A scoping review for building a criticality-based conceptual framework of altered states of consciousness — TEI-tagged view

Charles GervaispersName, Louis-Philippe BoucherpersName, Guillermo Martinez VillarpersName, UnCheol LeepersName, Catherine DuclospersName (2023date). “A scoping review for building a criticality-based conceptual framework of altered states of consciousness.” Frontiers in Systems Neurosciencetitle·j.

Object of study — empirical measures of brain criticalityterm across pathological and non-pathological altered states of consciousnessterm

Research question — whether altered states of consciousnessterm systematically deviate from brain criticalityterm and whether direction of deviationterm·var can support diagnosis or treatment

Contribution — maps heterogeneous evidence across seven altered-state categories and proposes a criticality-based framework for characterization and personalized neuromodulation

Full text — tagged

The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.

A scoping review for building a criticality-based conceptual framework of altered states of consciousnessterm Abstract The healthy conscious brain is thought to operate near a critical stateterm, reflecting optimal information processing and high susceptibility to external stimuli. Conversely, deviations from the critical stateterm are hypothesized to give rise to altered states of consciousnessterm (ASC). Measures of criticalityterm could therefore be an effective way of establishing the conscious state of an individual. Furthermore, characterizing the direction of a deviation from criticalityterm may enable the development of treatment strategies for pathological ASC. The aim of this scoping review is to assess the current evidence supporting the criticalityterm hypothesis, and the use of criticalityterm as a conceptual framework for ASC. Using the PRISMA guidelines, Web of Science and PubMed were searched from inception to February 7th 2022 to find articles relating to measures of criticalityterm across ASC. N = 427 independent papers were initially found on the subject. N = 378 were excluded because they were either: not related to criticalityterm; not related to consciousnessterm; not presenting results from a primary study; presenting model data. N = 49 independent papers were included in the present research, separated in 7 sub-categories of ASC: disorders of consciousnessterm (DOC) ( n = 5); sleepterm ( n = 13); anesthesiaterm ( n = 18); epilepsy ( n = 12); psychedelicsterm and shamanic state of consciousnessterm ( n = 4); deliriumterm ( n = 1); meditative state ( n = 2). Each category included articles suggesting a deviation of the critical stateterm. While most studies were only able to identify a deviation from criticalityterm without being certain of its direction, the preliminary consensus arising from the literature is that non-rapid eye movement (NREM) sleepterm reflects a subcritical stateterm, epileptic seizuresterm reflect a supercritical stateterm, and psychedelicsterm are closer to the critical stateterm than normal consciousnessterm. This scoping review suggests that, though the literature is limited and methodologically inhomogeneous, ASC are characterized by a deviation from criticalityterm, though its direction is not clearly reported in a majority of studies. Criticalityterm could become, with more extensive research, an effective and objective way to characterize ASC, and help identify therapeutic avenues to improve criticalityterm in pathological brain states. Furthermore, we suggest how anesthesiaterm and psychedelicsterm could potentially be used as neuromodulation techniques to restore criticalityterm in DOC.

Introduction Though scientists and philosophers have not established a single definition of consciousnessterm, it is generally agreed that conscious beings are aware of themselves, their surroundings, and their own perception. As such, consciousnessterm is thought to require the dual capacity for wakefulnessterm (arousal) and awarenessterm of oneself and the environment ( Blume et al., 2015 ). In healthy individuals, wakefulnessterm and awarenessterm covary across sleepterm and wake states, whereas pathological states of consciousnessterm, such as disorders of consciousnessterm (DOC), may reflect a dissociation of the two.

The assessment of human consciousnessterm generally relies on an individual’s willingness or capacity to reliably respond to the external environment at the time of assessment. However, consciousnessterm can exist in unresponsive individuals ( Owen et al., 2006 ; Sanders et al., 2012 ), and behavioral assessments have been shown to yield low diagnostic sensitivity in unresponsive individuals ( Schnakers et al., 2009 ) suggesting they are insufficient to i) detect capacity for consciousnessterm, and ii) probe the neural processes capable of sustaining consciousnessterm. Major theories of consciousnessterm disagree on the key neural substrates of conscious awarenessterm ( Lamme, 2010 ; Tononi et al., 2016 ; Mashour et al., 2020 ), and to this day, no single brain region has been unequivocally identified as the seed of consciousnessterm. Alternatively, consciousnessterm “might emerge through the complex interactions of spatially and temporally distributed brain functions” ( LeepersName and Mashour, 2018 ). In this perspective, the brain can be conceived as a complex network of structurally and/or functionally interacting components. Recent projects in neuroimaging and machine learning have produced highly accurate methods of assessing and predicting consciousnessterm in unresponsive individuals ( Logi et al., 2011 ; Casali et al., 2013 ; Stender et al., 2014 , 2015 ; Wu et al., 2015 ; Chennu et al., 2017 ; Song et al., 2018 ; DuclospersName et al., 2022 ). However, the majority of these indices rely on specialized technologies (e.g., fMRI, TMS) that have contraindications for many pathologically-unconsciousness individuals, preventing their widespread adoption for the assessment of consciousnessterm. Consequently, there is a critical need to understand which functional properties of the brain give rise to human consciousnessterm, and to develop accessible technologies for detecting consciousnessterm using bedside techniques.

The criticality hypothesis of consciousnessterm proposes that a healthy conscious brain self-organizes into a critical stateterm, or critical pointterm ( Wilting and Priesemann, 2019 ), which optimizes information processing ( Shew and Plenz, 2013 ; Scott et al., 2014 ; Fagerholm et al., 2016 ; Kim and LeepersName, 2019 ). Conversely, deviations from this critical stateterm are hypothesized to give rise to both pathological (e.g., coma) and non-pathological (e.g., sleepterm) altered states of consciousnessterm (ASC), as well as a reduction in the brain’s optimal information processing capacity. The criticalityterm hypothesis therefore offers a potentially unifying theory of the functional mechanisms underlying consciousnessterm and provides quantifiable empirical measures that can be applied to the brain. However, the application of criticalityterm to neuroscience remains recent, and it is unclear whether the criticalityterm hypothesis is currently supported by the scientific literature.

The concept of criticalityterm was originally introduced in statistical mechanics to study phase transitionsterm. A phase transitionterm is a change in the state of a systemterm, such as from liquid to gas, that occurs when a certain critical value of a parameter, such as a temperature, is reached. Criticalityterm is a state of a systemterm that is on the verge of a phase transitionterm ( Cocchi et al., 2017 ; Muñoz, 2018 ). At a critical pointterm, small fluctuations in the system are not damped, but instead, propagate throughout the system at all spatial and temporal scales. This results in scale-free (power-law) fluctuations in both the spatial and temporal domains, as well as high susceptibility to external stimuli. For example, when water is at its freezing point, a small change in temperature can cause the entire system to freeze or melt. This is because the system is at a critical pointterm, and small fluctuations can propagate throughout the system and cause a large-scale change in state. The state of a systemterm in which the control parameter is below a critical pointterm is called a subcritical stateterm. In contrast, the state of a systemterm in which the control parameter is above a critical pointterm is called a supercritical stateterm. Different phase transitionsterm can be investigated depending on the control parameter of interest, such as at the edges of chaos, synchrony, or avalanche. Criticality measuresterm, such as spatial or temporal correlation, complexity, synchronization, or neuronal avalancheterm, can be used to determine the system state (subcritical, critical, or supercritical). In the context of biological systems, criticalityterm in the brain confers many functional advantages that may establish a functional foundation for the emergence of consciousnessterm. These advantages include large computational capability with temporal correlation (memory) and long-range correlation (efficient brain integration), high flexibility to adapt to a changing environment from scale-invarianceterm, and wide repertoires of brain states from large fluctuations. Therefore, recent studies have proposed brain criticalityterm as a necessary condition for the emergence of consciousnessterm, and deviations from the critical stateterm (i.e., after a transition to one of the possible states in a sub- or supercritical stateterm) are hypothesized to give rise to ASC ( Kim and LeepersName, 2019 ; LeepersName et al., 2019 ; Kim M. et al., 2021 ).

An overview of measures of criticalityterm in the brain Quantitative evaluation of criticalityterm in the brain is crucial to determine whether a brain is operating near or far from a critical stateterm. The measures are developed primarily to quantify the typical system characteristics near criticalityterm: scale-invarianceterm (e.g., power-law of avalanche sizes/intervals, hurst exponent, branching parameter), large spatiotemporal correlations (e.g., long-range temporal correlationsterm, autocorrelation), and large fluctuations (e.g., detrended fluctuation analysis).

One of the most common methods used is the analysis of neuronal avalanchesterm. The basis of this method lies in the assumption that neuronal networks work in a similar fashion to nuclear chain reactions and earthquakes, where a simple unit reaches a threshold and then propagates its activity to the rest of the system, thereby initiating an “avalanche”. These avalanches are known to follow a power-law distribution which can be used to determine criticalityterm by evaluating the balance between rapidly dying activity and amplified activity over time ( Beggs and Plenz, 2003 ; Beggs and Timme, 2012 ). Subsequently, it is possible to extract the power-law exponent of the neuronal avalancheterm by plotting the power-law dataset into a log-log graphic that will then produce a straight line with the slope being the exponent.

Another group of methods consist of the analysis of long-range temporal correlationsterm, which focuses on the temporal structure of neuronal oscillations ( Linkenkaer-Hansen et al., 2001 ). Different sub-methods exist for measuring these long-range temporal correlationsterm, such as the autocorrelation function, which measures the degree at which a signal is similar to itself over time. Long-range temporal correlationsterm are confirmed if the autocorrelation function decreases, as a function of time, according to a power-law with an exponent between -1 and 0, showing that even though the signal is changing over time, a certain portion stays coherent to the past signal ( Meisel et al., 2017a ). However, autocorrelation function is said to be greatly affected by trends and shows a disturbingly high amount of noise, especially for large time lags ( Meisel et al., 2017a ).

The next method, detrended fluctuation analysis, was introduced to resolve the issues of the autocorrelation function. Detrended fluctuation analysis produces a scaling exponent from the amplitude envelope of an oscillatory signal that, when between 0.5 and 1, indicates the presence of long-range temporal correlationsterm in the time signal studied ( Hardstone et al., 2012 ). Another popular method for measuring long-range temporal correlationsterm is the Hurst exponent, which reflects the change in autocorrelation in a time series. A Hurst exponent between 0.5 and 1 confirms that long-range temporal correlationsterm are present ( Witton et al., 2019 ).

The last method described here for measuring long-range temporal correlationsterm, is the pair correlation functionterm, a measure of the variance of order parameters (i.e., phase coherence) between oscillators in a network ( Yoon et al., 2015 ). Pair correlation functionterm is maximal at a critical pointterm, presenting a long-range correlation and high network susceptibility to external and internal stimulation ( Yoon et al., 2015 ). However, the pair correlation functionterm diminishes as a system deviates from criticalityterm ( Kim and LeepersName, 2019 ).

Another method used in the criticalityterm literature is the branching parameter mterm , a stochastic description of the propagation of action potentials in a neuronal network. The branching parameter mterm stands for the mean number of subsequent spikes triggered after applying an extra spike to an excitatory neuron in the neuronal network. An m of 1 is found in the critical stateterm while a greater m suggests a supercritical stateterm and a lower m a subcritical stateterm ( Priesemann et al., 2019 ).

Aim The criticalityterm hypothesis could be tested to elucidate the functional brain dynamics that underly various states of pathological and non-pathological consciousnessterm. Considering that the brain of a healthy, awake subject is likely to operate in a lightly subcritical stateterm ( Priesemann et al., 2014 ), measuring the deviations from criticalityterm could also be a promising way to refine the diagnosis and prognosis of pathological ASC such as DOC ( Priesemann et al., 2013 ; Atasoy et al., 2017 ; Fekete et al., 2018 ; Hagemann et al., 2021 ; O’Byrne and Jerbi, 2022 ). A nuanced understanding of criticalityterm and its association to consciousnessterm may also provide insight into the potential therapeutic avenues that could help restore criticalityterm, and possibly consciousnessterm, in pathological ASC.

Our aim in this review is to determine the scope of the literature on criticalityterm in ASC, and to identify the evidence currently supporting the criticalityterm hypothesis in these states of consciousnessterm, in both humans and animals. Although some reviews already exist concerning brain criticalityterm and its applications to consciousnessterm ( Zimmern, 2020 ; Walter and Hinterberger, 2022 ), this review aims to evaluate how brain criticalityterm changes throughout alterations in consciousnessterm across humans and animals, and to provide a conceptual framework for the implementation of criticalityterm as a tool for evaluation and treatment of ASC.

Materials and methods Design Considering the broad and heterogeneous study subject, and the presence of multiple subtopics in the field of consciousnessterm, which was unamenable to a systematic reviewterm·mth, we conducted a scoping review of the literature. This study followed the Preferred Reporting Items for Systematic Reviewterm·mth and Meta-Analyses extension for scoping reviews (PRISMA-ScR) ( Tricco et al., 2018 ), and focused its search on published studies available through open access or academic research databases.

Search strategy Two databases (Web of Science, PubMed) were searched from inception to the 7th of February 2022 to find articles relating to measures of criticalityterm across different ASC. The following keywords were used during the search of the two databases: “criticalityterm sleepterm”, “criticalityterm coma”, “criticalityterm unresponsive wakefulnessterm syndrome”, “criticalityterm sedation”, “criticalityterm anesthesiaterm”, “criticalityterm disorders of consciousnessterm”, “criticalityterm epileptic seizureterm”, “criticalityterm meditationterm”, “criticalityterm deliriumterm”, “criticalityterm rem”, “criticalityterm nrem”, “criticalityterm minimally conscious state”, “criticalityterm consciousnessterm”, “criticalityterm vegetative state”, “criticalityterm psychedelicterm”, “criticalityterm eeg” and “criticalityterm fmri”. To ensure the maximum number of articles on the subject were found, the searches were automatically upgraded by the respective algorithm of each database (see Table 1 for the complete details of the search in PubMed). In all databases, the search was limited to studies published in English or French.

Complete electronic research of the PubMed database.

Inclusion and exclusion criteria All primary studies including case reports, case studies, clinical trials, cross-sectional studies and randomized controlled trials exploring criticalityterm in an altered state of consciousnessterm (i.e., DOC, sleepterm, anesthesiaterm, epilepsy, psychedelicsterm and shamanic state of consciousnessterm, meditationterm, deliriumterm) were considered for this review. Reviews, commentaries, opinion papers and book chapters were not considered for inclusion.

Considering that the degree of evidence of criticalityterm in ASC is still arguably low, a broad population of interest was agreed on for this study. Studies looking at brain criticalityterm in humans and animals were included. Studies on criticalityterm in healthy conscious wakefulnessterm or during cognitive tasks were excluded in both humans in animals. Studies looking at modelized brain data were also excluded. No other exclusion criteria were applied.

To be included in this review article, studies had to look at brain criticalityterm in ASC. No limitations were placed on the methods used to measure brain criticalityterm as long as the method was recognized and well documented to measure criticalityterm.

Study selection Following the search, all identified studies were uploaded into Covidence ® for duplicate elimination and further screening. Subsequently, article titles and abstracts generated from the initial search were independently assessed for eligibility by 3 members of the review team (CD, LPB, GMV) based on the inclusion and exclusion criteria mentioned previously. Two votes were necessary for an article to go through the next stage of selection. The full texts of remaining articles were independently examined by CD and LPB to reach a final list of articles. Disagreements at either screening stage were resolved through discussion with the supervisor of the study (CD). Consensus was reached for all included articles. The reasons for study exclusion were documented ( Figure 1 ).

PRISMA-ScR flow chart. This flow chart depicts the flow stages of the screening process and articles retrieved, screened and included at various stages of the review process.

Data extraction The review team identified the main areas of interest as follows: (1) title; (2) authors; (3) publication date; (4) DOI; (5) type of study; (6) goal of the study; (7) criticality measuresterm used; (8) methodology; (9) main results. For each study, relevant data were extracted using a customized data extraction form in a Microsoft Excel spreadsheet. All data were extracted by LPB and verified by CD and CG for accuracy and quality.

Results A total of N = 427 independent papers were initially found on the topic of interest. N = 378 were excluded because they were either: not related to criticalityterm; not related to consciousnessterm; not presenting results from a primary study; presenting modelized data rather than actual data from human or animal recordings. Finally, n = 49 independent papers were included in the present review, separated in 7 sub-categories of ASC: sleepterm ( n = 13); anesthesiaterm ( n = 18), DOC and coma, including vegetative state (or unresponsive wakefulnessterm syndrome) and minimally conscious state ( n = 5); epilepsy ( n = 12); psychedelicsterm and shamanic state of consciousnessterm ( n = 4); deliriumterm ( n = 1); meditationterm ( n = 2). Some articles ( Ribeiro et al., 2010 ; Liu et al., 2014 ; LeepersName et al., 2019 ; Fekete et al., 2021 ; Kim M. et al., 2021 ) were included in more than 1 sub-category as they presented results concerning various ASC. N = 33 studies were conducted in human participants, n = 15 were conducted in animal subjects and n = 1 study had both animal and human subjects. Reported results are from a human population unless mentioned otherwise.

Characteristics of included studies The literature search yielded articles published from October 2002 to February 7 th 2022, with the majority ( n = 27, 55%) being published in the last 5 years. The years 2020 and 2021 have seen the most publications ( n = 6 for 2020 and n = 7 for 2021, together representing 26.5% of all articles published) after a constant increase from the years prior, showing a growing interest in the application of criticalityterm to ASC ( Figure 2 ).

Year of publication of articles relating to criticalityterm in ASC. The number of articles on criticalityterm in an altered state of consciousnessterm (ASC) published per year is depicted by the vertical blue bars. The exponential trend line is depicted in dotted red.

The articles included in this review used various methods of measurement of criticalityterm. The two most common were the method of neuronal avalancheterm ( n = 21, 43%) and long-range temporal correlationsterm ( n = 13, 27%). Other methods used are the pair correlation functionterm ( n = 3, 6%), a stability parameter ( n = 2, 4%), a branching parameter mterm ( n = 1, 2%) and phase lag entropy ( n = 1, 2%). N = 8 articles (16%) reported methods that were less common and more personalized to the research team that used them ( Table 2 ).

Methods used to measure criticalityterm in articles included in this review.

In the following sections, we will briefly introduce each ASC and present articles found within this scoping review.

Sleepterm All animal species experience daily but reversible loss in their consciousnessterm through their sleepterm. During sleepterm, wakefulnessterm is entirely suspended and awarenessterm is partially, but not completely dissolved. For example, both internal (e.g., pain) and external (e.g., alarm clock) signals can interrupt sleepterm, reflecting some level of awarenessterm. Sleepterm is composed of rapid-eye movement sleepterm (REM), and non-rapid-eye movement sleepterm (NREM), which can further be subdivided into 3 stages (N1, N2, N3) reflecting sleepterm depth, and the degree to which awarenessterm is temporarily lost. Each sleepterm stage has distinct electrophysiological characteristics that can be detected by EEG. Sleepterm depth is generally quantified by the density of slow oscillations within the brain, reflecting synchronized activity between large populations of neurons ( Niethard et al., 2018 ). Although normal sleepterm is not considered to be a conscious state, some sleepers may experience some forms of consciousnessterm during sleepterm. This is true in lucid dreaming, where the sleeper becomes aware that their perception is a dream, and may have some degree of control over dream elements. It is also the case for somnambulism, where the sleeper appears awake, but sensory perceptions are drastically decreased, such that sight, sounds, smell and pain are not fully perceived and experiences are not carried into the waking state ( Popat and Winslade, 2015 ). Sleepterm is essential to overall health, learning, plasticity and neurogenesis ( Walker and Stickgold, 2004 ; Tononi and Cirelli, 2014 ), and is likely to play a direct role in recovery from illness and injury.

In human and animal studies, criticalityterm does not seem to be lost in sleepterm in general, except for NREM sleepterm, where data suggests a deviation from the critical stateterm ( Table 3 ). In three different studies that used detrended fluctuation analysis to measure long-range temporal correlationsterm, it was found that deep sleepterm comes with a deviation from criticalityterm ( Allegrini et al., 2013 , 2015 ; Paradisi et al., 2013 ), and, using Hurst exponent in fMRI, another study suggests that deep sleepterm is significantly associated with an exponentially decaying autocorrelation, meaning an even more subcritical brain than what is seen in wakefulnessterm ( Tagliazucchi et al., 2013 ). In rats, the same deviation from criticalityterm in NREM sleepterm was found multiple times, whether using detrended fluctuation analysis to measure long-range temporal correlationsterm ( Wang et al., 2019 ; Lombardi et al., 2020 ) or by looking at neuronal avalanchesterm ( Freeman et al., 2006 ; Ribeiro et al., 2010 ). These results were also reproduced in humans using the autocorrelation function to measure long-range temporal correlationsterm ( Meisel et al., 2017b ). Another evidence supporting that NREM sleepterm is associated with a deviation from criticalityterm comes from a fMRI study that looked at the differences in the power-law of neuronal avalanchesterm in every sleepterm stage and found that N2 induced the greatest deviation from power-laws found in wakefulnessterm ( Bocaccio et al., 2019 ). Although some studies seem to be converging on this topic, contradicting results were seen in a single study where NREM sleepterm was found to restore criticalityterm by bringing dynamics closer to the critical pointterm ( Priesemann et al., 2013 ). Also, while studying the effect of sleepterm deprivation on neuronal avalanchesterm and criticalityterm, prolonged wakefulnessterm was found to create a deviation from criticalityterm ( Freeman et al., 2006 ; Meisel et al., 2013 ). Sleepterm after these long bouts of wakefulnessterm was discovered to restore criticalityterm in the brain, pointing in the direction that sleepterm is involved in the self-organization of criticalityterm ( Freeman et al., 2006 ; Meisel et al., 2013 ). On the other hand, using local field potentials, an invasive recording method consisting of the insertion of micro-electrodes directly into the cerebral cortex, a study was not able to confirm any signs of power-law scaling while analyzing the neuronal avalanchesterm recorded in the motor cortex of cats and monkeys, as well as the premotor and temporal cortex of epileptic humans during slow wave sleepterm as well as REM sleepterm ( Dehghani et al., 2012 ).

Results of studies on criticalityterm in sleepterm.

Anesthesiaterm Anesthesiaterm is a pharmacologically-induced and reversible state of unconsciousness. For surgical procedures, balanced anesthesiaterm aims to create a synergistic state of analgesia, hypnosis, amnesia and muscle paralysis. Contrary to sleepterm, anesthesiaterm is thought to be marked by a complete loss of both wakefulnessterm and awarenessterm, with a complete absence of responsiveness to commands and to pain. On the EEG, anesthesiaterm is typically characterized by slow oscillations, similar to slow-wave (N3) sleepterm ( Chander et al., 2014 ; Purdon et al., 2015 ). When used in lighter doses, anesthetic agents can induce sedation, ranging from minimal (e.g., drowsiness and relaxation) to deep (e.g., purposeful response to painful stimulation only). Anesthesiaterm is often used as a model of unconsciousness and has been shown to reduce higher-order information processing, while not affecting primary sensory cortices ( Hudetz and Mashour, 2016 ). Though previous studies suggest that both sleepterm and anesthesiaterm may share common mechanisms, such as alterations in cortico-cortical and thalamocortical connectivity ( Larson-Prior et al., 2009 ; Stamatakis et al., 2010 ; Heine et al., 2012 ), a lack of consensus persists regarding the neural mechanisms by which anesthetic agents induce unconsciousness.

A wide range of anesthetics have been used to study brain criticalityterm in humans under anesthesiaterm, and all results point in the direction that anesthesiaterm induces a deviation from the critical stateterm ( Table 4 ). Studies using isoflurane and ketamine show a reduction in pair correlation functionterm ( Kim M. et al., 2021 ), as well as a reduction in phase lag entropy ( LeepersName et al., 2019 ), meaning a deviation from criticalityterm. Studies looking at general anesthesiaterm under propofol show similar results. It was found that propofol induces greater stability properties in the neuronal system, which can be interpreted as a deviation from criticalityterm considering that a recent model exhibiting complex spatio-temporal dynamics proposed that criticalityterm emerges from dynamical instability ( Alonso et al., 2014 ). Another study showed that the structure-function relation of the large scale intrinsic connectivity network is stronger under propofol anesthesiaterm than what is seen in an awake resting state without anesthesiaterm, which also indicates a deviation from criticalityterm, as it has been shown that the functional pattern is less dependent on the underlying structural network at the critical stateterm ( Stramaglia et al., 2017 ). Another study that characterized criticalityterm during propofol anesthesiaterm in humans showed a decrease in complexity (as measured by Jensen-Shannon divergence) and a reduction of long-range temporal correlationsterm at loss of consciousnessterm, still pointing in the direction of a deviation from criticalityterm ( Pullon et al., 2022 ). Lastly, while trying to create a parallel between criticalityterm and the Integrated Information Theory ( Tononi, 2004 ), one group showed that, as a subject loses consciousnessterm after injection of sevoflurane, criticalityterm measured by pair correlation functionterm and Φ, a surrogate metric of integrated information, both decrease, yielding further evidence that anesthesiaterm induces a deviation from the critical stateterm in humans ( Kim and LeepersName, 2019 ). Lastly, a single study was found on criticalityterm during propofol sedation. This study showed that the scale-free distributions of node size and degree (size being the number of contiguous voxels sharing a hemodynamic profile and degree being the number of node connections) were maintained before, during and after sedation, suggesting that criticalityterm is maintained during sedation ( Liu et al., 2014 ).

Results of studies on criticalityterm in anesthesiaterm.

Several studies have also investigated criticalityterm in anesthetized animals, and contradicting results seem to emerge. Three studies investigated neuronal avalanchesterm in mice under pentobarbital, and all showed a deviation from the critical stateterm, either through an absence of power-law dynamics in neuronal avalanchesterm ( Scott et al., 2014 ; Fagerholm et al., 2016 ), or a reduction of complexity ( Fagerholm et al., 2018 ). In studies investigating criticalityterm in cats, monkeys and rats under various anesthetics (cats: medetomidine and ketamine; rats: urethane; monkeys: ketamine, diazepam and remifentanil), neuronal avalanchesterm showed a deviation from criticalityterm by losing their power-law properties under anesthesiaterm ( Fekete et al., 2018 ). However, another study looking at the effects of isoflurane on cats showed dynamical changes in the visual cortex suggesting that anesthesiaterm may induce a state closer to criticalityterm than what is measured in wakefulnessterm ( Hahn et al., 2017 ). Contradicting results can also be found in studies on rats. While under the effects of anesthesiaterm using ketamine-xylazine and analyzing neuronal avalanchesterm, it was found that the avalanches were losing their power-law properties ( Ribeiro et al., 2010 ), while another study using desflurane showed the contrary: a preservation of power-law properties followed by a deviation from criticalityterm during the recovery of consciousnessterm ( Hudetz et al., 2016 ). While measuring criticalityterm using detrended fluctuation analysis and stability of neuronal dynamics in monkeys under propofol, ketamine or medetomidine, other studies have showed a decrease in long-range temporal correlationsterm ( Krzemiński et al., 2017 ) and a stabilization of neuronal dynamics ( Solovey et al., 2015 ), both indicators of a deviation from criticalityterm. Conversely, a more recent study on monkeys showed that propofol drastically reduced the size and duration of neuronal avalanchesterm, while ketamine induced more wake-like dynamics ( Varley et al., 2020b ). Finally, a study using a statistical models incorporating a scaling parameter on fMRI brain measures of monkeys under either propofol, ketamine or sevoflurane showed that brain dynamics under anesthesiaterm were supercritical when compared to resting brain activity ( Ponce-Alvarez et al., 2022 ). The authors conclude that anesthetics have a disconnecting effect on particular brain regions (i.e., insular, cingulate, and parietal cortices), making them less potent at transmitting information to the system as a whole. Overall, the current literature in animal models of anesthesiaterm shows contradictory results, and the anesthetics used in these studies seems to have varying effects on brain criticalityterm.

Coma and DOC Coma and DOC are typically caused by a severe brain injury or brain dysfunction. Coma is a deep state of unconsciousness that lacks both wakefulnessterm and awarenessterm, and in which responses to stimulation and pain are completely absent ( Blume et al., 2015 ). In DOC, however, awarenessterm is lacking despite the preserved capacity for wakefulnessterm, embodying a disconnect between wakefulnessterm and awarenessterm. In unresponsive wakefulnessterm syndrome (also known as vegetative state), patients are able to awaken (eye opening), but show no behavioral signs of being aware of themselves or their surroundings, therefore lacking goal-oriented or willful behaviors ( Multi-Society Task Force on PVS, 1994 ). As such, patients in unresponsive wakefulnessterm syndrome are considered to be unconscious. Minimally conscious state presents with eye opening (wakefulnessterm) and some reproducible, though minimal, oriented and/or willful behaviors (e.g., visual tracking, inconsistent command following) ( Giacino et al., 2002 ). Coma and DOC can be temporary, but the neural and functional mechanisms by which their reversal is possible are unknown. In this population, assessing conscious awarenessterm and establishing a prognosis for recovery in the absence of behavioral responsiveness are fundamental shortcomings of clinical practice. Ultimately, improving DOC patient diagnosis and prognosis are necessary to meaningfully contribute to clinical management and decision-making, leading to better-adapted rehabilitative strategies, and improved outcomes.

Even though criticalityterm could be an interesting way to refine the diagnosis and prognosis of DOC, a very restricted number of studies have investigated this subject ( Table 5 ). Most studies regrouped DOC as a single category, all of which showed a deviation from criticalityterm using neuronal avalanchesterm ( Fekete et al., 2021 ), pair correlation functionterm ( Kim M. et al., 2021 ), phase lag entropy ( LeepersName et al., 2019 ) or dimensionality (rate of decay in information transfer in regard to distance) ( Abeyasinghe et al., 2020 ). A single study focused specifically on unresponsive wakefulnessterm syndrome, and described the power-law distributions of network hubs. This study showed that the network loses its power-law properties in unresponsive wakefulnessterm syndrome, indicating a deviation from criticalityterm. This also represents a potential marker to be included in clinical practices relating to diagnosis of unresponsive wakefulnessterm syndrome ( Liu et al., 2014 ).

Results of studies on criticalityterm in coma and DOC.

Seizures and epilepsy Epilepsy is a brain condition characterized by a lasting predisposition to generate spontaneous seizures ( Thijs et al., 2019 ). The development of epilepsy is said to result from an imbalance between excitatory and inhibitory activity in a particular neuronal network, disrupting neuronal processing and potentially affecting other neuronal networks ( Fisher et al., 2005 ). Epilepsy can be defined as generalized when the epileptogenic networks are distributed across the whole brain and involve thalamocortical structures bilaterally, or can be defined as focal, where the epileptogenic network is usually located at the limbic or neocortical level in a single hemisphere ( Fisher et al., 2005 ). A considerable part of research on epilepsy focuses on the localization of the epileptic focus of debilitating, pharmaco-resistant epilepsy to eventually proceed to a surgical intervention aiming at removing the epileptic focus ( Rugg-Gunn et al., 2020 ). Considering that these gravely epileptic patients often have to go through an EEG evaluation prior to their operation ( Chen and Koubeissi, 2019 ), several studies have investigated criticalityterm in epilepsy ( Table 6 ).

Results of studies on criticalityterm in epileptic seizuresterm.

Results are presented in 3 categories: intraictal (during seizure), interictal (between seizures), and preictal (before seizure) recordings. Firstly, studies investigating intraictal criticalityterm properties unanimously suggest that the brain is in a supercritical stateterm during a seizure. Greater long-range temporal correlationsterm, measured by the Hurst exponent, were seen in intraictal brain activity, suggesting a supercritical stateterm ( Yan et al., 2016 ). While looking at the power-law of phase-locking intervals (PLI), a measure of synchronicity in oscillators, it was shown that the distributions were deviating from their power-law properties during a seizure in the direction of a supercritical stateterm ( Meisel et al., 2012 ). A study on rats has shown similar results, pointing out that seizures could be interpreted as “a generalized kind of phase transitionterm analogous to a kind of critical pointterm” ( Li et al., 2005 ) considering the presence of strong long-range temporal correlationsterm and an increase in neuronal avalanchesterm.

During interictal recordings, two studies showed that the brains of epileptic patients operate in a critical stateterm, considering that neuronal avalanchesterm follow a power-law ( Cranstoun et al., 2002 ; Worrell et al., 2002 ). Another study that investigated the branching parameter obtained from intracranial electrodes showed the absence of significant differences in criticalityterm between the epileptic focus and its contralateral counterpart in the brain during interictal recordings ( Hagemann et al., 2021 ). Conversely, while comparing the LRTCs located in the epileptic focus to the long-range temporal correlationsterm present in the symmetrically contralateral region of the brain, two studies showed that long-range temporal correlationsterm, measured by detrended fluctuation analysis, were stronger in the epileptic focus during interictal recordings, thus indicating a supercritical stateterm in that region ( Parish et al., 2004 ; Monto et al., 2007 ). An important concept in epilepsy that was also studied under the scope of criticalityterm is interictal epileptiform activity (IEA), an abnormal brain activity without apparent seizure ( Arviv et al., 2016 ). Using MEG, it was found that neuronal avalanchesterm were greatly increased during these periods of IEA, suggesting that the brain, and especially the potential epileptic focus, was in a supercritical stateterm ( Arviv et al., 2016 ).

In terms of preictal activity, one study showed that preictal activity shows slightly subcritical traits, similar to that of wakefulnessterm in the healthy brain ( Hagemann et al., 2021 ). Another study that used a parameter of fluctuation of electropathophysiological activity to measure criticalityterm suggests that preictal brain activity was in a critical stateterm ( Cerf et al., 2004 ). However, authors did not compare preictal activity to interictal activity for all patients, making it difficult to establish whether the critical stateterm they claim to observe is distinct from the interictal brain or the healthy brain. Another study showed that preictal neuronal avalanchesterm deviate from a power-law distribution, suggesting a deviation from criticalityterm without specifying the direction ( Fekete et al., 2021 ).

Psychedelicsterm and shamanic state of consciousnessterm Psychedelicsterm are psychoactive drugs known for their capacity to trigger deeply profound existential experiences ( Carhart-Harris, 2019 ). Though psychedelicsterm seem to have had their greatest impact on Western culture starting in the 1960s, potential new medical applications of psychedelicsterm in the fields of psychiatry and neurology have generated a revival of interest for these substances in the last decade. As agonists of 5-HT2A serotoninergic receptors, psychedelicsterm are said to induce greater plasticity at the cortical level, where 5-HT2A receptors are most densely concentrated ( Beliveau et al., 2017 ; Kraehenmann et al., 2017 ). The fact that they operate at the cortical level, said to be the basis of higher-level operations, is coherent with the fundamental changes in consciousnessterm reported by psychedelicterm users. Interestingly, the shamanic state of consciousnessterm is said to induce behavioral and conscious aspects similar to those seen under the influence of psychedelicsterm, such as mystical experiences and feelings of disembodiment ( Krippner, 2000 ; Winkelman, 2013 ; Sweeney et al., 2022 ).

Though only a few studies have investigated the effects of psychedelicsterm on criticalityterm using varying methodological approaches, the current state of the literature seems to suggest that psychedelicsterm improve criticalityterm in the brain ( Table 7 ). One study looking at the ensemble of individual harmonic brain states under LSD found that the power-law properties of some of the frequency specific harmonics were closer to criticalityterm than in the slightly subcritical resting brain ( Atasoy et al., 2017 ). Another study, looking this time at both LSD and psilocybin, used the fractal dimension of cortical brain activity in spatial and temporal domains to establish criticalityterm, considering that supporting evidence suggests that as a system approaches the critical pointterm, it starts to organize itself as a fractal system ( Watanabe et al., 2015 ; Varley et al., 2020a ). This research group showed that both psychedelicsterm drugs increased the fractal dimension of functional connectivity networks, therefore bringing the brain closer to criticalityterm ( Varley et al., 2020a ). A study looking at the effects of ketamine as a psychedelicterm (a dose insufficient for ketamine to be considered an anesthetic) found a significant reduction in the pair correlation functionterm, suggesting a deviation from criticalityterm ( Kim M. et al., 2021 ). A single study was found to describe the shamanic state and its impact on brain criticalityterm. Concerning brain criticalityterm during the shamanic state, results were strikingly similar to those seen during psychedelicterm usage; that is, pair correlation functionterm was higher in beta and gamma bands when compared to controls, indicating a state closer to criticalityterm than normal wakefulnessterm ( Huels et al., 2021 ).

Results of studies on criticalityterm in psychedelicterm and shamanic states consciousnessterm.

Deliriumterm Deliriumterm is an acute state of reduced awarenessterm with disturbances in attention and cognition, which usually emerges over a short period of time in vulnerable patients ( Hshieh et al., 2018 ). Even though the pathophysiological mechanisms of deliriumterm are still poorly understood, several biological and psychological factors such as dementia, history of alcohol abuse, psychoactive medication use and advanced age (> 70 years) are often involved in the development of this medical condition ( Hshieh et al., 2018 ). Deliriumterm can present various behavioral characteristics, such as in the hypoactive form where the patient is lethargic, in the hyperactive form where the patient is agitated and can sometimes present hallucinations, and a mixed form where the patient fluctuates between agitation and lethargy. The fact that deliriumterm presents such a broad range of behaviors makes it a difficult condition to diagnose making the study of criticalityterm in this ASC particularly interesting as it could be used as a neurophysiological diagnostic biomarker. Furthermore, given that deliriumterm is a serious medical complication with rapid onset, objective measures are necessary to better predict and prevent its onset.

Only a single criticalityterm study was found relating to deliriumterm ( Table 8 ), which investigated long-range temporal correlationsterm with autocorrelation function in delirious patients in post-operative care ( Kim H. et al., 2021 ). Authors suggest a deviation from criticalityterm in deliriumterm, due to reduced autocorrelation function in the alpha band in delirious patients compared to non-delirious patients. Interestingly, this study also showed that caffeine helped to prevent deliriumterm by bringing the cortical dynamics closer to the critical stateterm ( Kim H. et al., 2021 ).

Results of studies on criticalityterm in deliriumterm.

Meditationterm Meditationterm is a mindfulness practice, originally seen in the Buddhist and Hindu cultures and introduced more recently in the Western culture, where the individual intentionally increases their awarenessterm to thoughts and feelings ( Basso et al., 2019 ). Several forms of meditationterm exist, two of the most popular being the open monitoring meditationterm, where the practitioner tries to monitor in a non-judgmental way everything that spontaneously comes into conscious awarenessterm ( Davidson and Lutz, 2008 ; Basso et al., 2019 ), as well as the focused attention meditationterm, where the individual tries to concentrate its attention on a single object or movement (ex: its own breath) while ignoring internal and external distractions ( Yoshida et al., 2020 ). As described earlier, consciousnessterm is said to be composed of arousal as well as awarenessterm, making meditationterm (a practice where awarenessterm is increased), a sort of “super-conscious” state.

Research on brain criticalityterm in meditationterm is sparse ( Table 9 ), and all studies found in this review investigated criticalityterm in focused attention meditationterm. One study used detrended fluctuation analysis to measure long-range temporal correlationsterm in experienced practitioners of meditationterm and found a reduction in long-range temporal correlationsterm during meditationterm when compared to an eyes-closed resting control condition ( Irrmischer et al., 2018 ). Still, it is difficult to establish that this deviation from criticalityterm is purely due to the fact that the person is in a meditative state, as various studies have stated that experienced practitioners of meditationterm have permanent functional changes in their brain ( Cahn et al., 2010 ; Dürschmid et al., 2020 ). To answer this blind spot in the literature, another study isolated the effects of the meditative state by looking at the brain criticalityterm of novice practitioners of focused attention meditationterm who were instructed and guided through a meditationterm by an expert in the field. While using MEG data filtered in high frequency bands (> 100Hz) to look at neuronal avalanchesterm, they stated that the neuronal avalanchesterm of novices in a meditative state followed a power-law distribution with an exponent closer to the critical stateterm compared to mind-wandering controls and compared to a rest state ( Dürschmid et al., 2020 ).

Results of studies on criticalityterm in meditationterm.

Discussion This scoping review aimed to map the current state of the literature on criticalityterm across ASC, in order to appraise how the criticality hypothesis of consciousnessterm is currently supported in the literature. Firstly, we found that the majority of studies published to date focus on criticalityterm in anesthesiaterm, sleepterm and epilepsy. Importantly, our scoping review of the literature confirmed that scientific evidence remains relatively scarce and methodologically inhomogeneous. Despite this major limitation, we also found evidence suggesting that most ASC presented show a deviation from the critical stateterm, strengthening the hypothesis that deviations from criticalityterm accompany ASC. Even if the direction of this deviation was rarely characterized in the articles presented in this review, growing evidence supports that NREM sleepterm reflects a subcritical stateterm, epilepsy reflects a supercritical stateterm, and psychedelicsterm bring cortical dynamics closer to the critical stateterm than what is observed in the healthy awake brain, which is slightly subcritical (refer to Figure 3 for the visual presentation of the results of this review). We also found that neuronal avalanchesterm and long-range temporal correlationsterm were the most commonly used methods to measure brain criticalityterm in human and animal studies.

Directions of deviation from criticalityterm in various ASC. This figure is a synthesis and schematic representation of the current findings from the literature on criticalityterm in altered states of consciousnessterm (ASC). Colored circles illustrate the various ASC and their hypothetic position relating to the critical stateterm (or “critical pointterm”). ASC located on the left side of the critical pointterm have been shown (in minimally one study) to reflect a subcritical stateterm, whereas ASC on the right side of the critical pointterm have been shown (in minimally one study) to reflect a supercritical stateterm. Distance from the critical pointterm is schematic and not reflective of specific comparisons across ASC. Colored arrows represent a potential neuromodulation of criticalityterm that a specific ASC could induce on another ASC. The bell-shaped curve is used to demonstrate a perfect balance between integration and segregation at the critical pointterm. In the legend, question marks following the name of a particular ASC reflect that contradicting results have been found in the literature with regards to the subcritical/supercritical nature this ASC, and/or that findings remain sparse in this regard. Created with BioRender.com .

Although the study of brain criticalityterm in ASC has seen a surge of publications in recent years, some limitations still persist in the current literature. First, knowing that brain criticalityterm is a somewhat new concept in neuroscience, there is not yet a single definition or conceptual framework used to study criticalityterm, making comparison between articles difficult. Second, a plethora of methods exist to characterize criticalityterm in the brain, also rendering difficult the comparison of various studies. Indeed, two articles investigating criticalityterm in the same ASC but using a different measure could come to divergent results, simply due to varying methodological approaches. In order to create a reliable criticality-based conceptual framework of ASC, future studies should assess criticalityterm across a wide range of ASC using consistent methodology that can capture the amplitude and direction of deviationsterm·var from the critical stateterm. Third, the terminology related to the field of criticalityterm is both complex and vast, making it difficult to scope the entirety of the work related to the concept in the literature. Some studies have investigated criticality-related measures or concepts without ever using the term “criticalityterm”, rendering their papers less visible to the field, and undetected by the search strategy employed in the present review. Efforts should be made in the field of criticalityterm to use more uniform terminology, in order to make it easier to find studies of interest and to ensure more efficient discourse across studies. Fourth, the intricate terminology related to criticalityterm comes with downsides, such as the conflation of different concepts into a single term (e.g., avalanches used for neuronal avalanchesterm as observed in local field potentials, but also used in the context of power-law of oscillations, as seen in more “whole-brain” approaches such as EEG and fMRI). Mixing up data coming from whole-brain methodologies, such as EEG and fMRI, to more spatially localized methods such as local field potentials can also be somewhat problematic. Indeed, when using whole-brain approaches, criticalityterm becomes a one-dimensional measure applied to an extremely complex system. Evidently, this comes with a loss of information at both the spatial and temporal levels. It would be interesting to study criticalityterm as a fluctuating metric across these dimensions as it would be a more accurate depiction of the brain’s complex dynamics. Finally, the application of criticalityterm to the brain remains somewhat debated. Some have called into question the current methods used to assess criticalityterm in the brain, claiming they are insufficient to establish or distinguish criticalityterm in a multitude of systems, and pointing out the ubiquity of critical properties in varying neural systems with radically distinct properties ( Beggs and Timme, 2012 ; Beggs, 2015 , 2022 ; Destexhe and Touboul, 2021 ).

Criticalityterm as a framework for the diagnosis and neuromodulation of ASC The healthy awake brain seems to operate very near the critical stateterm ( Pearlmutter and Houghton, 2009 ; Beggs and Timme, 2012 ; Beggs, 2015 , 2022 ; Massobrio et al., 2015 ; Shi et al., 2022 ) rendering it most flexible and adaptable to its changing environment and processing demands. More specifically, it is thought to operate in a slightly sub-critical state, closer to chaoticity than synchrony ( Priesemann et al., 2013 , 2014 ; Fekete et al., 2018 ; Toker et al., 2022 ). Proximity to criticalityterm would enable the brain to benefit from most of the functional characteristics of criticalityterm (e.g., optimal information processing, ability to reorganize efficiently) whilst protecting itself from tipping off toward the supercritical side, associated with large-scale synchronized activity often seen in epileptic seizuresterm ( Shew and Plenz, 2013 ; Priesemann et al., 2014 ; Beggs, 2022 ). As seen in this review, the ability of psychedelicsterm to increase proximity to criticalityterm ( Beggs and Timme, 2012 ; Pullon et al., 2022 ), also suggests that healthy wakefulnessterm may in fact not be exactly at criticalityterm ( Atasoy et al., 2017 ; Varley et al., 2020b ).

Considering that the study of brain criticalityterm under anesthesiaterm has sprouted what seems to be the greatest amount of evidence across all ASC, it is reasonable to think that this has been done with various applications in mind. Indeed, in a scoping review on clinical applications of criticalityterm, Zimmern points out two distinct clinical applications of criticalityterm in anesthesiaterm. The first being a marker of depth of anesthesiaterm and the second being a predictive marker of recovery from a comatose state ( Zimmern, 2020 ). Given that having a reliable marker of unconsciousness under anesthesiaterm could help reduce the prevalence of intraoperative awarenessterm, a traumatizing event where a patient undergoing surgery is still aware of what is happening and can recall the events of the surgery ( Mashour et al., 2012 ), various studies have begun emerging relating criticalityterm to prediction of anesthetic depth ( Alonso et al., 2014 ; Krzemiński et al., 2017 ; Thiery et al., 2018 ). The prediction of recovery in DOC or comatose patients using criticalityterm is also an emerging area of research ( Zimmern, 2020 ). Future studies should look into more ways to clearly distinguish criticality-related traits between anesthesiaterm and DOC, to 1) help with the diagnosis and prognostication of DOC and 2) explore how sedation and/or anesthesiaterm could potentially improve criticalityterm in coma and DOC patients. Indeed, though highly counter-intuitive, anesthesiaterm (generally conceptualized as inducing a subcritical stateterm) could potentially be used to improve criticalityterm in DOC patients whose brains are in a supercritical stateterm. Given that recent evidence suggests that DOC patients can be either in subcritical or supercritical stateterm ( Toker et al., 2022 ), treatment would have to be personalized according to each patient’s deviation from the critical stateterm. As such, substances that are hypothesized to induce subcritical statesterm, such as sedatives/anesthetics, could be used to treat supercritical statesterm, while substances that improve criticalityterm (e.g., psychedelicsterm) could be used to treat pathological subcritical statesterm. Evidently, additional studies are needed to further characterize the direction of deviationsterm·var from criticalityterm in various ASC before exploring such hypotheses.

Evidence in epileptic seizuresterm (intraictal) seems to suggest a supercritical stateterm, though results pertaining to interictal and preictal activity in epileptic patients are still unclear. Additional evidence supporting the fact that epileptic seizuresterm reflect a supercritical stateterm can also be deducted from the clinical applications and short-term treatments of epileptic seizuresterm. Indeed, antiepileptic medications tend to induce a decrease in neuronal avalanchesterm as well as in long-range temporal correlationsterm, which suggests that their ability to decrease seizure occurrence may come from their ability to shift the brain toward a subcritical stateterm ( Meisel, 2020 ), as a way to counteract the supercritical functional properties of seizures. Treatment of refractory status epilepticus with strong doses of propofol may also reflect a similar phenomenon ( Rai and Drislane, 2018 ). Though various biomarkers for epilepsy have been studied to this day, most of them showed contradicting and inconsistent results and almost none of them used brain criticalityterm as a potential marker ( Pitkänen et al., 2019 ). Further studies should investigate in greater depth how brain criticalityterm fluctuates in epileptic patients, across pre-, intra-, and interictal activity, and in a baseline, resting state, compared to healthy controls.

While sleepterm has been more widely studied under the scope of criticalityterm, the preliminary consensus that seems to emerge from the current literature is that NREM sleepterm presents a deviation from criticalityterm, with preliminary results suggesting a deviation toward the subcritical side ( Tagliazucchi et al., 2013 ). The fact that REM sleepsterm seems to preserve brain criticalityterm is coherent with the fact that EEG activity during REM sleepterm resembles that seen in wakefulnessterm ( Simor et al., 2020 ), and that this sleepterm stage supports conscious-like experiences in sleepers, including non-lucid and lucid dreams ( Jing et al., 2016 ). As REM sleepterm can be referred to as the most “conscious” stage of sleepterm, N3 could be referred to as the most “unconscious” stage, with high prevalence of slow-wave activity and minimal behavioral responsiveness. Though this has yet to be unanimously confirmed, N3 could therefore be expected to show the greatest deviation from criticalityterm. Elucidating the functional characteristics and critical behavior across sleepterm stages could shed light on the restorative properties of sleepterm and its association to learning, memory, and neuroplasticity ( Walker and Stickgold, 2006 ). Additionally, it could contribute to further our understanding of the detrimental effects of sleepterm deprivation, or sleepterm disorders, on the brain.

Research around psychedelicsterm and criticalityterm is still in its early phases, but results seem unanimous in suggesting that psychedelicsterm bring cortical dynamics closer to the critical pointterm than in healthy wakefulnessterm. This finding could have important implications for the treatment of pathological ASC in which the brain operates in subcritical statesterm. As previously mentioned, a recent article published after our review of the literature seems to suggest that DOC patients could either be in a super- or subcritical stateterm ( Toker et al., 2022 ). This highlights the importance of evaluating the brains’ critical stateterm prior to choosing a potential course of treatment. When the brain of a DOC patient is in a subcritical stateterm, psychedelicsterm could potentially be used as a pharmacological neuromodulation technique, bringing the cortical dynamics closer to the critical stateterm. Indeed, psychedelicsterm are said to increase conscious content in awake individuals who experience a “psychedelicterm state” (Scott and Carhart-Harris, 2019 ). Also, as mentioned by Scott and Carhart-Harris (2019) , brain complexity (measured by Lempel-Ziv compressibility, a measure of criticalityterm) is positively correlated with states of awarenessterm (Scott and Carhart-Harris, 2019 ), and their current hypothesis is that psychedelic-induced increases in brain complexity could also increase conscious content of DOC patients, thus inducing a faster recovery. Seeing as the standard conceptualization of consciousnessterm includes two distinct mechanisms (i.e., wakefulnessterm or “arousal” and awarenessterm or “conscious content”), it could be argued that substances that stimulate arousal, like methylphenidate and other dopaminergic agents, could also optimize consciousnessterm recovery in DOC patients. Indeed, some preliminary results linking dopamine to DOC have started to emerge ( Ciurleo et al., 2013 ; Edlow et al., 2020 ; Spindler et al., 2021 ), but we argue that psychedelicsterm could be used either as a complement or as an alternative to these dopaminergic agents. Considering that arousal alone is not sufficient to induce a recovery from DOC, by potentially improving criticalityterm and awarenessterm, psychedelicsterm could be the key to effectively treating DOC patients. Further characterization of the criticalityterm properties of both psychedelicsterm and DOC are necessary prior to such investigations, but future experiments comparing the effects of dopaminergic stimulants to psychedelicsterm in DOC patients could help us identify whether drugs that increase criticalityterm and conscious content, like psychedelicsterm are hypothesized to do, are better at treating DOC than drugs that stimulate arousal (Scott and Carhart-Harris, 2019 ).

Meditationterm and shamanic states of consciousnessterm present interesting results in the way that they also bring the brain network closer to the critical stateterm. Benefits of meditationterm have been vastly studied in the current literature. Among others, benefits relating to stress reduction, anxiety, depression and pain improvement ( Brandmeyer et al., 2019 ) seem to consistently arise when introducing meditationterm protocols. While some studies have proposed a vast number of structural and functional correlates of the mechanisms by which meditationterm influences health-related outcomes, such as reduced activity in the Default Mode Network ( Garrison et al., 2015 ) or increased cortical thickness in prefrontal and insular regions ( Hölzel et al., 2011 ), we argue that brain dynamics converging toward the critical stateterm could be a potential explanatory factor of beneficial outcomes of meditationterm. Considering that the brain working in a critical stateterm is optimized for information processing as well as flexibility of thought, future research should aim to characterize the direct contribution of brain criticalityterm on health-related positive outcomes of meditationterm.

Though our review did not seek articles on the effects of caffeine on the brain functional network, caffeine was found to bring the cortical dynamics closer to criticalityterm in a study on deliriumterm ( Irrmischer et al., 2018 ). This is an interesting result, as it could possibly explain the increases in cognitive capacities, such as concentration and memory enhancement ( Cappelletti et al., 2015 ), reported while under this substance. Various clinical applications could also emerge from this finding, especially considering that caffeine is such an easily accessible drug and that is widely accepted to be safe for consumption in various populations. Some researchers are already looking at the applications of caffeine on anesthesiaterm emergence and preliminary results are promising, showing faster emergence from anesthesiaterm in subjects who were injected with intravenous caffeine ( Hölzel et al., 2011 ; Cappelletti et al., 2015 ). These results have been reported in rats ( Fox et al., 2020 ) as well as in humans ( Fong et al., 2018 ) without any adverse effects. When looking at these findings within the scope of criticalityterm, it is logical to think that a substance bringing cortical dynamics closer to criticalityterm (caffeine) could be used to restore consciousnessterm in an ASC generally conceptualized to induce a subcritical stateterm (anesthesiaterm). With this logic in mind, applications of caffeine as a treatment for subcritical DOC could also be plausible, though further research is necessary to characterize the criticalityterm properties of caffeine across a range of ASC.

Study limitations This scoping review should be interpreted in light of certain limitations. First, since the term “criticalityterm” is a somewhat new concept in neuroscience, some studies looking at measures related to criticalityterm (e.g., long-range temporal correlationsterm, neuronal avalanchesterm, detrended fluctuation analysis, Hurst exponent, and others) do not explicitly use the term “criticalityterm”, making it possible that these articles were not found during the literature search. In future reviews on the topic of criticalityterm, the search strategy should be adapted to reflect the state of the literature and include these measurement methodsterm·var as key search terms to make sure that all related articles are found.

Second, the fact that different measurement methodsterm·var (e.g., EEG, fMRI) and different types of subjectsterm·var (humans and animals) were compared to one another in this review implies prudence in the interpretation of the results. In the light of this limitation, we suggest that future projects explore brain criticalityterm in various ASC while using the same measurement methodsterm·var of criticalityterm. This will lead to a stronger conceptualization of criticalityterm, making it clearer which states are sub-critical, near criticalityterm or super-critical.

Third, this review did not include recent theoretical advances on criticalityterm that focus on predicting a critical pointterm and modulating brain criticalityterm. It is theoretically known that the underlying mechanisms of criticalityterm, such as the type of phase transitionterm [e.g., first order (abrupt) and second order (gradual)] or the bifurcation structure (e.g., Trans-critical (smooth transition), Hopf (onset of oscillation), Fold (abrupt transition) bifurcation, etc.) at a critical pointterm shape the transition pattern and the system stability, which are directly associated with loss and recovery of criticalityterm under external stimuli ( Scheffer et al., 2009 ; Bury et al., 2020 ) as well as the signal characteristics at a critical pointterm. However, because of the generality of the criticality measuresterm currently considered in this review, which are observable regardless of the type of phase transitionterm/bifurcation structure, criticality measuresterm alone cannot differentiate criticalityterm that originated from different types of phase transitionterm/bifurcation. Therefore, it may lack a precise evaluation of the deviation from criticalityterm and result in inconsistent and variable outcomes in application of criticality measureterm to ASC.

Finally, criticalityterm has been mainly applied to ASC to characterize and differentiate the brain states quantitatively and qualitatively. However, it does not explain how the brain can recover from ASC and return to a normal state, near the critical pointterm. In the future, the scope of reviews should be expanded to include principle and mechanism-based studies such as dynamical system theory and critical transition that describe the state transition behaviors during the loss and recovery of criticalityterm in the brain.

Conclusion This scoping review showed that the study of criticalityterm in ASC is still in its infancy, and that the current literature lacks homogeneity in its assessment tools as well as it analytical models and approaches. In the future, by directly mapping brain criticalityterm using homogeneous measurement methodsterm·var across a range of ASC, it would be easier to establish clear and widely accepted directions of deviation from the critical stateterm. Further research is particularly needed in the fields of sleepterm, DOC, anesthesiaterm and deliriumterm. Establishing these clear directions of deviation would not only help in the potential development of new diagnostic and treatment tools for pathological ASC, but also in helping understand the functional and neural basis of consciousnessterm and its alterations.

Author contributions L-PB completed the literature search, wrote a report of the results in French, and completed Table 1 . GV screened the first set of article and gave his insight on the review. CG revised the literature search and adapted L-PB’s report into a complete English article. CG wrote the first draft of the discussion and created all tables (except Table 1 ) and figures. UL gave his insight and recommendations on the review and helped with writing the manuscript. CD screened the first and second set of article, as well as managed and validated the entirety of the first and second phase of the manuscript. All authors revised the manuscript.