The impact of recreational cannabis use on neuropsychological function in epilepsy

Highlights • Recreational cannabis use is reported more commonly in males than females.• Cannabis use is associated with lower baseline intellectual reserves.• Cannabis use in people with epilepsy amplifies deficits in new learning.• Enhanced susceptibility of recall to distraction is also evident in this group.


Introduction
Cannabis is the most widely used recreational drug in the world and tops the table of 'the most used drug' in countries across the five continents [1].However, whilst widespread, only eight countries have legalised its use for recreational purposes (Canada, Georgia, Luxembourg, Malta, Mexico, South Africa, Thailand, and Uruguay).Recreational use has also been legalised in some states in the USA.Other legislative approaches to the drug have been to decriminalise possession and users are not actively pursued by criminal law enforcement.Whilst it is still illegal to possess and use cannabis, civil penalties rather than criminal convictions will apply.This approach is more common in Europe, with Germany, Italy, Netherlands, Spain, Portugal and Switzerland adopting this approach.
Cannabis is classified as a class B drug in the UK.Penalties for possession include up to 5 years in prison, an unlimited fine or both [2], Nevertheless, cannabis is the most widely available and commonly used recreational drug in the UK with 7.6 % of adults reporting that they regularly or sometimes use the drug [3].
These widely disparate percentages are likely to reflect a number of factors including the legal status of the drug in the country, and associated ease of availability.Methodological differences in recording cannabis use will also impact reported rates (for example some studies ask about regular use, whilst others ask about use within the past year).There are also likely to be cultural differences in the rate of reporting what is an illegal activity for many, in a clinical or research setting.
A recent review by Li et al. (2023) [20] presents a comprehensive overview of 'non-medical cannabis' use in people with epilepsy.In their review of the literature, the median number of adults and children with epilepsy using cannabis was 24.5 %.Factors associated with nonmedical cannabis use included male sex, younger adult age and lower education status, with cannabis use contributing to stigma and higher levels of depression.The reviewers suggest the evidence-base is sparse and heterogeneous with mostly small cross-sectional studies, and there are significant gaps in the literature including determining the neuropsychological effects of recreational cannabis use in people with epilepsy.
In non-patient populations, chronic cannabis use can lead to cognitive impairment including detrimental effects on learning, memory, executive function, global cognition, attention, and decision-making [21][22][23][24][25][26][27].The most commonly reported cognitive impact is in the memory domain with encoding, storage, manipulation, and retrieval of verbal information affected [28][29][30].Some studies report neutral or mildly beneficial effects of medicinal cannabidiol on neuropsychological function in people with epilepsy [31][32][33] which may be secondary to improved seizure control.These inconsistent findings may reflect, at least in some part, the significant methodological challenges of studies in this field.In non-experimental settings it is difficult to control for the heterogeneity of recreational cannabis use with respect to duration and frequency of use, potency, age of onset and periods of abstinence [25,30,[34][35][36].
THC is the psychoactive component of cannabis acting through the endocannabinoid system which regulates mood, learning, and memory [37] and has a high content within current strains of cannabis in the UK [38].Strains with higher Δ 9 -tetrahydrocannabinol (THC) content are thought to result in greater cognitive impairment and mood disturbance [39].
Little is known about the additive impact of recreational cannabis use on neuropsychological function in adults with epilepsy, with this population already at-risk of memory and cognitive difficulties due to the essential comorbidities of the condition.The current study aimed to examine the impact of cannabis use on cognitive function in this group.

Methods
The records of a consecutive series of 800 patients who underwent a neuropsychological assessment in the epilepsy service at the University College London Hospital between 2019 and 2022 were examined for references to recreational cannabis use.All patients who attend for a neuropsychological assessment in our department are routinely asked about recreational drug use in their clinical interview, prior to their formal assessment with standardised tests.Information regarding recreational drug use is also recorded in the medical and neuropsychiatric assessments they undergo when assessed by the multidisciplinary team.We identified seventy patients who admitted using cannabis for recreational purposes, either in the past or at the time of the assessment, comprising 8.75 % of our referrals.Forty-seven of these patients reported current use of cannabis.Three of the patients who reported current cannabis use had an exclusive diagnosis of non-epileptic attack disorder and were excluded from the analyses of cognitive performance.Two patients had undergone surgery prior to the assessment and were also excluded from these analyses to avoid the confound of the impact of surgery on cognitive performance.Following these exclusions, our search strategy identified 42 patients (27 males, 15 females) with epilepsy who reported current recreational cannabis use at the time of their neuropsychological assessment.

Comparator group
The neuropsychological test scores of the cannabis group were compared to those of a consecutive series of 254 (144 males, 110 females) age-matched, non cannabis using people with epilepsy who underwent an assessment over the same time period.Since none of the cannabis group had an intellectual disability, patients with an intellectual disability were excluded from the comparator group.Intellectual disability was established as part of the clinical neuropsychological assessment which consisted of medical and educational review and the completion of formal tests of cognitive function.As with the cannabis group, patients who were referred for a postoperative assessment were excluded.

Neuropsychological measures
The neuropsychological assessment included a standardised measure of pre-morbid ability, the National Adult Reading Test [40] and tests of intellectual, memory, language and executive function, in addition to questionnaire data.The scaled scores from the Digit Span and Coding tests from the Wechsler Adult Intelligence Scale (WAIS-IV) [41] were used as measures of working memory and processing speed respectively.Measures of memory included the List 1-5 score from the BMIPB II [42] list learning task (a measure of learning over 5 trials) and the List A6 score (a measure of retention following distraction).The design learning score from the BMIPB was used as our measure of visual memory.The total number of items correctly named on the Graded Naming Test was used as a measure of language function.The number of words generated in 60 s beginning with 'S' and number of animals generated were our measures of phonetic and sematic fluency.Self-report measures included the Hospital Anxiety and Depression Scale (HADS) [43] and the Subjective Memory Questionnaire (SMQ).All of these measures have been described in detail previously [44].
The tests that comprise a clinical neuropsychological assessment are determined on an individual basis and are selected in response to the specific referral received.Not all patients undertake all tasks.This is reflected in the numbers in each analyses.

Ethical Approval
All data was fully anonymised prior to the analyses in order to conform to ethical approvals granted for the analysis of routinely collected clinical data in an audit of recreational cannabis use in patients referred for a neuropsychological assessment.(Hospital Board Approval: 20-202223-SE).

Data availability statement
The clinical data analysed in the current study are not publicly available due to patient privacy and restricted access, but further information about the database is available from the corresponding author on reasonable request.

Statistical analyses
All analyses were conducted using SPSS v27.Differences between the cannabis group (n = 42) and the comparator group on the neuropsychological measures were explored with independent t tests.Linear regression models were used to examine the relative contributions of cannabis status and mood to measures of cognitive function, where preliminary analyses indicated significant group differences in function.

Demographic & clinical characteristics of cannabis users
8.75 % of the consecutive series of 800 patients who attended our department for a neuropsychological assessment reported past or present recreational cannabis use.Approximately one in 17 patients (5.9 %) reported current cannabis use at the time of the assessment (n = 47).In the consecutive series of 800 patients, cannabis use (current/past/ never) was not associated with epilepsy type (Focal, Generalised, Unclassified, Non Epileptic Attack Disorder, Dual Diagnosis) Chi square = 14.9, df 8, p > 0.05.
Typically more females are referred for an assessment in our department than males (56.3 %).However, in our comparator group, which was age-matched to the cannabis group and which excluded surgical patients, males represented 56 % of the group.In the cannabis group, males represented 64 % of the group (chi square 0.8, df 1, p > 0.05).
Males were approximately 1.8 times more likely to use cannabis compared to females although this just failed to reach statistical significance (Males n = 27; Females n = 15, χ 2 = 3.4, df 1, p = 0.06).
Patients in our series who reported current cannabis use were also significantly younger than past and non-cannabis users (f (2,521) = 8.4, p < 0.001) with a small-moderate effect size (η2 = 03).

Neuropsychological test scores
The clinical and demographic characteristics of the cannabis group and the age matched comparator group are presented in Table 2. Cannabis use was associated with lower intellectual reserve (NART IQ) reduced verbal learning (List Learning) and enhanced susceptibility to distraction on a subsequent recall task (List A6).The cannabis group also reported significantly higher levels of anxiety and depression on the HADS.No significant differences between the cannabis group and the comparator group were evident on tests of working memory (Digit Span), processing speed (Coding), immediate or delayed prose recall, visual learning, naming or verbal fluency scores.See Table 3.
Since elevated levels of anxiety and low mood can contribute to poor scores on neuropsychological tests we examined the relative contributions of cannabis use and the HADS Anxiety and HADS depression scores to the List learning and List Recall scores via linear regression.

List learning: Regression analyses
A univariate linear regression analysis was conducted to examine the relationship between verbal learning as the dependent variable (List Learning A1-5) and the independent variables: "HADS anxiety score," "HADS depression Score", "premorbid level of function" (NART IQ) and "cannabis use group" (Yes/No).

Recall following distraction: Regression analyses
A second univariate linear regression analysis was performed to investigate the relationship between the dependent variable "List A6″ and the independent variables "HADS Anxiety," "HADS Depression," "NART IQ," and cannabis use group (Yes/No).The results of the analysis are presented in Table 5.The corrected model, including the independent variables "HADS Anxiety," "HADS Depression," "NART IQ," and "Cannabis Group," showed statistical significance, F(4, 198) = 5.667, p < 0.001.This suggests that the model as a whole had a significant impact on the dependent variable "List A6."As with the predictors of verbal learning (List 1-5) "NART IQ" (F = 11.934,p < 0.001) and "Cannabis group" (F = 5.316, p = 0.022) made statistically significant contributions to the model, whereas self-reported mood scores on the HADS did not: "HADS Anxiety" (F = 0.012, p = 0.912); "HADS Depression" (F = 0.294, p = 0.588).Whilst the overall model accounted for a statistically significant portion of the variance in "List A6″ (R 2 = 0.103, Adjusted R 2 = 0.085), the portion was small.

Discussion
In our tertiary referral centre, 8.75 % of patients referred for a neuropsychological assessment reported past or present cannabis use, with 5.9 % reporting current use.Cannabis use was more commonly reported by males than females and the cannabis users were typically younger than the wider cohort.The cannabis group displayed lower intellectual reserve, reduced verbal learning abilities, and increased susceptibility to distraction on a verbal recall task compared to the matched comparator group, in a pattern consistent with that reported in a recent comprehensive review of cannabis use in this population [20].The patients in the cannabis group reported higher levels of anxiety and depression on the HADS.Linear regression analyses revealed that whilst premorbid level of intellectual function was a significant predictor of performance, cannabis use was also independently associated with poorer verbal learning and recall following distraction, even after controlling for elevated levels of anxiety and depression.
Our results are consistent with previous studies that have shown associations between cannabis use and cognitive impairments in healthy individuals, particularly in the verbal learning and memory domains [25,28,30].Whilst several studies have reported that cannabis use, particularly heavy and long-term use, is associated with deficits in various cognitive domains, including attention, memory, executive function, and processing speed [21][22][23][24]26,45] the deficits recorded in our sample were primarily limited to the verbal memory and attention domains.The increased susceptibility to distraction on a recall task among cannabis users aligns with research suggesting that cannabis use can impair attentional processes and is associated with difficulties in maintaining attention and inhibiting distractions [22,28].However it is also possible that poor performance on recall following distraction was a hangover from the original, shallow encoding across the five trials of the list learning task in out sample.We did not have the final trial score disaggregated from the total list learning score in this dataset available to test this hypothesis further.Further work is underway to examine the relationship between the final trial score on the list learning task and the subsequent recall score in the people with epilepsy and we plan to include cannabis use as a variable of interest in this wider analysis.
It is interesting that it is the language dominant memory networks that appear to be primarily impacted by cannabis use in our sample.We have observed a similar pattern when we have looked at the exacerbatory impacts of anxiety and low mood on neuropsychological test performance in people with epilepsy [46].These functions appear to be particularly vulnerable to factors that impact on the cognitive reserves of people with epilepsy, which may already be compromised due to organic factors associated with seizures and the impacts of anti-seizure medications.However, whilst this pattern is seen in relation to other factors that impact cognitive function, it is unclear why language dominant networks would be primarily affected.It may reflect something specific to the language dominant networks themselves or may be an artefact of our test construction.Our tests of verbal ability may tap into more discreet networks, whilst performance on tests of non-verbal ability may be more widely distributed, allowing for more pressure on the system before a deficit becomes evident.Insights from functional imaging paradigms may be able to elucidate this further.
There are many reasons why people with epilepsy experience difficulties in multiple cognitive domains, including the underlying pathology responsible for their seizures, the effects of seizures and sub-clinical EEG abnormalities, the impacts of antiseizure medications and common comorbidities associated with the condition [47,48].The fact that we did not find impacts of cannabis use on a wider range of cognitive functions, as would be predicted by the literature in healthy individuals may reflect the fact that these functions are already often compromised by epilepsy related variables [49].It is also worth noting that the cognitive impairments associated with cannabis use can be subtle and may not be evident in all individuals or under all conditions [50].Factors such as age, duration and frequency of use, potency of cannabis, and individual differences in vulnerability are likely to influence the magnitude and persistence of cognitive deficits.
The elevated levels of anxiety and depression in our cannabis group is consistent with previous research linking cannabis use to increased psychological distress [51].Self medicating with cannabis for anxiety and low mood is often reported as a primary motivation by people who use the drug on a recreational basis [52].Our study design does not allow for casual inferences and we can only note the increased association of cannabis use in our group in young men with limited cognitive reserve.It is well established that men are less likely to access mental health services as readily as women, with stigma and help-seeking attitudes identified as key facets contributing to males seeking less support for mental health difficulties [53,54].In the UK, a disproportionately low number of males compared to females are referred to mental health services and receive mental health treatments [55].As such, a tendency for males to self-medicate with alcohol and drugs to ease emotional distress has been recognised [53,54] which may explain the association between males and 'self-medicating' cannabis use within the present study.Additional barriers to accessing mental services are associated with limited cognitive reserve.In addition to cognitive factors, other barriers to accessing mental health services may include long waiting lists and inadequate provision resulting in individuals turning to alternative remedies for their difficulties.
However it is also possible that cannabis use has a bidirectional relationship with lower cognitive reserve.Whilst those with more limited cognitive reserve may be more likely to self-medicate due to the barriers they face accessing mental health services, cannabis use may also result in more limited cognitive reserves, particularly if people  [27].Due to the existing vulnerability of brain function in people with epilepsy, the additional neuropsychological burden of cannabis use during young adulthood may have negative impacts on the development of a broad range of neuropsychological functions.This could possibly account for our finding that the cannabis users within this study had significantly lower reading IQ than non-cannabis users.Further data establishing the age of onset of cannabis use will be needed to investigate these possibilities.

Limitations
There are several limitations of the study.The data relied on selfreport measures of recreational cannabis use.Given that cannabis use remains illegal in the UK, patient reports are likely to underestimate true cannabis use in this sample.Moreover, the sample consisted of patients who were referred for a clinical assessment of their cognitive function.Whilst the reasons for their referral varied, all patients will have been seen due to concerns about cognition, arising from the patient, their family or the treating clinician or in the context of a presurgical assessment for medically intractable epilepsy.These patients are not representative of the majority of people with epilepsy whose seizures are well controlled on medication and who are never referred for a formal neuropsychological assessment.The pragmatic sample and observational design limit our ability to establish causal relationships or assess the long-term or cumulative effects of cannabis use.Caution should therefore be used in generalising these findings to the broader population.

Future research
Whilst the legal status of recreational cannabis currently prohibits any kind of controlled trial to investigate the impact of recreational, non-prescribed use on cognitive function in many countries, such research may be possible in jurisdictions where the use of cannabis has been legalised, although ethical considerations may prevent any study designs involving randomisation, even in these settings.In other regions where recreational cannabis remains proscribed, follow-up assessments tracking the cognitive performance of those who report cannabis use to examine potential recovery or reversibility of cognitive deficits with cessation, would provide valuable insights to share with this patient population.Whilst we did not find any association between epilepsy type and cannabis use, further investigations into the interactions between cannabis use and different anti-seizure medications and impact on cognitive function may be a useful area of research.In a larger sample, the exploration of potential dose-response relationships and the impact of different cannabis strains and modes of consumption could also further our understanding of the cognitive effects of cannabis in this patient group.

Conclusions
In summary, we found that approximately one in 17 of the patients referred to our service reported recreational cannabis use.Usage was most commonly reported by young males with elevated levels of anxiety and depression.Recreational cannabis use was associated with lower intellectual reserve, reduced verbal learning, and increased susceptibility to distraction on a recall task.These effects remained significant even after controlling for elevated levels of anxiety and depression.The findings underscore the importance of considering cannabis use as a potential factor influencing cognitive function in neuropsychological assessments, particularly in presurgical settings where deficits in verbal memory may be misattributed.

Ethical statement
This study examined patient data collected on a routine clinical basis.All data was fully anonymised prior to the analyses in order to conform to the ethical approvals granted for the analysis of this data an audit of recreational cannabis use in patients referred for a neuropsychological assessment.(Hospital Board Approval: 20-202223-SE).No individuals can be identified from the information presented in this study.The clinical data analysed in the current study are not publicly available due to patient privacy and restricted access, but further information about the database is available from the corresponding author on reasonable request.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 1
Self-Reported Recreational Cannabis Use in Epilepsy Populations.
effects of cannabis use included impaired thinking (17 %), anxiety (16 %) and altered hunger (15 %).60 % found cannabis was 'somewhat' to 'very' effective in reduce seizure frequency.Cannabis users main concerns about use included financial strain (37 %), lack of recommendation from a doctor (30 %) and a lack of information around cannabis use (19 %).

Table 2
Demographic and clinical characteristics of the sample.

Table 3
Neuropsychological Test Scores in the Cannabis and Control Groups.Not all patients undertake all tasks.This is reflected in the numbers in each analyses.Shaded rows indicate significant group differences.** z scores based on the healthy population mean.These are not available for the questionnaire measures. *

Table 4
Linear Regression Results; Tests of Between-Subjects Effects List Learning (List A1-5).

Table 5
Linear Regression Results; Tests of Between-Subjects Effects Recall following distraction (LIST A6).