedelic dose can vary widely across participants, and this variance may be related to the structural volume of the rACC, previously noted in subjective scores on emotional subcategories of the 5D-ASC. rACC volume correlated with interindividual subjective effects, suggesting that its structural variation across volunteers may predict feelings of unity, bliss, spiritual experience, and insightfulness (Lewis et al., 2020). Structural increases of the ACC were noted in long-term ayahuasca users and may be related to long-term psychedelic use that is thought to preserve psychedelic-induced neuropsychological changes (Bouso et al., 2015). This sample also showed reduced cortical thickness in the PCC. Notably, these long-term users of ayahuasca displayed high levels of mindfulness, which may suggest a relationship between structural change of the SN and DMN following psychedelics that is related to the increased levels of mindfulness. These structural differences may have existed prior to exposure to ayahuasca and disposed participants toward mindfulness and psychedelic use. This highlights the potential bias in recruitment samples in psychedelic studies.

The ACC and PCC structural differences do, however, align with evidence of the reliance of mindfulness capacity on SN control over the DMN. Although the DMN is responsible for mind wandering, awareness of mind wandering can be regulated by the SN. This is reflected in functional connectivity increases during awareness of mind-wandering in meditation examined in fMRI (Hasenkamp et al., 2012). SN hierarchical dominance over the DMN has been validated using spectral DCM analysis (Friston et al., 2014; Zhou et al., 2018), and its control over the DMN is cited as a crucial component in cognitive health and performance (Putcha et al., 2016). Its breakdown results in dysregulated DMN function (Bonnelle et al., 2012). SN role in performance monitoring and emotional mediation may account for the influence of psychedelics on self-awareness and affective regulation and encourage a more mindful or present-centered focus.

8. Frontoparietal Control Network

The frontoparietal control network (FPCN) is an attention control network, also known as the central executive network or the frontoparietal central executive network (Menon, 2011). It encompasses the dorsolateral and anterior prefrontal cortices, inferior parietal lobes, AI, and ACC (Vincent et al., 2008; Rens et al., 2017). Notably, the latter two regions (AI and ACC) overlap with the SN. The lateral prefrontal cortex and inferior parietal lobule are associated with cognitive control and decision making (Lord et al., 2018). The FPCN is also associated with access to conscious information (Dehaene and Naccache, 2001) and is inhibited in states of deep sleep (Boly et al., 2008; Tagliazucchi et al., 2016). The functional role of the FPCN can be summarized as evaluations over time, involving uncertainty or abstraction in voluntary choices (Andrews-Hanna et al., 2014; Fan, 2014; Fan et al., 2014; Rens et al., 2017) and includes the dorsal prefrontal and posterior parietal cortices as key regions tasked with negotiating acute uncertainty (Huettel et al., 2005). Abnormalities in this network, following brain injury, have been associated with impaired self-awareness (Ham et al., 2014), which may indicate its relevance to ego dissolution.

Decreases of connectivity within the FPCN are detected under both LSD and psilocybin (Lord et al., 2018, 2019; Barnett et al., 2019). However, the decrease of within-network connectivity is accompanied by increases in between-network connectivity (i.e., deviation of functional connections from network pathways). This pattern of reduced-within and increased-between connectivity—coined as disintegration and desegregation (Carhart-Harris et al., 2016b)—is documented across resting-state networks under psychedelics (Barnett et al., 2019; Varley et al., 2019) and alters the functional integration of brain regions (Petri et al., 2014) (Fig. 5). Cognitive function depends on the dynamic, context-sensitive regulation of functional segregation and integration in the brain and is mediated by neural gain modulation. Neural gain by projections from ascending neuromodulatory nuclei has the computational capacity to elicit global fluctuations, altering the balance between segregation and integration in the brain (Shine et al., 2018; Shine, 2019). Disintegration and desegregation highlight the role of the sensitization (i.e., the neuromodulation) of high-level serotonergic populations under psychedelics underwriting changes to cognition. FPCN desegregated connectivity is highlighted by several studies (Tagliazucchi et al., 2016; Lord et al., 2018, 2019) and speaks to the importance of connectivity in the FPCN in psychedelic experiences.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5

Illustration of desegregated connectivity under psilocybin, inspired by Petri et al. (2014). (A) Integration between communities—organized by color—observed in healthy adults. (B) Greater integration and reduced constraint of connections between communities observed under psilocybin. For original schematic and methods, see Petri et al. (2014).

“Connectome harmonics” research has revealed that patterns of psychedelic brain dynamics preserve new, complex forms of coherence and stability (Atasoy et al., 2017; Lord et al., 2019; Varley et al., 2019). Increased power and energy of connectome harmonic brain states, observed under psilocybin, display higher spatial and temporal variability yet maintain novel forms of stability and have been related to an increased repertoire of brain states (Atasoy et al., 2017). Links between these changes and their relationship to disintegrated and desegregated connectivity across resting-state networks are not fully understood; however, theories of their role upon the FPCN can be surmised. For example, evidence of FPCN desegregated connectivity (Tagliazucchi et al., 2016; Lord et al., 2018, 2019) may influence consciousness by enabling access to connectivity usually restricted from the FPCN. Desegregation may also be mechanistic to increased repertoires of brain states made available under psychedelics (Tagliazucchi et al., 2014; Atasoy et al., 2017). Both observations might also relate to the phenomenological richness of psychedelic experience and therapeutic outcomes thought to depend on increased access to (precise) bottom-up prediction errors necessary for the revision of priors. The coinciding disintegration of the FPCN (Lord et al., 2018, 2019; Barnett et al., 2019) and relaxation of priors (Carhart-Harris and Friston, 2019) would also be mechanistic in this process; however, they may contribute to cognitive aspects of ego dissolution. Untangling the mechanistic relationship between FPCN changes under psychedelics and their role in ego dissolution needs more research; however, existing evidence points to the potential importance of its role in cognitive aspects of the psychedelic experience.

Correlating cortical oscillatory changes with psychedelic experiences and therapeutic introspection may also help elucidate the role of the FPCN under psychedelics. Under psilocybin, an MEG study reported decreased (bilateral) β-band activity in the FPCN (Muthukumaraswamy et al., 2013). β-Band connectivity is suggested to encode long-term priors of behavior and environmental stimuli (Betti, et al., 2021). Decreased oscillatory power of the FPCN may be a measure related to change and revision of the association between behavior and stimuli. θ-band connectivity is proposed to support internally directed attention between the DMN and FPCN (Kam et al., 2019) and thus important in psychedelic research regarding brain oscillations. However, FPCN connectivity changes are not always reported under psychedelics (Müller et al., 2018), and the source of inconsistencies—which are discussed later in detail—remain uncertain, although are not entirely unusual in imaging research. Added caution is further suggested by the variable functional topography of the FPCN between individuals (Marek and Dosenbach, 2018). The importance of psychedelic effects upon the self can be explored through large-scale integrated dynamics.

1. The Self in Networks

The SN is related to the minimal aspects of conscious self-awareness (Limanowski and Blankenburg, 2013), also known as the embodied aspect of self (Seth, 2013; Lebedev et al., 2015). The minimal or embodied aspect of self-awareness requires bottom-up processing to define physical boundaries, proprioception, interoception, and the experience of oneself being rooted in (sensory) motor processes (Blanke and Metzinger, 2009; Legrand and Ruby, 2009). This aspect of self may be important for ego dissolution, indicated by decreased within-SN connectivity associated with ego dissolution in an fMRI psilocybin study (Lebedev et al., 2015). The DMN, in contrast, is related to the narrative sense of self commensurate with personal identity, described in Bayesian terms as the brain predicting its own contribution to sensory input by top-down inferences about the causes (i.e., latent or hidden states of the world) of sensory information (Metzinger, 2003; Friston, 2010; Clark, 2013).

Aberrant DMN-SN connectivity is important in clinical and theoretical models of various mood disorders, including depression, anxiety, and schizophrenia (Menon, 2011). Mood disorders associated with this connectivity suggest its importance in therapeutic outcomes, and change to this connectivity under LSD is associated with positive mood and arousal (Atasoy et al., 2017). However, DMN-SN functional connectivity is more commonly linked to ego dissolution and shows reduced anticorrelation under psilocybin (Carhart-Harris et al., 2013) and decreased segregation under LSD (Carhart-Harris et al., 2016b). The relationship of DMN-SN connectivity to respective aspects of self suggests that psychedelic influence on these networks may contribute to ego dissolution. Furthermore, evidence of their relationship to psychopathology speaks to the overlap between networks involved in the self and mental well-being. Further analysis of this connectivity and psychologic changes under psychedelics may help determine the contribution of this network to ego dissolution and normal waking consciousness. These possibilities invite further investigation of the well established anticorrelation between these networks.

2. Coordinated Balance between Anticorrelated Networks

An important feature of intrinsic brain networks is their coordinated and balanced activity. Patterns of correlations and anticorrelations between networks constitute an aspect of functional integration supporting consciousness (Bak et al., 1987; Demertzi et al., 2019). For example, the DMN focuses attention inward in a task-free exploratory manner that—upon stimulation from the environment—rapidly deactivates in deference to externally focused networks like the DAN. The dorsal attention network (DAN) is composed of the frontal eye fields and inferior parietal sulcus and is responsible for orientation to behaviorally salient cues (Ptak and Schnider, 2010). Altered DAN connectivity has been reported in the psychedelic literature and has been related to an increased repertoire of brain states under psilocybin (Tagliazucchi et al., 2014).

The SN mediates the anticorrelation of the DMN and DAN (Fox et al., 2005) through dynamic mechanisms responsible for the coordination of attention between bottom-up perceptual features of the environment and top-down goals (Menon, 2015). The SN is hierarchically above the DAN and DMN, suggesting that it may be a control network capable of selecting between these anticorrelated networks (Zhou et al., 2018). SN-meditated switches enable attentional resources to transition efficiently from internal focus of the DMN to an external focus (i.e., sensory attention) of DAN. Psilocybin reduces the anticorrelation between DMN-DAN connectivity deployment (Carhart-Harris et al., 2013). Reduced anticorrelation indicates that attentional resources may not be divided between external and internal realms of attention in the usual manner. Descriptions of ego dissolution that describe the dissolution of boundaries between the objective and subjective world signify the importance of investigations of the DMN-DAN anticorrelation (Stoliker et al., 2021). Like SN-DMN connectivity, competition between the DMN and DAN is correlated with disorders such as schizophrenia (Whitfield-Gabrieli et al., 2009), highlighting the potential importance of DMN-DAN anticorrelation in mental health and perceptual synthesis. A certain form of meditation, known as nondual awareness, tries to reduce the separation between external and internal reality, which, intriguingly, reduces the anticorrelation between brain areas that respond to intrinsic and extrinsic stimuli (Josipovic et al., 2012). This finding reinforces the suggestion that DMN-DAN anticorrelation may be crucial to the loss of boundaries between the subjective and the objective world (Grof, 1980) that characterize ego dissolution.

Alterations to anticorrelated brain networks may also be involved in therapeutic outcomes following psychedelics; however, evidence is scarce. Whether therapeutic outcomes represent a direct result of altered anticorrelation or changes to connectivity between other systems such as the MTL under acute effects is unclear. It may be possible that the altered anticorrelations contribute to ego dissolution and promote neuroplasticity, leading to psychologic changes. Furthermore, the role of these networks in higher-order processing and attentional orientation could indicate access to information that is usually suppressed and may contribute to therapeutic processes that reorganize patterns of thought. Evidence of DMN-DAN connectivity changes associated with enhanced mindfulness capacities may represent a link to reduced reactivity, judgmental thinking and enhanced self-kindness (Sampedro et al., 2017). Moreover, anticorrelations under control of the SN also involve the FPCN (Zhou et al., 2018), which is known to be anticorrelated with the DMN (Menon, 2011, 2018; Andrews-Hanna et al., 2014; Chand et al., 2017; Bolton et al., 2020).

Control of the FPCN is managed by the SN recruitment of the AI, previously noted to be involved in the interactions between internal and external mental processes (Sridharan et al., 2008; Menon and Uddin, 2010) and detect behaviorally relevant stimuli (Menon, 2015). Specifically, activation of the right AI engages the FPCN, resulting in coactivation of the SN and FPCN (Raichle et al., 2001; Greicius and Menon, 2004; Menon, 2015) [Recent evidence suggests that nodal measurements are more sensitive to the severity of disorders of consciousness than global measurements (Martínez et al., 2020)]. Although normally anticorrelated, correlated activation of the FPCN and DMN has been documented to correlate with symptoms of psychosis that relate to a disturbed sense of self and uncertainty of what is and is not real (Leptourgos et al., 2020). Changes to the functional integration of the FPCN, DMN, and SN are associated with meditation practice (Doll et al., 2015) and the control of attention (Corbetta and Shulman, 2002). These observations speak to the putative role of anticorrelations in ego dissolution that involve networks that orient attention and contribute to self and perspectival belief updating.

Subcortical brain regions that evince anticorrelation under psychedelics have also been investigated. For example, the claustrum is a subcortical region that has strong connectivity with the prefrontal cortex (Brown et al., 2017). It has been shown to activate with demanding cognitive tasks and the engagement of the FPCN (Krimmel et al., 2019). A recent study showed that under psilocybin, connectivity between the right claustrum and the DMN decreased, whereas connectivity to the FPCN increased (Barrett et al., 2020b). Claustrum connectivity changes may contribute to functional changes in cortical networks that support attention and self-consciousness. Further investigations comparing anticorrelation of brain networks in psychosis, meditation, and classic psychedelics may help determine the functional importance of brain anticorrelations in ego dissolution, therapeutic outcomes, and the consciousness of self.

1. Conflicts within Psychedelics Findings

The association between subjective effects and brain connectivity is often inconsistent among psychedelic studies. A brief overview of these discrepancies characterizes the present state of understanding in psychedelic neuroscience. Ego dissolution is often associated with DMN hub regions under psychedelics. This is demonstrated by mPFC-PCC decoupling in fMRI (Smigielski et al., 2019b) and decreased PCC α power in MEG under psilocybin (Muthukumaraswamy et al., 2013). More generally, subjective effects under psilocybin are associated with decreases across the ACC/mPFC (Carhart-Harris et al., 2012b). However, brain-wide connectivity changes are also associated with ego dissolution and extend beyond the DMN under LSD (Tagliazucchi et al., 2016) and have been linked to connectivity between the anterior PHC and cortical communities, including visual, somatosensory, salience, prefrontal, and control areas under psilocybin (Lebedev et al., 2015). Furthermore, not all psychedelic investigations find changes to the DMN. For example, the expected changes in complexity of brain signals—measured with the fractal dimension—from DMN nodes in fMRI, under LSD and psilocybin, were not found (Varley et al., 2019).

An LSD study employing fMRI, MEG, and arterial spin labeling identified decreased RSC and PHC connectivity that correlated with ego dissolution and noted that activity of resting-state networks such as the DMN was not correlated with ego dissolution (Carhart-Harris et al., 2016b). These findings challenge the reliability of DMN changes associated with ego dissolution. An EEG study also found cortical oscillations in the RSC, PHC, and lateral orbitofrontal areas; however, in this study, these changes were associated with spiritual experience and insightfulness and not with ego dissolution (Kometer et al., 2015). Other resting-state networks suspected to have a role in psychedelic effects such as the FPCN are also sometimes reported as absent: in this case, in fMRI scans under LSD (Müller et al., 2018). Furthermore, effective connectivity involving the PCC was found to correlate with subjective effects of psilocybin in a DCM analysis of MEG (Muthukumaraswamy et al., 2013), yet evidence of its alteration in an fMRI study of LSD was absent (Müller et al., 2018). See Table 2 for a summary of connectivity results associated with ego dissolution (Millière et al., 2018).

2. Conflicts between Psychedelic Findings and Nonpsychedelic Findings

Occasionally, psychedelic-induced changes in connectivity—cited in association with subjective effects—are seen in nonpsychedelic studies without subjective effects. For example, decreased activation within the DMN is cited in association to ego dissolution (Carhart-Harris et al., 2016b). However, this also resembles the patterns of activity following administration of sertraline, a nonpsychoactive drug (Klaassens et al., 2015; Müller et al., 2018). Within the DMN, the deactivation and decoupling of the PCC and mPFC has been suggested to underlie ego dissolution (Carhart-Harris et al., 2012b; Smigielski et al., 2019b; Carhart-Harris et al., 2014b). Although, this deactivation is also identified in response to antidepressants, with no noticeable psychoactive effects (Goldstein-Piekarski et al., 2018), and a similar activation pattern is also elicited by anxiety (Zhao et al., 2007). If DMN connectivity changes underlie alterations to perception reported under psychedelics, then we might expect similar connectivity changes to produce psychedelic-like subjective effects. Psychedelic experimental designs are at risk of inducing anxiety, especially when rapid onset of a psychedelic through intravenous or intranasal route is employed. It is, therefore, important to note known confounds when trying to identify neural mechanisms.

3. Signal from the Noise

The results of neuroimaging studies suggest connectivity associations with ego dissolution—and other subjective effects of psychedelics—that have yet to be clearly determined. Examining and integrating cellular, regional, network, and whole-brain connectivity changes may best account for observed phenomenological changes in consciousness (Varela et al., 2001; Palva et al., 2005; Melloni et al., 2007; Busch et al., 2009; Dehaene and Changeux, 2011; Hipp et al., 2011; Fahrenfort et al., 2012). Psychedelic subjective effects will likely involve widespread changes to hierarchical connectivity subtending associative and sensory processing across the cortex and subcortex. Holistic accounts of brain changes under psychedelics—that identify reliable correlations between connectivity and subjective experience—may be followed by more thorough and advanced examinations, such as those using effective connectivity analyses. However, the accurate modeling of psychedelic brain connectivity is encumbered by challenges.

The anatomic coordinates of regions—such as the PCC—may be specified using various methods for subsequent region of interest–based analysis. For example, independent component analysis can be used to detect the peak activity of cardinal regions and their corresponding coordinates, or these coordinates may be borrowed from previous literature. Recent alternate methods have also been proposed, involving probabilistic mapping to improve consensus and inferences (Dworetsky et al., 2021). These strategies may produce different connectivity results when referring to the same regions. Variability due to individual differences in brain anatomy may further complicate accurate determination of anatomic coordinates. Commonly-referred-to networks may also be composed of different regions. Moreover, the application of global signal regression has been considered a cause of anticorrelated brain connectivity, and debate over its application questions interpretations of anticorrelations (Murphy et al., 2009). Concerns also exist with the interpretation of the blood oxygen level–dependent signal and whether the blood oxygen level–dependent signal is sensitive to changes in conscious experience (Laumann et al., 2017).

These concerns aside, understanding the mechanistic unfolding of processes after psychedelics can be advanced by imaging and analytic approaches. This requires increased sample size and the standardization of image processing pipelines, including the methods of drug administration, timing of imaging, environmental setting, balance of participant group–averaged neuroanatomy, and data preprocessing pipelines to clean (physiologic) signals. Variation in preprocessing pipelines in particular can be a significant barrier to the accurate comparison of findings across neuroimaging studies. For comparison of motion correction pipelines, see Parkes et al. (2018); for global signal regression see Almgren et al. (2020) and Aquino et al. (2020); and for motion correction in diffusion imaging pipelines, see Oldham et al. (2020). The optimization and standardization of these pipelines—in conjunction with replication and comparison between psychedelics—may be essential. Fortunately, growing interest in psychedelic therapeutic potential may motivate standardization and replication.

Lastly, we consider the measurement of ego dissolution in imaging studies. Ego dissolution is a phenomenon that varies across subjects provided a standard dose of a classic psychedelic. Therefore, the available connectivity findings do not precisely indicate the changes underlying psychedelic ego dissolution. Retrospective self-report measures also limit the measurement of ego dissolution. Self-reporting, typically taken after the effects have subsided, is inherently subjective and an issue of reliability. Subjects may have no prior basis for the experience of ego dissolution. This suggests variability in the self-scoring of ego dissolution across subjects who may have experienced similar subjective effects. An additional measurement consideration is that the onset of subjective effects can vary across subjects. This means that the time at which imaging occurs may not align with psychedelic peak effects. To offset these issues, we recommend subjects complete qualitative behavior reports detailing their experience with special attention to the imaging portion of the session. Researchers may also record observations as an additional means of assessing the degree of ego dissolution and the time of its occurrence. Standardized timing of imaging procedures may also be reconsidered by initiating procedural “time zero” at the point that subjective effects first become noticeable. This may perhaps be achieved by relying on subject reports, for example, of closed eye visual alterations. Development of such onset measures may help ensure imaging reliably records optimal subjective effects. Furthermore, researchers may wish to discard data from outlier subjects who experience very mild subjective effects from the research dose of a psychedelic to improve analyses and consider building a registry of subjects with high subjective effects for future imaging studies.

4. Unifying Neuroimaging Evidence of Psychedelic Mechanisms

Closing gaps in the understanding of the neural mechanisms of subjective effects of psychedelics and formulating a unifying theory of 5-HT2R activity across the brain remains a goal of psychedelic research. Psychedelics have been suggested to drive the brain away from order toward a state of disorder in the entropic brain theory (Carhart-Harris et al., 2014). The flexibility of conscious states, spatial and temporal changes in connectivity, and increased sensitivity to intrinsic and extrinsic perturbations under psychedelics supports this theory (Carhart-Harris et al., 2014a; Carhart-Harris, 2018; Lord et al., 2018, 2019). Observed criticality characteristics have been used to understand the stable and coherent organization of brain dynamics under psychedelics (Atasoy et al., 2017). Criticality is a widespread phenomenon in multiscale and complex systems such as the brain. Criticality arises between states of order and disorder. It is marked by scale-free fluctuations that stretch from the finest to the coarsest scale and that can spontaneously jump between diverse spatiotemporal patterns (Cocchi et al., 2017). Although criticality has also been challenged as an unnecessarily complex explanation of drug-induced states (Muthukumaraswamy and Liley, 2018), it illustrates that psychedelics may temporarily exercise neuronal and behavioral flexibility.

An empirical approach to measuring psychedelic mechanisms across the brain comes from a recent imaging study explaining broad FC patterns leading to subjective effects. This study measured time-dependent changes of psilocybin over the course of onset to peak effects at three timepoints (20, 40, and 70 minutes post–oral administration) (Preller et al., 2020). Resting-state FC was quantified using a data-driven global brain connectivity method, which suggested that psilocybin reduces the connectivity of associative regions and increases connectivity in sensory regions across subcortical and cortical areas (Preller et al., 2020; Vollenweider and Preller, 2020). The authors suggest that the disintegration of associative activity and synchronization of sensory activity may explain subjective sensory and self-related experiences under psychedelics.

Amounting basic and clinical evidence also indicates that subcortical alteration of thalamo-cortical gating alters the bottom-up flow of information (Vollenweider and Geyer, 2001; Geyer and Vollenweider, 2008). Thalamic gating is controlled by glutamatergic cortico-striatal and cortico-thalamic pathways and is modulated by serotonergic and dopaminergic projections (Vollenweider and Smallridge, 2022). Vollenweider and Smallridge (2022) and work from their laboratory have established a working model of thalamic cortical circuitry identifying psychedelic stimulation of 5-HT2A receptors on CTSC cortical pyramidal cells and GABA interneurons on thalamocortical information flow. For illustrations of their working hypothesis of CTSC and cortico-cortical circuits, see Vollenweider and Smallridge (2022). The result is altered cortico-cortical integration of the distributed neuronal activity underlying the integration of external and internal information reaching the cortex. Thalamic changes are proposed to underwrite psychedelic perceptual alterations and suggest a neural mechanism of disintegrating subject-object boundary (Vollenweider and Smallridge, 2022). The thalamic gating account suggests that psychedelics increase bottom-up prediction errors, which may share mechanisms with the disinhibition of top-down associative connectivity and related desegregated patterns of connectivity. These theoretical and empirical accounts offer useful summaries of psychedelic subjective effects that complement an evaluation of psychedelics offered by the framework of hierarchical predictive coding.

5. Hierarchical Predictive Coding

Hierarchical predictive coding is a mechanistic account of how the brain processes information (Friston, 2008). It is supported by empirical evidence (Mumford, 1992; Rao and Ballard, 1999) and informed by our understanding of synaptic communication (Adams et al., 2013; Friston, 2020; Hobson et al., 2021). Noting the neuromodulatory role of serotonergic neurotransmission (Picard and Friston, 2014) and the influence of psychedelics upon these receptors, it provides a unifying account of psychedelic subjective and therapeutic effects. The premise of hierarchical predictive coding is that the brain is a constructive organ, explaining (i.e., inferring) sensory impressions or the causes of input on sensory epithelia. Sensory signals from the outer world are transduced into electrical signals that ascend the brain’s hierarchical architecture — in the form of prediction errors—to update (Bayesian) beliefs about hidden states in the world. These (subpersonal) beliefs form the basis of perception and can be regarded as the brain making predictions about the causes of its sensations. As new impressions from the sensory epithelia ascend the hierarchy, the accuracy of prediction is tested against the new sensory evidence in a cyclical process described variously as predictive coding, Bayesian filtering, or Bayesian belief updating. In short, bottom-up sensory evidence carries information (i.e., prediction errors) that updates expectations about states of affairs (i.e., posterior perceptions based on prior beliefs). These expectations then provide revised predictions that descend through the hierarchy to resolve prediction errors at lower levels.

The computational formulation of hierarchical predictive coding goes on to consider neuromodulation in the construction of predictions, suggesting that gain control (i.e., sensitivity) is a key player in selecting the most reliable or precise prediction errors for belief updating at higher hierarchical levels. Synaptic gain is thought to encode precision, and synaptic gain control reflects predictions of precision that implement this (attentional) selection. Computationally, precision is the predictability of some random variable and may be interpreted as a level of confidence placed in a prediction or prediction error. At high levels of the hierarchy that may entail (declarative or experienced) conscious processing, precision refers to confidence in an inference or belief. The neuromodulation or sensitization of neural populations is thought to determine the attenuation or augmentation of precision (i.e., confidence) in a source of information. Precision control, therefore, provides a synaptic mechanism describing the selection and attenuation of sensory channels and is proposed as a neurologic explanation for many psychopathologies (Brown et al., 2013). For example, in psychosis, abnormal neuromodulation or sensitization of neural populations may confound selection and attenuation of sensory channels, resulting in false inferences (Adams et al., 2013; Kanai et al., 2015).

The role of psychedelics in neuromodulation derives from binding to serotonergic 5-HT2A receptors. Serotonergic receptors serve a crucial role in precision control underlying sentience and attention that is necessary to infer both the cause of sensory impressions and the consequences of self-initiated actions (Brown et al., 2013; Parr and Friston, 2017, 2019). Please see (Parr and Friston, 2017, 2018; Parr et al., 2018) for recent theoretical treatments of neuromodulation and precision under active inference. Imaging evidence and brain connectivity findings are consistent with the preferential stimulation of 5-HT2AR on deep pyramidal cells, which are densely distributed in visual and associative areas of the cortical hierarchy (Carhart-Harris et al., 2012b; Saulin et al., 2012; Roseman et al., 2016; Carhart-Harris et al., 2016b; Preller et al., 2018b; Carhart-Harris and Friston, 2019; Preller et al., 2020; Vollenweider and Preller, 2020). Under predictive coding, these cells are thought to be the cells-of-origin of descending predictions to subcortical systems or lower hierarchical levels.

Consistent connectivity findings across visual and associative areas have helped account for psychedelic hallucinogenic effects and ego dissolution, respectively, and can be framed within the hierarchical predictive coding framework. For example, subcortical connectivity to sensory and associative areas is dependent on synaptic gain control. Furthermore, links between pharmacology and psychology can be considered in terms of coordinated message-passing of belief-updating cortical hierarchies. Expression of 5-HT2AR agonism serves a role in the active coping with stress and uncertainty (Carhart-Harris and Nutt, 2017; Murnane, 2019). Increased entropy and changes to synaptic gain control at the top of the hierarchy may influence the executive functions of cognition and priors. In short, hierarchical predictive coding provides a biologic starting point to understand the effect of pharmacological interventions at the synaptic level in terms of sentience and planning.

The leading synthesis of hierarchical predictive coding and psychedelic effects to date is the relaxation of beliefs under psychedelics (REBUS) model. REBUS describes the relaxation of precision at high levels of the hierarchy, leading to a greater influence of ascending prediction errors. In other words, a rebalancing of prior and sensory precision that enables new prior hypotheses and narratives to be engaged. In virtue of these expectations being at the apex of a hierarchical generative model, they are necessarily concerned with agency and active sentience of a multimodal sort and, on one reading, entail an implicit or explicit notion of self as agent.

Physiologically, the relaxation of high-level precision corresponds to a reduction in the synaptic gain or efficacy of particular regions in the cortical hierarchy. These changes are relatively consistent with neuroimaging findings, which show decreased within-network connectivity and altered oscillatory rhythms. The REBUS model may explain both the mechanisms of psychedelic subjective (Stoliker et al., 2022) and therapeutic effects; however, it remains to be validated (Fig. 6). Understanding that synaptic efficacy depends upon neuromodulation and that neuromodulation (of synaptic gain) changes the inhibition-excitation balance of synaptic integration suggests how changes in serotonergic agonist activity can influence internalizing disorders—disorders that are characterized by rigid ways of thinking and unduly precise prior beliefs. This is particularly relevant to psychiatry by suggesting psychedelics as a neuromodulator that affects the selection of ascending information—and thereby may be capable of subverting overly precise beliefs (i.e., ego resistance).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6

(A1) Sensory input is compared with top-down predictions to form prediction errors that are passed onto higher levels of the hierarchy to revise Bayesian beliefs. These beliefs or representations then supply top-down predictions, which resolve the prediction errors at the lower level. This process is repeated to minimize the prediction error at each level. The predictive coding hierarchy tries to construct the best top-down explanation for bottom-up sensory input at each level of the hierarchy. (B1) Psychedelics are thought to reduce precision and flatten the energy landscape of beliefs generated in high levels of the hierarchy supporting self-related beliefs, thereby producing the dissolving of self-related priors (i.e., ego dissolution). (A2 and B2) Dissolution of precision of high-level priors flatten the curvature of the free energy landscape, enabling neural dynamics to escape their local minima or basins of attraction, allowing greater attention to the sensory input and prediction errors (computationally expressed as a free energy landscape). The cognitive-therapeutic result of ego dissolution is the reduced precision or commitment to higher-level beliefs in the high levels of the hierarchy that affords an opportunity to explore a landscape of alternative hypotheses of the causes of sensory impressions and the consequences of self-initiated actions. Change to these explanations can be therapeutic by enabling new ways to make sense of the world and lived exchanges with it. This notion of free energy landscapes is endorsed by empirical studies of electrical physiologic responses and functional anatomy (Bastos et al., 2012). Adapted from Carhart-Harris and Friston (2019).

Clearly, we have somewhat simplified the account of how psychedelics mediates its effects through changing precision or postsynaptic gain. The effects of psychedelics are complex, and the precise pathways and synaptic mechanisms remain unclear. For a detailed account of the specific signaling pathways engaged by psychedelics, see Marek (2018) and López-Giménez and González-Maeso (2018). The particular 5-HT2AR–expressing populations also need to be considered. For example, 5-HT2AR–expressing deep-layer pyramidal neurons appear to be those that are most likely to be directly activated by psychedelics (Martin and Nichols, 2016). Psychedelics affect many types of neurons and produce complex transcriptional responses within the brain (Martin and Nichols, 2016). Generally speaking, activation of 5-HT2ARs induces an increase in excitatory postsynaptic currents and discharge rates in pyramidal neurons (Celada et al., 2008), causing an asynchronous mode of glutamate release (Aghajanian and Marek, 1999; Aghajanian, 2009) and spike-field decoherence (Celada et al., 2008). This decoherence has also been observed in vivo using dynamic causal modeling of MEG data, implicating psilocybin-induced increases in deep pyramidal cell excitability as the cause of broadband decreases in oscillatory power (Muthukumaraswamy et al., 2013). This effect of psilocybin implicates inhibitory interneurons, in the sense that oscillatory dynamics rest upon the interaction between pyramidal cells and fast-spiking inhibitory interneurons (Lisman and Buzsáki, 2008; Sohal et al., 2009; Lisman, 2012).

Similar adaptations of the REBUS model are described by the altered beliefs under psychedelics (ALBUS) model and related strengthened beliefs under psychedelics (SEBUS) model, suggesting that the strength of a dose calibrates the flexibility or rigidity of prior beliefs (Safron, 2020). The ALBUS model details possible dose-dependent differential effects of psychedelics on prior belief through their influence on inhibitory interneurons and the transmission of hippocampal/entorhinal system signals, (Safron, 2020). We suggest the alteration and strengthening of beliefs at specific doses may also be mediated by interaction with psychologic and environmental context, which is a point acknowledged by the ALBUS model. Furthermore, the mechanisms underlying altered beliefs may also be informed by work investigating reality monitoring (Dijkstra et al., 2022). The REBUS model is also reflected in neural annealing as reduced energy sinks (https://opentheory.net/2019/11/neural-annealing-toward-a-neural-theory-of-everything/) and are surmised to enable new directions for thought patterns under psychedelics without a complete breakdown of cognition (Petri et al., 2014; Atasoy et al., 2017). On this view, precision plays the role of an inverse temperature, where decreasing the precision of prior beliefs (i.e., increasing their temperature) makes it easier for higher-level beliefs to jump from one to another. In other words, the energy landscape is flattened. A useful analogy that captures the flattened landscape of priors under ego dissolution—and subsequent promotion of flexibility and plasticity—is that of a skier descending a hill after a fresh snow fall: Tracks (i.e., ruts), which were preferentially traveled previously (likened to prior belief updating), are leveled by the snowfall (i.e., the grooming action of psychedelics on high-level precision), resulting in a flattened landscape that enables travel in new directions without prior resistance. In short, psychedelics may simply flatten the “rut” and stop belief updating getting “stuck in a rut.”

Finding itself contained in a dynamic environment, the brain must possess a hierarchical generative model that can represent the world's contents, causes, and contexts (Bastos et al., 2015). We view ego dissolution as a breakdown of the generative model, particularly at the highest levels involved in our sense of identity and identification with the world. Psychedelics' net-excitatory effect on layer 5 pyramidal neurons at the upper levels of the hierarchy may, at its extreme, strip down the precision of the hierarchical generative model's priors of the world's contents, causes, and contexts. This makes sense because neurons in deep cortical layers represent perceptual hypotheses. We suggest this temporary breakdown alters the observer's sense of agency (i.e., own contribution to sensory input) and manifests a sense of heightened existential experience and primordial awareness.

Stated in alternative terms, the hierarchical generative model places boundaries on the physiologic states of living systems to support the behavior that ensures biologic survival (Friston, 2008). These boundaries need to be flexible enough to adapt to the dynamic environment. Belief updating supports adaptation by resolving mismatches in bottom-up prediction errors and top-down predictions and is itself supported by the coordination of hierarchical signals and the balance of integrated activity across distributed networks and regions that respond to internal and external contexts. At doses sufficient to induce ego dissolution, increased entropy may elicit a state of criticality wherein the ordinary cognitive boundaries of biologic system states are temporarily exceeded. This suggests reduced Bayesian model evidence and the extension of the distribution of hypotheses that underlie beliefs that make sense of the self and world. Cognitively, this may generate alternative models of the world and illustrate how our evolved, specialized faculties (i.e., connectivity) can construct it.

Top-down disinhibition may also factor into alternative models and may enable the desegregation of connectivity, which we suggest may further extend the probability distribution and cognitive flexibility at the level of consciousness and sentience. Together, beliefs organized along the hierarchical generative model may be afforded alternate hypotheses, and mentation may be afforded the exploration of alternate states of thought, feeling, and perception. Therefore, serotonergic psychedelics may expose the possible constructions of the world, and these possibilities may offer alternative therapeutic pathways for beliefs about the relationships within it.

This gives some indication of how the effect of serotonergic psychedelics may alter the generative model's role in sentience and planning, which speaks to the function of serotonin and the layer 5 apical dendrites in the emergence of consciousness. Advancements in psychedelic imaging analyses are uncovering more specific mechanisms underwriting connectivity changes proposed by unifying theories. For example, entropy changes in the brain, once described as uniformly elevated across the brain, are now suggested to vary and decrease across brain regions (Herzog et al., 2020). Similarly, recent work using “template” independent component analysis revealed the mediodorsal and pulvinar thalamic nuclei connectivity between visual connectivity and the DMN in association with subjective effects under psilocybin (Gaddis et al., 2022). The thalamus is composed of as many as 60 nuclei (Cassel and de Vasconcelos, 2015) and provides one example of the precision required to identify the neural mechanisms of psychedelics. As research moves forward and the need for more precise estimations of connectivity changes are demanded, combining methods that isolate highly specific regions with the precision DCM can advance our understanding of the brain-behavior relationship under psychedelics.

6. Future Research Priorities

Neuroimaging research provides a basis for interpreting psychedelic effects on consciousness and lasting behavioral outcomes. As has been described above, psychedelic changes span the hierarchical organization of the brain—from binding to serotonergic receptors to the downstream effects of oscillatory patterns on specialized regions and network connectivity. Measurement of correlates of psychedelic changes to neurons that form brain connectivity can contribute to our understanding of the biologic underpinnings of psychedelic subjective effects. Analysis methods such as measuring the cytoarchitecture of the brain’s hub regions may be a focus of future research. Cytoarchitecture research suggests that rich links between brain hubs can be explained by genetic factors, whereas peripheral links are better explained by environmental influences (Fulcher and Fornito, 2016). The desegregation of FC from network pathways under psychedelics (Lord et al., 2018; Varley et al., 2019; Luppi et al., 2021) also demonstrates deviation of connectivity mediating rich links that is consistent with the observations of increased sensitivity to intrinsic and extrinsic perturbations (i.e., environmental influences) under psychedelics.

Hierarchical interactions are necessarily directed (i.e., bottom-up and top-down). Functional connectivity measures are blind to directed connectivity. This suggests that more focus should be placed on the analysis of effective connectivity changes under psychedelics. Correlation-based whole-brain FC techniques can be complemented by the ability of DCM to infer the directed causes of neuronal responses. This technique is capable of parcellating the brain hierarchy and identifying the direction and strength of message-passing across subcortical and cortical levels. Crucially, unlike functional connectivity, the effective connectivity estimates supplied by DCM include intrinsic self-connections, namely the intrinsic excitability of different regions or neuronal populations. This is important in terms of estimating the effective precision or synaptic gain of a population under a hierarchical predictive coding formulation. In short, DCM can link synaptic level mechanisms described in hierarchical predictive processing to the directed connectivity changes of region level neural substrates.

b. Determinants of dose-response and behavioral measures of ego dissolution

Uncertainty exists whether ego dissolution involves a gradual onset or a binary (i.e., all or nothing) experience. Evidence suggesting the binary (i.e., nonlinear) dose-response hypothesis is supported by PET investigation of plasma psilocin occupancy (the active metabolite of the prodrug psilocybin) and the reported subjective intensity of experience (Madsen et al., 2019). However, visual schematics illustrating subjective responses of oceanic boundlessness (i.e., ego dissolution) counter this assertion by providing evidence of a gradual onset (Fig. 7) (Hirschfeld and Schmidt, 2020).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7

Ego dissolution rating by body weight–adjusted psilocybin dose, adapted from Hirschfeld and Schmidt (2020)’s review of psilocybin studies using the 5D-ASC. Psilocybin doses assigned by varying body weights suggest ego dissolution (oceanic boundlessness) may be amplified in a linear, dose-dependent manner (i.e., gradual) (Hirschfeld and Schmidt, 2020).

We approach the conflict between gradual and binary dose-response curves by proposing that alterations to high-level observer constructs (i.e., self or ego and meaning) occur as a binary shift. In contrast, lower-level observed constructs encounter gradual dose-dependent changes, for example, in visual alterations. See Safron (2020) for related views. Paradoxically, psychedelic desegregation of connectivity pathways and the collapse of the predictive hierarchy may reduce differentiation between hierarchical levels and contribute to blurred subject-object boundaries. Our review characterizes ego dissolution as the emergent shift in the observer. Like the mystery of consciousness, ego dissolution is suggested to be a phenomenon that is greater than the sum of its parts. Those parts represent psychedelic altered object phenomena such as visual, temporal, and emotional alterations. We also recognize the limitations of comparing self-reports across subjects with different prior experiences and references to base their responses. To measure the binary onset of ego dissolution, researchers conducting imaging studies may wish to note momentary metacognitive lapses that subjects report in their stream of consciousness—for example, experiences of disembodiment.

Time-dependent connectivity mechanisms of ego dissolution may be helpful to answer the dose-response question and have been used in the work of Preller et al. (2020). However, this study used a body-weight dose equivalent to 14 mg per 70 kg, whereas the standard “full” dose used in therapeutic trials is typically 25 mg. Sequential time-dependent scanning and behavioral measures across high-dose onset may require psychedelic-experienced participants and enable better measurement of the correspondence between connectivity changes and the onset of ego dissolution. Alternatively, imaging efforts may attempt to demarcate shifts in consciousness using the high temporal specificity of EEG and MEG modalities. Follow-up analysis may then observe whether a transition in brain connectivity accompanies reports of ego dissolution during brain imaging.

A related controversy pertains to psychedelic doses adjusted by body weight. Weight-adjusted doses have been the standard form of dose administration across neuroimaging trials; however, recent evidence suggests no influence of body weight on subjective effects of psilocybin (Garcia-Romeu et al., 2021). This research also marks the need for greater understanding of the mechanisms controlling the relationship between dose and response. Moreover, subjective responses can vary significantly following a standard or weight-adjusted dose of a psychedelic (Preller et al., 2019). Several approaches to the dose-response question may be entertained. For example, greater understanding of relationship between subjective effects and the density and distribution of serotonin receptors may be informative. Measures of the volume and functional connectivity profiles of individual brain structures may also predict the dose-response relationship and warrants future investigation. For example, baseline global connectivity may be associated with subjective responses to psilocybin (Preller et al., 2020) and intensity of ego dissolution has been linked to a low diversity of connections in the anterior parahippocampus (Lebedev et al., 2015). Structural predictors of dose responses are also intriguing, following evidence that psychedelics can influence particular region volumes in long-term users (Bouso et al., 2015) and rACC thickness can predict emotional subjective effects (Lewis et al., 2020). To date, no method has been established to predict dose-response relationships in psychedelic naive participants. Neuroimaging studies sensitive to associations between the density and distribution of receptors, region structural volume, measures of connectivity and the subjective response of participants may be valuable in determining predictors of dose-responses and enable the tailoring of doses for clinical applications. Post-hoc analyses can also contribute to answering the dose-response question by determining how effective connectivity pathways between key brain regions and hierarchical strength (calculated by summing efferent and afferent region connectivity) (Zhou et al., 2018) between networks predict aspects of the experience reported on behavioral measures. Understanding the connectivity mechanisms of the dose-response relationship may provide biomarkers that determine the optimal dose in therapeutic contexts and, in turn, identify the mechanisms of subjective effects and ego dissolution.

Furthermore, improvements in existing behavioral measures administered after psychedelic ingestion are crucial to advancing research. At present, aspects of the psychedelic experience such as ego dissolution are lumped into a single measure of changes to sense of self and the relationship between subject and object. However neuroscientific and consciousness research suggest that the self can be dichotomised into different aspects related to distinct networks, such as the SN (i.e., minimal self) and DMN (i.e., narrative self). In associating connectivity patterns to ego dissolution, it may be essential to implement subjective measures sensitive to discriminations between minimal and narrative aspects of the self. These independent behavioral measures may then be associated with underlying connectivity and inform existing theories of SN and DMN contributions to ego dissolution. This will help determine network-dependent changes to consciousness under psychedelics with greater accuracy.