Individuals with this condition often face a wide range of debilitating comorbidities, including delayed psychomotor development, abnormal gait, hyperactivity, attentional difficulties, autism spectrum features, disrupted sleep, anxiety, depression, language delays, and profound cognitive impairments, all of which significantly diminish quality of life [55]. Current therapeutic approaches typically involve polytherapy with ASMs, such as valproate and clobazam [56]. Nevertheless, seizure control remains inadequate for many patients, and these treatments frequently lead to serious neurological and psychiatric side effects, including increased anxiety, depressive symptoms, and cognitive deficits such as memory loss [57].

ASs in children and teenagers were in the past considered relatively benign because of their nonconvulsive nature and high remittance rate in early adulthood. However, recent studies on large cohorts of drug-naïve CAE patients have now conclusively demonstrated that 30% of children with AS are pharmacoresistant [82,85], and 60% suffer from various neuropsychiatric comorbidities. These include anxiety, depression, attention-deficit/hyperactivity disorder (ADHD), and learning deficits that may precede epilepsy diagnosis, persist, and even be aggravated after full pharmacological control of the seizures (see [69] and references within).

For instance, 61% of children with CAE had a psychiatric diagnosis, with 43% experiencing linguistic difficulties and 25% exhibiting subtle cognitive deficits. The severity of these comorbidities was associated with factors such as epilepsy duration, seizure frequency, and antiepileptic drug treatment. However, only 23% of the affected children were receiving interventions for these issues [86]. These findings underscore the importance of the early identification and management of neuropsychiatric comorbidities in children with CAE, highlighting the need for holistic treatments that address both seizures and associated behavioral and cognitive impairments [69].

CBD medication, Epidiolex^®, has been approved for patients aged 1 year and older with DS or LGS and TSC, in conjunction with clobazam [18]. CBD approval was based on six randomized controlled trials (RCTs) demonstrating a 40 to 50% reduction in seizure frequency among Epidiolex-treated patients [56,142,143,144,145]. A recent retrospective multicenter chart review in Germany observed a reduction in seizure frequency and sustained treatment retention for up to 12 months across different age groups, among patients with severe, treatment-refractory LGS or DS receiving adjunctive CBD and clobazam simultaneously [146]. Several meta-analyses have been published on CBD use in DRE among the adult [147] and pediatric population [148,149,150]. Adding CBD benefits most children with pharmacoresistant epilepsy. While 20 mg/kg/day offers better seizure control than 10 mg/kg/day, the lower dose remains effective and a viable treatment option [148,149,150]. Common adverse effects of CBD include drowsiness, lethargy, diarrhea, loss of appetite, and weight loss [151,152]. DS is a channelopathy mainly linked to SCN1A and SCN1B mutations, with additional contributions from potassium and calcium channel gene variants. These disrupt excitatory/inhibitory balance through altered channel function [153]. Thus, the modulation of these channels may underlie CBD’s therapeutic mechanism, although current support comes primarily from preclinical studies [153]. CBD may also exert therapeutic effects in DS and LGS by modulating dysregulated components of the ECS, although ECS alterations in these conditions remain poorly characterized. For instance, Rubio et al. [154] reported an increased expression of CB2Rs in lymphocytes derived from DS patients, potentially indicating a compensatory or pathological upregulation in peripheral immune cells. Notably, no significant changes were detected in plasma levels of key endocannabinoid, such as AEA or 2-AG, suggesting that peripheral eCB levels may not accurately represent central ECS dysfunction in these syndromes [154].

CBD has been shown to attenuate hyperthermia-induced seizures in Scn1a+/− mice, which closely mirrors the clinical features of DS [54,155], and reduce the frequency and severity of spontaneous seizures [54]. CBD’s anticonvulsant effects were initially thought to stem from pharmacokinetic and pharmacodynamic interactions with clobazam [56]. However, the preclinical evidence suggests that CBD’s efficacy goes beyond elevated clobazam levels, likely involving the positive modulation of GABA_ARs and multimodal engagement of anticonvulsant pathways in DS [155].

Several other cannabinoids have shown efficacy against thermal-induced seizures in Scn1a+/− mice. Treatment with phytocannabinoids, such as CBCA and CBCVA [156] CBGA, CBDVA, and CBGA [157], as well as the cannabinoid-like compound magnolol from non-cannabis plant Magnolia officinalis [158], reduced spontaneous seizures and improved survival in the Scn1a+/− mouse model of DS.

Impairments in the ECS have been proved in the same Scn1a+/− mouse model [159,160]. Anderson et al. (2022) [161] identified Cnr1, the gene encoding CB1R, as a genetic modifier of epilepsy in Scn1a+/− mice, with a deficiency of eCBs serving the pathological background. Reduced hippocampal CB1R expression was observed in this model, while cortical CB1 levels and brain concentrations of major endocannabinoids (AEA and 2-AG) remained unchanged (Anderson et al., 2022). However, lesser-studied monoacylglycerols, such as 2-linoleoylglycerol (2-LG) and 1-linoleoylglycerol (1-LG), were significantly elevated following hyperthermia-induced seizures in Scn1a+/− mice [160]. The delayed timing of sampling (5 min post-seizure) in these studies [160,162] may have limited the detection of transient changes in eCB levels. For instance, in vivo two-photon imaging demonstrated significant increases in 2-AG concentrations in the hippocampal CA1 region contextually to electrically induced seizures, while AEA levels remained unchanged [163]. In support of a deficiency of ECS in DS, the positive allosteric modulation (PAM) of CB1 receptor GAT229 and enhancement of brain 2-AG concentrations using ABX-1431, a MAGL inhibitor, both produced significant anticonvulsant effects against hyperthermia-induced seizures in Scn1a+/− mice [161].

A novel DS model involving heterozygous, conditional, knock-in mice with a missense mutation (A1783V) in the SCN1A gene, expressed exclusively in CNS neurons (Syn-Cre/SCN1AWT/A1783V), revealed elevated CB2R levels in the hippocampus, particularly in the dentate gyrus [164]. This upregulation may represent an endogenous protective response aimed at reducing neuroinflammation in DS. Additionally, this model exhibited other alterations in the ECS, including the downregulation of CB1Rs in the cerebellum and hippocampus, changes that may reflect impaired synaptic function and correlate with the motor and memory deficits observed in these mice. Notably, reductions in eCB-inactivating enzymes were detected in specific brain regions: both MAGL and FAAH in the cerebellum, MAGL alone in the prefrontal cortex, and FAAH alone in the hippocampus. These enzymatic changes suggest region-specific elevations in endocannabinoid levels, particularly 2-AG, given the prominent downregulation of MAGL [164]. These findings underscore the potential of targeting the ECS for next-generation anticonvulsant therapies. However, the therapeutic utility of MAGL inhibition may be constrained by its narrow therapeutic range, as subchronic treatment with high doses of ABX-1431 exacerbated spontaneous seizures. This highlights the need for dose optimization and further investigation into the long-term safety and efficacy of this approach [161].

Animal studies on LGS are instead limited, but a promising model has recently emerged: the Gabrb3+/D120N mouse, a heterozygous knock-in for the β3 subunit of the GABA_AR [165]. Acute administration of magnolol significantly reduced the number and duration of atypical ASs in the Gabrb3+/D120N mouse model, consistent with an antiseizure effect [158]. In vitro, magnolol inhibited all T-type calcium channel subtypes, similar to other phytocannabinoids [94], but showed no activation of CB1 or CB2Rs [158].

In conclusion, cannabinoid efficacy in DS extends beyond interactions with clobazam, involving the direct modulation of GABA_ARs and multiple anticonvulsant mechanisms. Additionally, PAM and elevating eCBs show antiseizure effects through CB1 and CB2R modulation and T-type calcium channel inhibition, underscoring the therapeutic potential of targeting the broader endocannabinoid system.

No direct data are available on the involvement of alterations in ECS in epilepsy-associated comorbidities. However, indirect evidence comes from studies investigating the long-term effects (followed for up to 2 years) of CBD treatment in patients with DRE, where CBD improved mood [166,167] without impairing cognition [168]. Furthermore, in Scn1a−/− and Scn1a+/− mouse models of DS, CBD administered at anticonvulsant doses improved their welfare by reducing pain and enhancing general health [54,169]. In heterozygous mice, it also alleviated several complex behavioral comorbidities, including motor dysfunction, anxiety-like behavior (assessed in the elevated plus maze), and depression-like behavior (evaluated with the sucrose preference test) [169]. Furthermore, CBD improved spatial learning and memory in the eight-arm radial maze [169] as well as deficits in autism-like social interactions [54,169].

CBD exerts dual actions in the hippocampus of DS mice, contributing to the restoration of excitatory–inhibitory balance. First, it lowers the rheobase of interneurons, thereby enhancing their excitability and addressing deficits in GABAergic firing [50,170]. Second, CBD acts through the GPR55 receptor to enhance inhibitory synaptic transmission onto dentate gyrus granule cells (DGCs), leading to a reduction in their spontaneous firing. Additionally, CBD directly reduces DGC excitability, which may underlie its capacity to decrease both seizure frequency and severity [54].

These findings underscore the complex role of the ECS system in DS and LGS, highlighting its potential as a therapeutic target. With a favorable side-effect profile, emerging evidence supports CBD’s efficacy in treatment-resistant epilepsies and their associated comorbidities. However, further research is required to better understand the underlying mechanisms and the broader therapeutic implications (Figure 1).

As far as CAE is concerned, the evidence regarding the involvement of the ECS in human ASs remains conflicting. However, a recent prospective pilot study [191] demonstrated the effect of pharmaceutical-grade CBD on a small cohort of 14 patients with typical CAE. CBD treatment (8–20 mg/kg/day) administered over 90 days appeared ineffective for typical absence seizures and may have even exacerbate them, as several patients exhibited a clinically significant increase in SWDs. Despite this, CBD was generally well tolerated, with only mild side effects reported. However, an anti-absence effect of CBD was observed in 5 out of the 14 (35%) patients enrolled, with 3 patients receiving CBD in combination with ETX and 2 patients on CBD monotherapy [191]. This anti-absence effect was also reported in the pooled analysis of the CBD Expanded Access Program (EAP), which provided treatment to 892 patients with DRE for up to 33 months [192], as well as in 4 out of 5 patients enrolled in a separate EAP conducted in Massachusetts involving 50 patients treated for up to 60 months [193]. However, definitive conclusions from these EAP studies cannot be drawn, as they did not distinguish between typical and atypical CAE, and changes in seizure frequency were not confirmed by EEG monitoring [192].

Two clinical trials (ClinicalTrials.gov Identifiers: NCT03355300 and NCT03336242) were initiated in 2017 to evaluate the antiepileptic efficacy of a CBD oral solution in pediatric patients with treatment-resistant childhood absence seizures. However, both trials were terminated before completion, and no conclusive results were obtained [194]. From the limited information available, it appears that an effect of CBD was observed only at low concentrations (approximately 20 mg/kg/day) in the small cohort of patients enrolled [194]. No studies have specifically investigated the effect of ∆9-THC on absence seizures (ASs), except for a small prospective study on childhood epilepsy that included three patients with idiopathic generalized epilepsy treated with oral cannabis extracts containing both CBD and ∆9-THC [32]. The study yielded inconclusive results regarding whether the combination of ∆9-THC and CBD offers better seizure control, or at what ratio, compared to CBD alone, but it highlighted that the risk–benefit profile may be less favorable than that of pharmaceutical-grade CBD [32]. The major outcomes are summarized in Table 3.

Although the experimental evidence remains inconsistent, more data are available on the role of cannabinoids in ASs from animal experimental models of the disease compared to other pediatric epilepsies (Table 3). One of the earliest investigations on ∆⁹-THC and CBD SWDs photically evoked showed a lack of effect by CBD and conversely pro-absence activity by ∆⁹-THC [195]. Of note, by analyzing cannabinoids in other animal models of epilepsy, it was proposed that cannabinoids were useful against convulsive seizures but not CAE [195]. For the last few decades, the cannabis research has been hindered by law restrictions and its criminalization in the USA Controlled Substances Act of 1970 [204] determining that cannabis therapeutic potentials were forgotten until recently. In 2010, a study on epileptic WAG/Rij rats [198] showed a complex biphasic modulation of SWDs by a synthetic CB1/2R agonist, WIN 55,212-2 [124]. The authors showed that the acute general administration of high doses (3–12 mg/kg, s.c.) of cannabinoids biphasically modulated SWDs. The doses of 6 and 12 mg/kg decreased the incidence of SWDs in the first 2 and 3 h, respectively, but produced an elongation of the seizures after 3 h, with single SWDs reaching up to 100 s or more at the 6 h mark. The CB1R antagonist/inverse agonist AM251 was ineffective at 6 and 12 mg/kg, but the latter dose blocked the early inhibition in SWDs, the late increase in mean SWD duration, and produced at the 4th and 5th hours an increase in the incidence of SWDs. These data are puzzling, especially since incredibly high doses of cannabinoids were used. WIN 55,212-2 loses affinity for CB1Rs at high doses, and an inhibition of motor activity was shown, which might have caused the early decrease in SWDs. A 12 mg/kg dose of WIN 55,212-2 completely suppressed SWDs in a WAG/Rij rat that developed generalized convulsive seizures after cannabinoid treatment [205], suggesting that CB1R activation may facilitate the transition from nonconvulsive to generalized convulsive seizures.

Nevertheless, more recent investigations did not replicate the anti-absence effect of WIN 55,212-2 [199,205]. Indeed, both acute and subchronic administrations of 6 mg/kg WIN 55,212-2 did not modify the incidence of SWDs, but similarly to the previous study [198] elongated the duration of the SWDs, likely by inhibiting NRT neurons known to be involved in the stopping mechanism of SWDs [206].

Sysoeva et al. (2023) explored the effect of cannabinoids on absence epileptic networks, revealing that WIN55,212-2 modulated the network strength in the frontal and hippocampal areas. It decreased SWD frequency but increased its duration, indicating a complex interaction with epileptic networks and neuronal coupling [205].

Recently, Howland and colleagues reported a contrasting effect of agonist PAMs (ago-PAMs) for CB1Rs [201,203] and the phytocannabinoids ∆⁹-THC and CBD [196] in GAERS. Systemic administrations of ago-PAMs GAT229 and GAT211 reduced SWD incidence and duration by 50% and 40%, respectively [201], while the more powerful ago-PAM GAT591 and GAT593 only decreased the total time spent in seizures by 36% and 34%, respectively, without affecting SWD incidence, average duration, or oscillatory frequency [203]. The anti-absence effects of CB1R PAMs appear CB1R-dependent, as SR141716A, a CB1R antagonist/inverse agonist, blocked GAT211’s effect [201]. In contrast, purified ∆⁹-THC (3–10 mg/kg) exerted a pro-absence effect on male and female GAERS, increasing SWD incidence, duration, and total seizure time, while reducing SWD frequency [196]. Conversely, purified CBD (30–100 mg/kg) did not alter SWD incidence, but reduced seizure duration by shortening the average SWD length. Inhalation of smoke from a high-∆9-THC cannabis strain increased the total and average SWD duration, while leaving the frequency unchanged. These effects were likely not CB1R-dependent, as they were not blocked by SR141716A. On the other hand, exposure to smoke from a CBD-rich cannabis strain had no effect on any SWDs, likely because the CBD plasma concentrations did not reach effective levels [196] (Table 3).

In WAG/Rij rats, both intracerebroventricular (i.c.v.) and intraperitoneal (i.p.) administrations of AEA and its analog N-palmitoylethanolamine (PEA) reduced SWD incidence and duration. PEA’s effects were blocked by pre-treatment with SR141716 and the PPAR-α antagonist GW6471, whereas AEA’s i.c.v. anti-absence effects were blocked only by SR141716A [202]. Further studies showed that AEA and the CB1R agonist WIN55,212-2, when injected into the thalamic nuclei (NRT, VB), the somatosensory cortex (S1po), or administered i.c.v., also suppressed SWDs. Interestingly, SR141716A alone did not induce a significant pro-absence effect [200].

In GAERS, the infusion of the ago-PAM GAT229 into the motor cortex, a region not directly involved in SWD generation, reduced SWD incidence, average duration, and total duration; this effect was prevented by systemic SR141716A, which alone had no effect [201] (Table 3).

eCB concentrations in specific brain regions were assessed in 2- and 6-month-old WAG/Rij rats, with ACE rats and normal Wistars as controls [202]. In 6-month-old WAG/Rij rats, the amygdala showed reduced levels of AEA, 2-AG, and PEA, while the cortex exhibited higher levels of PEA, and no significant changes were detected in the thalamus compared to ACE rats [202]. In contrast, adult, symptomatic GAERS exhibited increased 2-AG levels in the cortex, hippocampus, and cerebellum, but not in the thalamus, and elevated AEA only in the cerebellum [201].

In situ hybridization in age-matched WAG/Rij and ACI rats revealed reduced CB1R mRNA expression in the cortex, NRT, hippocampus, and caudate/putamen of WAG/Rij rats, with no detectable signal in the VB [198]. However, Western blot (WB) analysis indicated decreased CB1R protein levels only in the NRT and VB [198]. In GAERS, CB1R expression was examined in the cortex, thalamus, hippocampus, and cerebellum, showing reduced protein levels in the cortex and hippocampus [201]. This cortical downregulation was further confirmed by [³H]SR141716A binding analysis via liquid scintillation spectrometry [201].

The two major eCBs, AEA and 2-AG, are both synthesized postsynaptically but act on presynaptic CB1Rs to inhibit neurotransmitter release [207]. Given the adverse effects observed with direct-acting CB1R agonists and antagonists in preclinical rodent models of seizures, several studies have investigated whether enhancing endogenous cannabinoid tone might reduce seizure occurrence and severity [14]. This type of pharmacological intervention has not yet been investigated in CAE; however, our preliminary studies using PF-04457845, a novel FAAH inhibitor [208], have shown a promising anti-absence effect in GAERS (unpublished observations).

Alterations in the ECS have been implicated in anxiety and mood disorders, and the modulation of ECS components may provide therapeutic benefits [184]. Therefore, the neuropsychiatric comorbidities observed in CAE [69] may be responsive to cannabinoid modulators. Nevertheless, studies on GAERS may have limited the ability to provide further insights, as recent evidence shows that NEC rats exhibit higher anxiety-like behavior and neophobia compared to GAERS [209,210,211], contrary to previous assumptions [69]. Additionally, the effects of cannabinoid receptor agonists, such as WIN 55,212-2, appear to be strain-dependent: this compound reduces anxiety in NEC rats but not in GAERS, where it instead produces sedative effects [209]. These behavioral outcomes were paralleled by monoaminergic alterations. GAERS exhibited lower baseline 5-HT levels in the hippocampus and substantia nigra, as well as reduced NA levels in the entopeduncular nucleus, potentially contributing to their hypermotility and reduced anxiety phenotype. Moreover, ECS activation modulated monoamine levels, further implicating ECS involvement in the observed behavioral differences [209]. However, other evidence suggests that the effect of cannabinoid treatment may not be uniform and could be dependent on the behavioral test used. For example, WIN 55,212-2 induced hyperlocomotion in GAERS and did not improve anxiety in the elevated plus maze in NEC rats and GAERS [212] both of which exhibited comparable levels of anxiety-like behavior [212,213]. Consistently, Roebuck et al. [213] also found that CB1R PAM GAT211, while ineffective on anxiety-like behavior in elevated plus-maze and open-field tests or locomotory activity, showed increase sociability and reduced the elevated startle response, particularly in female GAERS [213].

Together, these findings suggest that the GAERS model does not fully replicate the anxiety observed in CAE, and cannabinoids do not significantly alter the emotional state of GAERS. This limits the ability to draw definitive conclusions about the efficacy of cannabinoids in treating comorbid anxiety in children with AS. While the anxiolytic effects of cannabinoids have been consistently observed in normal rats [184], the apparent unresponsiveness in GAERS may be due to impairments in ECS signaling, complicating interpretation. However, the observed effects on sociability and startle response in female GAERS [213] are intriguing, suggesting a potential therapeutic role for cannabinoids in addressing other neuropsychiatric comorbidities. These findings highlight the importance of investigating sex differences in ECS modulation. Further studies are needed to clarify the mechanisms and therapeutic windows of ECS-targeted interventions for both seizures and associated psychiatric symptoms. Critically, the development or selection of more appropriate animal models that better reflect both the seizure phenotype and comorbidities is essential.

Comparing GAERS to WAG/Rij rats reveals a different form of cannabinoid control of SWDs [198,201,205], both in terms of their generation and termination, further increasing the number of differences already described between these two epileptic strains [214].

CB1Rs at GABAergic synapses act as strategically placed control points for the activity-dependent regulation of dynamically changing normal and pathological oscillatory network activity [215]. The proepileptic effects of CB1Rs may also involve other brain regions, such as the output structures of the basal ganglia, particularly the substantia nigra pars reticulata (SNr), as shown in a study on normal mice [197]. CB1R activation induced thalamocortical high-voltage spindles (HVSs) (putative SWDs) by selectively increasing the activity in the nigro-thalamic pathway [197]. This may lead to increased GABA release in the thalamus and enhanced GABA_AR receptor-mediated tonic inhibition, contributing to ASs [72], highlighting a potential risk of CAE induced by cannabis abuse (Figure 3).

Both NRT and TC VB neurons can synthesize and release 2-AG; moderate DAGLα expression in the NRT and VB of adult Wistar rats was observed [216]. Nevertheless, a synapse target-dependent effect was observed in the thalamus, as DSI occurs only at intra-NRT synapses, not to NRT to VB [217]. This 2-AG-mediated DSI may increase NRT neuronal discharges, reducing burst firing due to membrane depolarization [218], or Ca²⁺ influx during a low-threshold spike (LTS) may elevate intracellular 2-AG via DAGLα, depressing LTS burst spikes itself [219]. Summing up, eCBs modulate synaptic strength at specific inhibitory thalamic pathways, dynamically influencing thalamic–cortical synchrony. NRT neurons inhibit either VB neurons, promoting rhythmic oscillations, or neighboring TRN neurons, limiting synchrony. Therefore, the emergence of ASs may result from reduced eCB signaling in the NRT, supported by a reduction of both CB1R mRNA and protein expression [198] (but not in GAERS, in which no thalamic ECS alterations were detected [201]), which weakens intra-TRN inhibition and enhances thalamic synchrony [217]. The long-lasting increase in long SWDs observed after acute and subchronic WIN administrations in adult WAG/Rij rats is likely driven by CB1R activation and a consequent reduction in GABA availability within the NRT [199]. Another brain region implicated in the ictogenesis of ASs is the peri-oral region of the somatosensory cortex (poS1) [66,80], where the dysregulation of the ECS has been observed. In GAERS, reduced CB1R levels in the cortex likely create a permissive environment for increased excitability and seizure propagation [201], whereas no changes in CB1R mRNA expression were detected in WAG/Rij rats [198]. These ECS alterations may contribute to seizure onset during development and their exacerbation during adolescence [201]. The hippocampus (HPC) shows decreased CB1R protein [201] and mRNA expression levels [198], whereas 2-AG levels are elevated in GAERS compared to the controls [201]. These findings are relevant considering the recent evidence of HPC involvement in CAE and the marked decrease in HPC frontal-cortex unidirectional coupling exclusively during SWDs, which was reversed by WIN55,212-2 treatment that effectively reduced ASs [205]. These changes in the ECS in the HPC may help explain the cognitive and behavioral alterations seen in GAERS [210,211,220,221,222,223] and patients with CAE [69].

The effects of CB1R-dependent activation, whether through agonists or ago-PAMs, or CB1R-independent effects mediated by CBD administration on modulating SWDs remain unclear. The pro-absence effects observed after acute ∆9-THC exposure in GAERS [196], and following both acute and subchronic administrations of WIN55,212-2 in WAG/Rij rats [199] limit the therapeutic potential of CB1R orthosteric agonists. In contrast, CB1R ago-PAMs appear more promising, as they do not lead to the overactivation of CB1Rs or associated pro-absence, such as elongated SWD duration [203], nor do they produce psychotropic effects. Since ago-PAMs require the presence of eCBs to exert their action, they provide an intrinsic ceiling effect that reduces the risk of overstimulation.

Although ago-PAM anti-absence effects are blocked by CB1R antagonism, the eCB may also have mixed effects. These include the modulation of TRPV1 channels, inhibition of L- and T-type Ca²⁺ channels [224,225], or direct activation of extrasynaptic GABA_ARs [226,227] by binding to their β₂ subunit [228]. Synaptic GABA_AR-mediated inhibition has also been reported [227]. Notably, enhanced tonic GABA_AR inhibition via the δ-containing extrasynaptic GABA_A receptor in VB TC neurons is both necessary and sufficient for the expression of ASs in different rodent models, while phasic GABA_AR inhibition remains unaltered [72]. These observations support the hypothesis of a potential hyper-eCB signal contribution to the pathophysiology of ASs and suggest limitations in the use of cannabinoids, given their contrasting effects on GABA_ARs and other targets, for this type of childhood epilepsy. However, further research is needed to explore ECS involvement in both seizures and neurobehavioral comorbidities in CAE (Figure 2 and Figure 3).