Cannabidiol treatment is associated with broadband spectral electrophysiological changes and improved cognition in boys with low-functioning autism spectrum disorder
=====================================================================================================================================================================

* C. Cazares
* A. Hutton
* G. Paez
* D. Trauner
* B. Voytek

## Abstract

Oral cannabidiol (CBD) treatment has been suggested to alleviate negative symptoms of autism spectrum disorder (ASD). While many CBD preparations have been studied in randomized clinical trials involving ASD, none have used purified CBD preparations or preparations approved by the U.S. Food and Drug Administration, nor have they focused on low-functioning children with ASD. Previous studies have identified several candidate electrophysiological biomarkers for the cognitive and behavioral disabilities in ASD, with one emerging biomarker being aperiodic neural activity. Here we examined whether periodic (oscillatory) and/or aperiodic electroencephalography (EEG) features are predictive of any symptomatic changes in ASD following pharmacological CBD intervention. To do this, we leveraged resting-state EEG from children with low-functioning ASD (24 boys, aged 7-14 years) using data obtained during a prior double-blind, placebo-controlled, crossover Phase III Clinical Trial ([NCT04517799](http://medrxiv.org/lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT04517799&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom)) that investigated using cannabidiol to treat severe behavior problems in children with ASD. Using linear mixed effect models, we found that aperiodic EEG signal features varied directly with 7-COOH-CBD metabolite levels in blood, as evidenced by a larger aperiodic offset (p < 0.001) and decreased aperiodic exponent (p < 0.05) across the scalp. Furthermore, 7-COOH-CBD metabolite levels in blood had a positive association with nonverbal intelligence and visuomotor coordination (p < 0.05). Finally, changes in visuomotor coordination attributed to occipital oscillatory EEG activity were mediated by changes in 7-COOH-CBD metabolite levels in blood, with distinct effects observed for the delta frequency band (p < 0.05). Our analytical results suggest that this daily CBD preparation and administration schedule exerted some benefits, with improvements to cognitive and behavioral abilities in a low-functioning ASD children population. Our findings support the inclusion of resting-state, aperiodic signal features as candidate biomarkers for tracking the clinical impact of CBD treatment, in addition to traditional oscillatory EEG measures, within a neurodevelopmental context.

## 1. Introduction

Neurodevelopmental disorders (NDDs) characterized by severe cognitive and behavioral dysfunction impose significant burdens on caregivers, families, and finances [1]. Current pharmacotherapies have limited efficacy, and progress in developing new treatments is hindered by a lack of reliable biomarkers that can identify NDD patients who are most likely to benefit from investigational therapies in early life [2]. One such condition is Autism Spectrum Disorder (ASD), a NDD whose core symptoms can include deficits in social communication and stereotyped repetitive behaviors [3]. The vast heterogeneity of behavioral features in ASD, which can vary widely in presentation and severity among affected individuals, further complicates the search for unbiased, non-invasive, and clinically relevant biomarkers that could be sensitive to the effects of investigational therapies.

For decades, non-invasive scalp electroencephalography (EEG) features have served as biomarkers for cognitive and behavioral abilities in both health and disease. Local field potential signals, such as those recorded by scalp EEG, comprise mixed periodic and aperiodic components. These components are thought to reflect synchronized and asynchronous neuronal firing in cortical networks, respectively, and can serve as putative indices of cortical excitation-inhibition (E:I) balance in health and disease [4–9]. Periodic and aperiodic components of EEG signals correlate with cognitive processes and show alterations in a wide variety of neurological and psychiatric conditions associated with E:I imbalance, including NDDs [7,10–15]. Indeed, resting state EEG measures have provided a suitable and well-validated approach to investigate neural correlates of cognition and repetitive behaviors in intellectually impaired and minimally verbal pediatric populations [10,14,16,17]. It’s believed that EEG abnormalities and cognitive impairments observed in patients with NDDs likely stem from dysfunctional neuronal communication, resulting from impaired neuronal maturation during early development [18–20]. However, work on relating EEG signal features to ASD in children have largely focused on the periodic component. In particular, the oscillatory sensorimotor mu rhythm and visual cortical alpha rhythms – both in the ∼8-12 Hz frequency range – have been found to be disrupted or altered in both children and adults with ASD [21–23]. In addition, peak alpha frequency in the 6-12 Hz range has been proposed as a biomarker for non-verbal cognition in ASD children [24]. Furthermore, disruptions in delta rhythms in the 1-4 Hz range have also been shown to be disrupted in individuals with ASD [16,25]. Thus, the search for more effective pharmacotherapies and candidate biomarkers in severe forms of ASD early in neurodevelopment may greatly benefit from identifying characteristics of EEG activity that reflect estimates of E:I balance and treatment outcomes.

Medications such as risperidone and aripiprazole have shown some efficacy in improving core behavioral deficits in children with ASD [26]. Unfortunately, these atypical antipsychotics can induce unintended adverse effects, including cardiac symptoms and metabolic disturbances, which necessitate close medical monitoring [27]. As a result, interest in cannabis-derived compounds has grown in pediatric ASD medical research, particularly its non-psychoactive component, cannabidiol (CBD). CBD has shown promising anecdotal and clinical trial outcomes for treating children with intractable epilepsy, including those with co-morbid ASD, with mild or tolerable side effects [28–32]. However, these studies used CBD concurrently with other anticonvulsant therapies, leaving unknown whether the reported side-effects had any direct relation to CBD alone. The efficacy of CBD in treating intractable epilepsy has led to the U.S. Food and Drug Administration (FDA) approval of Epidiolex®, a plant-based purified CBD medication, for the treatment of intractable epilepsies in children. This approval has coincided with investigative avenues for the use of CBD-based medications in other conditions, including retrospective studies in children with ASD [33,34]. Single doses of CBD have been shown to modulate GABA levels in prefrontal regions in humans [35]. This presents an opportunity to relate pharmacological treatments, including CBD, to changes in E:I balance, as GABAergic inhibition has long been thought to drive network synchrony that can be partially captured by power spectral parameterization in EEG [36–38]. While CBD has not been directly studied in animal models of ASD, its use in treating seizures in a mouse model of Dravet syndrome, a severe genetic epilepsy syndrome, also resulted in improvements to social deficits [39]. Furthermore, CBD administration during adolescence in C57BL/6J male mice did not produce substantial negative impacts on a range of locomotor or anxiety-like behaviors in adulthood, but instead yielded modest improvements in spatial memory [40]. Altogether, CBD-based medication presents an ideal candidate for studying cognitive and behavioral improvements in children with ASD due to: the availability of a formulation with regulatory precedent, (2) a promising safety profile, and (3) preliminary data in both humans and mouse models that show these improvements. Still, to the best of our knowledge, there has been no active, longitudinal, and placebo-controlled monitoring of brain activity and cognitive-behavioral abilities in children with low-functioning ASD in response to CBD treatment.

As CBD is thought to improve negative ASD symptoms, we aimed to explore the potential of resting-state, aperiodic activity measures as a promising biomarker for monitoring the indirect effects of CBD treatment on brain electrical activity in a neurodevelopmental context. We focused on investigating these relationships via 7-COOH-CBD, a major metabolite of CBD, due to its longer half-life in blood compared to CBD and 7-OH-CBD [41–43]. This metabolite’s levels in blood remain relatively stable over multiple days, making it more likely to reflect CBD metabolism at the time when our measures of interest - EEG recordings and cognitive-behavioral assessments - were obtained. Thus, the 7-COOH-CBD metabolite could be a more reliable assessment of CBD treatment effects near the end of the dosing regimen, even when blood draws weren’t collected on the same day as those other measures. We hypothesized that EEG signal features and cognitive-behavioral assessment outcomes are predictive of 7-OH-CBD metabolite blood levels as a reflection of positive CBD treatment responses. To test this, we used state-of-the-art analytical tools to parameterize electrophysiological periodic and aperiodic features of non-invasive scalp EEG and correlate them to CBD metabolite blood levels and cognitive-behavioral assessments from the same individuals before and after an 8-week-long oral CBD dosing regimen. Here, we provide novel evidence that daily CBD treatment is associated with broadband spectral electrophysiological changes and improves cognition in boys with low-functioning autism spectrum disorder. Our spectral parameterization approach could thus offer a more sensitive method for tracking treatment outcomes above and beyond more traditional oscillatory measures that are typically studied in isolation within clinical settings.

## 2. Patients and Methods

### 2.1. Experimental recordings and cognitive-behavioral testing

#### 2.1.1. Participants

Our data was sourced from a clinical trial performed at the University of California. The trial was approved by the institutional review board of the University of California, San Diego and registered with [www.clinicaltrials.gov](http://www.clinicaltrials.gov) ([NCT04517799](http://medrxiv.org/lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT04517799&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom)) to investigate the effect of daily oral CBD (Epidiolex®) on problem behaviors in boys with severe symptoms of autism. Due to the subjects’ substantially limited cognitive capabilities, a waiver of child assent was approved for the study from which our data was sourced. Parental or guardian permission was obtained for each child’s enrollment and participation prior to their inclusion in the study from which our data was sourced. The parent or guardian also provided consent for completing the questionnaires required for the study from which our data was sourced. The dataset included in this study included 24 male participants aged 7-14 years that had a confirmed diagnosis of autism based on Autism Diagnostic Observation Schedule (ADOS) testing and were free of other neurological conditions. Included participants were reported to demonstrate severe stereotypies, aggressive and/or self-injurious behaviors, and/or pervasive hyperactivity impairing daily-life functioning.

#### 2.1.2 Experimental design

Our dataset consisted of assessments made via a randomized, two-arm design (CBD-Placebo or Placebo-CBD) with treatment periods lasting 8 weeks, separated by a 4-week washout period. Cognitive-behavioral tests and EEG recordings were conducted at four time points: Baseline, Post-placebo, Post-CBD, and Post-wash. The trial used Epidiolex®, an FDA-approved oral purified CBD solution developed by GW Research Ltd. for treating certain rare pediatric epilepsies. The dosage was adjusted based on participant weight, with a gradual increase over the first two weeks to reach the maximum dose (20 mg/kg/day, divided into two oral doses) for the remainder of the study period. Our analyses focused solely on the EEG data and cognitive-behavioral assessments collected during this trial, aiming to investigate the relationship between these measures and CBD metabolite blood levels.

CBD is metabolized to form the major active metabolite 7-hydroxy-cannabidiol (7-OH-CBD) and non-active metabolite 7-carboxy-cannabidiol (7-COOH-CBD) [41,44]. The dataset used in this study also included quantifications of whole blood samples near EEG and cognitive-behavioral assessments. For Post-CBD study time point, quantifications were taken an average of 2-3 days prior to assessments. The lower limits of quantification (LLOQ) in whole blood were: 0.5 ng/mL for CBD, for 7-OH-CBD, and for 7-COOH-CBD. The upper limits of quantification (ULOQ) in whole blood were: 500 ng/mL for CBD, 500 ng/mL for 7-OH-CBD, and 3000 ng/mL for 7-COOH-CBD. Values below the LLOQ of metabolites were implemented as 0 in our analyses, while those exceeding the ULOQ were analyzed as the max value of each metabolite.

While this approach may have slightly underestimated very low concentrations and capped the highest concentrations, it provided us with a standardized way to deal with values outside the quantifiable range.

#### 2.1.3 Cognitive-behavioral testing

Our dataset included cognitive and behavioral assessments from each time point: (1) Test of Nonverbal Intelligence, Fourth Edition (TONI-4), a language-free test that assesses intelligence in individuals with limited language abilities [45], (2) Repetitive Behavior Scale – Revised (RBS-R), a rating scale administered to the participant’s primary caregiver to report the frequency and severity of the repetitive and restricted behaviors that are a common feature in individuals with ASD [46], (3) Peabody Picture Vocabulary Test version 4 (PPVT-4), a measure of receptive vocabulary for Standard American English [47], (4) Expressive One Word Picture Vocabulary Test – Fourth Edition (EOWPVT-4), an assessment of the participant’s ability to name a variety of objects, actions, and concepts [48], (5) Beery-Buktenica Test of Visual Motor Integration (Beery VMI-6), a composite assessment of the participant’s ability to integrate their visual and motor abilities and includes supplemental test scores for visual perception (Beery VP) and motor coordination (Beery MC) [49]. We used raw scores instead of standardized scores for all cognitive-behavioral assessments, largely due to the severe intellectual disability of participants. Specifically, raw scores offered a more direct measure of absolute ability in each assessment, rather than relative standing compared to typically developing peers. Indeed, as standardized scores are normalized for typically developing populations, individuals with severe intellectual disabilities may largely score at the floor of these standard scores. This floor effect could reduce the sensitivity to detect changes in cognitive-behavioral assessments in our dataset due to the more limited range of values in standardized scores compared to raw scores.

#### 2.1.4 EEG recording setup

Our dataset included recordings made using a 20-channel dry EEG wireless headset device configured in the 10-20 montage (CGX, formerly Cognionics, Quick-20 Dry EEG Headset, [https://www.cgxsystems.com](https://www.cgxsystems.com)). Electrodes were referenced online to the left ear clip (A1), and the right ear clip (A2) was recorded as an active electrode. EEG data were digitized with 24 bits of resolution at 500 Hz with a bandwidth of 0-131 Hz with true DC coupling. When possible, electrode impedance was kept under 500 kΩ prior to recordings. EEG recordings were made during awake, resting state periods with the goal of capturing 5 to 10 minutes during which the child remained calm by viewing a youtube video of their choice.

### 2.2. Data analysis

Data analysis was performed with python using numpy [1.26.3] [50], scipy [1.11.4] [51], autoreject [0.4.3] [52], MNE [1.6.0] [53], pingouin [0.5.4] [54], and specparam [1.1.0] [55] for EEG, and statsmodel for calculating linear mixed effects models [0.14.1] [56]. A total of 19 EEG channels (F7, Fp1, Fp2, F8, F3, Fz, F4, C3, Cz, P8, P7, Pz, P4, T3, P3, O1, O2, C4, T4) were included in the analysis. Channels were also subdivided into frontal (F7, F3, Fz, F4, F8), central (Cz, C3, C4) and occipital (O1, O2) groups for post-hoc testing. The analysis code needed to reproduce the analysis and figures is provided here: [https://github.com/voytekresearch/cbd-asd-eeg](https://github.com/voytekresearch/cbd-asd-eeg)

#### 2.2.1. Preprocessing

First, EEG data was digitally re-referenced offline to a linked ears reference composed of both A1 and A2 for analysis. We then implemented an automated cross-validation procedure via the Autoreject python package [52] to estimate a within-session, “global” peak-to-peak voltage rejection threshold shared across all channels. This rejection threshold was then used to remove channels contaminated by excessive noise artifacts. The surviving channel data was then bandpass filtered between 1-100 Hz per specified recommendations of the ICLabel independent component classifier integrated with MNE [57]. In brief, to detect and remove ocular artifacts and eyeblinks, independent component analysis (ICA) was applied to the filtered data using the “extended infomax” method and to decompose the signal into the smallest number of components required to explain the 99% of the cumulative variance of the data. Components labeled as “brain” or “other” (i.e., non-classifiable components) were kept while all others, including labels that indicated ocular artifacts, were removed. Due to the small number of channels, ICA could not always remove ocular and muscular artifacts. Therefore, we epoched EEG data into 10-second segments and calculated “local” peak-to-peak rejection thresholds within these epochs using the same Autoreject python package procedures described above. This “local” approach estimates a rejection threshold for each channel within epochs, and depending on the number of highly contaminated channels, this epoch is repaired by interpolation or excluded from subsequent analysis. Surviving epochs then had a hamming window applied to account for epoch edge artifacts and discontinuity before undergoing concatenation. The preprocessing procedures used in this study were devised to account for the difficulty in maintaining a high quality of EEG data in ASD children who are prone to movement.

Exclusion criteria were: (1) Sessions with post-processed EEG data of < 30 seconds in duration (2) in cases where multiple EEG recording attempts were made due to disruption from the participant within a study time point, only the session with the highest average alpha power signal-to-noise-ratio estimated from channels O1 and O2 was used for further analysis. In the end, a total of 87 EEG testing sessions (Baseline: 20, Post-CBD: 23, Post-Placebo: 22, Post-Wash: 22) were included in this study, with 3-4 recordings per child corresponding to each of the 4 study time points.

#### 2.2.2. Calculation of EEG signal features

PSDs were estimated from EEG signals using Welch’s method [58] using 1.0 s Hamming windows with 0.5 s overlap within the frequency range 0.1-50 Hz. High levels of muscle noise are indicated by elevated power levels above 10 Hz in the spectral domain. To minimize the contamination of the exponent model fit due to muscle noise, we parameterized the spectra within the 0.5-13 Hz frequency range [5]. Thus, the settings for the spectral parameterization algorithm were: peak width limits: (0.5, 13.0); maximum number of peaks: 6; minimum peak amplitude exceeding the aperiodic fit: 0.0; peak threshold: 2.0; and aperiodic mode: ‘fixed’. This resulted in estimates of the aperiodic exponent, aperiodic offset, and aperiodic adjusted delta (1-4 Hz), theta (4-8 Hz), and alpha (8-13 Hz) power. Only channels with model fits satisfying a minimum R2 value of 0.95 were kept for further analysis. To quantify the strength of neural oscillations, we calculated the Signal-to-Noise Ratio (SNR) for each frequency band of interest (Delta, Theta, and Alpha). In brief, for each channel and frequency band, we computed an aperiodic-corrected PSD by subtracting the modeled aperiodic fit from the original PSD. Within each frequency band, we identified the maximum value of the corrected PSD, which we defined as the SNR for that channel and band. To obtain a global measure, we calculated the mean SNR across all valid channels for each frequency band. This approach allowed us to quantify the strength of oscillations in each frequency band relative to the aperiodic background, providing a robust measure of band-specific neural activity across the scalp. The resulting SNR values are expressed in units of power spectral density (µV2/Hz), representing the difference between the peak oscillatory power and the aperiodic background power in each frequency band.

#### 2.2.3. Statistical analysis

We z-scored metabolite blood levels across subjects to account for non-normality caused by floor and ceiling effects from LLOQ and ULOQ detection thresholds. For statistical evaluation, Linear Mixed Effects (LME) models were built to investigate the predictive relationship between z-scored CBD metabolite levels (7-COOH-CBD) in blood samples and (1) EEG signal features or (2) cognitive-behavioral testing outcomes. Linear mixed-effects models were fitted using the smf.mixedlm function from the Statsmodels library [0.14.1], with a subject grouping variable as a random intercept term to account for within-subject correlations across study time-points. To account for variance explained by individual differences and treatment conditions, fixed terms included the randomization order (CBD first or placebo first) and the time point of EEG and cognitive-behavioral assessments (Baseline, Post-placebo, Post-CBD, Post-wash). Additionally, to account for subject characteristics that might affect CBD metabolism or cognitive-behavioral assessments, we included fixed terms for age and ADOS module.

To test the predictive relationships between 7-COOH-CBD metabolite levels in blood samples and EEG signal features, we included fixed terms for aperiodic offset, aperiodic exponent, and aperiodic-adjusted signal-to-noise ratios in the delta, theta, and alpha bands. Thus, the LME model for 7-COOH-CBD metabolite levels in blood against EEG signal features consisted of the following formula: ![Formula][1]</img>  

Where *Mz* is the z-score of the CBD metabolite (7-COOH-CBD), *T* the study time point (Baseline, Post-placebo, Post-CBD or Post-wash), *Rand* is the randomization condition (CBD first or placebo first), *Off* is the aperiodic offset, *Exp* is the aperiodic exponent, *Delta* is the aperiodic-adjusted delta SNR, *Theta* is the aperiodic-adjusted theta SNR, *Alpha* is the aperiodic-adjusted alpha SNR, *Age* is the age of the participant, and *ADOS* is the subject’s corresponding ADOS module. *βx* is the linear regression coefficient for each corresponding fixed effect term (β0 being the intercept term). *(1*|*SUBJ)* represents the random intercept term for each participant, allowing for individual variation in the overall blood level of the metabolite. εi is the error term. This model was applied separately to each CBD metabolite and repeated for each electrode group (Frontal, Central and Occipital). This approach allowed us to investigate whether the relationships between EEG signal features and CBD metabolite levels differed across the scalp.

To test the predictive relationships between metabolite levels in blood samples and cognitive-behavioral assessments, we adapted our model to include fixed terms corresponding to the score of each assessment. Thus, the LME model for 7-COOH-CBD metabolite levels in blood against the cognitive-behavioral assessment scores consisted of the following formula: ![Formula][2]</img>  

Where *Mz* is the z-score of the CBD metabolite (7-COOH-CBD), *T* is the study time point (Baseline, Post-placebo, Post-CBD or Post-wash), *Rand* is the randomization condition (CBD first or placebo first), *Score* is the assessment score, *Age* is the age of the participant, and *ADOS* is the subject’s corresponding ADOS module. *βx* is the linear regression coefficient for each corresponding fixed effect term (β0 being the intercept term). *(1*|*SUBJ)* represents the random intercept term for each participant, allowing for individual variation in the overall blood level of the metabolite. εi is the error term. This model was applied separately for each assessment (TONI-4, RBS-R, PPVT-4, EOWPVT-4, Beery VMI-6, Beery VP, and Beery MC). This approach allowed us to investigate whether the relationships between cognitive-behavioral assessment scores and 7-COOH-CBD metabolite levels differed across various domains of cognition and behavior being assessed.

All model fitting was performed using the Nelder-Mead optimization method, with a maximum of 1000 iterations to ensure convergence. A Variance Inflation Factor (VIF) was calculated for each explanatory variable to assess multicollinearity and determine whether correlated variables should be combined or removed from the model (e.g. variables above a VIF of 10). We extracted coefficients, standard errors, z-values, z-statistic associated p-values, and 95% confidence intervals from model outputs to determine post-hoc linear regression analyses on each significant predictor’s estimated marginal means in the Post-CBD time point.

Mediation analyses were performed using the mediation_analysis function from the pingouin library (Vallat 2018) across electrode groups to assess the impact of metabolite levels on the relationship between EEG signal features and cognitive-behavioral measures. Separate analyses were performed for each combination of cognitive-behavioral assessment (TONI-4, RBS-R, PPVT-4, EOWPVT-4, Beery VMI-6, Beery VP, and Beery MC), CBD metabolite (7-COOH-CBD), electrode group (Frontal, Central, Occipital) and EEG signal feature (Offset, Exponent, Alpha SNR, Theta SNR, Delta SNR). Each model included the randomization order (CBD first or placebo first), the time point of EEG and cognitive-behavioral assessments (Baseline, Post-placebo, Post-CBD, Post-wash), age and ADOS module as covariates. Each model was estimated using 1000 bootstrap samples to calculate confidence intervals and p-values for the indirect, direct, and total effects. To account for multiple comparisons, we implemented a hierarchical False Discovery Rate (FDR) correction. First, we applied FDR correction across electrode groups (Occipital, Frontal, Central). Then, within each electrode group that showed a significant effect after the first-level correction, we applied FDR correction to all significant indirect effects within that group. Only mediation relationships that remained significant after the hierarchical FDR correction (corrected p < 0.05), were included in our visualizations. We estimated the mediation effect by calculating the product of β1 from the path “Metabolite ∼ X” and β2 from the path “Y ∼ Metabolite”, which represents the indirect effect of EEG feature on cognitive-behavioral assessments through its effect on CBD metabolite levels.

## 3. Results

### 3.1. Broadband and aperiodic adjusted resting state EEG features across the scalp had a positive association with 7-COOH-CBD metabolite levels in blood

The design of the trial from which our dataset was sourced allowed us to evaluate the effects of oral CBD administration compared to placebo across the study duration (Figure 1A). Resting state, non-invasive scalp EEG recordings, blood sample collection, and cognitive-behavioral assessments were made at baseline, after each 8-week treatment period, and following the washout phase. CBD is known to be metabolized by the human body to form 7-OH-CBD and 7-COOH-CBD, with 7-COOH-CBD concentrations reported to be highest of these three as a function of oral CBD consumption [41]. Analysis of CBD metabolite concentrations in these samples revealed levels comparable to those found in a prior report using orally administered CBD (Epidiolex®) for four days in healthy adults [41]. We found a clear increase across all CBD metabolites in the blood following only the 8-week CBD treatment period, regardless of whether CBD was administered in the first or second treatment phase (One-way ANOVA, p’s < 0.001) (Figure 1B–D).

![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/30/2024.09.27.24314448/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2024/09/30/2024.09.27.24314448/F1)

Figure 1. Overview of study design and spectral parameterization approach.
**(A)** Schematic representation of data included in our analyses, comprising four study time points for EEG, blood draws, and cognitive-behavioral assessments. **(B)** Concentration of surveyed metabolite levels in blood, including CBD, **(C)** 7-OH-CBD and **(D)** 7-COOH-CBD. **(E)** Schematic of electrodes and electrode groups used in this study. **(F)** Representative traces of segmented EEG before and after artifact rejection. **(G)** Representative power spectrum of channel ‘O2’ alongside extracted aperiodic exponent and offset values for all electrode channels in a single session. **(H)** Duration of post-processed EEG traces across study time points. **(I)** Mean and standard error of participant age alongside total sessions included from each data modality across study time points. \***|p < 0.001.

EEG signal features were calculated for each channel, recording session and participant prior to calculating averages for each corresponding electrode group (Figure 1E–F). We used the Python package specparam [55] to parameterize the periodic and aperiodic power spectral density components of EEG signals. Extracted features included the broadband offset and exponent of the aperiodic signal model fit, as well as the strength of oscillations relative to the aperiodic background (referred to as the SNR, see Methods) within canonical frequency bands: delta (1-4 Hz), theta (4-8 Hz), and alpha (8-13 Hz) (Figure 1G). Prior EEG studies on intellectually disabled and minimally verbal neurodevelopmental disorder populations have shown heterogeneity in neural spectral power and aperiodic activity across the scalp [10,14,16]. We used a similar approach and grouped channels into frontal, central, and occipital areas to investigate how different brain regions may be affected by CBD treatment. Post-processed EEG durations were similar across study time points (Figure 1H). A total of 87 resting state scalp EEG recordings were analyzed across the four study timepoints (Figure 1I).

Our linear mixed effect models revealed significant associations between aperiodic neural activity and CBD treatment while accounting for participant age, ADOS module, and treatment randomization order. As expected, we found a positive relationship between the CBD time point and normalized 7-COOH-CBD metabolite levels (β = 2.16, 97.5% CI = [0.035, 0.193], p < 0.001, Figure 2A). We also found a significant negative relationship between the aperiodic exponent and normalized 7-COOH-CBD metabolite blood levels (β = -0.416, 97.5% CI = [-0.714, -0.118], p = 0.006). In contrast, we found a significant positive relationship between the aperiodic offset and normalized 7-COOH-CBD metabolite blood levels (β = 0.114, 97.5% CI = [0.035, 0.193], p = 0.005). To further interpret our findings, we calculated the Post-CBD timepoint estimated marginal means (EMMs) derived from the significant model predictions (rsEEG n = 23). In frontal electrodes, we confirmed that the aperiodic exponent showed a significant negative relationship with normalized 7-COOH-CBD metabolite blood levels (R2 = 0.252, p = 0.015) (Figure 2B), while the aperiodic offset showed a significant positive relationship (R2 = 0.388, p = 0.002) (Figure 2C).

![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/30/2024.09.27.24314448/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2024/09/30/2024.09.27.24314448/F2)

Figure 2. Relationships between EEG power spectrum features and levels of 7-COOH-CBD metabolite in blood.
Results from linear mixed models and linear regressions relating EEG features to levels of 7-COOH-CBD metabolite in blood across groups of electrodes correspondingly located above (A–C) frontal, (D–G) central, and (H–J) occipital regions of the brain. Visualization of each linear mixed model shows the z-statistic for each coefficient and the two-tailed p-value associated with the z-statistic. *p<0.05, **p<0.01, \***|p<0.001.

We found a positive relationship between the CBD time point and normalized 7-COOH-CBD metabolite levels (β = 2.133, 97.5% CI = [1.890, 2.375], p < 0.001, Figure 2D). We also found a significant negative relationship between the aperiodic exponent and normalized 7-COOH-CBD metabolite blood levels (β = -0.290, 97.5% CI = [-0.566, -0.013], p = 0.040), and a significant positive relationship between the aperiodic offset and normalized 7-COOH-CBD metabolite blood levels (β = 0.106, 97.5% CI = [0.028, 0.184], p = 0.008). There was also a weak, yet non-significant, positive relationship between Theta SNR in central electrodes only (β = 0.094, 97.5% CI = [-0.004, 0.192], p = 0.060). Using EMMs, we confirmed that in central electrodes, the aperiodic exponent showed a weak negative relationship with normalized 7-COOH-CBD metabolite blood levels (R2 = 0.147, p = 0.071) (Figure 2E). The aperiodic offset showed a significant positive relationship (R2 = 0.363, p = 0.002) (Figure 2F), and theta SNR showed a weak, yet significant, positive relationship (R2 = 0.194, p = 0.035) (Figure 2G).

We found a positive relationship between the CBD time point and normalized 7-COOH-CBD metabolite levels (β = 2.161, 97.5% CI = [1.898, 2.424], p < 0.001, Figure 2H). We also found a significant negative relationship between the aperiodic exponent and normalized 7-COOH-CBD metabolite blood levels (β = -0.331, 97.5% CI = [-0.585, -0.078], p = 0.010), and a significant positive relationship between the aperiodic offset and normalized 7-COOH-CBD metabolite blood levels (β = 0.105, 97.5% CI = [0.027, 0.184], p = 0.008). Using EMMs, we confirmed that in occipital electrodes, the aperiodic exponent showed a weak negative relationship with normalized 7-COOH-CBD metabolite blood levels (R2 = 0.115, p = 0.113) (Figure 2I), while the aperiodic offset showed a significant positive relationship (R2 = 0.476, p < 0.001) (Figure 2J).

### 3.2. Measures of visuomotor coordination and nonverbal intelligence had a positive association with 7-COOH-CBD metabolite levels in blood

Our linear mixed effect models revealed significant associations between cognitive performance measures and CBD treatment while accounting for participant age, ADOS module, and treatment randomization order. In our model for Beery-Buktenica test performance, we found a significant positive relationship between the CBD time point and normalized 7-COOH-CBD metabolite levels (β = 2.192, 97.5% CI = [1.940, 2.443], p < 0.001, Figure 3A). We also observed a significant positive relationship between VMI scores and normalized 7-COOH-CBD metabolite blood levels (β = 0.035, 97.5% CI = [0.017, 0.053], p < 0.001). In contrast, we found a weak, yet non-significant, negative relationship between ADOS module and normalized 7-COOH-CBD metabolite blood levels (β = -0.170, 97.5% CI = [-0.344, 0.003], p = 0.054). Using EMMs, we confirmed that for the Beery-Buktenica tests (n = 21, Post-CBD), VMI scores showed a significant positive relationship with normalized 7-COOH-CBD metabolite blood levels (R2 = 0.472, p = 0.001) (Figure 3C).

![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/30/2024.09.27.24314448/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2024/09/30/2024.09.27.24314448/F3)

Figure 3. Relationships between cognitive-behavioral assessments and levels of 7-COOH-CBD metabolite in blood.
Results from linear mixed models and linear regressions relating **(A, C)** visuomotor coordination and **(B, D)** non-verbal intelligence test scores to levels of 7-COOH-CBD metabolite in blood. Visualization of each linear mixed model shows the z-statistic for each coefficient and the two-tailed p-value associated with the z-statistic. #p<0.06, *p<0.05, \***|p<0.001.

In our model for non-verbal intelligence performance, we also found a significant positive relationship between the CBD time point and normalized 7-COOH-CBD metabolite levels (β = 2.201, 97.5% CI = [1.845, 2.557], p < 0.001, Figure 3B). Furthermore, we observed a significant positive relationship between TONI-4 scores and normalized 7-COOH-CBD metabolite blood levels (β = 0.015, 97.5% CI = [0.002, 0.028], p = 0.027). We also found an unexpected significant negative relationship between ADOS module and normalized 7-COOH-CBD metabolite blood levels (β = -0.192, 97.5% CI = [-0.377, -0.008], p = 0.410). Using EMMs, we confirmed that TONI-4 scores (n = 15, Post-CBD) showed a positive, yet non-significant, relationship with normalized 7-COOH-CBD metabolite blood levels (R2 = 0.227, p = 0.072) (Figure 3D). Our analysis revealed no significant effects of CBD treatment on repetitive behaviors (via RBS assessment), expressive vocabulary (via EOWPVT-4 assessment), or receptive vocabulary (via PPVT-4 assessments) (Figure 3 Supplement).

### 3.3. Changes in visuomotor coordination were linked to occipital delta frequency oscillatory activity and mediated by 7-COOH-CBD metabolite levels in blood

By use of mediation analysis, we found that changes in visuomotor coordination attributed to occipital oscillatory neural activity were mediated by decreases in normalized 7-COOH-CBD metabolite blood levels, with distinct effects observed for the delta frequency bands (Figure 4). This analysis revealed that an increase in Delta SNR is associated with a significant increase in Beery VMI Score (Cdirect = 1.98, 97.5% CI [0.885, 3.066], p = 0.001). In contrast, the indirect effect through 7-COOH-CBD metabolite blood levels was negative (Cindirect = -0.41, 97.5% CI [-0.988, -0.061], p < 0.036). This suggests that while increased Delta SNR directly predicts higher Beery VMI scores, it also leads to decreased 7-COOH-CBD metabolite levels (β1 = -0.25, 97.5% CI [0.476, -0.023], p = 0.031), which in turn are positively associated with Beery VMI Scores (β2 = 1.17, 97.5% CI [0.080, 2.261], p = 0.036). In essence, 7-COOH-CBD metabolite levels acted as a significant mediator in the relationship between Delta SNR and Beery VMI score, partially suppressing the overall positive effect of Delta SNR on visual-motor integration performance.

![Figure 4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/30/2024.09.27.24314448/F4.medium.gif)

[Figure 4.](http://medrxiv.org/content/early/2024/09/30/2024.09.27.24314448/F4)

Figure 4. Mediating effect of 7-COOH-CBD metabolite levels in blood on the relationship between aperiodic adjusted occipital delta power and visuomotor coordination.
**(A)** Results from mediation analysis describing how the effect of aperiodic adjusted occipital delta power on visuomotor coordination is mediated by levels of 7-COOH-CBD metabolite in blood. β1 and β2 indicate the coefficients associated with each path, with Cdirect and Cindirect indicating the effect of 7-COOH-CBD metabolite levels on these relationships.

## 4. Discussion

The double-blind, placebo-controlled design of the trial from which our dataset was sourced provided rigorous, longitudinal measures of therapeutic CBD effects on EEG features, CBD-related metabolism, and cognitive-behavioral assessments in children with low-functioning forms of ASD. Findings from our linear mixed modeling approach suggest that this CBD treatment regimen is associated with significant changes in aperiodic neural activity across the scalp. The scalp-wide negative relationship between 7-COOH-CBD metabolite levels and the aperiodic exponent, coupled with the positive relationship with the aperiodic offset, indicated a shift towards a “flatter” power spectrum and increased broadband neural power. Our results also suggested that this CBD treatment regimen is associated with improvements in both visual-motor integration skills, as assessed by the Beery-Buktenica VMI test, and nonverbal intelligence, as assessed by the TONI-4. In contrast, this CBD treatment regimen did not have a notable impact on repetitive behaviors or expressive and receptive vocabulary abilities in the sampled population included in our analyses. The larger magnitude in the positive relationship observed with the VMI score and 7-COOH-CBD metabolite levels, compared to the TONI-4 score, may indicate that this CBD treatment regimen has a more pronounced effect on visual-motor skills than on general nonverbal intelligence. Additionally, the negative relationship with participants’ ADOS module across both cognitive-behavioral assessment scores suggests that individuals with higher autism severity may show less improvement in these domains following this CBD treatment regimen. Finally, our mediation analyses suggest that increased Delta SNR was generally associated with better visual-motor integration performance, and that this relationship is partially mediated by levels of 7-COOH-CBD in blood. These findings align with the notion that the effects of CBD treatment on brain activity and cognitive-behavioral abilities may involve a complex mixture of physiological pathways, indirectly via 7-COOH-CBD (as it does not cross the blood-brain barrier), and perhaps directly via CBD and 7-OH-CBD metabolites and their downstream molecular targets.

Conducting longitudinal studies in children with severe forms of ASD presents unique obstacles that imposed clear limitations to our analyses. Due to the challenges in maintaining consistent engagement at rest in this patient population, children viewed videos of their choice during EEG recordings. Differences in video content were not available, leaving unknown whether individual preferences could have affected attention levels and, consequently, brain activity across participants and EEG recording sessions. Yet prior work in similarly intellectually disabled populations that watched videos of their choice during resting state EEG did not suggest concern for confounded interpretations of the holistic measures of network neural activity used in our analyses [10]. Furthermore, the sensory sensitivities commonly found in ASD could have also potentially affected EEG recording quality in the trial as a function of motor artifacts due to restless behaviors. We believe our unbiased, automated approach using ICA and autoReject for artifact removal prior to spectral parameterization partially accounted for this confound, as shown in other studies [52].

In addition, our dataset had variability in sample sizes across these assessments, presumably due to difficulties encountered in administering several cognitive-behavioral assessments, experimenter error, or low participant cooperation. We believe these variations were accounted for in our linear mixed modeling approach and through the inclusion of a minimum of three study time points per participant in our analyses. Importantly, our findings are limited to male children, presumably due to recruitment availability from local, severe ASD patient populations. This limitation reduces the generalizability of our findings to the broader ASD population, including severe forms of the condition known to primarily affect females, like Rett syndrome. However, there have been few differences reported between male and female metabolism of CBD, with one report indicating that adult females show increased max levels of 7-COOH-CBD metabolites in blood compared to adult males after normalizing by body-weight [41]. While participants included in our dataset varied in age, all were in late childhood or early adolescence and were determined to have minimal verbal and cognitive abilities as measured by a diverse battery of assessments. Finally, our analyses lacked a comparison to an age-matched typically developing group, which leaves unknown whether the effects of CBD on EEG measures are unique to ASD. Yet, in light of these limitations, our use of resting-state EEG measures provided a suitable and well-validated approach to relating cognitive-behavioral performance to neural activity, as shown in similar studies on intellectually impaired and minimally verbal children populations [10,14,16,17].

There is an urgent need for the development and validation of novel, systemic drugs for treating the negative symptoms of ASD that show more favorable long-term side effect profiles compared to existing treatments. A retrospective study demonstrated that children with ASD, aged 5-17 years, exhibited improvements in disruptive behavior when treated with an oral CBD:THC (tetrahydrocannabinol) formulation at a ratio of 20:1, with CBD dosage not exceeding 10 mg/kg/day [33]. Another retrospective study showed that parents reported improvements to comorbid symptoms including aggression, hyperactivity and anxiety following a treatment regimen with a similar CBD:THC formulation of 20:1 for a median duration of 66 days across participants [34]. In contrast, our analyses provide the first active, longitudinal monitoring of treatment outcomes via objective assessment tools, including a battery of well-validated cognitive-behavioral tests, resting state EEG, and importantly, a biometric readout of CBD metabolism.

Our analysis’ predictive effects predominantly revolved around the 7-COOH-CBD metabolite. Prior work on acute CBD administration has shown that CBD and 7-OH-CBD metabolite levels in blood decrease rapidly beyond 8 hours post-administration, whereas 7-COOH-CBD exhibits a half-life exceeding 48 hours [43]. A report that used a similar chronic dosing regimen as this study found that CBD levels in blood are detectable in the majority of patients after 2-3 days from treatment cessation; however, this was in an adult Huntington’s Disease patient population, and levels maxed at 2.5 ng/mL, much lower than those reported in this study [59]. Given that assessments were not guaranteed to be conducted within a 8-12 hour period following blood draws, our focus on 7-COOH-CBD offered a more reliable measure of any sustained influence of CBD treatment on our measures of interest. Furthermore, the longitudinal study design from which our data were sourced used a daily administration regimen, which was reflected in the substantially higher concentrations of 7-COOH-CBD metabolite blood levels observed in our analyses, compared to those of CBD and 7-OH-CBD. However, the 7-COOH-CBD metabolite is non-active and does not cross the blood-brain barrier. Consequently, any relationships observed between the measures included in our dataset and 7-COOH-CBD metabolite levels in blood likely reflect indirect mechanisms through the periphery, possibly involving other metabolites, rather than direct modulation of neural activity, as one would expect from CBD’s active metabolites. In light of these considerations, we recommend that future work should conduct blood draws a few hours prior to assessments to directly relate CBD and 7-OH-CBD levels to neural activity and cognitive-behavioral assessments. If blood draws cannot occur during this time-window, we propose that 7-COOH-CBD could serve as a more appropriate biomarker for assessing the neural and behavioral effects of daily CBD administration in ASD children.

The need for novel treatments for negative ASD symptoms is coupled with the need to characterize brain activity changes in response to such systemic drugs. This characterization could inform active treatment progression in cognition and behaviors of children with ASD across longitudinal studies and in the clinic. Candidate biomarkers have included resting state EEG measurements, which have been shown to be associated with ASD symptom severity, even when collected as early as the first 3 months of life [24,60–63]. However, these early life studies did not include explicit parameterization of neural power spectra as we did in our current analyses, which means they are mixing both periodic and aperiodic components, making it unclear which neural process is altered in ASD. Studies that have addressed this issue have yielded mixed results depending on the age of the patient population. Adult ASD populations show mixed differences in resting-state periodic and aperiodic EEG features compared to non-ASD individuals [24,64]. In contrast, in preterm children with ASD, aperiodic features are associated with subsequently increased autism risk [65]. Furthermore, aperiodic activity in 6-8 year old children with ASD showed high test-retest reliability compared to healthy controls [8], underscoring its potential as a candidate biomarker to be tracked over developmental windows and between longitudinal study treatment periods. Our analytical study has addressed these gaps by using spectral parameterization on resting-state EEG from near-adolescent children with well-defined, severe ASD phenotypes. This approach revealed that changes in aperiodic, rather than periodic, neural activity are primarily influenced by this 8-week CBD treatment regimen. Specifically, we observed a negative relationship between levels of 7-COOH-CBD metabolite in blood and the aperiodic exponent, indicating a “steeper” slope of the power spectrum model fit after this CBD treatment regimen, unlike the “flattening” observed in typically developing children [5,66,67]. In contrast, we saw a positive relationship between levels of 7-COOH-CBD metabolite in blood and the aperiodic offset, indicating an increase in broadband power, which does reflect what has been observed in typically developing children [66,67]. The contrast in these findings warrant further investigation and suggest that this CBD regimen may not be able to induce all of the expected trajectories of periodic activity throughout neurodevelopment. Nonetheless, our findings support the use of both features of aperiodic activity as a candidate resting-state EEG biomarkers to track neural and cognitive-behavioral responses to non-psychoactive medical treatments for negative ASD symptoms, in parallel with other known ‘oscillatory’ measures involving mu, alpha, and as we saw in our own analytical study, delta and theta rhythmic activity [16,21–25].

In conclusion, we report that an 8-week daily CBD regimen using a purified formulation with regulatory precedent improved select domains of cognitive and behavioral ability, such as visuomotor coordination and nonverbal intelligence, and that these improvements coincided with neural activity changes primarily in the aperiodic component of non-invasive scalp EEG signals. These attributes align well with the current research needs in ASD pharmacotherapy for a safer and more effective alternative to existing treatments for improved cognitive and behavioral abilities in this patient population. Future studies should explore different dosing strategies and longer treatment durations to fully capture the potential of CBD in managing negative ASD symptoms.

## Data Availability

The data presented in this study will be available on request once data analyses are completed from UC San Diego Center for Medicinal Cannabis Research (CMCR) through contacting Doris Trauner, M.D.

## Data Availability Statement

The data presented in this study will be available on request once data analyses are completed from UC San Diego Center for Medicinal Cannabis Research (CMCR) through contacting Doris Trauner, M.D. The EEG and cognitive-behavioral assessment data are not publicly available in order to protect the children’s and their families’ privacy. The analysis code needed to reproduce the analysis and figures is provided here: [https://github.com/voytekresearch/cbd-asd-eeg](https://github.com/voytekresearch/cbd-asd-eeg)

## Author Contributions

C.C.: Formal analysis, Investigation, Methodology, Visualization, Writing - original draft, Writing - review and editing.

G.P.: Methodology, Data acquisition, Writing - review and editing.

A.H.: Writing - review and editing.

D.T.: Conceptualization, Project administration, Funding acquisition, Investigation and Writing - review and editing.

B.V.: Supervision, Project administration, Funding acquisition, and Writing - review and editing.

## Funding Statement

The research study was supported by a grant from the Ray and Ty Noorda Foundation and the Wholistic Foundation To Dr. Trauner. Purified CBD (Epidiolex®) and identical placebo compounds were provided by GW Pharmaceuticals (Cambridge, U.K.; now Jazz Pharmaceuticals). C.C. was supported by a NIH/NIMH Blueprint D-SPAN Award (K00 MH132569) and a NIH/NIGMS IRACDA Award (K12 GM068524). A.H. was supported by a NIH Blueprint BRAIN-ENDURE Award (R25NS119707). Research space for the EEG study was provided by the UC San Diego Center for Medicinal Cannabis Research, I. Grant, Director. Blood for cannabinoid levels was drawn at the UC San Diego Altman Clinical and Translational Research Institute (ACTRI). The project was partially supported by the National Institutes of Health, Grant UL1TR001442 (to the ACTRI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

## Conflicts of Interest

The authors declare no conflict of interest.

## Figure Legends

![Figure 3 Supplement.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/30/2024.09.27.24314448/F5.medium.gif)

[Figure 3 Supplement.](http://medrxiv.org/content/early/2024/09/30/2024.09.27.24314448/F5)

Figure 3 Supplement. Non-significant relationships between cognitive-behavioral assessments and levels of 7-COOH-CBD metabolite in blood.
Results from linear mixed models and linear regressions relating (**A)** repetitive behavior, **(B)** expressive vocabulary and **(C)** receptive vocabulary test scores to levels of 7-COOH-CBD metabolite in blood. Visualization of each linear mixed model shows the z-statistic for each coefficient and the two-tailed p-value associated with the z-statistic. *p<0.05, \***|p<0.001.

## Acknowledgements

We’d like to thank Dillan Cellier, Sydney Smith, and Dr. Maribel Patiño for their thoughtful discussion.

*   Received September 27, 2024.
*   Revision received September 27, 2024.
*   Accepted September 30, 2024.


*   © 2024, Posted by Cold Spring Harbor Laboratory

This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/)

## References

1.  1.GBD 2019 Mental Disorders Collaborators. Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry. 2022;9:137–150.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/s2215-0366(21)00395-3&link_type=DOI) 

2.  2.Varcin KJ, Nelson CAI. A developmental neuroscience approach to the search for biomarkers in autism spectrum disorder. Curr Opin Neurol. 2016;29:123.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/WCO.0000000000000298&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26953849&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

3.  3.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013.
    
    

4.  4.Gao R, Peterson EJ, Voytek B. Inferring synaptic excitation/inhibition balance from field potentials. NeuroImage. 2017;158:70–78.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neu-roimage.2017.06.078&link_type=DOI) 

5.  5.Schaworonkow N, Voytek B. Longitudinal changes in aperiodic and periodic activity in electrophysiological recordings in the first seven months of life. Dev Cogn Neurosci. 2020;47:100895.
    
    

6.  6.Li W. Excitation and Inhibition Imbalance in Rett Syndrome. Front Neurosci. 2022;16:825063.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3389/fnins.2022.825063&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35250460&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

7.  7.Wilkinson CL, Nelson CA. Increased aperiodic gamma power in young boys with Fragile X Syndrome is associated with better language ability. Mol Autism. 2021;12:17.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35064716&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

8.  8.Levin AR, Naples AJ, Scheffler AW, Webb SJ, Shic F, Sugar CA, et al. Day-to-Day Test-Retest Reliability of EEG Profiles in Children With Autism Spectrum Disorder and Typical Development. Front Integr Neurosci. 2020;14:21.
    
    

9.  9.Ostlund BD, Alperin BR, Drew T, Karalunas SL. Behavioral and cognitive correlates of the aperiodic (1/f-like) exponent of the EEG power spectrum in adolescents with and without ADHD. Dev Cogn Neurosci. 2021;48:100931.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.dcn.2021.100931&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33535138&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

10. 10.Roche KJ, LeBlanc JJ, Levin AR, O’Leary HM, Baczewski LM, Nelson CA. Electroencephalographic spectral power as a marker of cortical function and disease severity in girls with Rett syndrome. J Neurodev Disord. 2019;11:15.
    
    

11. 11.Fries P. A mechanism for cognitive dynamics: neuronal communication through neuronal coherence. Trends Cogn Sci. 2005;9:474–480.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.tics.2005.08.011&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16150631&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000232739000009&link_type=ISI) 

12. 12.Schnitzler A, Gross J. Normal and pathological oscillatory communication in the brain. Nat Rev Neurosci. 2005;6:285–296.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nrn1650&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15803160&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000228092300012&link_type=ISI) 

13. 13.Kahana MJ. The Cognitive Correlates of Human Brain Oscillations. J Neurosci. 2006;26:1669–1672.
    
    [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjk6IjI2LzYvMTY2OSI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDI0LzA5LzMwLzIwMjQuMDkuMjcuMjQzMTQ0NDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

14. 14.Sysoeva O, Maximenko V, Kuc A, Voinova V, Martynova O, Hramov A. Abnormal spectral and scale-free properties of resting-state EEG in girls with Rett syndrome. Sci Rep. 2023;13:12932.
    
    

15. 15.Ostlund B, Donoghue T, Anaya B, Gunther KE, Karalunas SL, Voytek B, et al. Spectral parameterization for studying neurodevelopment: How and why. Dev Cogn Neurosci. 2022;54:101073.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35074579&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

16. 16.Neo WS, Foti D, Keehn B, Kelleher B. Resting-state EEG power differences in autism spectrum disorder: a systematic review and meta-analysis. Transl Psychiatry. 2023;13:389.
    
    

17. 17.Van der Molen MJW, Van der Molen MW. Reduced alpha and exaggerated theta power during the resting-state EEG in fragile X syndrome. Biol Psychol. 2013;92:216–219.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.biopsycho.2012.11.013&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23182872&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

18. 18.Rubenstein JLR. Three hypotheses for developmental defects that may underlie some forms of autism spectrum disorder. Curr Opin Neurol. 2010;23:118–123.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/WCO.0b013e328336eb13&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20087182&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000276196700005&link_type=ISI) 

19. 19.Sohal VS, Rubenstein JLR. Excitation-inhibition balance as a framework for investigating mechanisms in neuropsychiatric disorders. Mol Psychiatry. 2019;24:1248–1257.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/S41380-019-0426-0&link_type=DOI) 

20. 20.Voytek B, Knight RT. Dynamic network communication as a unifying neural basis for cognition, development, aging, and disease. Biol Psychiatry. 2015;77:1089–1097.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.biopsych.2015.04.016&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26005114&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

21. 21.Bender A, Voytek B, Schaworonkow N. Resting-state is not enough: alpha and mu rhythms change shape across development, but lack diagnostic sensitivity. 2023:2023.10.13.562301.
    
    

22. 22.Freschl J, Azizi LA, Balboa L, Kaldy Z, Blaser E. The development of peak alpha frequency from infancy to adolescence and its role in visual temporal processing: A meta-analysis. Dev Cogn Neurosci. 2022;57:101146.
    
    

23. 23.Javitt DC, Siegel SJ, Spencer KM, Mathalon DH, Hong LE, Martinez A, et al. A roadmap for development of neuro-oscillations as translational biomarkers for treatment development in neuropsychopharmacology. Neuropsychopharmacology. 2020;45:1411–1422.
    
    

24. 24.Dickinson A, DiStefano C, Senturk D, Jeste SS. Peak alpha frequency is a neural marker of cognitive function across the autism spectrum. Eur J Neurosci. 2018;47:643–651.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/ejn.13645.Peak&link_type=DOI) 

25. 25.Cornew L, Roberts TPL, Blaskey L, Edgar JC. Resting-State Oscillatory Activity in Autism Spectrum Disorders. J Autism Dev Disord. 2012;42:1884–1894.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10803-011-1431-6&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22207057&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

26. 26.Ghanizadeh A, Sahraeizadeh A, Berk M. A Head-to-Head Comparison of Aripiprazole and Risperidone for Safety and Treating Autistic Disorders, a Randomized Double Blind Clinical Trial. Child Psychiatry Hum Dev. 2014;45:185–192.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10578-013-0390-x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23801256&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

27. 27.Politte LC, McDougle CJ. Atypical antipsychotics in the treatment of children and adolescents with pervasive developmental disorders. Psychopharmacology (Berl). 2014;231:1023–1036.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s00213-013-3068-y&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23552907&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

28. 28.Porter BE, Jacobson C. Report of a parent survey of cannabidiol-enriched cannabis use in pediatric treatment-resistant epilepsy. Epilepsy Behav EB. 2013;29:574–577.
    
    

29. 29.Devinsky O, Marsh E, Friedman D, Thiele E, Laux L, Sullivan J, et al. Cannabidiol in patients with treatment-resistant epilepsy: an open-label interventional trial. Lancet Neurol. 2016;15:270–278.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S1474-4422(15)00379-8&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26724101&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

30. 30.Devinsky O, Patel AD, Cross JH, Villanueva V, Wirrell EC, Privitera M, et al. Effect of Cannabidiol on Drop Seizures in the Lennox-Gastaut Syndrome. N Engl J Med. 2018;378:1888–1897.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJMoa1714631&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29768152&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

31. 31.Kuester G, Vergara K, Ahumada A, Gazmuri AM. Oral cannabis extracts as a promising treatment for the core symptoms of autism spectrum disorder: Preliminary experience in Chilean patients. J Neurol Sci. 2017;381:932–933.
    
    

32. 32.Thompson MD, Martin RC, Grayson LP, Ampah SB, Cutter G, Szaflarski JP, et al. Cognitive function and adaptive skills after a one-year trial of cannabidiol (CBD) in a pediatric sample with treatment-resistant epilepsy. Epilepsy Behav EB. 2020;111:107299.
    
    

33. 33.Aran A, Cassuto H, Lubotzky A, Wattad N, Hazan E. Brief Report: Cannabidiol-Rich Cannabis in Children with Autism Spectrum Disorder and Severe Behavioral Problems-A Retrospective Feasibility Study. J Autism Dev Disord. 2019;49:1284–1288.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10803-018-3808-2&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30382443&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

34. 34.Barchel D, Stolar O, De-Haan T, Ziv-Baran T, Saban N, Fuchs DO, et al. Oral Cannabidiol Use in Children With Autism Spectrum Disorder to Treat Related Symptoms and Co-morbidities. Front Pharmacol. 2019;9:1521.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3389/fphar.2018.01521&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30687090&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

35. 35.Pretzsch CM, Freyberg J, Voinescu B, Lythgoe D, Horder J, Mendez MA, et al. Effects of cannabidiol on brain excitation and inhibition systems; a randomised placebo-controlled single dose trial during magnetic resonance spectroscopy in adults with and without autism spectrum disorder. Neuropsychopharmacology. 2019;44:1398–1405.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41386-019-0333-8&link_type=DOI) 

36. 36.Isaacson JS, Scanziani M. How Inhibition Shapes Cortical Activity. Neuron. 2011;72:231–243.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuron.2011.09.027&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22017986&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000296224000007&link_type=ISI) 

37. 37.Salvatore SV, Lambert PM, Benz A, Rensing NR, Wong M, Zorumski CF, et al. Periodic and aperiodic changes to cortical EEG in response to pharmacological manipulation. J Neurophysiol. 2024. 7 February 2024. doi:10.1152/jn.00445.2023.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1152/jn.00445.2023&link_type=DOI) 

38. 38.Brake N, Duc F, Rokos A, Arseneau F, Shahiri S, Khadra A, et al. A neurophysiological basis for aperiodic EEG and the background spectral trend. Nat Commun. 2024;15:1514.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38374047&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

39. 39.Kaplan JS, Stella N, Catterall WA, Westenbroek RE. Cannabidiol attenuates seizures and social deficits in a mouse model of Dravet syndrome. Proc Natl Acad Sci U S A. 2017;114:11229–11234.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoicG5hcyI7czo1OiJyZXNpZCI7czoxMjoiMTE0LzQyLzExMjI5IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMDkvMzAvMjAyNC4wOS4yNy4yNDMxNDQ0OC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

40. 40.Kaplan JS, Wagner JK, Reid K, McGuinness F, Arvila S, Brooks M, et al. Cannabidiol Exposure During the Mouse Adolescent Period Is Without Harmful Behavioral Effects on Locomotor Activity, Anxiety, and Spatial Memory. Front Behav Neurosci. 2021;15:711639.
    
    

41. 41.Zhang Q, Melchert PW, Markowitz JS. Pharmacokinetic Variability of Oral Cannabidiol and Its Major Metabolites after Short-Term High-Dose Exposure in Healthy Subjects. Med Cannabis Cannabinoids. 2024;7:1–9.
    
    

42. 42.Tayo B, Taylor L, Sahebkar F, Morrison G. A Phase I, Open-Label, Parallel-Group, Single-Dose Trial of the Pharmacokinetics, Safety, and Tolerability of Cannabidiol in Subjects with Mild to Severe Renal Impairment. Clin Pharmacokinet. 2020;59:747–755.
    
    

43. 43.Millar SA, Stone NL, Yates AS, O’Sullivan SE. A Systematic Review on the Pharmacokinetics of Cannabidiol in Humans. Front Pharmacol. 2018;9:1365.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3389/fphar.2018.01365&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30534073&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

44. 44.Ujváry I, Hanuš L. Human Metabolites of Cannabidiol: A Review on Their Formation, Biological Activity, and Relevance in Therapy. Cannabis Cannabinoid Res. 2016;1:90–101.
    
    

45. 45.Brown L, Sherbenou RJ, Johnsen SK. Test of nonverbal intelligence: TONI-4. 2010.
    
    

46. 46.Bodfish JW, Symons FJ, Parker DE, Lewis MH. Varieties of repetitive behavior in autism: comparisons to mental retardation. J Autism Dev Disord. 2000;30:237–243.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1023/A:1005596502855&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=11055459&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000089674100006&link_type=ISI) 

47. 47.Dunn LM, Dunn DM. Peabody Picture Vocabulary Test--Fourth Edition. 2012.
    
    

48. 48.Martin, Nancy A., and  Rick Brownell. Expressive one-word picture vocabulary test-4 (EOWPVT-4). Academic Therapy Publications; 2011.
    
    

49. 49.Beery, K. E., Buktenica, N. A., & Beery, N. A. Beery-Buktenica Developmental Test of Visual-Motor Integration, Sixth Edition (Beery VMI). Pearson; 2010.
    
    

50. 50.Harris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D, et al. Array programming with NumPy. Nature. 2020;585:357–362.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41586-020-2649-2&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32939066&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

51. 51.Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17:261–272.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41592-019-0686-2&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32015543&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

52. 52.Jas M, Engemann DA, Bekhti Y, Raimondo F, Gramfort A. Autoreject: Automated artifact rejection for MEG and EEG data. NeuroImage. 2017;159:417–429.
    
    

53. 53.Gramfort A, Luessi M, Larson E, Engemann DA, Strohmeier D, Brodbeck C, et al. MNE software for processing MEG and EEG data. NeuroImage. 2014;86:446–460.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuroimage.2013.10.027&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24161808&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000330335300046&link_type=ISI) 

54. 54.Vallat R. Pingouin: statistics in Python. J Open Source Softw. 2018;3:1026.
    
    

55. 55.Donoghue T, Haller M, Peterson EJ, Varma P, Sebastian P, Gao R, et al. Parameterizing neural power spectra into periodic and aperiodic components. Nat Neurosci. 2020;23:1655–1665.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41593-020-00744-x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33230329&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

56. 56.Skipper S, Perktold J. Statsmodels: Econometric and Modeling with Python. 9th Python in Science Conference, Austin, 28 June-3 July, 2010, 57–61. Proc 9th Python Sci Conf.
    
    

57. 57.Pion-Tonachini L, Kreutz-Delgado K, Makeig S. ICLabel: An automated electroencephalographic independent component classifier, dataset, and website. NeuroImage. 2019;198:181–197.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/J.NEUROIMAGE.2019.05.026&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

58. 58.Welch P. The use of fast Fourier transform for the estimation of power spectra: A method based on time averaging over short, modified periodograms. IEEE Trans Audio Electroacoustics. 1967;15:70–73.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1109/TAU.1967.1161901&link_type=DOI) 

59. 59.Consroe P, Kennedy K, Schram K. Assay of plasma cannabidiol by capillary gas chromatography/ion trap mass spectroscopy following high-dose repeated daily oral administration in humans. Pharmacol Biochem Behav. 1991;40:517–522.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/0091-3057(91)90357-8&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=1666917&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1991GR80400012&link_type=ISI) 

60. 60.Levin AR, Varcin KJ, O’Leary HM, Tager-Flusberg H, Nelson CA. EEG power at 3 months in infants at high familial risk for autism. J Neurodev Disord. 2017;9:34.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/s11689-017-9214-9&link_type=DOI) 

61. 61.Bosl W, Tierney A, Tager-Flusberg H, Nelson C. EEG complexity as a biomarker for autism spectrum disorder risk. BMC Med. 2011;9:18.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/1741-7015-9-18&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21342500&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

62. 62.Bosl WJ, Tager-Flusberg H, Nelson CA. EEG Analytics for Early Detection of Autism Spectrum Disorder: A data-driven approach. Sci Rep. 2018;8:6828.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41598-018-24318-x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29717196&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F09%2F30%2F2024.09.27.24314448.atom) 

63. 63.Wilkinson CL, Levin AR, Gabard-Durnam LJ, Tager-Flusberg H, Nelson CA. Reduced frontal gamma power at 24 months is associated with better expressive language in toddlers at risk for autism. Autism Res. 2019;12:1211–1224.
    
    

64. 64.Li Q, Weiland RF, Konvalinka I, Mansvelder HD, Andersen TS, Smit DJA, et al. Intellectually able adults with autism spectrum disorder show typical resting-state EEG activity. Sci Rep. 2022;12:19016.
    
    

65. 65.Shuffrey LC, Pini N, Potter M, Springer P, Lucchini M, Rayport Y, et al. Aperiodic Electrophysiological Activity in Preterm Infants is Linked to Subsequent Autism Risk. Dev Psychobiol. 2022;64:e22271.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/dev.22271&link_type=DOI) 

66. 66.Hill AT, Clark GM, Bigelow FJ, Lum JAG, Enticott PG. Periodic and aperiodic neural activity displays age-dependent changes across early-to-middle childhood. Dev Cogn Neurosci. 2022;54:101076.
    
    

67. 67.Favaro J, Colombo MA, Mikulan E, Sartori S, Nosadini M, Pelizza MF, et al. The maturation of aperiodic EEG activity across development reveals a progressive differentiation of wakefulness from sleep. NeuroImage. 2023;277:120264.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuroimage.2023.120264&link_type=DOI)

 [1]: /embed/graphic-1.gif
 [2]: /embed/graphic-2.gif