Available via license: CC BY 4.0
Content may be subject to copyright.
This work is shared under a CC BY license unless otherwise noted
Complex slow waves radically reorganise human brain
dynamics under 5-MeO-DMT
George Blackburne1,2*, Rosalind G. McAlpine3, Marco Fabus4,5, Alberto Liardi2,
Sunjeev K. Kamboj3, Pedro A. M. Mediano1,2, and Jeremy I. Skipper1
1Department of Experimental Psychology, University College London
2Department of Computing, Imperial College London
3Clinical Psychopharmacology Unit, University College London
4School of Medicine and Biomedical Sciences, University of Oxford
5Wellcome Centre for Integrative Neuroimaging, University of Oxford
*Corresponding authors: George Blackburne george.blackburne.18@ucl.ac.uk
October 4, 2024
Abstract
5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT) is a psychedelic drug known for its
uniquely profound effects on subjective experience, reliably eradicating the perception of
time, space, and the self. However, little is known about how this drug alters large-scale
brain activity. We collected naturalistic electroencephalography (EEG) data of 29 healthy
individuals before and after inhaling a high dose (12mg) of vaporised synthetic 5-MeO-DMT.
We replicate work from rodents showing amplified low-frequency oscillations, but extend
these findings with novel tools for characterising the organisation and dynamics of complex
low-frequency spatiotemporal fields of neural activity. We find that 5-MeO-DMT radically
reorganises low-frequency flows of neural activity, causing them to become incoherent,
heterogeneous, viscous, fleeting, nonrecurring, and to cease their typical travelling forwards
and backwards across the cortex compared to resting state. Further, we find a consequence
of this reorganisation in broadband activity, which exhibits slower, more stable, low-
dimensional behaviour, with increased energy barriers to rapid global shifts. These findings
provide the first detailed empirical account of how 5-MeO-DMT sculpts human brain
dynamics, revealing a novel set of cortical slow wave behaviours, with significant implications
for extant neuroscientific models of serotonergic psychedelics.
1
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
1 Introduction
Psychedelic drugs are reentering scientific research as unique and reliable tools for interrogating
the neural processes that entail alterations in the structure of subjective experience [1, 2]. In
tandem, they are being explored as potential treatments for an array of mental health conditions
[3, 4]. However, 5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT), likely the most profound
agent in this class [5], has remained understudied in human neuroscience. This is despite its
current involvement in clinical trials for the treatment of depression [6, 7], bipolar [8], and
alcohol use disorders [9].
5-MeO-DMT is a naturally occurring, potent, rapid-acting tryptamine, with use likely dating
back over 1000 years in South American Tiwanaku ritual practices via insufflation of the trace
concentrations contained in Anadenanthera seeds [10, 11]. Contemporary usage however, is
markedly different, predominantly occurring through either the synthetic form, or via the venom
secreted from the parotid glands of the Incilius alvarius toad [12]. Indeed, it is the high-dose
combustion or vaporisation and inhalation of these formats through which 5-MeO-DMT has
attained its unofficial status as the world’s most powerful psychedelic, or the ‘God Molecule’ [5].
The experience elicited by inhaled 5-MeO-DMT, like its chemical relative DMT, is char-
acterised by a radical progression, or ‘breakthrough’, into what is felt as ‘another dimension’,
such that sensation becomes detached from consensus reality [13]. However, unlike DMT, this
environmental disconnection generally does not typically involve immersion into an overwhelm-
ingly visually intricate geometric space filled with seemingly sentient entities, but rather a world
entirely lacking the usual axes of variation to structure it. At the peak of the 5-MeO-DMT
experience, rudimentary structures of subjective experience such as time, space and the self are
felt to be abolished, and a dramatic lack of differentiation ensues [14]. Fully awake and aware,
individuals feel they have entered into an ineffable realm of ‘everything and nothing’, where
even the inference that they are the subject of the experience is protracted [15]. 5-MeO-DMT is
therefore a perturbation that renders which of a creature’s consciousness-related capacities can
be online, but does not in the classical sense ‘reduce’ conscious level [16]. Thus, it may be most
productively thought of as a global mode of deconstructed consciousness.
Preclinical work in awake rodents indicates that 5-MeO-DMT elicits its most pronounced
effects in augmenting low-frequency rhythms (<4Hz) [17–20]. 5-MeO-DMT has been shown
to increase the power of low-frequency oscillations in local field potentials, with concurrent
multi-unit recordings showing neural activity to alternate between periods of high-firing and
generalised silence, while mice freely behave. This hybrid state was therefore dubbed ‘paradoxical
wakefulness’, since mice exhibited the canonical neurophysiological signatures of slow-wave sleep
(SWS) while awake [19]. The amplification of low-frequency oscillations is thought to be a
domain-general indicator of loss of consciousness beyond SWS, such as the period of slow-wave
activity saturation (SWAS) in general anaesthesia [21], and coma patients being diagnosed with
unresponsive wakefulness syndrome [22, 23]. Thus, these results with 5-MeO-DMT challenge the
idea that amplified low-frequency rhythms have a one-to-one correspondence with reductions in
conscious level [24].
Simple oscillatory phenomena in the brain are increasingly being empirically understood
as univariate symptoms of more complex spatiotemporal propagating patterns, with SWS
and SWAS exemplifying this shift in understanding [25–28]. Indeed, there have been recent
theoretical and empirical calls to define brain states more generally via wave-like processes
[29–31], though the idea that the brain is a non-stationary system composed of trajectories
2
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
inextricably structured in space and time is not itself a recent development [32–34]. Complex
flows, or metastable waves, are transient activity patterns that sweep across the cortex and
converge around points of stationarity, termed singularities. These can take on a host of forms,
such as sources or sinks, which expand or contract around a point, spirals which rotate around
a point, and saddles which are usually the superposition of different interacting waves. These
patterns have been consistently observed in mesoscopic and macroscopic neural recordings
[30, 35–39], and are thought to play crucial roles in cortical computation [40, 41], modulating
excitation-inhibition balance [42], memory [43, 44], and perception [45–47].
Thus, a productive approach for the empirical characterisation of the effects of 5-MeO-DMT
on macroscopic brain dynamics might lie in understanding how complex low-frequency flows
are reorganised. We hypothesised that, while 5-MeO-DMT may look analogous to states of
unconsciousness from a univariate oscillatory perspective, it may look strikingly different from a
more comprehensive spatiotemporal standpoint. Instead of inducing more coherent, durable and
stereotyped global low-frequency flow structures, like in anaesthesia [25, 27, 48], we predicted
that 5-MeO-DMT would induce a state of resolute complex low-frequency incoherence, with
disjointed and viscous flows limiting the emergence of global propagating patterns able to
spread across the cortex. We hypothesised that there would be a greater occurrence of localised
low-frequency patterns with differing directions and speeds, possessing short lifetimes, and
bearing little resemblance to each other over time. In sum, we expected that 5-MeO-DMT
would upset the backbone of cortical dynamics with ‘broken’, rather than tidal, waves.
Importantly, one should ask what such low-frequency changes functionally imply for statistical
motifs of broadband (wide frequency) brain dynamics. Owing to the expected spectral alterations,
we predicted that the intrinsic timescales of the cortex would shift towards a slower regime, and
that excursions from similar activity states would depart from each other more tardily. Crucially,
if slow spatiotemporal flows are disorganised, the brain may be less able to effectively orchestrate
large rapid global amplitude reconfigurations, thus increasing the steady-state residence and
stability of broadband cortical dynamics [49]. Given this, broadband responses may be forced
from their large population space onto an unusual and simpler sub-space. Put simply, these flows
may curb the degrees of freedom of the brain. This bounding of the system’s dimensionality
could distort and ablate the typical construction of neural representations [50], as such, the
individual’s model of the world should collapse in parameters, resulting in the deconstruction of
subjective experience.
We investigate these predictions by collecting naturalistic electroencephalography (EEG)
data of individuals (N=29) before and immediately after inhaling a high-dose (12mg) of vaporised
synthetic 5-MeO-DMT. We first replicate recent work in rodents on the power of low-frequency
oscillations, but extend this by performing a suite of analyses characterising alterations in the
organisation and dynamics of complex low-frequency spatiotemporal flows of neural activity.
We then complement this by assessing the statistical behaviour of broadband signals across
the cortex in terms of their intrinsic timescales, separation rates, underlying energy landscape,
and dimensionality. In doing so, we provide the first detailed account of the neuroscience of
5-MeO-DMT in humans.
3
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
2 Results
Slow oscillations
Given the lack of 5-MeO-DMT research in humans, we begin by assessing the frequency content of
our 64-channel EEG data. The signature effect of 5-MeO-DMT in mice has been the amplification
of low-frequency oscillations (
<
4
Hz
). A peculiar effect, since augmentation of low-frequency
content has hitherto been considered a signature of loss of wakefulness and awareness. Moreover,
rodent studies have shown no clear effect of 5-MeO-DMT on the power of alpha oscillations, the
deflation of which is currently considered the neurophysiological signature of psychedelic drug
effects in humans [51–64]. To enable robust time-frequency representation of cortical signals,
we use Empirical Mode Decomposition (EMD). We do so since canonical methods such as the
Fourier Transform (FT), can provide unreliable estimates when signals are noisy, non-sinusoidal,
and non-stationary, as is routinely seen in in neural recordings [65–70]. As an illustration, a
slightly non-linear high-delta (4
Hz
) oscillation can be mistaken to contain a low-alpha (8
Hz
)
rhythm if the FT is applied due to harmonic effects (Supplementary Fig.1). Therefore, we use
EMD to iteratively sift out the intrinsic basis functions of cortical signals without assuming they
are sinusoidal or static. Here, we apply EMD and the Hilbert-Huang Transform (HHT) to the
epoched data in the range of 0.5-50Hz (n=19, 10 exlcuded to movement artifacts), to maximise
the number of viable participants. The non-linear power spectrum is taken as the sum across
instantaneous frequencies over intrinsic mode functions (IMFs), normalised by their density.
5-MeO-DMT shifts the power spectra of cortical signals at its antipodes. Slow (<1.5Hz) and
fast (>20Hz) oscillations rapidly rise in power within 20s after drug inhalation and decay after
8-10 minutes (Fig.1a,d), coincident with the expected rise and decay of drug effects. Alterations
in spectra can be seen to track the expected duration of drug effects, with the first 4 minutes
being marked by a particularly constant spectral profile (Fig.1b).
Examining the estimated peak of the drug effect (1.5-2.5 minutes) reveals a stark effect at
low-frequencies. We find that the mean power of the lowest frequency bin (0.5Hz), increases by
415% compared to baseline (Fig.1c). These amplified slow oscillations (0.5-1.5Hz) are spatially
distributed, with increases occurring over frontal, parietal, temporal, central and occipital
electrodes (Fig.1e; max: PO3 T=5.104, pF D R =0.011, BF10 =365.623, d=1.472).
Inspecting higher-frequency changes, we observe distributed but weaker increases in gamma
oscillations at the estimated peak (Fig.1e max: Pz
T
=3.750,
pF DR
=0.001,
BF10
=26.542,
d
=1.025). As the expected end-point of the experience is arriving, between 7.5-9 minutes,
there is a sudden jump in high-frequency oscillations (Fig.1a). Notably, we see that alpha
oscillations are not robustly reduced across every area of the cortex with the large effect
sizes seen with other psychedelic compounds [63] (Fig.1c). In fact, when parameterising the
peaks of the average power spectra with Gaussians [71], we find no significant change in alpha
(
pF DR
=0.411), but a significant reduction in the peak frequency (
pF DR
=0.023). Examining the
topography of alpha changes we find significant reductions over the right parietal-occipital cortex
(Pz:
pF DR
=0.030), but also increases in alpha power over the temporal lobes approaching
significance (T8: pF DR =0.052).
Finally, we investigate whether fast oscillations can themselves be understood in terms
of slow oscillations. Specifically, we test if there is greater amplitude modulation of high-
frequency content by low-frequency content under 5-MeO-DMT, as evidenced by changes in
their holospectrum. Indeed, we find that high-frequency rhythms exhibit greater fluctuations in
4
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
power at low-frequencies under 5-MeO-DMT (Fig.1f; p=.005 ANOVA cluster-level permutation).
Figure 1: 5-MeO-DMT induces diffuse high-amplitude slow oscillations. a Average time-
frequency power spectral density (PSD) for the 20 minutes post-inhalation of 5-MeO-DMT, derived via
Empirical Mode Decomposition and the Hilbert-Huang transform. bEuclidean distance matrix for
average spectra vectors over time. cTime-averaged power spectra for eyes-closed baseline period and
estimated peak 5-MeO-DMT at 1.5-2.5 minutes (mean
±
SEM). dFocused low-frequency spectrogram.
eScalp topographic statistical maps for T-test comparing the two conditions per frequency band and
electrode, with electrodes representing significant differences (p<.05 FDR-corrected Benjamini–Hochberg
procedure). Purple indicates increases under 5-MeO-DMT, green indicates reductions. fHolospectrum
statistical map for ANOVA comparing the two conditions, with significant clusters coloured (p<.05
cluster-corrected).
In sum, these results provide evidence for increases in the power of slow-oscillations in the
human brain under 5-MeO-DMT, replicating results from rodents [17–19], as well as providing
evidence that substantial whole-brain loss of alpha power may not be as essential for the
5-MeO-DMT experience, in contrast to other classical psychedelics.
Complex slow waves
Moving past characterising the power of univariate rhythmic fluctuations in neural activity
in the two conditions, we characterise multivariate macroscopic propagating patterns with
particular temporal motifs. We do this by adopting the mathematics of flow velocity from
fluid dynamics. We first perform EMD on the continuous signals from the peak window for
those participants who have it artefact free (1.5-2.5 minutes post-ingestion, n=13) to maximise
detection of the low-frequency cycles, obtaining six time-evolving oscillatory intrinsic mode
functions (IMFs), for each electrode [72]. These modes have a mean instantaneous frequency (IF)
of 16.398
±
1.104, 6.797
±
0.290, 2.270
±
0.093, 0.741
±
0.028, 0.319
±
0.007, 0.133
±
0.004 respectively
(all in
Hz
). Since our interest is low-frequency modes, we only analyse the third, fourth and
5
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
fifth modes, excluding the sixth as a residual artefactual drift, and refer to these as Delta, Slow
and Ultra-Slow throughout. The mean IF of these modes do not significantly change between
conditions (Fig.2a; Delta:
T
=3.012,
pF DR
=0.065; Slow:
T
=0.462,
pF DR
=1; Ultra-Slow:
T
=-0.112,
pF DR
=1). After interpolation of the instantaneous amplitude and phase of the
modes onto a uniform 32x32 scalp grid, mimicking the pixels from optical imaging voltage grids
[48], we obtain the velocity field at each sample for each mode and condition (see Methods).
As expected, we detect continuous and complex propagating spatiotemporal patterns in both
conditions, but with markedly different structure and dynamics (Fig.2b-m).
Specifically, assessing alterations in phase interactions of the field via a Kuramoto-like
formalism (see Methods), we find that 5-MeO-DMT reduces the phase synchrony (global
coherence) of low-frequency phase fields Fig.2c (Delta:
pF DR
=0.010; Slow:
pF DR
=2.07
×
10
−5
;
Ultra-Slow:
pF DR
=8.73
×
10
−7
), and also results in less variance in this synchrony over time
(global metastability) (Fig.2d Delta:
pF DR
=0.080; Slow:
pF DR
=5.96
×
10
−4
; Ultra-Slow:
pF DR
=8.73
×
10
−7
). We find that this is also the case locally, across radii ranging from 1.75-
7.875 cm in increments of 0.875 cm for both coherence (Fig.2e), and metastability (Fig.2f)
(see Supplementary Table.2 for statistics) Thus, 5-MeO-DMT induces a state of persistent
low-frequency desynchronisation.
Examining the velocity fields further, we find that 5-MeO-DMT causes waves to travel faster
on average (Fig.2h; Delta:
pF DR
=0.002; Slow:
pF DR
=1.49
×
10
−4
; Ultra-Slow:
pF DR
=2.61
×
10
−4
).
However, this largely reflects disorganised and spatially circumscribed flow, since globally the
field exhibits less directional alignment and collective motion, as shown by reductions in
the fields normalised velocity for Delta and Slow waves (Fig.2i; Delta:
pF DR
=0.003; Slow:
pF DR
=7.19
×
10
−5
). However, the opposite is true for Ultra-Slow waves (Fig.2i
pF DR
=0.026).
Supporting the idea that flow becomes more fragmented, we find that velocity fields become
significantly more spatially heterogeneous (Fig.2j; Delta:
pF DR
=0.001; Slow:
pF DR
=1.49
×
10
−4
;
Ultra-Slow:
pF DR
=1.49
×
10
−4
). We predicted that this would be the result of decreases in
the area occupied by flow sources, and increases in the strength (divergence) exerted by flow
sources. Indeed, we show that such decreases in the area of sources and increases in the strength
of sources strongly correlate with the observed changes in field heterogeneity, with the lowest
rs =.74 (Fig.2l).
Finally, we further confirm the ‘provincialising’ effects of 5-MeO-DMT on low-frequency flow
by showing it to become more viscous, revealing that the diffusion of neural activity across the
cortex at this timescale is significantly hampered by the drug (Fig.2k; Delta:
pF DR
=5.23
×
10
−4
;
Slow:
pF DR
=2.34
×
10
−4
; Ultra-Slow:
pF DR
=7.35
×
10
−4
). In sum, 5-MeO-DMT induces a state
of disorganised low-frequency flow, giving rise to greater local patterns with diverse shapes,
directions and speeds. These flows collectively diminish the emergence of global propagating
patterns with specific paths across the cortex.
6
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
Figure 2: 5-MeO-DMT reorganises patterns of complex low-frequency flow. a Mean instanta-
neous frequency for each intrinsic mode function (IMF). bInstantaneous global phase coherence for each
IMF (mean
±
SEM). cGlobal phase coherence for each IMF. dLocal phase coherence for each IMF
across radii (mean
±
SEM). eGlobal phase metastability for each IMF. fLocal phase metastability
for each IMF across radii (mean
±
SEM). gExample velocity field for Mode 2 from a representative
participant over the course of 560ms during estimated peak of 5-MeO-DMT. hMean wave speed for
each IMF. iAverage normalised velocity of the velocity field for each IMF. jAverage heterogeneity of
the velocity field for each IMF. kViscosity of the velocity field for each IMF. lCorrelations between
changes in field heterogeneity with changes in source area (top row) and source strength (bottom row)
for each IMF (columns). Change scores are 5-MeO-DMT - Rest. mChanges in the direction of flow
for each IMF across the cortex. Bin position indicates direction, and bin orientation indicates change.
Inward pointing bins indicate reduced travel under 5-MeO-DMT. T-test p<.05 FDR-corrected bins are
indicated with black borders. (*,**,*** indicates p<.05,.01,.001 FDR-corrected Benjamini–Hochberg
procedure)
7
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
To directly assess changes in the path of waves across the cortex, we obtain the angle of
each grid element on the scalp via the inverse of the tangent of their X and Y directional
vectors. We bin the 2
π
complete angle of directions into 50 bins, and find that 5-MeO-DMT
significantly alters wave travel directionality (Fig.2m). We find that Slow and Ultra-Slow waves
travel significantly less in true anterior, posterior directions, and lateral directions, instead
preferring to travel in an intercardinal fashion (Fig.2m). The loss of canonical forward and
backward travelling waves is particularly striking in the case of Ultra-Slow waves (Fig.2m;
Anterior travel:
T
=-10.375,
pF DR
=3.61
×
10
−5
,
BF10
=6.11
×
10
4
,
d
=2.639; Posterior travel:
T
=-9.065,
pF DR
=7.67
×
10
−5
,
BF10
=1.66
×
10
4
,
d
=2.648). Thus, 5-MeO-DMT reorganises
wave travel and specifically diminishes forwards and backward travelling slow waves.
Visual inspection of velocity fields at baseline indicated that field patterns were persistent
and frequently reoccurred. However, under 5-MeO-DMT patterns appeared to have shorter dwell
times, with future patterns noticeably more dissimilar to past ones. To specifically test such
alterations in the temporal dynamics of low-frequency flows, we compute recurrence matrices
for the velocity fields (see Methods). Fig.3a-b show example network plots for the first 1
/
4
(15s) of the recurrence network for Mode 2 from a representative participant. Fig.3c shows the
same for the remaining participants. The 5-MeO-DMT recurrence networks can be seen to be
scattered and less densely interconnected, as evidenced by the average degree distributions for
each condition in (Fig.3d). The mean degree of the recurrence network is significantly reduced
by 5-MeO-DMT for all three modes (Delta:
pF DR
=1.38
×
10
−4
. Slow:
pF DR
=4.21
×
10
−6
.
Ultra-Slow:
pF DR
=4.21
×
10
−6
). Indeed, we also see that the number of communities in the
recurrence network significantly increases under 5-MeO-DMT (Fig.3e Delta:
pF DR
=6.89
×
10
−5
;
Slow:
pF DR
=0.003; Ultra-Slow:
pF DR
=0.033;), indicating an increase in the number of
unique field patterns. Finally we see that the global efficiency of the network, representing how
effectively flow patterns are exchanged between the past and the future, is significantly reduced
by 5-MeO-DMT for all three modes (Delta:
pF DR
=0.001. Slow:
pF DR
=6.61
×
10
−5
. Ultra-Slow:
pF DR
=2.32
×
10
−5
). In sum, 5-MeO-DMT makes low-frequency flow fields less recurrent, and
less structured over time.
8
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
Figure 3: 5-MeO-DMT dampens low-frequency field recurrence and pattern endurance. a
Field recurrence network with example velocity fields from Mode 2 in the Rest condition over 15s of a
representative participant. Plotted with communities determined by greedy modularity maximisation
(25 modules). bField recurrence network plot with same parameters for 5-MeO-DMT (1513 modules).
cField recurrence network with same parameters for remaining participants. dMean histograms for
degree distribution of field recurrence networks for each mode eNumber of communities in recurrence
networks for each mode. fGlobal efficiency for recurrence networks for each mode. g. Illustrations
of canonical singularity patterns. hAverage persistence of a flow pattern as the dominant pattern in
the brain. S indicates a stable singularity, U indicates unstable. Sources (U-Nodes), sinks (S-Nodes),
spiral-out (U-Focus), spiral-in (S-Focus) and saddles. (*,**,*** indicates p<.05,.01,.001 FDR-corrected
Benjamini–Hochberg procedure)
9
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
To test for changes in specific patterns, we classify singularities as either unstable node
(sources), stable node (sinks), unstable focus (spiral-out), stable focus (spiral-in) or saddle waves
(see Methods) (Fig.3g) and calculate their dynamical properties. We find that 5-MeO-DMT
significantly reduces the average lifetime of every pattern of Delta and Slow waves, while
engendering a mixed effect on Ultra-Slow waves, reducing the endurance of nodes but increasing
the endurance of spirals and saddles (Fig.3h Delta max: S-Focus:
T
=-5.880,
pF DR
=3.31
×
10
−4
,
BF
=371.185; Slow max: S-Node:
T
=-8.757,
pF DR
=6.63
×
10
−5
,
BF
=1.20
×
10
4
; Ultra-Slow
max: Saddle:
T
=5.463,
pF DR
=5.42
×
10
−4
,
BF
=209.142) (See Supplementary Table.7 for
extended statistics). Overall, these results show that 5-MeO-DMT disrupts canonical patterns
of low-frequency flow.
Broadband stability
We predicted that a potential consequence of these low-frequency alterations, and a potential
cause of the structurally constrained experience induced by the drug, would be that broadband
neural activity would be forced onto an unusual manifold, where large global shifts are pro-
hibited. We expected that broadband signals would on average exhibit slower, more stable,
low-dimensional behaviour, with increased energetic costs for major deviations.
First, we assessed the average decay time of the automutual information (
AMI
) function,
a measure of the ‘intrinsic timescale’ of neural activity, and find that this doubles under 5-
MeO-DMT from 221.9 ms
±
72.6 to 495.3ms
±
94.7 (Fig.4a;
pF DR
=0.018), an effect robust to
parameter choices (Supplementary Fig.6). Thus, 5-MeO-DMT makes broadband activity exhibit
more regular trends and long-term dependence. We then compute the maximum Lyapunov
exponent, a measure of the average separation rate of neighbouring points in phase space,
and find a weak but significant reduction (
λmax
) (Fig.4b;
pF DR
=0.046). Thus, 5-MeO-DMT
makes cortical dynamics more stable by decreasing its sensitivity to initial conditions. We
next tested whether cortical dynamics would be easier to explain via principal components
analysis with singular value decomposition. We indeed show that the variance explained (the
eigenvalue) is significantly greater for 5-MeO-DMT for each component (eigenvector) (Fig.1c;
P C1pF DR
=0.027), indicating that the brain under 5-MeO-DMT indeed exhibits more low-
dimensional behaviour.
10
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
Figure 4: 5-MeO-DMT pushes cortical dynamics toward a low-dimensional steady-state.
aAverage decay time of auto mutual information (
AM I
) function on broadband signals. bAverage
separation rate of neighbouring points in the phase space of broadband signals. cEigenvalues for the
first 10 eigenvectors from PCA of broadband signals. d2D average energy landscape for Rest (top),
5-MeO-DMT (middle) and their difference (bottom). e3D average energy landscape for Rest (left)
and 5-MeO-DMT (right). fEnergy required for each MSD bin averaged across all lags. ’Energy well’
illustrations are included to depict energetic costs. (* indicates p<.05 FDR-corrected)
.
Finally, we examine whether this slow, stable, low-dimensional broadband behaviour can
be cashed out in the language of statistical mechanics, as alterations in the topography of a
putative energy landscape of cortical dynamics. Here, energy represents the ability of the brain
to move from one activity state to another at a particular timescale, and is operationalised as
the inverse probability of a mean-squared displacement in global neural activity of a particular
size as a function of time [49]. Given the hypothesis that 5-MeO-DMT pushes the brain into an
unusual functional sub-space, we expected that 5-MeO-DMT would ’protect’ the brain from
major deviations in neural activity. Indeed, we find that 5-MeO-DMT significantly increases
the energy requirement of large scale deviations (>3 MSD) (max
T
=5.527,
pF DR
=0.001,
BF10
=228.686,
d
=1.193), while also making minor shifts (<1.5 MSD) easier to attain (max
T
=-4.628,
pF DR
=0.002,
BF10
=62.859,
d
=1.307) Fig.4d. We also use non-linear least squares
to fit
em·x
curves to our 2D landscape functions and find that 5-MeO-DMT significantly increases
the slope of the estimated curve (
p
=1.58
×
10
−4
). Thus, 5-MeO-DMT constrains the global
structure of neural dynamics.
11
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
3 Discussion
We demonstrate that 5-MeO-DMT, a short-acting psychedelic drug, produces a unique state in
the human brain. This state is marked by amplified diffuse slow rhythmic activity, that coalesces
into spatiotemporally disorganised and dismantled wave patterns unable to travel up and down
the putative cortical hierarchy. Furthermore, we find that this disruption pushes broadband
neural activity towards a low-dimensional steady-state, paralleling the subjective quality of the
experience.
Our results undermine the hypothesis that the induction of diffuse high-amplitude slow
oscillations are a universal signature of loss of consciousness. They do, however, provide
evidence that augmented low-frequency oscillations are related to environmental disconnection,
where subjective experience becomes independent of sensory variables imported from the
external world [54, 73–75]. This result therefore reinforces the need for establishing more
robust neuroscientific methods for discriminating conscious states. We offer the theoretical and
empirical contribution that a fruitful strategy will be found by moving beyond one-size-fits-all
univariate metrics such as oscillatory power, towards comprehensive frameworks that permit
the detailed characterisation of complex spatiotemporal activity structures.
By tracking the extended spatiotemporal patterns of low-frequency flows of neural activity,
we discovered crucial characteristics of the 5-MeO-DMT brain state that distinguish it from
states of unconsciousness. Rather than triggering simple, coherent, fluid, persistent, recurrent
propagating global waves like in anaesthesia [25, 27, 48], we found that 5-MeO-DMT induces
complex, incoherent, viscous, fleeting, unique wave patterns. While these diverse local flows
constrain the cortex-wide propagation, it may be the case that they enhance a form of regional
collision-based distributed dynamical computation [40]. The 5-MeO-DMT state may therefore
be marked by enhanced parallel, but not necessarily integrated, information processing [40].
Future work should investigate how spatiotemporal flow structures relate to taxonomies of
multivariate information dynamics, including measures of integrated information [76].
We found that 5-MeO-DMT instigates more stable low-dimensional broadband behaviour
with a decreased likelihood of major rapid global activity reconfiguration. These results are
consistent with work in mice showing that cortical dynamics become less chaotic under 5-MeO-
DMT [20]. This overall simplicity may reasonably occur as a consequence of, rather than in
spite of, the complexity seen in low-frequency flows. The lack of collective spatiotemporal
coordination of the low-frequency flow fields implies that cortical dynamics are sub-served by a
more fragmented network architecture, which hinders the recruitment of multiple regions to
effectively orchestrate simultaneous large global amplitude deviations. In short, a segregation
across spatiotemporal scales occurs, deconstructing the brain’s canonical functional organisation.
Our results contest a number of key concepts in the developing literature on the neuroscience
of serotonergic psychedelic drugs. First, whole-brain suppression of alpha power is considered a
central feature of the psychedelic state [77]. However, we found that alpha oscillations are not
robustly reduced across the cortex, with no deflation of the average power, and only right parietal-
occipital electrodes reaching significance. Future studies should investigate how alterations
in properties such as waveform shape and rhythmicity may be key. Second, psychedelics are
thought to ‘liberate’ the bottom-up flow of neural activity [78, 79]. By contrast, we found a
striking reduction in the probability that low-frequency flows travel in the anterior direction,
as well as the posterior direction. Finally, psychedelics are thought to reduce the curvature
of the energy landscape constraining neural activity, allowing more facile activity shifts, and
12
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
higher dimensional dynamics [78, 80–82]. Yet, we found that 5-MeO-DMT steepened the brain’s
energy landscape, increasing the barriers for major rapid activity shifts, and diminishing the
dimensionality of neural dynamics. Our results therefore suggest that 5-MeO-DMT has a unique
effect on the human brain compared to other classic tryptamine psychedelics.
Neurobiologically, we speculate that our results may represent the consequences of a unique
shift in the balance of distinct thalamocortical subsystems. The presence of diffuse high-
amplitude slow rhythmic activity, reduced communication through coherence, and approach
toward a low-dimensional steady state, suggest a reduction in diffuse coupling in the brain.
Non-specifically projecting thalamic structures, such as the mediodorsal and centromedian
nuclei, may be relieved of their canonical tonic firing patterns orchestrating supragrangular
cortical dynamics and global synchronisation patterns [83, 84]. Instead, cyclical bursting and
quiescence of these nuclei may occur, resulting in functional deafferentiation, and cortical activity
to become dominated by the intrinsic slow non-stationary bistability prescribed the anatomy of
local cortical circuits [85–87]. However, as internal awareness and arousal is maintained, the
level of overall thalamocortical resonance expected by wakefulness is likely maintained. Indeed,
the presence of fleeting viscous and heteregenous local flow hints there could be markedly greater
driving gain by structures with targeted projections, like the pulvinar and ventral lateral nucleus,
disrupting local excitation-inhibitory balance and overwhelming wave pattern dynamics. We
note that this general hypothesis is distinct from existing corticothalamic models of psychedelic
action which posit that there is an indiscriminate reduction in thalamic gating [1, 88].
Acknowledging some limitations, the study suffers from a small number of participants
with a high exclusion rate. This was due to movement artefacts, a common issue for human
neuroimaging studies with psychedelic drugs. The control condition is a resting-state baseline,
rather than a blinded placebo group, meaning that expectancy factors are uncontrolled for. A
standardised single high-dose (12mg) dose was administered to each participant, and blood
samples were not taken throughout, meaning that inter individual differences in pharmacokinetics
are likely present. Methodologically, our analyses assume the cortex to be a relatively sparse flat
surface, and therefore does not account for the high-spatial resolution, nor the gyri and sulci,
inherent to the human brain. Subsequent studies should therefore investigate 5-MeO-DMT
with high-density electrophysiological methods with effective cortical surface modelling. Lastly,
the present work does not integrate neural measures with first-person reports. Accordingly,
future work should combine neuroimaging with rigorous time-resolved measures of subjective
experience, such that inferences about both can mutual constrain each other.
In conclusion, we present the first detailed neuroscientific investigation in humans of perhaps
the most radically altered state of consciousness known. We report novel changes in the
spatiotemporal organisation of low-frequency waves across the cortex, revealing a disruption
in the congruity of the dynamical processes upon which the rest of cortical activity is built.
This work not only underscores the need for a revitalised understanding of the role of slow
waves in orchestrating subjective experience, but emphasises the need for more comprehensive
spatiotemporal methods in neuroscience to better understand the full diversity of human brain
states.
13
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
4 Methods
Ethical statement
This study was conducted in accordance with the Declaration of Helsinki and approved by the
University College London (UCL) Research Ethics Committee (ID: 19437/004). The study was
conducted in collaboration with Tandava Retreat Centre (TRC), who provided trained facilitators
and facilities. All participants provided informed consent via a secure online platform after
reviewing comprehensive study information. Participation was voluntary and uncompensated.
Participants were informed of their right to withdraw at any stage of the study without penalty
and were provided with information about sources of support in case of any distress related to
the study.
Participants
Participants were recruited globally through the study advertisement on the F.I.V.E. (5-MeO-
DMT Information & Vital Education) platform (www.five-meo.education), social media, news
outlets, and word-of-mouth. Potential participants underwent a four-stage screening process:
(1) TRC online screening, (2) TRC facilitator interview, (3) UCL online screening, and (4) UCL
researcher interview (RM, GB). Exclusion criteria included: age <18 years, no prior 5-MeO-DMT
experience, significant physical illness (e.g epilepsy, heart disease), psychiatric diagnoses, use of
psychiatric medications of any kind, family history of psychosis, history of adverse reactions
psychedelic drugs, and known physical movement under the influence of 5-MeO-DMT. Eligible
participants (N = 32) were initially assigned to one of six three-day retreats, with assignments
based on availability and balanced distribution. However, three participants withdrew before
the retreats began. The final cohort (N = 29; 16 male, 13 female; mean age = 48.52 years,
SD = 10.44, range = 34–75) was predominantly White/Caucasian (86.21%). The majority of
participants held a Bachelor’s degree (58.62%) and identified as non-religious (93.10%). Mean
lifetime 5-MeO-DMT use was 39.03 occasions (SD = 72.91, range = 1–300). Detailed participant
characteristics are provided in Supplementary Table.1.
Experimental procedures
The three-day retreat protocol included a preparation day consisting of orientation and par-
ticipant briefing (Day 1), a dosing day for 5-MeO-DMT administration and primary data
collection (Day 2), and an integration day featuring facilitator-guided group discussions (Day
3). This setup was consistent across all six retreats. Participants were instructed to refrain
from consuming drugs, with the exception of nicotine containing products, for two weeks prior
to dosing. On dosing day, participants fasted and abstained from caffeine. Individual dosing
sessions commenced at 08:00 AM in 90-minute intervals, supervised by two lead facilitators. Each
session began with a 7-minute baseline eye-closed resting-state EEG recorded with participants
seated in a self-selected comfortable position conducive to minimal movement. Participants were
instructed to remain relaxed but wakeful, and to alert experimenters if they felt like they may
fall asleep. Participants then moved to a centrally located padded recliner, where they assumed
a semi-supine position equipped with a standardised neck pillow and opaque eye mask. Synthetic
5-MeO-DMT (12 mg) was vaporised using an argon gas piston vaporiser (203-210°C,
∼
120s
14
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
heating). Following a standardised 2-3 minute relaxation exercise, participants inhaled the
vapour in a single breath and were instructed to maximise breath retention before exhalation, as
per protocols established during preparation. EEG recording began at the moment of completed
inhalation, and continued for 20 minutes post-inhalation. The EEG system was a saline-based
64-channel ANT Neuro Waveguard Net connected to an NA-261 EEG amplifier, recorded at
500 Hz, with Ag electrode placement following the 10-20 system (reference: CPz, ground:
AFz, impedance <20 kΩ). Ambient non-percussive music was played throughout the baseline
and drug period, and a Koshi bell was rung when EEG recording was complete. Post-session,
participants rested for a self-determined period.
EEG preprocessing
Preprocessing was performed using custom scripts using MNE-Python [89].Participants were
excluded from analysis if their physiological artifacts (ocular, muscular, respiratory, or cardio-
vascular) were substantial enough to necessitate the removal of over 50% of the data collected
within the first 10 minutes post-drug administration (10 participants). Data was filtered using
a finite impulse response (FIR) one-pass, zero-phase, non-causal bandpass filter at 0.1-50 Hz.
Data was segmented into non-overlapping 3s epochs, producing 140 epochs for the baseline
condition and 400 epochs for the drug condition. Electrolyte bridging was computed using the
intrinsic Hjorth algorithm with a 16
µV 2
cut-off [90, 91]. Participants with
≥
1%of potential
total bridges, or significant bridging that localises such that interpolation would depend on
sensors which are also bridged, were excluded from further analysis (1 participant). For those
with <1%, spherical spline interpolation was performed via the generation of virtual channels
between bridged electrodes (Rest: 3.684
±
2.076, 5-MeO-DMT: 1.158
±
0.785 bridges). Sensor
time series were manually inspected to identify bad channels, participants with >20 bad channels
were excluded from further analysis (0 participants), and those with <20 bad channels had
these interpolated from surrounding sensors (Rest: 7.579
±
1.049, 5-MeO-DMT: 7.368
±
1.06
channel). Extended infomax independent components analysis (ICA) was performed with 40
components, and those manually deemed to be equipment noise or physiological artefacts were
zeroed out and excluded from sensor signal reconstruction (Rest: 4.842
±
0.766, 5-MeO-DMT:
5.053 ±0.89 components). Epochs corrupted by noise or physiological arefacts were manually
removed (Rest: 3.611
±
2.267, 5-MeO-DMT: 38.105
±
10.718 epochs). Data was baseline
corrected and rereferenced to an average reference. This resulted in 19 subjects, 13 of which
had an artifact-free continuous minutes at the peak of the drug’s effects.
Empirical Mode Decomposition
A central challenge in time series analysis is creating robust representations of the frequency
content of complex signals. The standard approach in neuroscience for doing so is based on the
Fourier Transform (FT). The FT is a powerful and versatile technique that seeks to represent a
signal as a combination of strictly linear sinusoidal basis functions. However, brain rhythms are
noisy, exhibit distinctly non-sinusoidal waveforms, and enter into transient bursting behaviour
[65–70]. This presents a challenge for the FT, since it cannot fully and directly represent the
non-sinusoidal forms at their intrinsic frequency, instead assuming them to be partly the result
of a higher frequency harmonic [92]. This is problematic for neuroscience, since the canonical
electrophysiological frequency bands are all harmonics of a lower band, and the power in each
15
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
specific band is a variable of interest in its own right for distinguishing different brain states.
This problem is exacerbated by the fact that brain signals comprise a myriad of multiplexing
oscillations, each with their own harmonics. Furthermore, the linear basis functions of the FT
are static, meaning that in order to characterise evolving brain activity analysts must use a
windowing technique that inherently blurs the ongoing dynamics.
An elegant solution to these problems comes from Empirical Mode Decomposition (EMD),
which permits the automatic decomposition of signals into a finite number of distinct dynamic
non-linear oscillatory modes, known as Intrinsic Mode Functions (IMFs) [93]. EMD sifts the
IMFs out of a signal, from fastest to slowest, in an iterative process of peak-trough detection,
amplitude envelope interpolation and subtraction. Here, we use the recently developed iterated-
masking EMD (itEMD), as it has been shown to be an automated solution to the mode mixing
problem [72, 94, 95].
We then compute the instantaneous phase, frequency and amplitude of each IMF via the
Normalized Hilbert-Huang transform (HHT) [96, 97], with the signal’s time-resolved power
spectrum being an instantaneous amplitude weighted representation of frequency content [98].
Finally, to investigate the specific frequency content of amplitude modulations in our signals,
we perform a second level sift on the instantaneous amplitudes of each IMF, producing a set of
higher-order IMFs which represent the frequency content (amplitude modulation frequency) of
dynamical changes in oscillatory power (carrier frequency) [97]. Carrier frequencies are placed
into 50 linearly spaced histogram bins 0.5-4Hz and amplitude-modulation frequencies are placed
into 50 linearly spaced bins 8-50Hz, producing the holospectrum.
Velocity fields
Velocity vector fields are mathematical tools from physical theories of fluid dynamics that
include descriptions of vortices and eddies in collective fluid motion [99]. Recently, velocity fields
have been appropriated in neuroscience to characterise the structure of large-scale propagating
patterns of neural activity, adapting techniques from optical flow estimation in computer science
[36, 100].
First, the instantaneous amplitude of an IMF is pushed onto a 32x32 scalp grid, using radial
basis function interpolation and a multidimensional gaussian kernel of width 1 S.D, creating
the sequence
ϕ
(
x, y, t
). The velocity field v
ϕ
(
x, y, t
) = (u(
x, y, t
)
,
v(
x, y, t
)) , of element-wise x
and y-magnitudes, is then simply recovered from the spatial and temporal derivatives of the
amplitude via solving the constancy equation [27]:
Dϕ
Dt =∂ϕ
∂t +v· ∇ϕ⇒v=−
∂ϕ
∂t
/∥∇ϕ∥2∇ϕ(1)
Order parameters
Synchrony
As is done with the classical Kuramoto order parameter [101], we compute the synchrony of
instantaneous phases θas:
Rt=1
NX
xy
eiθ(x,y,t)
.(2)
16
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
Coherence is then the mean of this parameter across time, and metastability is the variance
across time [102]. We compute this both globally, and locally across radii ranging from 1.750 cm
to 7.875 cm in increments of 0.875. The minimum corresponds to one grid element and anything
significantly larger than the maximum would entail areas that are predominantly outside of the
scalp.
Normalised Velocity
To characterise the collective motion of the field, we calculate the norm of the sum of the
velocities divided by the sum of the norm of the velocities:
Φt=
PN
i=1 −→
vi
PN
i=1
−→
vi
,(3)
where
N
is the number of spatial elements in the field and
−→
vi
is the velocity the i-th element. If
Φ
t
is close to 1, the vectors in the field are aligned and there is collective motion in a particular
direction [41, 103].
Heterogeneity
To characterise the diversity of flows in the field we compute the heterogeneity [48], as the mean
over time of the standard deviation of wave propagation speeds. More formally:
H=1
T
T
X
t=1
1
µ(t)v
u
u
t 1
Nt
i=Nt
X
i=1 |−→
vi(t)| − µ(t)2!,(4)
where
T
is the total time,
Nt
the number of vectors at time t, and
µ(t) = 1
NtPN
i=1 −→
vi(t)
the
average speed at time t.
Complex wave patterns
Complex spatiotemporal waves are organised around critical points of stationarity [104]. These
singularities are of two main flavours, sources where flow arises, and sinks where flow gathers.
First, we compute the divergence of the velocity field as the sum of the partial derivatives of each
field component
∂u
∂x
+
∂v
∂y
. Singularities were then identified as extremal nested closed contours
divergence [36]. Contours were run over 10 levels equally spaced between the 1st and 99th
percentile of the divergence values [27]. From this we compute the centroid of the singularity,
the area of the contour occupied by the singularity, and the strength of the singularity as the
magnitude of velocity divergence.
This allows us to calculate viscosity of flow, which is taken simply as the product of the area
and divergence of the singularities [27], giving us a cm2s−1unit, more formally:
ν=A∇ · −→
v.(5)
Lower values indicate that waves can flow across space with greater ease, whereas higher
values indicate their distributed diffusion is restricted.
17
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
The nature of the singularities is then further characterised by the values of the trace (
τ
)
and the determinant (∆) of the Jacobian
J
[36], bilinearly interpolated from the neighbouring
grid elements on the scalp, which correspond to the divergence and curl of the field respectively:
J=∂u
∂x
∂u
∂y
∂v
∂x
∂v
∂y (6)
Unstable singularities (sources) have
τ <
0and stable singularities (sinks) have
τ >
0. Nodes
have ∆>0and τ2>4∆, focuses have ∆>0and τ2<4∆. Stable focuses are spiral-in waves,
and unstable focuses are spiral-out waves. Lastly, saddles are characterised by ∆
<
0, and are
usually formed by interactions between other waves [36].
Velocity field recurrence
In order to characterise the recurrence of velocity fields we calculated their alignment matrix
[30] via their mutual information (MI):
I(vi;vj) = H(vi) + H(vj)−H(vi,vj).(7)
Where
H
is the classic Shannon entropy function. We then shuffle the velocities spatially,
compute the alignment matrix over 100 permutations and take the maximum
MI
value from
each permutation and chose the 99th percentile of these values as a threshold for our empirical
matrix. We treat this as the binary adjacency matrix
A
of a temporal graph, and characterise
its structure [105].
Recurrence network topology
We first cluster nodes using Clauset-Newman-Moore greedy modularity maximization [106],
which finds the network partition that maximises the generalised modularity Q:
Q=
n
X
c=1 "Lc
m−γkc
2m2#.(8)
Where the sum iterates over all communities
c
,
Lc
is the number of intra-community edges,
m
is the total number of edges,
kc
is the sum of the degrees in the community, and
γ
is the
resolution parameter which we set to 1 [107].
We also characterise the compactness of this recurrence network by computing the average
of the inverse distances between nodes, i.e. its global efficiency:
E=1
N(N−1) X
i=j∈V
1
dij
(9)
, where
dij
is the distance between the i-th and j-th nodes. In this case, higher global efficiency
indicates that the network is compact and velocity fields are robust over time, whereas a lower
global efficiency indicates that velocity fields exhibit less recurrence.
18
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
Intrinsic timescales
To identify the intrinsic timescale of the cortex we compute the average decay time of the
auto-mutual information (
AMI
) function. That is, we find the mean lag at which the mutual
information between neural signals and a delayed copy of themselves reaches a stable minima.
To do so, we fit a straight line to the last third of the
AMI
function, and find the first time
point that falls below the
AMI
value equal to the line’s y-intercept. Longer intrinsic timescales
indicate that a time series contains greater information about its own future, and shorter intrinsic
timescales indicate that a timeseries is less history dependent [108].
Separation rate
To characterise the dynamical sensitivity of the system, we characterise the rate of divergence
of similar points in phase space with the maximum Lyapunov exponent:
λmax = lim
t→∞ lim
|δZ0|→0
1
tln |δZ(t)|
|δZ0|,(10)
which we derive via Rosenstein’s algorithm [109]. Here
δZ(t)
is the trajectory separation at
time
t
. Briefly, the broadband amplitude envelope is produced by the Hilbert transform and its
dynamics are reconstructed using Taken’s embedding. The nearest neighbour of each time point
is identified using the Euclidian distance, and their trajectories tracked. The separation rate is
then the exponent of the curve of the mean logarithmic distance between these trajectories as a
function of time, normalised by the sampling rate.
Energy landscape
To characterise the shifts in the ability of the brain to orchestrate rapid global broadband
reconfigurations, we follow previous work [49, 110, 111], and characterise the energy landscape
of neural dynamics through the statistical inference of the probability of an effect of a particular
size. Effect is operationalised as mean-squared-displacement (MSD) in the Fisher z-scored
amplitude envelope of broadband signals across the cortex:
MSDt,t0=|xt0+t −xt0|2r,(11)
where
r
is the number of electrodes. We compute the MSD with maximum lags of 1s, with 200
equally spaced initialisation points. As is done in statistical mechanics, we can then compute
the energy
E
of the electrophysiological attractor basin via the natural logarithm of the inverse
probability, estimated via Gaussian kernel density estimation K. Following [111], we compute
this over MSD values between 0 and 5.
E= ln
1
1
4n Pn
i=1 KMSDt,t0
4
(12)
Thus, a highly probable relative change in neural activity corresponds to a putatively low energy
requirement, and vice versa.
19
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
5 Data availability
Data will be made publicly available upon acceptance of the manuscript for publication.
6 Code availability
Code will be made publicly available upon acceptance of the manuscript for publication.
7 Acknowledgements
We are grateful to Tandava Retreats for their generous support in assisting the data collection
process, specifically Victoria Wueschner and Joel Brierre for their pivotal roles in facilitating
the 5-MeO-DMT sessions. Additional thanks go to Luis Fabian Rodriguez, Otto Maier, James
Sanders, Milly Sellers and George Deane for providing support during data collection. We
extend our deepest thanks to the courageous participants who volunteered for this study. Lastly,
we thank supporters of the crowdfunding campaign for helping secure the funding required to
perform this study. G.B is supported by the Leverhulme Trust. R.G.M is supported by the
Wellcome Trust. J.I.S. is supported by Wellcome Leap.
8 Contributions
G.B; Conceptualization, Investigation, Methodology, Software, Data curation, Formal analysis,
Supervision, Project administration, Writing. R.G.M; Conceptualization, Investigation, Project
Administration, Writing – Review and Editing. M.S.F; Methodology, Software, Writing. A.L;
Writing – Review and Editing. S.K.K; Resources, Supervision, Writing – Review and Editing.
P.A.M.M; Supervision, Writing – Review and Editing. J.I.S; Supervision, Writing – Review and
Editing.
9 Ethics declarations
Competing interests
The authors declare no competing interests.
References
1.
Vollenweider, F. X. & Preller, K. H. Psychedelic drugs: neurobiology and potential for
treatment of psychiatric disorders. Nat. Rev. Neurosci. 21, 611–624 (2020).
2. Nichols, D. E. Psychedelics. Pharmacol. Rev. 68, 264–355 (2016).
3.
Aday, J. S., Mitzkovitz, C. M., Bloesch, E. K., Davoli, C. C. & Davis, A. K. Long-term
effects of psychedelic drugs: A systematic review. Neurosci. Biobehav. Rev. 113, 179–189
(2020).
20
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
4.
Nutt, D., Erritzoe, D. & Carhart-Harris, R. Psychedelic Psychiatry’s Brave New World.
Cell 181, 24–28 (2020).
5.
Davis, A. K., Barsuglia, J. P., Lancelotta, R., Grant, R. M. & Renn, E. The epidemiology of
5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT) use: Benefits, consequences, patterns
of use, subjective effects, and reasons for consumption. J. Psychopharmacol. 32, 779–792
(2018).
6.
European Union Clinical Trials Register
https://www.clinicaltrialsregister.eu/
ctr-search/trial/2022-000574-26/IE. 2022.
7.
US National Library of Medicine
https://clinicaltrials.gov/study/NCT05660642
.
2023.
8.
European Union Clinical Trials Register
https://www.clinicaltrialsregister.eu/
ctr-search/trial/2021-006861-39/NL. 2022.
9.
US National Library of Medicine
https://clinicaltrials.gov/study/NCT05674929
.
2023.
10.
Miller, M. J., Albarracin-Jordan, J., Moore, C. & Capriles, J. M. Chemical evidence
for the use of multiple psychotropic plants in a 1,000-year-old ritual bundle from South
America. Proc. Natl. Acad. Sci. U. S. A. 116, 11207–11212 (2019).
11.
Torres, C. M. & Repke, D. B. Anadenanthera: Visionary Plant of Ancient South America
(Routledge, New York, 2006).
12. Most, A. Bufo Alvarius: the Psychedelic Toad of the Sonoran Desert (1983).
13.
Shulgin, A. & Shulgin, A. TiHKAL: Tryptamines I Have Known and Loved (Transform
Press, 1997).
14.
Metzner, R. The Toad and the Jaguar: A Field Report of Underground Research on a
Visionary Medicine Bufo Alvarius and 5-methoxy-dimethyltryptamine (Four Trees Press,
2022).
15.
Milliere, R. The varieties of selflessness. Philosophy and the Mind Sciences 1, 1–41 (2020).
16. Mckilliam, A. K. What is a global state of consciousness? PhiMiSci 1(2020).
17.
Riga, M. S., Lladó-Pelfort, L., Artigas, F. & Celada, P. The serotonin hallucinogen
5-MeO-DMT alters cortico-thalamic activity in freely moving mice: Regionally-selective
involvement of 5-HT1A and 5-HT2A receptors. Neuropharmacology 142, 219–230 (2018).
18.
Souza, A. C. et al. 5-MeO-DMT induces sleep-like LFP spectral signatures in the hip-
pocampus and prefrontal cortex of awake rats. Sci. Rep. 14, 11281 (2024).
19.
Bréant, B. J. B., Mengual, J. P., Bannerman, D. M., Sharp, T. & Vyazovskiy, V. V.
“Paradoxical wakefulness” induced by psychedelic 5-methoxy-N,N-dimethyltryptamine in
mice. bioRxiv, 2022.12.11.519886 (2022).
20.
Toker, D. et al. Criticality supports cross-frequency cortical-thalamic information transfer
during conscious states. Elife 13 (2024).
21.
Ní Mhuircheartaigh, R., Warnaby, C., Rogers, R., Jbabdi, S. & Tracey, I. Slow-wave
activity saturation and thalamocortical isolation during propofol anesthesia in humans.
Sci. Transl. Med. 5, 208ra148 (2013).
21
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
22.
Schiff, N. D., Nauvel, T. & Victor, J. D. Large-scale brain dynamics in disorders of
consciousness. Curr. Opin. Neurobiol. 25, 7–14 (2014).
23.
Rosanova, M. et al. Sleep-like cortical OFF-periods disrupt causality and complexity in
the brain of unresponsive wakefulness syndrome patients. Nat. Commun. 9, 4427 (2018).
24.
Frohlich, J., Mediano, P. A. M., Bavato, F. & Gharabaghi, A. Paradoxical pharmacological
dissociations result from drugs that enhance delta oscillations but preserve consciousness.
Commun Biol 6, 654 (2023).
25.
Bhattacharya, S. et al. Propofol Anesthesia Alters Cortical Traveling Waves. J. Cogn.
Neurosci., 1–13 (2022).
26.
Massimini, M., Huber, R., Ferrarelli, F., Hill, S. & Tononi, G. The sleep slow oscillation
as a traveling wave. J. Neurosci. 24, 6862–6870 (2004).
27.
Fabus, M. S., Woolrich, M. W. & Warnaby, C. E. Brainwave viscosity in propofol
anaesthesia. Br. J. Anaesth. 128, e61–e62 (2022).
28.
Siclari, F. et al. Two distinct synchronization processes in the transition to sleep: a
high-density electroencephalographic study. Sleep 37, 1621–1637 (2014).
29.
Foster, M. & Scheinost, D. Brain states as wave-like motifs. Trends Cogn. Sci. 28, 492–503
(2024).
30. Roberts, J. A. et al. Metastable brain waves. Nat. Commun. 10, 1056 (2019).
31.
Pang, J. C. et al. Geometric constraints on human brain function. Nature 618, 566–574
(2023).
32.
Kampis, G. Self-Modifying Systems in Biology and Cognitive Science (Pergamon Press,
1991).
33.
Maturana, H. R. & Varela, F. J. On Machines and Living Beings (Editorial Universitaria
S.A, 1972).
34.
Adrian, E. D. & Matthews, B. H. The interpretation of potential waves in the cortex. J.
Physiol. 81, 440–471 (1934).
35.
Mohajerani, M. H. et al. Spontaneous cortical activity alternates between motifs defined
by regional axonal projections. Nat. Neurosci. 16, 1426–1435 (2013).
36.
Townsend, R. G. et al. Emergence of complex wave patterns in primate cerebral cortex.
J. Neurosci. 35, 4657–4662 (2015).
37.
Rubino, D., Robbins, K. A. & Hatsopoulos, N. G. Propagating waves mediate information
transfer in the motor cortex. Nat. Neurosci. 9, 1549–1557 (2006).
38. Huang, X. et al. Spiral wave dynamics in neocortex. Neuron 68, 978–990 (2010).
39.
Xu, Y., Long, X., Feng, J. & Gong, P. Interacting spiral wave patterns underlie complex
brain dynamics and are related to cognitive processing. Nat Hum Behav 7, 1196–1215
(2023).
40.
Gong, P. & van Leeuwen, C. Distributed dynamical computation in neural circuits with
propagating coherent activity patterns. PLoS Comput. Biol. 5, e1000611 (2009).
41.
Chen, G. & Gong, P. Computing by modulating spontaneous cortical activity patterns as
a mechanism of active visual processing. Nat. Commun. 10, 4915 (2019).
22
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
42.
Heitmann, S., Rule, M., Truccolo, W. & Ermentrout, B. Optogenetic Stimulation Shifts
the Excitability of Cerebral Cortex from Type I to Type II: Oscillation Onset and Wave
Propagation. PLoS Comput. Biol. 13, e1005349 (2017).
43.
Pinotsis, D. A. & Miller, E. K. In vivo ephaptic coupling allows memory network formation.
Cereb. Cortex 33, 9877–9895 (2023).
44.
Das, A., Zabeh, E., Ermentrout, B. & Jacobs, J. Planar, Spiral, and Concentric Traveling
Waves Distinguish Cognitive States in Human Memory. bioRxiv, 2024.01.26.577456 (2024).
45.
Muller, L., Reynaud, A., Chavane, F. & Destexhe, A. The stimulus-evoked population
response in visual cortex of awake monkey is a propagating wave. Nat. Commun. 5, 3675
(2014).
46.
Lee, S.
-
H., Blake, R. & Heeger, D. J. Traveling waves of activity in primary visual cortex
during binocular rivalry. Nat. Neurosci. 8, 22–23 (2005).
47.
Davis, Z. W., Muller, L., Martinez-Trujillo, J., Sejnowski, T. & Reynolds, J. H. Spontaneous
travelling cortical waves gate perception in behaving primates. Nature 587, 432–436
(2020).
48.
Liang, Y. et al. Complexity of cortical wave patterns of the wake mouse cortex. Nat.
Commun. 14, 1434 (2023).
49.
Munn, B. R., Müller, E. J., Wainstein, G. & Shine, J. M. The ascending arousal system
shapes neural dynamics to mediate awareness of cognitive states. Nat. Commun. 12, 6016
(2021).
50.
Skarda, C. A. & Freeman, W. J. How brains make chaos in order to make sense of the
world. Behav. Brain Sci. 10, 161–173 (1987).
51.
Valle, M. et al. Inhibition of alpha oscillations through serotonin-2A receptor activation
underlies the visual effects of ayahuasca in humans. Eur. Neuropsychopharmacol. 26,
1161–1175 (2016).
52.
Pallavicini, C. et al. Neural and subjective effects of inhaled N,N-dimethyltryptamine in
natural settings. J. Psychopharmacol. 35, 406–420 (2021).
53.
Pallavicini, C. et al. Spectral signatures of serotonergic psychedelics and glutamatergic
dissociatives. Neuroimage 200, 281–291 (2019).
54.
Timmermann, C. et al. Neural correlates of the DMT experience assessed with multivariate
EEG. Sci. Rep. 9, 16324 (2019).
55.
Timmermann, C. et al. Human brain effects of DMT assessed via EEG-fMRI. Proc. Natl.
Acad. Sci. U. S. A. 120, e2218949120 (2023).
56.
Muthukumaraswamy, S. D. et al. Broadband cortical desynchronization underlies the
human psychedelic state. J. Neurosci. 33, 15171–15183 (2013).
57.
Schenberg, E. E. et al. Acute Biphasic Effects of Ayahuasca. PLoS One 10, e0137202
(2015).
58.
Carhart-Harris, R. L. et al. Neural correlates of the LSD experience revealed by multimodal
neuroimaging. Proc. Natl. Acad. Sci. U. S. A. 113, 4853–4858 (2016).
23
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
59.
Riba, J. et al. Topographic pharmaco-EEG mapping of the effects of the South American
psychoactive beverage ayahuasca in healthy volunteers. Br. J. Clin. Pharmacol. 53, 613–
628 (2002).
60.
Riba, J., Anderer, P., Jané, F., Saletu, B. & Barbanoj, M. J. Effects of the South
American psychoactive beverage ayahuasca on regional brain electrical activity in hu-
mans: a functional neuroimaging study using low-resolution electromagnetic tomography.
Neuropsychobiology 50, 89–101 (2004).
61.
Don, N. S. et al. Effects of Ayahuasca on the human EEG. Phytomedicine 5, 87–96 (1998).
62.
Ort, A. et al. TMS-EEG and resting-state EEG applied to altered states of consciousness:
oscillations, complexity, and phenomenology. iScience 26, 106589 (2023).
63.
Barnett, L., Muthukumaraswamy, S. D., Carhart-Harris, R. L. & Seth, A. K. Decreased
directed functional connectivity in the psychedelic state. Neuroimage 209, 116462 (2020).
64.
Kometer, M., Pokorny, T., Seifritz, E. & Volleinweider, F. X. Psilocybin-induced spiritual
experiences and insightfulness are associated with synchronization of neuronal oscillations.
Psychopharmacology 232, 3663–3676 (2015).
65.
Van Ede, F., Quinn, A. J., Woolrich, M. W. & Nobre, A. C. Neural Oscillations: Sustained
Rhythms or Transient Burst-Events? Trends Neurosci. 41, 415–417 (2018).
66.
Cole, S. R. & Voytek, B. Brain Oscillations and the Importance of Waveform Shape.
Trends Cogn. Sci. 21, 137–149 (2017).
67.
Uddin, L. Q. Bring the Noise: Reconceptualizing Spontaneous Neural Activity. Trends
Cogn. Sci. 24, 734–746 (2020).
68.
Quinn, A. J. et al. Within-cycle instantaneous frequency profiles report oscillatory wave-
form dynamics. J. Neurophysiol. 126, 1190–1208 (2021).
69.
Hazarika, N., Tsoi, A. C. & Sergejew, A. A. Nonlinear considerations in EEG signal
classification. IEEE Trans. Signal Process. 45, 829–836 (1997).
70.
García-Rosales, F., Schaworonkow, N. & Hechavarria, J. C. Oscillatory Waveform Shape
and Temporal Spike Correlations Differ across Bat Frontal and Auditory Cortex. J.
Neurosci. 44 (2024).
71.
Donoghue, T. et al. Parameterizing neural power spectra into periodic and aperiodic
components. Nat. Neurosci. 23, 1655–1665 (2020).
72.
Fabus, M. S., Quinn, A. J., Warnaby, C. E. & Woolrich, M. W. Automatic decomposition
of electrophysiological data into distinct nonsinusoidal oscillatory modes. J. Neurophysiol.
126, 1670–1684 (2021).
73.
Frohlich, J., Toker, D. & Monti, M. M. Consciousness among delta waves: a paradox?
Brain 144, 2257–2277 (2021).
74.
Vesuna, S. et al. Deep posteromedial cortical rhythm in dissociation. Nature 586, 87–94
(2020).
75.
Acosta-Urquidi, J. QEEG Studies of the Acute Effects of the Visionary Tryptamine DMT.
Cosmos and History: The Journal of Natural and Social Philosophy 11, 115–129 (2015).
24
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
76.
Mediano, P. A. M. et al. Towards an extended taxonomy of information dynamics via
Integrated Information Decomposition. arXiv [q-bio.NC] (2021).
77.
Erritzoe, D. et al. Exploring mechanisms of psychedelic action using neuroimaging. Nature
Mental Health 2, 141–153 (2024).
78.
Carhart-Harris, R. L. & Friston, K. J. REBUS and the Anarchic Brain: Toward a Unified
Model of the Brain Action of Psychedelics. Pharmacol. Rev. 71, 316–344 (2019).
79.
Alamia, A., Timmermann, C., Nutt, D. J., VanRullen, R. & Carhart-Harris, R. L. DMT
alters cortical travelling waves. Elife 9(2020).
80.
Hipólito, I., Mago, J., Rosas, F. E. & Carhart-Harris, R. Pattern breaking: a complex
systems approach to psychedelic medicine. Neurosci Conscious 2023, niad017 (2023).
81.
Girn, M. et al. A complex systems perspective on psychedelic brain action. Trends Cogn.
Sci. 27, 433–445 (2023).
82.
Singleton, S. P. et al. Receptor-informed network control theory links LSD and psilocybin
to a flattening of the brain’s control energy landscape. Nat. Commun. 13, 5812 (2022).
83.
Shine, J. M. The thalamus integrates the macrosystems of the brain to facilitate complex,
adaptive brain network dynamics. Prog. Neurobiol. 199, 101951 (2021).
84.
Shine, J. M., Lewis, L. D., Garrett, D. D. & Hwang, K. The impact of the human thalamus
on brain-wide information processing. Nat. Rev. Neurosci. 24, 416–430 (2023).
85.
Sanchez-Vives, M. V. & Mattia, M. Slow wave activity as the default mode of the cerebral
cortex. Arch. Ital. Biol. 152, 147–155 (2014).
86.
Sanchez-Vives, M. V., Massimini, M. & Mattia, M. Shaping the Default Activity Pattern
of the Cortical Network. Neuron 94, 993–1001 (2017).
87.
Sanchez-Vives, M. V. & McCormick, D. A. Cellular and network mechanisms of rhythmic
recurrent activity in neocortex. Nat. Neurosci. 3, 1027–1034 (2000).
88.
Vollenweider, F. X. & Geyer, M. A. A systems model of altered consciousness: integrating
natural and drug-induced psychoses. Brain Res. Bull. 56, 495–507 (2001).
89.
Gramfort, A. et al. MNE software for processing MEG and EEG data. Neuroimage 86,
446–460 (2014).
90.
Tenke, C. E. & Kayser, J. A convenient method for detecting electrolyte bridges in multi-
channel electroencephalogram and event-related potential recordings. Clin. Neurophysiol.
112, 545–550 (2001).
91.
Greischar, L. L. et al. Effects of electrode density and electrolyte spreading in dense array
electroencephalographic recording. Clin. Neurophysiol. 115, 710–720 (2004).
92.
Fabus, M. S., Woolrich, M. W., Warnaby, C. W. & Quinn, A. J. Understanding Harmonic
Structures Through Instantaneous Frequency. IEEE Open J Signal Process 3, 320–334
(2022).
93.
Huang, N. E. et al. The empirical mode decomposition and the Hilbert spectrum for
nonlinear and non-stationary time series analysis. Proceedings of the Royal Society A:
Mathematical, Physical and Engineering Sciences (1998).
25
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
This work is shared under a CC BY license unless otherwise noted
94.
Yang, Y., Deng, J. & Wu, C. Analysis of Mode Mixing Phenomenon in the Empirical
Mode Decomposition Method in 2009 Second International Symposium on Information
Science and Engineering (IEEE, 2009), 553–556.
95.
Bueno-López, M., Giraldo, E., Molinas, M. & Fosso, O. The Mode Mixing Problem and
its Influence in the Neural Activity Reconstruction. IAENG International Journal of
Computer Science (2019).
96. Huang, N. E. Hilbert-Huang Transform and Its Applications (World Scientific, 2014).
97.
Huang, N. E. et al. On Holo-Hilbert spectral analysis: a full informational spectral
representation for nonlinear and non-stationary data. Philos. Trans. A Math. Phys. Eng.
Sci. 374, 20150206 (2016).
98.
Quinn, A. J., Lopes-Dos-Santos, V., Dupret, D., Nobre, A. C. & Woolrich, M. W. EMD:
Empirical Mode Decomposition and Hilbert-Huang Spectral Analyses in Python. J Open
Source Softw 6(2021).
99.
Fazle Hussain, A. K. Coherent structures and turbulence. J. Fluid Mech. 173, 303–356
(1986).
100.
Horn, B. K. P. & Schunck, B. G. Determining optical flow. Artif. Intell. 17, 185–203
(1981).
101.
Wildie, M. & Shanahan, M. Metastability and chimera states in modular delay and
pulse-coupled oscillator networks. Chaos 22, 043131 (2012).
102.
Shanahan, M. Metastable chimera states in community-structured oscillator networks.
Chaos 20, 013108 (2010).
103. Vicsek, T. & Zafeiris, A. Collective motion. Phys. Rep. 517, 71–140 (2012).
104.
Perry, A. E. & Chong, M. S. A description of eddying motions and flow patterns using
critical-point concepts. Annu. Rev. Fluid Mech. 19, 125–155 (1987).
105.
Varley, T. F. & Sporns, O. Network analysis of time series: Novel approaches to network
neuroscience. Front. Neurosci. 15, 787068 (2021).
106. Clauset, A., Newman, M. E. J. & Moore, C. Finding community structure in very large
networks. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 70, 066111 (2004).
107.
Newman, M. E. J. Equivalence between modularity optimization and maximum likelihood
methods for community detection. Phys Rev E 94, 052315 (2016).
108.
Sorrentino, P. et al. Dynamical interactions reconfigure the gradient of cortical timescales.
Netw Neurosci 7, 73–85 (2023).
109.
Rosenstein, M. T., Collins, J. J. & De Luca, C. J. A practical method for calculating
largest Lyapunov exponents from small data sets. Physica D 65, 117–134 (1993).
110.
Müller, E. J. et al. The non-specific matrix thalamus facilitates the cortical information
processing modes relevant for conscious awareness. Cell Rep. 42, 112844 (2023).
111.
Taylor, N. L. et al. Causal evidence for cholinergic stabilization of attractor landscape
dynamics. Cell Rep. 43, 114359 (2024).
26
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted October 7, 2024. ; https://doi.org/10.1101/2024.10.04.616717doi: bioRxiv preprint
