Autistic Neural ‘Shiftability’: The Distinct Pharmacological Landscape of the Autistic Brain
================================================================================================

* Tobias P. Whelan
* Lucas G. S. França
* Mihail Dimitrov
* Charlotte M. Pretzsch
* Hester Velthuis
* Andreia C. Pereira
* Claire L. Ellis
* Glynis Irvin
* James L. Findon
* Robert H. Wichers
* Francesca M. Ponteduro
* Johanna Kangas
* Naoise Mulcrone
* Nichol M. L. Wong
* Dafnis Batalle
* Nicolaas A. J. Puts
* Eileen Daly
* Declan G.M. Murphy
* Gráinne M. McAlonan

## Abstract

Functionally inter-connected large-scale brain networks regulated by multiple neurotransmitter systems underpin complex human behaviour. A wide range of alterations in this neurofunctional landscape have been associated with autism. However, this evidence is mostly derived from cross-sectional analyses and thus assumes that any differences are ‘static’. Indeed, large sample sizes are required to identify reproducible baseline differences in functional connectivity between autistic and non-autistic individuals. Instead, we propose what is different in autism is not necessarily the baseline functional connectivity of resting-state networks, but rather their responsivity or regulation by neurotransmitter systems. We tested this hypothesis using a uniform analytical framework to capture the modulation of functional connectivity by single doses of pharmacological probes targeting three major neurotransmitter systems implicated in autism, namely gamma-amino-butyric-acid (GABA), serotonin (5HT) and mu opioid systems. Every drug challenge altered resting-state network connectivity differentially in the autistic compared to the non-autistic group. However, regardless of the neurotransmitter system probed, drug challenge elicited increases in between-network connectivity in autistic participants but minimal or decreased between-network connectivity in their non-autistic peers. There was no group difference in the responsivity of within-network connectivity. Thus, there is altered *responsivity* of neurotransmitter systems in the autistic brain. This has important implications for pharmacotherapy in autism because these neurotransmitter systems are the targets of several medications commonly prescribed to manage mental health conditions that frequently co-occur in autism. Investigating if these drug-induced ‘shifts’ in functional connectivity can help provide better targeted clinical interventions, will be important next translational steps.

**One Sentence Summary** Resting-state network connectivity is differentially modulated in autism by pharmacological probes targeting the GABAA, GABAB, serotonin and mu opioid neurotransmitter systems.

## Introduction

Autism is the result of biological differences in the brain. It is, however, complex and commonly complicated by the co-occurrence of mental health challenges. To date, no biological measure reliably separates autistic from non-autistic people1. This has been explained as a consequence of i) too broad a diagnostic definition of autism2 and/or ii) cohorts that are inherently heterogeneous; and/or iii) differences masked by mixing ages and genders3 and/or iv) a lack of harmonized methods4. Yet, perhaps more importantly, the brain is a dynamic organ and doesn’t simply function ‘at rest’; rather it *responds* to the modulatory effect of a *variety* of neurotransmitter systems. This is rarely examined in humans. Instead, most studies examine the brain at rest and not in response to the challenge of key neurotransmitter pathways. In our recent studies, we have begun to examine neural responsivity of the human brain to pharmacological challenge by targeting a specific neurotransmitter system, and found that autistic individuals respond differently to non-autistic individuals on selected neuroimaging parameters5. However, we have not examined brain responsivity *across more than one neurotransmitter system (i.e. a variety of challenges)* using a uniform analytical framework.

Here we focus on functionally inter-connected large-scale brain networks which are generally accepted to underpin complex human behaviour6. Their within- and between-network connectivity can be captured using functional MRI (fMRI) techniques during ‘resting-state’, a condition uncomplicated by task-based demands. This systems-level communication arises through neuronal activity across several organisational scales. At the circuit-level, binding of receptors alters neuronal firing rate and modifies the ‘blood oxygen level dependent’ (BOLD) fMRI signal. Various alterations in functional connectivity have been associated with autism. However, identifying reproducible baseline differences in autism, for example incorporating the default mode and somatomotor networks7, has required extremely large sample sizes (e.g. in excess of n = 18007), often pooling across multiple studies and cohorts for meta-analyses. Even then, these connectivity differences are only moderately associated with clinical traits of autism8. Moreover, to achieve such large samples, participants are often combined across age groups and baseline differences are especially limited when analyses are limited to autistic adults3. This demand for large sample sizes to identify only modest differences in baseline functional connectivity differences in autism relative to controls, is also impractical for translating findings to clinical settings.

Cross-sectional designs also limit investigation of how neurotransmitter systems modulate functional connectivity in the human brain and prior studies have relied upon indirect correlational approaches to link measures of brain chemistry (for example, using PET or MR spectroscopy) to function. To provide direct experimental evidence for altered responsivity of functional networks in autism, we need to *change* the target system and observe *shifts* in connectivity which are distinct from non-autistic individuals. This concept underpins a series of studies adopting our ‘shiftability’ paradigm5, which involves pharmacologically challenging specific neurotransmitter systems and measuring the resultant shift in brain function across multiple organisational levels.

In our prior ‘shiftability’ studies of autistic and non-autistic adults we applied this paradigm to investigate several neurotransmitter systems previously implicated in autism with a range of compounds: acting at GABA receptors A (targeted by positive allosteric modulator AZD7325)9 and B (targeted by agonist arbaclofen)10, serotonin (5HT; targeted serotonin re-uptake inhibitor citalopram)11 and mu opioid receptor (targeted by agonist tianeptine)12 pathways. In this study, we used the fMRI data collected across these studies to set-up a uniform analytical framework to test the hypothesis that the pharmacological responsivity of the autistic and non-autistic human brain is different.

First, we assessed the baseline group differences in within- and between-network functional connectivity which, based on extant literature, we predicted to be minimal between autistic and non-autistic participants at this sample size (non-autistic, n = 95; autistic, n = 79)3. Next, we investigated within- and between-network responsivity in autistic and non-autistic participants by examining each pharmacological probe under a uniform analytical framework. Finally, we pooled within- and between-network shifts following drug challenge to establish any generalised response profile in autistic and non-autistic people.

## Results

A comprehensive neuropharmacological profile of changes in the functional connectivity of cortical networks was produced by collecting resting-state fMRI data using four different pharmacological probes of the GABAA, GABAB, serotonin and mu opioid systems. All fMRI data was parcellated into the same 100 cortical regions16. The resulting correlation matrices were Fisher-transformed (r-to-z values) and average connectivity calculated for all within- and between-network interactions using a previously defined atlas of seven cortical functional networks: the default-mode (DMN), frontoparietal (FPN), ventral attention (VAN), dorsal attention (DAN), limbic (LN), somatomotor (SMN) and visual networks (VN). An effect size (Cohen’s D) was then obtained for within and between each functional network to represent the magnitude of the ‘shift’ from baseline (i.e. inactive placebo).

### Autistic and non-autistic network connectivity is not significantly different at baseline

First, we examined the main effect of group to ascertain if there were baseline differences between autistic and non-autistic participants in our sample. Prior to correction for multiple comparisons, only somatomotor-visual network between-network connectivity was significantly greater in the autistic group (punc = 0.007) (**Supplementary Fig. 1**). After correction for multiple comparisons, we observed no statistically significant differences between the groups (For the full main effect of group result at baseline (i.e. placebo) see **Supplementary Table 1**).

### Neurotransmitter systems differentially ‘shift’ network connectivity in the autistic brain

Below we report the within-network group x drug interactions (**Fig. 1**) and then the between-network connectivity group x drug interactions for each drug challenge (**Fig. 2**). Key findings of differential shifts in within- and between-network connectivity between the non-autistic and autistic group for each drug challenge are also summarised in **Table 1**.

![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/08/2024.11.08.24316969/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/F1)

Figure 1. 
Within-network functional connectivity ‘shifts’ by neurotransmitter system targeted for the non-autistic and autistic groups. Values (non-autistic, n = 95 & autistic, n = 79) represent the effect sizes (Cohen’s d) and significant drug x group interactions are shown with an asterisk (uncorrected p < 0.05*, p obtained with a two-sided permutation test), those that survive FDR correction are shown with a dagger (p < 0.05†).

![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/08/2024.11.08.24316969/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/F2)

Figure 2. 
Between-network functional connectivity ‘shifts’ by neurotransmitter system targeted for the non-autistic and autistic groups. Values (non-autistic, n = 95 & autistic, n = 79) represent the effect sizes (Cohen’s d) and significant drug x group interactions are shown with an asterisk (uncorrected p < 0.05*, p obtained with a two-sided permutation test), those that survive FDR correction are shown with a dagger (p < 0.05†).

View this table:
[Table 1.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/T1)

Table 1. Key group x drug interaction findings for within- and between-network connectivity across all drug challenges.

### Targeting the GABA system

#### Targeting GABAA

##### Within-network connectivity

Within-network connectivity was differentially shifted by GABAA activation in each group across several networks, in particular by low dose which increased connectivity of the visual network in non-autistic controls, but decreased it in autistic participants.

##### Between-network connectivity

In general, shifts in between-network connectivity elicited by GABAA activation were more pronounced in the autism group than controls.

In controls, GABAA activation by low and high dose AZD7325 elicited broadly similar shifts in between-network connectivity apart from between-network connectivity of the visual network, which was significantly reduced by the high dose only.

In autistic participants, GABAA activation by low and high dose AZD7325 elicited a broadly similar pattern of shifts in between-network connectivity but again, the exceptions were mostly in between-network connectivity of the visual network, specifically, increases in connectivity with the higher-order default mode and frontoparietal networks at the higher dose.

Together, GABAA activation primarily (but not exclusively) shifted within and between-network connectivity of the visual network. The most marked shifts in connectivity were present in the autistic group. Distinct decreases in between-network of higher-order networks (e.g. DMN) responses in the autism were evident at the low dose. By contrast, there were increases in connectivity between these networks at the high dose.

#### Targeting GABAB

##### Within-network connectivity

Within network responses to GABAB agonism were qualitatively similar in autistic and control groups and at low and high doses. Lower-order (visual) network connectivity tended to increase and higher-order network connectivity tended to decrease in response to arbaclofen. However, within network responses were more pronounced in the control group, particularly in the frontoparietal network.

##### Between-network connectivity

There was a differential response in between network connectivity in autistic and non-autistic groups across the whole brain. In non-autistics, GABAB activation by low and high doses of arbaclofen predominantly decreased between-network connectivity. In autistics, low and high doses of arbaclofen predominantly increased between-network connectivity.

Thus, GABAB activation had a more prominent impact within-networks in controls versus autistics but had opposite effects on between-network connectivity across the brain in autistic (increased connectivity) compared to non-autistic participants (decreased connectivity). These group differences were more marked at the higher dose and particularly pronounced in higher-order and attentional/sensory networks.

### Targeting the serotonin system

#### Within-network connectivity

Within-network connectivity responses to serotonin activation were broadly similar in each group with the exception of visual and ventral attention network connectivity which was increased in the non-autistic group, and decreased in the autistic group.

#### Between-network connectivity

Activation of serotonin signalling primarily elicited a decrease in between-network connectivity of higher order networks (such as default mode and frontoparietal) in non-autistic but not and autistic participants; indeed, serotonin activation significantly increased between-network connectivity of the default mode network in autism and did the opposite in non-autistic participants.

### Targeting the opioid system

#### Within-network connectivity

The direction of response to mu opioid agonism within-networks was broadly similar in direction in both groups but of a greater magnitude in autistics. There was especially more prominent increase in somatomotor and decrease in limbic network connectivity in autistics compared to non-autistic participants. In contrast, ventral attention network connectivity was increased by mu opioid activation in non-autistic participants groups, but decreased in the autism.

#### Between-network connectivity

Again, there were differential effects in each group on between-network connectivity. In general, in the autism group, mu opioid activation increased between-network connectivity of somatomotor and limbic networks. Whereas, it decreased connectivity of the higher order default mode and frontoparietal networks in controls.

Thus, activation of the mu opioid system has similar effects on within-network connectivity and differential effects on between-network connectivity, with differences in autism concentrated upon lower- and middle-order network (somatomotor, limbic and attentional) within and between-network connectivity.

### The autistic brain shifts more in response to drug challenge

Next, we investigated whether there was an overarching pattern to the way in which the autistic brain responded to a drug challenge, by pooling together the effect sizes for each drug as reported above. Overall, the autistic brain had a greater magnitude of shift in response to any drug, compared to the non-autistic brain (median: non-autistic, −0.035 & autistic, 0.08; interquartile range: non-autistic, 0.40 & autistic, 0.48; p = 0.0003). A net positive shift in connectivity was observed in autism (mean effect size 0.11), and a slight decrease in connectivity in controls (mean effect size −0.03).

Post-hoc analyses confirmed that this group-level difference in network responsivity was primarily driven by *increases* in *between-network* connectivity in autism compared to a minimal decrease in the non-autistic group (median: non-autistic, −0.09 & autistic, 0.11; interquartile range: non-autistic, 0.41 & autistic, 0.45; p = <0.0001); and there was no difference in within-network connectivity shifts between groups (median: non-autistic, 0.07 & autistic, −0.01; interquartile range: non-autistic, 0.39 & autistic, 0.38; p = 0.20). Ventral attention, frontoparietal, somatomotor and (to a lesser extent) dorsal attention network connectivity increases are the largest contributors to shifts in between-network connectivity in autism; whereas, default mode, limbic and visual network connectivity remained unchanged (or were slightly increased) (**Fig. 3**).

![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/08/2024.11.08.24316969/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/F3)

Figure 3. 
Aggregate between-network connectivity shifts (also represented by as effect size) for each resting-state network separately for both groups. A distinct pattern of increased connectivity in autism is observed, and largely reduced connectivity in non-autistic controls across networks in response to drug challenge.

## Discussion

We have used a uniform analytical framework to capture the change in functional connectivity within and between seven brain networks following activation of several known targets of the GABAA, GABAB, serotonin and mu opioid neurotransmitter systems in autistic and non-autistic adults. As predicted, despite minimal baseline group differences, functional connectivity measures were ‘shifted’ differently by each drug probe in autistic compared to non-autistic individuals.

There were some target-specific patterns; modulation of GABAA shifted visual network connectivity; GABAB shifted between-network connectivity across the brain; serotonin shifted sensory and higher order networks (e.g. default mode network); and mu opioid shifted somatomotor and limbic network connectivity. However, we also found that, regardless of neurotransmitter system targeted, in general drugs elicited significantly greater increases in between-network connectivity in autistic participants compared to their non-autistic peers; especially between-network connectivity of the VAN, FPN and SMN, which are primarily networks which ‘allow’ bottom-up information to shift attention and control the initiation of behaviour21.

### Target-specific patterns of connectivity shift are different in autistic and non-autistic participants

We observed target-specific differences in functional connectivity responses in autism depending on the neurotransmitter system probed. The networks which shifted in response to each drug broadly aligned with the known expression pattern of their receptors but the direction and/or extent of shift in connections within and between these brain regions was different in autism.

#### GABA response differences

Autistic differences in network responsivity following GABAA activation were observed in lower-order sensory networks. The effect was most prominent in the visual cortex where GABAA receptor expression is reported to be particularly high22. Altered responsivity in autism may, in part, be explained by reduced levels of GABAA receptors in the autistic brain23,24.

However, this evidence comes from post-mortem data, and the reported decrease in GABA receptors has been less consistent in the living human brain using PET imaging. Recent PET data suggests that cortical GABAA receptor levels may be no different in autistic and non-autistic adults25,26.

There were more widespread autistic differences in between-network connectivity in response to GABAB activation, and this may reflect the broader extent of GABAB expression across the whole brain22,27. Again, this may, in part, be explained by reduced levels of GABAB receptors in the autistic brain23,24. For example, GABAB receptors may be differently expressed in prefrontal and parietal regions that comprise higher-order and attentional networks in which we see the largest differential shifts in response to arbaclofen here28. However, there are no reports to-date of altered GABAB receptor expression in autism from PET data.

Thus, minimal baseline differences in GABA receptor availability may not entirely explain the diverging direction of shifts in connectivity of autistic and non-autistic resting-state networks in response to GABAA/B activation.

We suggest that the explanation may be functional differences in the GABA system. This is consistent with other evidence from preclinical animal work and humans. For example, agonism of GABAA and GABAB receptors ameliorates autistic features in an autism-relevant mouse model29; and we have shown that visual and auditory sensory processing in the same autistic and non-autistic adults who participated in the present study are also differentially modulated by GABAB receptor agonism30,31.

#### Serotonin response differences

There were paradoxical responses to citalopram in the autistic brain. In contrast to non-autistic controls, within-network connectivity of the visual and ventral attention networks was reduced in the autistic brain, but connectivity between the visual and default mode network was increased.

In the visual network, there is evidence for lower SERT levels in the visual cortex in autism32, several regions of the DMN also express SERT including the medial prefrontal cortex and posterior cingulate cortex (PCC)33 and lower SERT binding has been reported in these regions in autistic adults34. However, it is difficult to align receptor availability differences with the range of within and between-network alterations elicited by citalopram in autism. The increase in between-network connectivity of the DMN in autism in response to citalopram is noteworthy given that altered DMN connectivity is implicated in depression, which is more common in autistic people35. The augmented responses to SERT activation in autistic individuals across the default mode and sensory (e.g. visual) networks observed here aligns with evidence that autistic patients respond at lower doses to SSRIs35 and SSRIs can be poorly tolerated by autistic people36.

#### Mu opioid response differences

Lastly, in autistic participants, agonism at the mu opioid receptor by tianeptine altered within and between-network differently in autistics. Within-network connectivity was increased in somatomotor networks and decreased in limbic and ventral attention networks in autistics but not in non-autistics. Between-network connectivity of the ventral attention and limbic network was increased in autism and decreased in controls.

Mu opioid receptors are expressed in several components of these networks with particularly high expression in the cingulate cortex of the limbic network37 and in somatomotor regions38. Recent human PET data suggests that baseline differences in cortical mu opioid receptor expression in autism are restricted to upregulation in the precuneus only39, so again differences in receptor levels do not explain responsivity differences. Thus, as for other drug probes, what sets the autistic group apart is a distinct functional response of the brain to neurotransmitter perturbation.

These observed differences in the responsivity of resting-state networks across the neurotransmitter systems probed here may need to be understood in the context of neurodevelopment. Indeed, we have evidence for early alterations in functional connectivity in newborn infants who have a higher likelihood of a neurodevelopmental outcome such as autism40. This work observed that local functional connectivity in neonates with a higher likelihood of autism was greater in sensory cortices and limbic areas responsible for face processing41. We have more recently reported that moment-to-moment dynamic functional connectivity profile of brain states that engage sensorimotor networks at birth predicts neurodevelopmental and autistic traits at 18 months42. Thus, by birth, there are already differences in the functional landscape of brain regions associated with autism. Although, capturing convincing baseline cross-sectional differences in functional connectivity in larger autistic and non-autistic cohorts assessed at later ages has been challenging3. We suggest that what persists in autism is not any average baseline difference in functional connectivity but a true functional difference as defined by altered response to perturbation (in this case by pharmacological probe).

Consistent with this, we see no substantial baseline difference in resting-state network connectivity between groups of adults in our sample. It could be argued that this is unsurprising given our relatively small sample size but one recent mega-analysis used a sample of 1824 individuals, 796 of whom were autistic to report effects which were still relatively small7. What is more striking therefore, is that significant differences in functional connectivity responses to drug challenge were observed even with our small sample size.

### Between network functional connectivity in autism increases in response to drug challenge

We pooled resting-state connectivity data across all the pharmacological probes to investigate if there was any generalised response profile common to the autistic group. We report that overall, drug challenge *increased between-network connectivity in autism*, but either had little impact or *decreased* between-network connectivity in non-autistic controls. The main drug-elicited increases in functional connectivity were mainly localised to between-network connectivity of the VAN, FPN and SMN, which are implicated in detecting unexpected stimuli to trigger attentional shift, flexible initiation of new tasks and the processes controlling motor output43, respectively. Thus, to some extent they can be thought of as allowing a flexible response to external stimuli. Consequently, how neurotransmitter systems in autistic individuals regulate the engagement and disengagement of functional networks in response to changing task demands may be altered. By facilitating connectivity between-networks the drugs used here may promote more flexible responses to external stimuli. We emphasise however that this is highly speculative and whether this profile has clinical meaning remains to be examined.

An additional or alternative explanation for the increased between-network shift induced by drug challenge in autistic participants may be that homeostasis is differently regulated in autism. For example, if networks remain in their shifted state for longer after drug challenge in autistic participants, we may simply be continuing to capture a functional shift in autistics which has moved back towards baseline in controls by the time of data acquisition. In support of this, we have previously reported that limbic system activation takes longer to return to baseline in autism following citalopram during a fMRI face processing task44. This result was restricted to *between*-networks, so following perturbation local systems may return to baseline quicker compared to those *between*-networks that are more distal and/or more functionally segregated. Our previous work using other modalities to capture brain dynamics supports this position. We have used the aperiodic (1/f) slope of EEG to show that, despite no baseline group differences, autistic participants ‘shifted’ their 1/f slope at lower doses than non-autistic controls45. Placed together with the present analyses, the homeostatic mechanisms which might be at play in controls to limit shifts in brain function are not so evident in autism. MRI does not allow us to address this question but there is some evidence from cellular models that synaptic scaling to adjust excitatory or inhibitory synaptic strength up or down to stabilise neural firing rates46 is disrupted in animal models of autism47.

We acknowledge a number of limitations to our work. First, our approach included different MRI acquisition sequences adopted between studies (SERT/mu opioid versus the GABA probes). The impact of this was somewhat mitigated by using a consistent preprocessing pipeline across all studies. Similarly, an atlas-based parcellation approach was used to provide a uniform analysis framework across studies/probes. Second, motion artifacts (i.e. due to participants moving in the scanner) can alter functional connectivity measures48. As a result, our preprocessing pipeline included several techniques focused on motion correction specifically (e.g. despiking, multi-echo ICA correction, censoring). There was also stringent post-processing quality control and datasets with significant movement (or other quality control issues) were omitted from analyses. Third, our sample size for each pharmacological probe is relatively small and the sexes are also not balanced (only males were recruited for the studies targeting SERT and the mu opioid receptor), therefore more work including female needs to be done. This is a recognised problem in past pharmacological research49 and in our current studies our goal is to recruit an equal number of men and women. Despite a small sample size for each probe our statistical power was increased due to the within-subject repeated-measures design of our studies.

In conclusion, our findings have implications for drug development and treatment of neurological and psychiatric conditions experienced by neurodivergent people. Our work indicates that the autistic brain is pharmacologically atypical. This is crucial because autistic people have a higher incidence of psychiatric and neurological conditions50,51 but drug development and trials for mental health problems do not consider neurodivergence. These results suggest this needs to change. Investigating if these drug-induced ‘shifts’ in functional connectivity can help provide better targeted interventions, will be important next translational steps.

## Materials and Methods

### Experimental Design

All the randomised, double-blind, placebo-controlled, cross-over, repeated-measures, case-control studies included here were conducted in accordance with the Declaration of Helsinki at the Institute of Psychiatry, Psychology and Neuroscience (IoPPN), King’s College London (KCL) in London, UK in different time periods. Our studies did not address safety or clinical efficacy and the UK Medicines and Health Regulatory Authority (MHRA) confirmed that these studies used a drug as a probe of neurochemical systems and none constituted a clinical trial of an Investigational Medicinal Product (IMP) as defined by the EU Directive 2001/20/EC. Nevertheless, as our design incorporated a ‘drug intervention’ our protocols were registered on [clinicaltrials.gov](http://clinicaltrials.gov) for transparency. Ethical approval had been received from an appropriate Institutional Ethics Board or National UK Health Research Authority. All studies acquired resting state data following either a placebo or a single acute oral dose of a pharmacological probe (at maximum plasma concentration, tmax) with administration order randomized across participants.

### Participants

Participants recruited into the autism group had IQ>70 and a clinically confirmed Autism diagnosis as per the recruitment criteria of the parent study. The control participants were recruited to match the group age and IQ of the autistic participants. All participants gave written informed consent prior to being included in any experimental procedures. **Table 2** summarizes the parent studies providing data for these analyses.

View this table:
[Table 2.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/T2)

Table 2. Studies included in the analyses. Values in parentheses indicate the number of female participants, note that only male participants were recruited for the serotonin and opioid study. ‘Time of scan’ refers to the approximate time of functional MRI scan post-administration of drug. ‘Washout time’ refers to the minimum time between repeat visits in which drug/placebo was administered and allows sufficient time for drug to leave the system.

### MRI Data Acquisition

MRI acquisition information is provided in **Table 3** and includes the MR scanner, sequences to acquire both structural and functional scans, repetition time and echo time for each study included in analyses. Further information regarding MRI acquisition is in **Supplementary Table 2**.

View this table:
[Table 3.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/T3)

Table 3. MRI acquisition information included in analyses. TR, repetition time; TE, echo time; ADNI GO, Alzheimer’s Disease Neuroimaging Initiative Grand Opportunities13 & echo planar imaging, EPI.

### MRI Data Preprocessing

For all studies the functional fMRI data was pre-processed using AFNI14 (AFNI v21.1.07 ([https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/#](https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/#)) and Anaconda v3.8 ([https://anaconda.org/](https://anaconda.org/)) and consisted of de-spiking; slice time correction; co-registration of fMRI to T1 structural scan; normalisation to standard (MNI152_T1_2009c) space; tissue segmentation; optimal combination of echoes and motion correction using multi-echo ICA (tedana v0.0.8)15; smoothing (6mm3); scaling (to percent signal change with µ = 0); band-pass filter (0.01 - 0.1Hz); regression of white matter and cerebrospinal fluid signal & censoring (if framewise displacement > 3mm and/or if motion outlier (i.e. >2 σ)). The raw and pre-processed MRI data was inspected and checked for quality issues using simple image visualisation in FSL’s FSLeyes in combination with examination of the reports that were automatically generated by AFNI.

### Functional Network Analysis

The pre-processed fMRI data for each drug was parcellated into 100 cortical brain regions of interest (ROIs) according to the Schaefer parcellation scheme16. The average time series in each parcel was used to construct functional connectivity matrices of Fisher-transformed Z values (or Fisher-transformed correlation coefficients) between every pair of ROIs for each individual scan using CONN v21a17. Each of the 100 ROIs were assigned into one of seven main functional networks according to the Yeo-7 atlas: default mode (DMN); frontoparietal (FPN); ventral attention (VAN) (or salience); dorsal attention (DAN); limbic (LN); somatomotor (SMN) & visual (VN) networks16. Based on this parcellation, within-network connectivity was defined as the mean connectivity between each ROI belonging to a Yeo-7 network. Between-network connectivity was defined as the mean connectivity between ROIs that comprises two different networks **(Fig. 4)**.

![Figure 4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/08/2024.11.08.24316969/F4.medium.gif)

[Figure 4.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/F4)

Figure 4. 
Schematic of the consistent resting-state network functional connectivity analysis pipeline.

### Statistical Analyses

First, we tested the prediction that there would be no statistically significant differences within and between networks in the placebo condition between the autistic and non-autistic group. Then, for each drug we tested our main hypothesis that there would be a difference in the brain response to drug challenge in autistic and non-autistic individuals i.e. a group x drug interaction. To evaluate the statistical significance of the interaction between dose and group we fit a general linear mixed effects model (GLME) for all networks with two-sided permutation tests with 10,000 repetitions. The GLME used was given by the expression z̄ ∼ B0 + Bdose*group + (1|Subject ID) – with Subject ID accounting for the random effect. Significance was assessed with two-sided permutation tests featuring 10,000 GLME repetitions. We report p-values obtained from the permutation tests uncorrected, highlighting those surviving multiple comparison correction using the False Discovery Rate method with α = 5%18. This approach accommodates missing data as not all participants had a complete dataset including placebo and all drug conditions. Autistic participants that took part in two or more studies (n = 7, 5 male and 2 female) were included in the analysis. We report the within-network results first, then the between-network results per pharmacological probe.

To investigate the extent to which any interactions were driven predominantly by a shift in the autistic group or non-autistic group, we estimated the Cohen’s D effect size19 for connectivity changes within- and between every functional network against placebo in each group separately; where our placebo baseline included every placebo fMRI scan from all studies (non-autistic controls, n = 95 & autistic, n = 79). This is represented by equation XX, where z̄ represents the average Fisher’s Z for a given network and *σ*(Z) the standard deviation of such network: ![Formula][1]</img>  Finally, we examined whether drug challenge preferentially increases or decreases within- and between-network connectivity in the autistic brain, irrespective of pharmacological probe. To do this, we pooled the within- and between-network effect sizes generated in the analysis described above and then performed non-parametric Mann Whitney U tests to compare the aggregate within- and between-network connectivity shift between groups.

All statistical analyses were conducted in R version 4.1.1 (2021-08-10) (R Core Team 2021) with supporting packages ggplot220, reshape2, dplyr, tidyr, tibble, shadowtext, lme4, and ptestr.

## Competing Interests

T.P.W. is an employee of Compass Pathfinder Ltd. and holds shares in the company. N.A.P. has consulted for Deerfield Discovery and Research. D.G.M.M. has consulted for Jaguar Gene Therapy LLC. G.M.M. has received funding for investigator-initiated studies from GW Pharmaceuticals and Compass Pathfinder Ltd.. G.M.M. has consulted for Greenwich Biosciences, Inc. L.G.S.F., M.D., C.M.P., H.V., A.C.P., C.L.E., G.I., J.L.F., R.H.W., F.M.P., J.K., N.M., N.M.L.W., D.B. and E.D. have no conflicts to declare.

## Author Contributions

T.P.W. performed data preprocessing and analysis, prepared and drafted the manuscript. L.G.S.F. contributed to data analysis and helped draft the manuscript. M.D. contributed to data collection and preprocessing, and helped drafted the manuscript. C.M.P., H.V., A.C.P., C.L.E., J.L.F., R.H.W., F.M.P., J.K., N.M.L.W. contributed to data collection. G.I. provided pharmacy services for the studies. N.M. contributed to preprocessing and data analysis. D.B. and N.A.P. edited the manuscript and provided input on data analysis. E.D., D.G.M.M. and G.M.M. conceived the idea for the studies; contributed to study design and drafted the manuscript. All authors reviewed and approved the final manuscript.

## Funding

The authors also receive support from EU-AIMS (European Autism Interventions)/EU AIMS-2-TRIALS, an Innovative Medicines Initiative Joint Undertaking under Grant Agreement No. 777394. In addition, this paper represents independent research part funded by the infrastructure of the National Institute for Health Research (NIHR) Mental Health Biomedical Research Centre (BRC) at South London and Maudsley NHS Foundation Trust and King’s College London, and the Medical Research Council Centre for Neurodevelopmental Disorders. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

## Data and Material Availability

Data from this study is available upon request.

## Supplementary Information

View this table:
[Supplementary Table 1.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/T4)

Supplementary Table 1. The main effect of group showing differences in connectivity of all resting-state networks between autistic and non-autistic participants at baseline (placebo).

View this table:
[Supplementary Table 2.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/T5)

Supplementary Table 2. Additional MRI acquisition information.

![Supplementary Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/08/2024.11.08.24316969/F5.medium.gif)

[Supplementary Figure 1.](http://medrxiv.org/content/early/2024/11/08/2024.11.08.24316969/F5)

Supplementary Figure 1. 
Baseline difference in somatomotor-visual network connectivity between the non-autistic and autistic groups.

## Acknowledgements

## Footnotes

*   * Joint senior authors

*   Received November 8, 2024.
*   Revision received November 8, 2024.
*   Accepted November 8, 2024.


*   © 2024, Posted by Cold Spring Harbor Laboratory

The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission.

## References

1.  1.Loth, E. et al. The meaning of significant mean group differences for biomarker discovery. PLoS Comput Biol 17, e1009477 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/JOURNAL.PCBI.1009477&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34793435&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

2.  2.Mottron, L. & Gagnon, D. Prototypical autism: New diagnostic criteria and asymmetrical bifurcation model. Acta Psychologica 237, 103938 (2023).
    
    

3.  3.Uddin, L. Q., Supekar, K. & Menon, V. Reconceptualizing functional brain connectivity in autism from a developmental perspective. Front. Hum. Neurosci. 7, (2013).
    
    

4.  4.Müller, R.-A. et al. Underconnected, but How? A Survey of Functional Connectivity MRI Studies in Autism Spectrum Disorders. Cerebral Cortex 21, 2233–2243 (2011).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/cercor/bhq296&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21378114&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000294808800004&link_type=ISI) 

5.  5.Whelan, T. P., et al. Editorial Perspective: Bridging the translational neuroscience gap in autism – development of the ‘shiftability’ paradigm. Child Psychology Psychiatry 65, 862–865 (2024).
    
    

6.  6.Mišić, B. & Sporns, O. From regions to connections and networks: new bridges between brain and behavior. Current Opinion in Neurobiology 40, 1–7 (2016).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.conb.2016.05.003&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27209150&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

7.  7.Ilioska, I. et al. Connectome-wide Mega-analysis Reveals Robust Patterns of Atypical Functional Connectivity in Autism. Biological Psychiatry 94, 29–39 (2023).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/J.BIOPSYCH.2022.12.018&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36925414&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

8.  8.Holiga, Š. et al. Patients with autism spectrum disorders display reproducible functional connectivity alterations. Sci. Transl. Med. 11, eaat9223 (2019).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTE6InNjaXRyYW5zbWVkIjtzOjU6InJlc2lkIjtzOjE1OiIxMS80ODEvZWFhdDkyMjMiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyNC8xMS8wOC8yMDI0LjExLjA4LjI0MzE2OTY5LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

9.  9.Fan, C. et al. Transcriptomics of Gabra4 knockout mice reveals common NMDAR pathways underlying autism, memory, and epilepsy. Molecular Autism 11, 13 (2020).
    
    

10. 10.Schuler, V. et al. Epilepsy, Hyperalgesia, Impaired Memory, and Loss of Pre- and Postsynaptic GABAB Responses in Mice Lacking GABAB(1). Neuron 31, 47–58 (2001).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0896-6273(01)00345-2&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=11498050&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000170101100010&link_type=ISI) 

11. 11.Muller, C. L., Anacker, A. M. J. & Veenstra-VanderWeele, J. The serotonin system in autism spectrum disorder: From biomarker to animal models. Neuroscience 321, 24–41 (2016).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuroscience.2015.11.010&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26577932&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

12. 12.Becker, J. A. et al. Autistic-Like Syndrome in Mu Opioid Receptor Null Mice is Relieved by Facilitated mGluR4 Activity. Neuropsychopharmacol 39, 2049–2060 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/npp.2014.59&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24619243&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

13. 13.Jack, C. R. et al. Magnetic resonance imaging in Alzheimer’s Disease Neuroimaging Initiative 2. Alzheimer’s & Dementia 11, 740–756 (2015).
    
    

14. 14.Cox, R. W. AFNI: Software for Analysis and Visualization of Functional Magnetic Resonance Neuroimages. Computers and Biomedical Research 29, 162–173 (1996).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1006/cbmr.1996.0014&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=8812068&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1996UV56700002&link_type=ISI) 

15. 15.Kundu, P. et al. Integrated strategy for improving functional connectivity mapping using multiecho fMRI. Proc. Natl. Acad. Sci. U.S.A. 110, 16187–16192 (2013).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoicG5hcyI7czo1OiJyZXNpZCI7czoxMjoiMTEwLzQwLzE2MTg3IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMDgvMjAyNC4xMS4wOC4yNDMxNjk2OS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

16. 16.Schaefer, A. et al. Local-Global Parcellation of the Human Cerebral Cortex from Intrinsic Functional Connectivity MRI. Cerebral Cortex 28, 3095–3114 (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10/gd738m&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28981612&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

17. 17.Whitfield-Gabrieli, S. & Nieto-Castanon, A. *Conn* : A Functional Connectivity Toolbox for Correlated and Anticorrelated Brain Networks. Brain Connectivity 2, 125–141 (2012).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1089/brain.2012.0073&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22642651&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

18. 18.Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series B: Statistical Methodology 57, 289–300 (1995).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.2307/2346101&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=WOS:A1995QE4&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1995QE45300017&link_type=ISI) 

19. 19.Sawilowsky, S. S. New Effect Size Rules of Thumb. J. Mod. App. Stat. Meth. 8, 597–599 (2009).
    
    

20. 20.Wickham, H. Ggplot2. (Springer International Publishing, Cham, 2016). doi:10.1007/978-3-319-24277-4.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/978-3-319-24277-4&link_type=DOI) 

21. 21.Rosazza, C. & Minati, L. Resting-state brain networks: literature review and clinical applications. Neurol Sci 32, 773–785 (2011).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10072-011-0636-y&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21667095&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

22. 22.Calver, A. R. et al. The expression of GABAB1 and GABAB2 receptor subunits in the cNS differs from that in peripheral tissues. Neuroscience 100, 155–170 (2000).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0306-4522(00)00262-1&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=10996466&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000089509800014&link_type=ISI) 

23. 23.Fatemi, S. H. et al. mRNA and Protein Levels for GABAAα4, α5, β1 and GABABR1 Receptors are Altered in Brains from Subjects with Autism. J Autism Dev Disord 40, 743–750 (2010).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10803-009-0924-z&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20066485&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

24. 24.Oblak, A., Gibbs, T. T. & Blatt, G. J. Decreased GABA A receptors and benzodiazepine binding sites in the anterior cingulate cortex in autism. Autism Research 2, 205–219 (2009).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=19650112&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

25. 25.Fung, L. K. et al. Thalamic and prefrontal GABA concentrations but not GABAA receptor densities are altered in high-functioning adults with autism spectrum disorder. Mol Psychiatry 26, 1634–1646 (2021).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32376999&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

26. 26.Horder, J. et al. GABA A receptor availability is not altered in adults with autism spectrum disorder or in mouse models. Sci. Transl. Med. 10, eaam8434 (2018).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTE6InNjaXRyYW5zbWVkIjtzOjU6InJlc2lkIjtzOjE1OiIxMC80NjEvZWFhbTg0MzQiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyNC8xMS8wOC8yMDI0LjExLjA4LjI0MzE2OTY5LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

27. 27.Billinton, A. et al. GABAB receptor heterodimer-component localisation in human brain. Molecular Brain Research 77, 111–124 (2000).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0169-328X(00)00047-4&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=10814837&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

28. 28.Fatemi, S. H., Reutiman, T. J., Folsom, T. D. & Thuras, P. D. GABAA Receptor Downregulation in Brains of Subjects with Autism. J Autism Dev Disord 39, 223–230 (2009).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10803-008-0646-7&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18821008&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000262566800004&link_type=ISI) 

29. 29.Yang, J.-Q., Yang, C.-H. & Yin, B.-Q. Combined the GABA-A and GABA-B receptor agonists attenuates autistic behaviors in a prenatal valproic acid-induced mouse model of autism. Behavioural Brain Research 403, 113094 (2021).
    
    

30. 30.Huang, Q. et al. GABA B receptor modulation of visual sensory processing in adults with and without autism spectrum disorder. Sci. Transl. Med. 14, eabg7859 (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1126/scitranslmed.abg7859&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34985973&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

31. 31.Huang, Q. et al. Exploratory evidence for differences in GABAergic regulation of auditory processing in autism spectrum disorder. Transl Psychiatry 13, 320 (2023).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37852957&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

32. 32.Nakamura, K. et al. Brain Serotonin and Dopamine Transporter Bindings in Adults With High-Functioning Autism. Arch Gen Psychiatry 67, 59 (2010).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/archgenpsychiatry.2009.137&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20048223&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000273262800007&link_type=ISI) 

33. 33.Beliveau, V. et al. A High-Resolution *In Vivo* Atlas of the Human Brain’s Serotonin System. J. Neurosci. 37, 120–128 (2017).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjg6IjM3LzEvMTIwIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMDgvMjAyNC4xMS4wOC4yNDMxNjk2OS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

34. 34.Andersson, M. et al. Serotonin transporter availability in adults with autism—a positron emission tomography study. Mol Psychiatry 26, 1647–1658 (2021).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32848204&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

35. 35.Howes, O. D. et al. Autism spectrum disorder: Consensus guidelines on assessment, treatment and research from the British Association for Psychopharmacology. J Psychopharmacol 32, 3–29 (2018).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29237331&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

36. 36.King, B. H. et al. Lack of Efficacy of Citalopram in Children With Autism Spectrum Disorders and High Levels of Repetitive Behavior: Citalopram Ineffective in Children With Autism. Arch Gen Psychiatry 66, 583 (2009).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/archgenpsychiatry.2009.30&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=19487623&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000266566800003&link_type=ISI) 

37. 37.Kantonen, T. et al. Interindividual variability and lateralization of μ-opioid receptors in the human brain. NeuroImage 217, 116922 (2020).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32407992&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

38. 38.Sun, L. et al. Mu-opioid receptor system modulates responses to vocal bonding and distress signals in humans. Phil. Trans. R. Soc. B 377, 20210181 (2022).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36126675&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

39. 39.Noppari, T. et al. Dopamine D2R and opioid MOR availability in autism spectrum disorder. Preprint at doi:10.1101/2024.04.09.588651 (2024).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoiYmlvcnhpdiI7czo1OiJyZXNpZCI7czoxOToiMjAyNC4wNC4wOS41ODg2NTF2MSI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDI0LzExLzA4LzIwMjQuMTEuMDguMjQzMTY5NjkuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

40. 40.Ciarrusta, J. et al. Emerging functional connectivity differences in newborn infants vulnerable to autism spectrum disorders. Transl Psychiatry 10, 131 (2020).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32376820&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

41. 41.Ciarrusta, J. et al. Social Brain Functional Maturation in Newborn Infants With and Without a Family History of Autism Spectrum Disorder. JAMA Netw Open 2, e191868 (2019).
    
    

42. 42.França, L. G. S. et al. Neonatal brain dynamic functional connectivity in term and preterm infants and its association with early childhood neurodevelopment. Nat Commun 15, 16 (2024).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38331941&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

43. 43.Seitzman, B. A., Snyder, A. Z., Leuthardt, E. C. & Shimony, J. S. The State of Resting State Networks. Topics in Magnetic Resonance Imaging 28, 189–196 (2019).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/RMR.0000000000000214&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31385898&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

44. 44.Wong, N. M. L. et al. Serotonin differentially modulates the temporal dynamics of the limbic response to facial emotions in male adults with and without autism spectrum disorder (ASD): a randomised placebo-controlled single-dose crossover trial. Neuropsychopharmacol. 45, 2248– 2256 (2020).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32388538&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

45. 45.Ellis, C. L. et al. The dynamically neurodiverse human brain: Measuring excitatory-inhibitory dynamics and identifying homeostatic differences in autistic and non-autistic people. Preprint at doi:10.1101/2023.06.19.23291507 (2023).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoibWVkcnhpdiI7czo1OiJyZXNpZCI7czoyMToiMjAyMy4wNi4xOS4yMzI5MTUwN3YxIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMDgvMjAyNC4xMS4wOC4yNDMxNjk2OS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

46. 46.Turrigiano, G. Homeostatic Synaptic Plasticity: Local and Global Mechanisms for Stabilizing Neuronal Function. Cold Spring Harbor Perspectives in Biology 4, a005736–a005736 (2012).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTE6ImNzaHBlcnNwZWN0IjtzOjU6InJlc2lkIjtzOjExOiI0LzEvYTAwNTczNiI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDI0LzExLzA4LzIwMjQuMTEuMDguMjQzMTY5NjkuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

47. 47.Tatavarty, V. et al. Autism-Associated Shank3 Is Essential for Homeostatic Compensation in Rodent V1. Neuron 106, 769–777.e4 (2020).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32199104&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

48. 48.Mahadevan, A. S., Tooley, U. A., Bertolero, M. A., Mackey, A. P. & Bassett, D. S. Evaluating the sensitivity of functional connectivity measures to motion artifact in resting-state fMRI data. NeuroImage 241, 118408 (2021).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34284108&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

49. 49.Sosinsky, A. Z. et al. Enrollment of female participants in United States drug and device phase 1– 3 clinical trials between 2016 and 2019. Contemporary Clinical Trials 115, 106718 (2022).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35247632&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

50. 50.Pan, P.-Y., Bölte, S., Kaur, P., Jamil, S. & Jonsson, U. Neurological disorders in autism: A systematic review and meta-analysis. Autism 25, 812–830 (2021).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32907344&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom) 

51. 51.Underwood, J. F. G. et al. Neurological and psychiatric disorders among autistic adults: a population healthcare record study. Psychol. Med. 53, 5663–5673 (2023).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36189783&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F08%2F2024.11.08.24316969.atom)

 [1]: /embed/graphic-8.gif