The paper introduces the "shiftability" paradigm to bridge the translational gap by experimentally manipulating neurotransmitter systems in humans to observe causal effects on brain function in autism. It presents PSILAUT, an exemplar protocol using psilocybin as an in vivo serotonin probe to test whether serotonergic modulation produces different neurobiological “shifts” in autistic versus non‑autistic adults.
Clinical trials of pharmacological approaches targeting the core features of autism have failed. This is despite evidence from preclinical studies, genetics, post-mortem studies and correlational analyses linking peripheral and central markers of multiple candidate neurochemical systems to brain function in autism. Whilst this has in part been explained by the heterogeneity of the autistic population, the field has largely relied upon association studies to link brain chemistry to function. The only way to directly establish that a neurotransmitter or neuromodulator is involved in a candidate brain function is to change it and observe a shift in that function. This experimental approach dominates preclinical neuroscience, but not human studies. There is very little direct experimental evidence describing how neurochemical systems modulate information processing in the living human brain. As a result, our understanding of how neurochemical differences contribute to neurodiversity is limited and impedes our ability to translate findings from animal studies into humans. Here, we begin by introducing our “shiftability” paradigm, an approach to bridge the translational gap in autism research. We then provide an overview of the methodologies used and explain our most recent choice of psilocybin as a pharmacological probe of the serotonin system in vivo . Finally, we provide a summary of the protocol for ‘PSILAUT’, an exemplar “shiftability” study which uses psilocybin to directly test the hypothesis that the serotonin system functions differently in autistic and non-autistic adults.
Papers cited by this study that are also in Blossom
Wojtas, A., Bysiek, A., Wawrzczak-Bargiela, A. et al. · International Journal of Molecular Sciences (2022)
Carhart-Harris, R. L., Nutt, D. J. · Journal of Psychopharmacology (2017)
Dinis-Oliveira, R. J. · Drug Metabolism Reviews (2017)
Halberstadt, A. L., Geyer, M. A. · Neuropharmacology (2011)
Autism spectrum disorder (ASD) is a lifelong neurodevelopmental condition characterised by differences in social communication, repetitive or restricted behaviours, and sensory atypicalities. Clinical trials seeking pharmacological treatments for core ASD features have largely failed, a problem the authors attribute not only to diagnostic heterogeneity but also to a lack of experimental paradigms that causally link neurochemical systems to brain function in humans. Most human neuroscience work in this area has relied on correlational approaches that cannot establish whether a given neurotransmitter system actually regulates a candidate brain mechanism. Whelan and colleagues introduce a “shiftability” paradigm intended to bridge this translational gap by deliberately perturbing neurochemical systems in vivo and measuring the resulting ‘shift’ in brain function across multiple organisational levels. The paper outlines the rationale for this experimental medicine approach, summarises methodological options, and presents an exemplar protocol (PSILAUT) that uses low doses of psilocybin as a serotonergic probe to test the hypothesis that 5HT-mediated regulation of sensory processing, local circuits and whole-brain networks differs between autistic and non-autistic adults. The broader goal is to create a pharmacological repository describing how different drugs change (or fail to change) brain mechanisms across individuals, thereby improving stratification for future clinical trials.
PSILAUT is described as a placebo-controlled, randomised, double-blind, repeated-measures, cross-over, case-control study. The investigators plan to recruit up to 70 adult participants (reporting 40 autistic and 30 non-autistic controls) who will each attend three dosing visits. On separate occasions participants will receive oral synthetic COMP360 psilocybin at one of two low doses (2 mg and 5 mg) or placebo. Dosing order will be pseudo-randomised to balance first-visit conditions but the low dose will always precede the higher dose to allow for safety-driven unblinding if needed. Screening will be via remote contact and written informed consent obtained; baseline characterisation will include an accepted expert ASD diagnosis (supplemented by standard diagnostic instruments where available) and questionnaires measuring core autistic features, sensory function and co-occurring psychopathology. Multimodal neurophysiological and neuroimaging measures form the core outcomes. MRI acquisitions on a 3.0 Tesla scanner will include structural imaging, resting-state functional MRI with multiband sequences (to permit dynamic functional connectivity analyses) and a face emotion processing task. Magnetic resonance spectroscopy using a HERMES sequence will target the dorsomedial prefrontal cortex to estimate GABA and glutamate–glutamine measures that index excitation/inhibition (E/I) balance. High-density (64-channel) EEG will be recorded at rest and during passive sensory paradigms: visual steady-state evoked potentials probing surround suppression, an auditory oddball paradigm measuring mismatch negativity (MMN) as an index of repetition suppression and deviance response, and a tactile paired-pulse inhibition task for sensory gating. Task-dependent EEG will measure event-related potentials such as the N170 to faces. Psychophysical tests (tactile detection threshold, amplitude discrimination), a cognitive battery (Reading the Mind in the Eyes test, probabilistic reversal learning, verbal fluency), the 5D-ASC questionnaire to index subjective drug effects, and peripheral measures including whole blood serotonin are also included. A wrist wearable will continuously record heart rate and heart-rate variability as autonomic arousal indices. Analytically, the design uses within-subject repeated measures so each participant serves as their own control. Both parametric and non-parametric statistics will test the primary hypotheses that psilocybin elicits modality-specific ‘shifts’ that differ by diagnostic group. The investigators plan to compute individual-level ‘shift’ metrics for each modality and examine associations with clinical measures (for example, questionnaire scores and whole blood serotonin), and to explore post-hoc multimodal relationships. Power considerations are based on prior pharmacological neuroimaging studies by the group that detected group differences with small samples (n < 20), and sample sizes for sensory tasks are justified with reference to medium effect detection at conventional power thresholds. Safety oversight and procedures for participant dropout and replacement are described at a protocol level; explicit exclusion criteria are not fully detailed in the extracted text.
The extracted text does not report empirical results; this paper presents a protocol and rationale rather than outcome data. Instead, the investigators state specific, testable expectations derived from prior literature and their own previous pharmacological challenge work. They predict that low-dose psilocybin will engage 5HT2A receptor-mediated mechanisms and produce measurable shifts in multiple modalities, including alterations in EEG oscillatory metrics (beta/gamma power, aperiodic activity) and local circuit proxies of E/I balance, changes in MRS-derived GABA and glutamate–glutamine markers, modulation of sensory suppression in visual (SSVEP surround suppression), auditory (repetition suppression/MMN) and tactile (paired-pulse inhibition) paradigms, and acute effects on resting-state and dynamic functional connectivity measured with fMRI and EEG. Task-dependent measures are expected to show altered dynamics of face-emotion processing (both fMRI and N170 ERP) and possible changes in cognitive tests sensitive to social cognition and flexibility. The authors also anticipate individual differences in pharmacological response and plan to relate observed shifts to baseline clinical and peripheral markers such as whole blood serotonin. No numerical effect sizes, confidence intervals or statistical outcomes are provided because no study data are reported.
Whelan and colleagues argue that the shiftability paradigm offers a way to move beyond correlational studies by directly testing whether perturbation of a neurochemical system produces predictable changes in brain mechanisms linked to autism. They position this experimental medicine approach as potentially more sensitive to individual neurobiological differences than standard case–control comparisons, and emphasise that measures targeting early or ‘foundational’ properties of brain organisation (sensory processing, local circuit E/I balance and network connectivity) may be stable across the lifespan and back-translatable to animal models. The authors acknowledge important limitations. The chosen pharmacological probe, psilocybin, is not entirely selective for 5HT2A receptors and binds multiple serotonin receptor subtypes, which complicates receptor-level interpretation. Practical constraints of the protocol—use of EEG caps, requirement to remain visually fixated, and repeat visits—may limit generalisability to autistic people with higher support needs or younger age groups; to mitigate this they include eye-tracking and focus on many passive tasks that require minimal behavioural response. The team also recognises participant drop-out risk in repeated-measures designs and states plans to replace participants as needed. They note that low doses (2 mg and 5 mg) are chosen to reduce the likelihood of marked psychedelic experiences while still engaging 5HT2A receptors, and that safety monitoring and a conservative dosing order are embedded in the protocol. In terms of implications, the investigators propose that pharmacological profiling using the shiftability approach could identify subgroups who do or do not show biologically meaningful responses to serotonergic modulation, thereby informing personalised-medicine strategies and improving the design of future clinical trials. They suggest that this approach could reduce the likelihood of failure in large-scale pharmacological trials that enrol participants on diagnostic criteria alone and do not assess mechanism.
The paper concludes that the shiftability paradigm aims to map how neurochemical systems modulate brain function across organisational scales and to identify individual differences between autistic and non-autistic adults. Using low-dose psilocybin as an exemplar serotonergic probe, the PSILAUT protocol seeks to reveal whether serotonergic regulation of sensory processing, local circuits and network dynamics differs by diagnostic group. The authors propose that results from such experimental medicine studies may improve understanding of neurodivergent brain biology, support more personalised approaches to pharmacological interventions, and help avoid costly, mechanism-insensitive clinical trials by enabling better stratification of likely responders and non-responders.
Autistic spectrum disorder (ASD), the overarching definition used by our current diagnostic guidelines, is a life-long neurodevelopmental condition characterised by differences in social interaction and communication, repetitive or restricted patterns of behaviour, and sensory differences. Clinical trials have failed to offer pharmacological options for those that seek support. This in part reflects the complex neurobiology and diversity of autism. However, clinical trial efforts have also been hampered because there is no reliable experimental bridge from 'basic' neuroscience to clinical application. This is not an isolated problem; it is common across neuropsychiatric conditions. We do not yet fully understand how neurochemical systems regulate information processing within and across different levels of brain organisation in neurotypical or neurodivergent people, and in particular, how this links to core traits of these conditions. A constraint is that human neuroscience research has largely relied on association or correlational (and cross-sectional) evidence linking genetic, preclinical, post-mortem and neuroimaging data to, for example, trait(s), symptom(s) or diagnoses. However, correlations are not causal. The only way to establish that a neurosignalling system is involved in a candidate brain mechanism is to change it and observe a 'shift' in that mechanism. This is basis of preclinical study designs, but there has been limited translation to humans. We have developed a direct experimental approach to examine neurochemical regulation of information processing across the organisational levels of the human brain which is sensitive to individual differences in neurotypical and neurodivergent people. We call this a "shiftability" paradigm. To date, our "shiftability" studies have demonstrated that the autistic brain responds differently when different neurochemical systems are perturbed by a single, low dose pharmacological challenge. Here, we begin by introducing our "shiftability" paradigm, a novel approach to bridge the translational gap in autism research. We then provide an overview of the methodologies used and explain our most recent choice of psilocybin as a pharmacological probe of the serotonin system in vivo. Finally, we provide a summary of the protocol for 'PSILAUT', our latest "shiftability" study which uses psilocybin to directly test the hypothesis that the brain's serotonin system functions differently in autistic and non-autistic adults. The end goal across our many shiftability studies is to build a pharmacological repository of how different drugs 'shift' brain processing mechanisms across multiple levels of organisation in autistic and nonautistic individuals.
There is currently no reliable 'biomarker' that separates people with an autism diagnosis from controls. For example, functional magnetic resonance imaging (fMRI) data from the EU-AIMS consortium indicates that there is no statistically significant difference in brain activation during facial emotion processing in autism. This work emphasises the message from a recent large-scale meta-analysis carried out by Mottron and colleagues. They found that, even if there are statistically significant case-control differences, the effect sizes reported from ASD casecontrol analyses are modest at best (the largest is around a standard deviation of 1.0 for Theory of Mind studies) and have gotten smaller over time, despite ever larger study sample sizes. One interpretation of this pattern is that we have too broad a diagnostic definition of autism; another is that the larger the sample size the more likely it is to revert towards the mean; and yet another is that we need to identify more biologically homogeneous subgroups (i.e. stratify). None of these approaches have significantly advanced the field. We propose a different strategy -that measures should be sensitive to individual diversity and capture more 'foundational' aspects of brain function; and that they should permit examination of how these functions 'shift' in response to a change in neurochemical pathways in autistic and non-autistic people. In selecting measures that capture 'shift', we must recognise that the genetic and environmental influences which increase the likelihood of neuropsychiatric conditions do not map directly to autism. Instead, they generate the neurobiological conditions for autism and other neurodevelopmental, neurological and mental health conditions, such as schizophrenia, attention deficit hyperactivity disorder, mood disorders, epilepsy and learning difficulties. These are life factors which act on the brain systems developing early, namely sensory systems and emerging interconnecting whole-brain networks. Altering these 'foundational' aspects of brain development may have cascading effects as the brain matures, generating diversity in cognition and behaviour. Secondary and compensatory mechanisms will also continue to act across the lifespan leading to heterogeneous clinical outcomes. However, early changes to the foundational properties of the brain should persist and could therefore be a measurement target of differences in brain function. In support of this principle, we and others have evidence that the regulation of sensory processing and the functional connectivity of brain networks is associated with ASD across the lifespan. For example, using EEG, Kolenik and colleagues have reported that infants later diagnosed with autism do not suppress neural responses to repeated auditory stimulation; similarly, Piccardi and colleagues identified that infants with elevated likelihood for autism show reduced tactile sensory gating. Using MRI, we have reported that the local functional connectivity of sensory systems is altered in newborns with a higher likelihood of developing autism; and the global functional connectivity of brain networks at full-term predicts emergence of later autistic traits. Moreover, we have reported similar profiles in autistic adults, namely reduced sensory suppression detected using EEGand altered functional connectivity of sensory systems and whole-brain networks using MRI. This is important because these measures not only capture brain function across age groups, they may also be back-translated to animal models to support a translational bridge. Also important is that selected measures of altered brain function are likely to be sensitive to modulation with a pharmacological challenge. Using fMRI, we have consistently found this to be the case. For example, we have reported that fMRI response can be modulated with single dose drug challenges which target components of the serotonin system (citalopram and tianeptine), glutamate-GABA system (riluzole)and endocannabinoid system (cannabidiol). We have also used EEG recordings during sensory stimulation to capture response to GABAB receptor agonist, arbaclofen. We have been able to quantify individual 'shifts' in these measures in response to drug challenge and differences at group-level between autistic and non-autistic participants. The range of measures used to capture 'shift' and their target level of brain organisation is shown in Figure.
Alterations in the serotonin system are consistently reported in autism and hence, whether there are functional differences in this system is an important focus for our "shiftability" studies. However, the evidence linking serotonin to autism to date, has mostly been indirect from associations and/or correlational studies. For example, polymorphisms in genes for serotonin synthesis, transporters and receptors are associated with autism. Elevated whole blood serotonin levels are also reported in one-third of autistic individuals. In contrast to these indirect approaches, direct alteration of serotonergic system function in vivo can be achieved using pharmacological probes. In animal model systems, this modulation of receptor targets of serotonin has included agonism of 5HT1A and 5HT7 receptors to increase social behaviour and reduce stereotypy; and activation of 5HT2A, 5HT1A and 5HT7 receptors to shift glutamate-GABA indices of excitation/inhibition (E/I) balance, proposed to be altered in autism. In humans, this approach is less frequently taken, but our team has previously reported that acutely elevating serotonin levels with a single dose of selective-serotonin reuptake inhibitor (SSRI) citalopram produces sustained activation of brain regions associated with facial expression processing in autistic adults, but not in controls. This finding suggests that acute non-selective activation of the serotonin system may influence behaviours associated with autism. There is also evidence implicating specific serotonin receptors in autism. The 5HT2A receptor, encoded by the HT2RA gene, is a robust functional candidate gene in autism, and lower cortical 5HT2A receptor binding has been reported to correlate with social communication differences in autism. The 5HT2A receptor is expressed throughout the cortex but especially in regions related to sensorimotor integrationand the so-called default mode network responsible for "self" and "other" processing. At the circuit level, 5HT2A receptor signalling is thought to enhance neural plasticityand increases cortical glutamate and thalamic GABA levels. However, it is not known whether there are autistic differences in 5HT2A regulation of sensorimotor processes, networks such as the default mode network, and regulation of glutamate-GABA levels is not known. To assess this, we need targeted pharmacological probes. Unlike animal studies, the choice of probe in humans is constrained by the safety and side effect profile of candidate compounds, and few (if any) neuropsychiatric drugs used in people are entirely selective. With this caveat in mind, we selected psilocybin (4phosphoryloxy-N.N-dimethyltryptamine) as a pharmacological probe of the serotonin system based on its affinity for 5HT2ARs. Psilocybin is a classic psychedelic compound produced by several species of mushrooms, including so-called "magic mushrooms". Psilocybin is rapidly metabolised into its active component psilocin. Psilocin is a 5HT2AR agonist but also binds several serotonin receptors, including 5HT7, 5HT2B, 5HT1D, 5HT6, 5HT5, 5HT2C & 5HT1B receptors in decreasing order of reported affinity. Prior studies have used relatively high doses of psilocybin to explore the effects of psychedelics on the brain. However, we will use lower doses (2 mg and 5 mg) in our "shiftability" protocol to assess whether 5HT2AR-mediated brain responses are uncomplicated by marked psychedelic experiences. We are confident that our dose range will generate a 'shift' in brain function based on evidence using positron emission topography that similar doses of psilocybin engage 5HT2AR receptorsand low dose psilocybin is sufficient to alter cognition and obsessivecompulsive behaviour. Low doses of the serotonergic psychedelic lysergic acid diethylamide (LSD) and psilocybin-containing mushrooms also acutely alter brain resting-state fMRI and EEG indices in the non-autistic population, as well as neural responses to sensory stimuli. Therefore, we expect that 2 mg and 5 mg of psilocybin will expose functional differences in the serotonin system targeted by psilocybin in autistic and non-autistic individuals. We aim to recruit up to 70 healthy adult participants including 40 autistic and 30 non-autistic control participants in our placebo-controlled, randomised, double-blind, repeated-measures, cross-over, case-control study. All participants will provide written informed consent.
Participants will receive either placebo or one of two single doses (2 mg or 5 mg) of oral synthetic COMP360 psilocybin on three separate visits. The order of administration of placebo and psilocybin will be pseudo-randomised to ensure balanced numbers of individuals have placebo or drug on their first visit, and to control for order effects. However, the lowest dose of psilocybin will always precede the higher dose to allow for unblinding should a participant experience unwanted side effects on either of the first two visits. If that visit is a low dose psilocybin visit, we will avoid exposing that participant to the higher dose on a subsequent visit. Interested participants will be sent an information sheet and screened via video call (eg. Zoom) or phone for eligibility. Written consent will be sought after inclusion criteria are confirmed and the participant is then assigned to a study schedule. Participants will be provided with log in details to an online platform (Delosis Ltd., London) to complete a battery of questionnaires remotely. Participants will visit the study site on three separate occasions. A blood sample will be collected on one of the three visits for quantification of whole blood serotonin levels. Each participant will complete an MRI scan session to acquire a structural, resting-state functional MRI scan and a face emotion processing task. Participants will wear a wrist wearable (ActiGraph, Pensacola, FL, USA) following the MRI scan session for the remainder of the study visit. The EEG paradigm will include resting-state and functional activation during a face processing, auditory oddball, visual processing and tactile paired-pulse inhibition tasks. Psychophysical tasks will be collected prior to a cognitive battery which will include the 'reading the mind in the eyes' (RMET), probabilistic reversal learning (PRT), both of which will be delivered using PsyTools (Delosis Ltd., London) and a semantic verbal fluency task. The 5dimensional altered states of consciousness (5D-ASC) questionnaire will then be completed to quantify any subjective effects experienced by participants.
Baseline characterisation: A comprehensive baseline characterisation will be obtained. An expert clinical diagnosis of ASD from a recognised UK assessment service will be accepted. This may be supported by the Autism Diagnostic Interview-Revisedwhere an appropriate informant is available. An Autism Diagnostic Observation Schedule 57 will be used to assess current symptom level, but if it has already been used to inform the diagnostic assessment in adulthood, it will not be repeated. Additional baseline questionnaires will quantify core autistic features (e.g. social behaviour or sensory differences), relevant cognitive domains (e.g. intolerance of uncertainty and behavioural flexibility) and the symptomology of cooccurring psychiatric conditions.
Reconstruction of MEGA-Edited Spectroscopy (HERMES) sequencewill be collected during the MRI scan for the dorsal medial prefrontal cortex region. HERMES permits the quantification of levels of metabolites in the living brain and is focused on estimating GABA and Glutamate-glutamine markers of E/I balance. For the purposes of our study, given the evidence that E-I pathways are modulated by 5HT2A receptor action in animal models, we will be able to examine the impact of psilocybin on these tissue level measures of E-I balance.
Resting-state: High-density (64-channel) EEG data will be collected during the resting-state, metrics from which local circuit activity can be derived such as beta and gamma band power/frequency. Oscillations in the beta frequency band at rest, for example, are associated with inhibitory neurotransmitter levels in sensorimotor cortex. This, and other EEG-derived metrics such as aperiodic activity, are considered a proxy measure for E/I balance in vivo. Hence, we will be able to examine the impact of psilocybin on these dynamic measures of E-I balance.
Visual domain: A visual processing task (contrast saturation) in which steady-state evoked potentials (SSVEPs) are elicited by passive surround suppression stimuli will be conducted. We have shown that SSVEPs during this task are altered in autism. 5HT2ARs are particularly highly expressed in the primary visual cortex, and their agonism alters visual response amplitudes and surround suppression in mouse primary visual cortex. In humans, we expect visual processing to be altered by 5HT2AR activation given that the marked visual perceptual changes robustly induced with higher doses of psychedelics are blocked by pretreatment with the 5HT2 receptor antagonist, ketanserin.
A conventional auditory oddball paradigm (mismatch negativity, MMN)will be used to passively measure 'repetition suppression' (or habituation) to repetitive auditory stimuli and response to an unexpected 'deviant' stimulus (the eventrelated mismatch negativity MMN response). We and others have observed less repetition suppression in both eight-month-old infants who go on to receive a diagnosis or autism, and adults with a diagnosis of autism. Thus, this signal appears linked to autism across infancy to maturity. The impact of autism on the event-related MMN is less consistent and varies with age. The latter may in part be due to differences in the serotonin system, as the MMN response can be modulated by acute elevation of serotonin levels by the highly selective SSRI escitalopram. In this study we will test the prediction that psilocybin alters both sensory suppression and MMN in autism differently compared to controls.
A tactile paired-pulse inhibition task will be used to passively investigate EEG responses as objective markers of sensory gating. Aligning with findings in the auditory and visual domain, less tactile neural repetition suppression in this task is associated with autism. As processing of tactile stimuli is also known to be perturbed by 5HT2AR agonism with psilocybin, we also expect to elicited functional differences following psilocybin in autistic and non-autistic individuals in this paradigm.
Resting-State: EEG: Oscillatory power will be assessed across multiple frequency bands during the restingstate. This will include electrodes over key brain regions implicated in autism such as those belonging to the default mode network. Reduced oscillatory power over DMN regions using electrophysiological approaches following 5HT2AR activation by psilocybin has been reported previously. Functional connectivity analyses (e.g. within and between brain networks) can also be derived from EEG, and this will complement connectivity analyses from resting-state fMRI.
Participants will undergo a structural and functional MRI scan. Scans will be acquired on a 3.0 Tesla MR Scanner (General Electric Premier). A fMRI scan with a multiband 4 sequence will be acquired during the resting-state, multiband 4 is preferable for connectivity analyses. In addition, multiband sequences will considerably reduce the repetition time (TR), therefore they have the advantage of allowing dynamic functional connectivity analyses. Autistic differences in both 'averaged' functional connectivity and dynamic functional connectivity have been reliably reported across different datasets. Functional connectivity of brain networks in neurotypical individuals has also be shown to be acutely modulated by 5HT2AR activation. Notably, psilocybin alters dynamic functional connectivity, mediated by 5HT2AR agonism. It facilitates state transitions and more temporally diverse brain activity in neurotypical individuals. Our study will be the first to examine the effects of psilocybin on conventional and dynamic functional metrics in autistic individuals.
MRI: fMRI studies of face emotion processing in autism have produced inconsistent results. In the largest study to date, no differences between autistic and non-autistic individuals in fMRI response to facial expressions of emotion were observed. However, we have recently examined the fMRI response to facial expressions of emotion in a social brain network before and after administration of the SSRI, citalopram. We reported that the dynamics of the response to faces is different in autism, in that 5HT reuptake inhibition slows habituation. Consistent with this, blockade of 5HT2ARs causes an 'opposite' effect and reduces neural responses to emotional faces during fMRI in neurotypical individuals. Therefore, we expect that psilocybin will alter the dynamics of face emotion processing, but differently in autistic individuals. EEG: Event-related potentials (ERPs) in response to face stimuli will also be assessed during EEG. The N170 component, a neural response present at 170ms following the presentation of facial stimuli and can be modulated by 5HT2AR activation. An altered N170 response is associated with social communication differences in autism and may have utility as a stratification marker that is amenable to support. It is also going to be the first prognostic biomarker for autism (or any neurodevelopmental or psychiatric condition) to be approved by regulatory agencies. The incorporation of this task in our protocol therefore will be an important test of whether an autism biomarker can be modified pharmacologically.
Psychophysical approaches are structured approaches in which stimulus characteristics are tightly controlled, and they provide robust, objective measures of sensory sensitivity by estimating perceptual metrics. Serotonin has been directly implicated in tactile perception, such as in affective touch, by studies using tryptophan depletion (which acutely reduces central serotonin levels) alongside psychophysical approaches. Tactile detection threshold and amplitude discrimination will be assessed, as differences in tactile thresholds have already been reported in autism and are associated with outcomes 82 .
Questionnaires: The 5-dimensional altered states of consciousness (5D-ASC) questionnairewill be completed on each visit following the completion of study procedures to quantify the subjective effects of psilocybin, which are primarily mediated by the 5HT2AR. Cognitive Battery: 'Theory of mind' (ie. cognitive empathy, the ability to understand and take into account the mental state of another individual) can be investigated using the 'reading the mind in the eyes' (RMET) task. Other cognitive processes in which differences are observed in autism such as language and executive and reward-related functioning (e.g. flexible choice behaviour) will be assessed with a verbal fluency task and probabilistic reversal learning task, respectively. Peripheral biochemistry: Participants will be asked to provide a blood sample on one visit prior to placebo/drug administration (their preference). Whole blood serotonin levels will be determined for each individual, given that elevated levels are present in one-third of individuals with autism. This will allow us to explore whether any 'shift' in brain function in response to psilocybin depends on overall serotonin 'tone' as indexed by proxy.
A wrist wearable will be used to passively collect heart rate data, from which autonomic nervous system function indices of heart rate and heart rate variability, can be derived. Heart rate variability is thought to reflect stress or arousal and is modulated by the serotonin system. These autonomic measures have also been reported to be different in autistic individuals compared to controls. Therefore, in this study we will examine the action of psilocybin on these indices of arousal and also consider if arousal interacts with other measurements acquired in the protocol.
The overarching goal of our analyses is to assess whether we see a 'shift' by psilocybin in autistic and non-autistic individuals for each modality. Both parametric and nonparametric statistical analyses will be used to test hypotheses that the serotonergic targets of psilocybin functioning differently in autistic individuals. Given the heterogeneity of the autistic population and our prior observations that there is a wide range of pharmacological responses in both autistic and non-autistic individuals, we will calculate individual 'shift' for each modality, and what characteristics (eg. clinical scores, questionnaire responses, whole blood serotonin) these are associated with. Although we will generate and analyse data from single modalities, post-hoc we will also explore multimodal metrics (ie. associations between modalities) to understand how 'shifts' detected across multiple organisational levels are interrelated.
We will use a within-subject, repeated-measures design with a placebo condition so that each subject is their own control, thus increasing statistical power. Results from our prior neuroimaging studies using pharmacological challenge were successful in detecting group differences in MRI metrics with sample sizes of n = < 20. This implies an effects size (ES, expressed in Cohen's d) in excess of 1.2. In sensory tasks a sample size of n = 16 per cell is estimated to achieve 80% power to detect a medium effect (0.5) at a = 0.05; this has been achieved even in mixed sex groups of 20 participants or fewer. This reflects the literature in which significant group differences are evident even in mixed sex groups of 20 participants or fewer (e.g. n = 16). Our design relies upon participants attending for repeat test sessions. Thus, there is a chance that participants may 'drop-out' and need to be replaced, this is accommodated with our sample size.
Our protocol requires active participation despite the passive nature of several of our tasks. For example, participants will wear an EEG cap and will be asked to remain focused on the screen during the presentation of sensory stimuli. This may limit generalisability across ages or to autistic individuals with higher support needs. Even in our very capable adult cohort, participants can get restless or become distracted. Hence, we have included concurrent eye tracking during EEG to control for the potential confound of participant variability in fixation on the screen. However, many of our tasks require minimal or no response from participants and so they are less likely to be impacted by confounds such as individual cognitive difficulties. We hope that this way, should our indices prove worthy of further investigation and/or incorporation in (for example) clinical trials, they will be more accessible for individuals who may otherwise be excluded from drug development studies.
Our "shiftability" paradigm aims to determine how different organisational scales of brain function are modulated by neurochemical systems; and identify differences between autistic and non-autistic adults at an individual level. Our most recent shiftability study will use psilocybin to test the hypothesis that the serotonin system targeted by psilocybin is different in autistic and non-autistic people. We believe that the results will expand our understanding of brain biology in neurodivergent and neurotypical (autistic and non-autistic) individuals. We also hope our work will inform a more personalized medicine approach to autism. Not everyone in our study will respond the same way to psilocybin. By capturing individual biology in both autistic and non-autistic people, this experimental medicine approach may help identify autistic individuals whose serotonin system functions no differently from nonautistic people, and who might therefore not be expected to show a clinical response in a clinical trial. And vice versa; those who respond biologically to psilocybin challenge might ultimately benefit clinically. To date, all clinical trials for the core features of autism have failed. This is in large part because they have included participants based on diagnosis alone and measured outcomes without evaluating mechanisms. In depth pharmacological 'profiling' adopting some of the methods described here may help avoid the unnecessary expense and likely failure of clinical trials and facilitate the discovery of novel pharmacological support options for those who would like that choice.
Create a free account to open full-text PDFs.
Madsen, M. K., Fisher, P. M., Burmester, D. et al. · Neuropsychopharmacology (2019)
Cavanna, F., Muller, S., de la Fuente, L. A. et al. · Translational Psychiatry (2022)
Bershad, A. K., Preller, K. H., Lee, R. et al. · Biological Psychiatry (2020)
Duerler, P., Brem, S., Fraga-González, G. et al. · Cerebral Cortex (2021)
Muthukumaraswamy, S. D., Carhart-Harris, R. L., Moran, R. J. et al. · Journal of Neuroscience (2013)
Carhart-Harris, R. L., Erritzoe, D., Williams, T. et al. · PNAS (2012)
Kringelbach, M. L., Cruzat, J., Cabral, J. et al. · PNAS (2020)
Singleton, S. P., Luppi, A. I., Carhart-Harris, R. L. et al. · Nature Communications (2022)
Bernasconi, F. ;., Schmidt, A. ;., Pokorny, T. ;. et al. · Cerebral Cortex (2013)
Schmidt, A., Kometer, M., Bachmann, R. et al. · Psychopharmacology (2012)
Vollenweider, F. X., Vollenweider-Scherpenhuyzen, M. F. I., Bäbler, A. et al. · NeuroReport (1998)