INTRODUCTION
Within the United States, cannabis (CAN) is the second most commonly used substance among adolescents and young adults (Schulenberg et al., Reference Schulenberg, Johnston, O’Malley, Bachman, Miech and Patrick2019). Approximately 52% of young adults (aged 18–25) (Han, Compton, Blanco, & Jones, Reference Han, Compton, Blanco and Jones2019) and 31% of adolescents (Johnston et al., Reference Johnston, Miech, O’Malley, Bachman, Schulenberg and Patrick2020) have used CAN within the past year. Repeated and regular CAN use within this age range is associated with adverse neurocognitive (Gonzalez, Pacheco-Colon, Duperrouzel, & Hawes, Reference Gonzalez, Pacheco-Colon, Duperrouzel and Hawes2017; Meier et al., Reference Meier, Caspi, Ambler, Harrington, Houts, Keefe and Moffitt2012) and brain structural and functional outcomes (Batalla et al., Reference Batalla, Bhattacharyya, Yucel, Fusar-Poli, Crippa, Nogue and Martin-Santos2013; Lisdahl, Gilbart, Wright, & Shollenbarger, Reference Lisdahl, Gilbart, Wright and Shollenbarger2013); however, structural findings are not always consistent (Lisdahl et al., Reference Lisdahl, Gilbart, Wright and Shollenbarger2013). Thus, there is a call to investigate potential moderating factors, which could prove to be influential in these associations (Lorenzetti, Chye, Silva, Solowij, & Roberts, Reference Lorenzetti, Chye, Silva, Solowij and Roberts2019).
Exogenous CAN acts on the endogenous cannabinoid system, primarily through binding to its cannabinoid receptor type 1 (CB1), which is principally involved in neuromodulation (Mechoulam & Parker, Reference Mechoulam and Parker2013) diffusely across the cerebral cortex (Eggan & Lewis, Reference Eggan and Lewis2007). Chronic and regular CAN use can affect CB1 downregulation (Hirvonen et al., Reference Hirvonen, Goodwin, Li, Terry, Zoghbi, Morse and Innis2012) and binding (Villares, Reference Villares2007) for at least a month. In relation to preclinical adolescent models, this developmental period influences the effects of CAN administration (Viveros, Llorente, Moreno, & Marco, Reference Viveros, Llorente, Moreno and Marco2005) with altered dopaminergic systems (Higuera-Matas et al., Reference Higuera-Matas, Botreau, Del Olmo, Miguens, Olias, Montoya and Ambrosio2010) and frontal circuitry (Eggan, Mizoguchi, Stoyak, & Lewis, Reference Eggan, Mizoguchi, Stoyak and Lewis2010) potentially resulting in structural brain changes (Renard, Rushlow, & Laviolette, Reference Renard, Rushlow and Laviolette2016). One primary brain morphological index of continued interest is regional gray matter volume. Gray matter volume is known to be at its largest during childhood and, due to pruning, decreases in adolescence and then plateaus into young adulthood (Mills et al., Reference Mills, Goddings, Herting, Meuwese, Blakemore, Crone and Tamnes2016). Introduction of repeated CAN exposure during this developmental period may be associated with abnormal structure and downstream effects on neuropsychological functioning.
Aberrations in these volumetric indices related to CAN use in this age range include smaller medial orbitofrontal and inferior parietal cortices (Price et al., Reference Price, McQueeny, Shollenbarger, Browning, Wieser and Lisdahl2015), smaller left rostral anterior cingulate cortex (Maple, Thomas, Kangiser, & Lisdahl, Reference Maple, Thomas, Kangiser and Lisdahl2019), larger cerebellar vermis (Medina, Nagel, & Tapert, Reference Medina, Nagel and Tapert2010), and smaller bilateral hippocampal volumes (Ashtari et al., Reference Ashtari, Avants, Cyckowski, Cervellione, Roofeh, Cook and Kumra2011). Some studies have reported that aberrant brain morphometry was linked to poorer executive functioning (Medina et al., Reference Medina, McQueeny, Nagel, Hanson, Yang and Tapert2009), long-delay recall (Jacobus et al., Reference Jacobus, Goldenberg, Wierenga, Tolentino, Liu and Tapert2012), complex attention (Price et al., Reference Price, McQueeny, Shollenbarger, Browning, Wieser and Lisdahl2015), working memory (Bava, Jacobus, Mahmood, Yang, & Tapert, Reference Bava, Jacobus, Mahmood, Yang and Tapert2010), and affect discrimination (Maple et al., Reference Maple, Thomas, Kangiser and Lisdahl2019) in CAN users. One potential reason underlying inconsistent findings is that potential moderators identifying at-risk or more resilient individuals are underspecified in the literature to date.
In healthy adolescent and young adult samples, regional gray matter volume development has exhibited trajectories that are sex specific (Vijayakumar et al., Reference Vijayakumar, Allen, Youssef, Dennison, Yucel, Simmons and Whittle2016). In addition, preclinical models have demonstrated sexual dimorphic CB1 diffusivity in the endocannabinoid system, with greater desensitization of these receptors shown in adolescent female rodents after tetrahydrocannabinol (THC) administration compared to males (Burston, Wiley, Craig, Selley, & Sim-Selley, Reference Burston, Wiley, Craig, Selley and Sim-Selley2010; Rodriguez de Fonseca, Ramos, Bonnin, & Fernandez-Ruiz, Reference Rodriguez de Fonseca, Ramos, Bonnin and Fernandez-Ruiz1993). Furthermore, investigations into effects of sex within humans have indicated differences between CAN-using males and females and their nonusing same-sex counterparts within use patterns (Cuttler, Mischley, & Sexton, Reference Cuttler, Mischley and Sexton2016; Khan et al., Reference Khan, Secades-Villa, Okuda, Wang, Perez-Fuentes, Kerridge and Blanco2013) and neuropsychological performance (Crane, Schuster, Fusar-Poli, & Gonzalez, Reference Crane, Schuster, Fusar-Poli and Gonzalez2013; Crane, Schuster, Mermelstein, & Gonzalez, Reference Crane, Schuster, Mermelstein and Gonzalez2015). Specifically, male CAN users exhibit impairments on psychomotor and visuospatial performance (Crane et al., Reference Crane, Schuster, Fusar-Poli and Gonzalez2013) and age of regular CAN onset is associated with poorer episodic memory among female users (Crane et al., Reference Crane, Schuster, Mermelstein and Gonzalez2015). Furthermore, previous studies have demonstrated sex differences in the impact of CAN use on brain structure with female users exhibiting larger right amygdala volume (McQueeny et al., Reference McQueeny, Padula, Price, Medina, Logan and Tapert2011), larger prefrontal volume (Medina et al., Reference Medina, McQueeny, Nagel, Hanson, Yang and Tapert2009), and larger cortical surface structure (Sullivan, Wallace, Wade, Swartz, & Lisdahl, Reference Sullivan, Wallace, Wade, Swartz and Lisdahl2020). Notably, most of the reviews examining outcomes indicate a majority of studies either skew or are predominantly male or did not examine sex differences (Crane et al., Reference Crane, Schuster, Fusar-Poli and Gonzalez2013; Lisdahl et al., Reference Lisdahl, Gilbart, Wright and Shollenbarger2013; Lisdahl, Wright, Kirchner-Medina, Maple, & Shollenbarger, Reference Lisdahl, Wright, Kirchner-Medina, Maple and Shollenbarger2014). Thus, there is an increased need to examine sex as a potential moderator to determine sex-specific associations linked to potential adverse outcomes related to CAN use (Rubino & Parolaro, Reference Rubino and Parolaro2015).
A novel factor to consider in CAN-related investigations is extent of aerobic fitness. Increased levels of aerobic fitness has been robustly related to positive brain outcomes in older adults (Bherer, Erickson, & Liu-Ambrose, Reference Bherer, Erickson and Liu-Ambrose2013; Thomas, Dennis, Bandettini, & Johansen-Berg, Reference Thomas, Dennis, Bandettini and Johansen-Berg2012), and converging lines of research indicate that fitness is additionally beneficial within healthy adolescents and young adults (Chaddock, Pontifex, Hillman, & Kramer, Reference Chaddock, Pontifex, Hillman and Kramer2011; Herting & Nagel, Reference Herting and Nagel2013; Pereira et al., Reference Pereira, Huddleston, Brickman, Sosunov, Hen, McKhann and Small2007; Schwarb et al., Reference Schwarb, Johnson, Daugherty, Hillman, Kramer, Cohen and Barbey2017; Voss, Vivar, Kramer, & van Praag, Reference Voss, Vivar, Kramer and van Praag2013; Whiteman, Young, Budson, Stern, & Schon, Reference Whiteman, Young, Budson, Stern and Schon2016). Its link to regional gray matter volume may be due to a number of mechanisms, including, but not limited to: brain-derived neurotropic factors (BDNFs; Huang, Larsen, Ried-Larsen, Moller, & Andersen, Reference Huang, Larsen, Ried-Larsen, Moller and Andersen2014), vascular growth factors (VGFs; Fleenor, Marshall, Durrant, Lesniewski, & Seals, Reference Fleenor, Marshall, Durrant, Lesniewski and Seals2010), and neurogenesis (Nokia et al., Reference Nokia, Lensu, Ahtiainen, Johansson, Koch, Britton and Kainulainen2016). As it pertains to CAN use, acute aerobic exercise releases endocannabinoids (Koltyn, Brellenthin, Cook, Sehgal, & Hillard, Reference Koltyn, Brellenthin, Cook, Sehgal and Hillard2014), and thus it has been theorized this may counteract downregulation of CB1 receptors among exogenous CAN use (Lisdahl et al., Reference Lisdahl, Gilbart, Wright and Shollenbarger2013). Supporting this notion, CAN users who were more aerobically fit demonstrated superior neuropsychological performance compared to less aerobically fit users (Wade, Wallace, Swartz, & Lisdahl, Reference Wade, Wallace, Swartz and Lisdahl2019), and an intervention utilizing aerobic exercise found decreased craving and use among otherwise-sedentary users (Buchowski et al., Reference Buchowski, Meade, Charboneau, Park, Dietrich, Cowan and Martin2011). Our group previously reported that aerobic fitness was positively related to temporal, parietal, and frontal cortical surface structure in a similar sample of CAN users and controls (Sullivan et al., Reference Sullivan, Wallace, Wade, Swartz and Lisdahl2020). Even though this emerging line of research presents fitness as a viable moderator, few studies have incorporated aerobic fitness into assessments of CAN-related consequences on neurocognition and doing so may elucidate aerobic fitness outcomes that potentially put adolescents and young adults at more or less of a risk for adverse effects of CAN use.
Determining the contribution of potential moderators (i.e., sex and aerobic fitness) to brain morphology of CAN users may elucidate subgroups at higher or lower risk for abnormal structural outcomes. Therefore, the aim of the present study is to examine the main effect of CAN use, and novel interactions with sex and aerobic fitness on volume and subcortical volume regions of interest (ROIs). We expect to see aerobic fitness associated with larger brain volume regardless of group, overall CAN group differences, and CAN-by-sex interactions. Exploratory analyses will examine the correlation between significant regions and neurocognitive performance (i.e., working memory, processing speed, and sustained attention) (Lisdahl & Price, Reference Lisdahl and Price2012; Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019).
METHODS
Participants
Seventy-four participants (CAN users = 36, controls = 38) were recruited through flyers and advertisements in the local community and college as part of a larger parent study, which examined the neurocognitive effects of CAN use in young adults (R01-DA030354; PI: Lisdahl). Participants in the present analysis were between the ages of 16 and 26 years (M = 21.1, SD = 2.6), were sex balanced (44.6% female), and racial identities consisted of predominantly: Caucasian (64.9%), Asian (10.8%), multiracial (10.8%), and African-American (8.1%) (see Table 1).
Demographic, substance use, and aerobic fitness characteristics
Note. CAN: Cannabis.
a Measured in standard uses on TLFB (Sobell & Sobell, Reference Sobell, Sobell, Litten and Allen1992).
b Measured at VO2 maximum testing.
c Measured in standard uses on Customary Drinking and Drug Use Record (CDDR) (Brown et al., Reference Brown, Myers, Lippke, Tapert, Stewart and Vik1998).
d Not applicable.
e Calculated from TLFB last CAN use date and date of sMRI.
f Maximum rate of oxygen consumption—measured in milliliters of oxygen consumed per kilogram of body weight, per minute.
* p < .01.
Participants in the parent study were included if they were right handed, spoke English, and were willing to abstain from substance use over a 3-week period. Exclusion criteria included having an independent DSM-IV Axis I (attention, mood, anxiety, or psychotic) disorder, current use of psychoactive medications, major medical or neurological disorders (including metabolic disorders), loss of consciousness >2 min, history of learning disability or intellectual disability, prenatal medical issues or premature birth (gestation <35 weeks), MRI contraindications (pregnancy, claustrophobia, and metal in body), reported significant prenatal alcohol exposure (≥4 drinks in a day or ≥6 drinks in a week), prenatal illicit drug exposure, or prenatal nicotine exposure (average >5 cigarettes per day longer than 1 month), elevated Physical Activity Readiness Questionnaire (Thomas, Reading, & Shephard, Reference Thomas, Reading and Shephard1992) scores screening eligibility for VO2 maximum (VO2 max) testing, or excessive other illicit drug use (>20 times of lifetime use for each drug category, including CAN use for nonusing control participants). Based on the International Physical Activity Questionnaire (Fogelholm et al., Reference Fogelholm, Malmberg, Suni, Santtila, Kyrolainen, Mantysaari and Oja2006), participants were additionally balanced based on being active versus inactive to increase likelihood of adequate range in aerobic fitness within both groups.
CAN users in the present analysis were categorized as current users who used CAN at least 44 times in the last year (i.e., nearly weekly) and at least 100 lifetime uses. Nonusing controls in the present analysis used CAN no more than five times in the past year and less than 20 times in their lifetime (Lisdahl & Price, Reference Lisdahl and Price2012; Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019; Wallace, Wade, Hatcher, & Lisdahl, Reference Wallace, Wade, Hatcher and Lisdahl2019).
Procedures
All aspects of the protocol were approved by local Institutional Review Board (IRB) and in accordance with the Helsinki Declaration. Potential participants who expressed interest in the parent study were consented and interviewed with a detailed phone screen, along with a parental informant phone interview (explained further in Supplementary Materials).
Eligible participants came in for five study sessions over the course of 3 weeks. The first three sessions occurred 1 week apart and consisted of drug toxicology and a brief neuropsychological battery (for details, see Wallace, Wade, & Lisdahl, Reference Wallace, Wade and Lisdahl2020). Sessions 4 and 5 occurred at least 1 week after Session 3 and consisted of ascertaining body composition, VO2 max testing, a full neuropsychological battery, and then an MRI that occurred within 24–48 hr of each other.
During the entire study period, participants were asked to remain abstinent from alcohol, CAN, and other drug use (other than tobacco), which was confirmed through breath, urine, and sweat toxicology screening. If positive for illicit drug use, showed an increase in 11-nor-9-carboxy-THC (THCCOOH) levels, or had a breath alcohol concentration greater than .000 at the start of any subsequent session after baseline, participants were allowed to continue their involvement in the study from Session 1. Participants who used tobacco were asked to abstain from use an hour before the MRI scan to prevent interference with functional task data.
Measures
Past-year substance use
A modified version of the Timeline Follow-Back (TLFB) interviews was conducted by trained research assistants (RAs) to measure substance use patterns on a weekly basis for the past year while providing memory cues such as holidays and personal events (Lisdahl & Price, Reference Lisdahl and Price2012; Sobell & Sobell, Reference Sobell, Sobell, Litten and Allen1992). Substances were measured by standard units [alcohol (standard drinks), nicotine (number of cigarettes; occasions for chew/pipe/cigar/hookah), CAN (smoked/vaped flower, concentrates, edibles were measured, and dosing was converted to joints-based grams), ecstasy (tablets), sedatives (pills), stimulants (mg), hallucinogens (hits), heroin/opium (hits), and inhalants (hits)]. Days of CAN abstinence at scan were calculated from date of last CAN use based on the TLFB and date of scan.
Verifying drug abstinence
As participants were expected to remain abstinent from all alcohol and drugs (other than nicotine) throughout the course of the study, abstinence was evaluated at each session with the following: urine samples were tested using the ACCUTEST SplitCup 10 Panel drug test, which measures amphetamines, barbiturates, benzodiazepines, cocaine, ecstasy, methadone, methamphetamines, opiates, phencyclidine (PCP), and delta-9-THC; in addition, urine samples were tested using NicAlert to test cotinine level—a metabolite of nicotine; participants also wore PharmChek Drugs of Abuse Patches, which continuously monitor sweat toxicology for the presence of cocaine, benzoylecgonine, heroin, 6-Monoacetylmorphine (6MAM), morphine, codeine, amphetamines, methamphetamine, THC, and PCP, and gave quantified values of THCCOOH (a metabolite of THC); and participants underwent breathalyzer screens to test for alcohol use at the start of each session.
Neuropsychological battery
Immediately prior to VO2 max testing, participants were administered a full neuropsychological battery (see Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019, for further information), which included the Paced Auditory Serial Addition Task (PASAT), Ruff 2&7 Selective Attention Task, and Delis-Kaplan Executive Function System (D-KEFS) Trails Making Test-4 (i.e., switching). PASAT total raw scores were used as a measure of processing speed, concentration, and working memory (Diehr, Heaton, Miller, & Grant, Reference Diehr, Heaton, Miller and Grant1998). Ruff 2&7 age-corrected total accuracy was used to measure selective and sustained attention (Lezak, Howieson, Loring, & Fischer, Reference Lezak, Howieson, Loring and Fischer2004; Ruff, Niemann, Allen, Farrow, & Wylie, Reference Ruff, Niemann, Allen, Farrow and Wylie1992). Trails switching total time was used as a measure of executive control and working memory (Arbuthnott & Frank, Reference Arbuthnott and Frank2010; Lezak et al., Reference Lezak, Howieson, Loring and Fischer2004; Sanchez-Cubillo et al., Reference Sanchez-Cubillo, Perianez, Adrover-Roig, Rodriguez-Sanchez, Rios-Lago, Tirapu and Barcelo2009).
Body fat percentage
An electrical bioimpedance analysis system was utilized to estimate body fat percentage [The Tanita Body Composition Analyzer, TBF-300 (Tanita Corporation, Tokyo, Japan)] with all pretesting requirements met, which was utilized to compare between-group differences to address attributions of adiposity on results within the present analysis (Schwartz et al., Reference Schwartz, Dickie, Pangelinan, Leonard, Perron, Pike and Paus2014).
VO2 maximum
Participants were asked to refrain from food and caffeine for 4 hr prior to the exercise test. Exercise testing was completed using a calibrated ParvoMedics TrueOne 2400 metabolic measurement system (ParvoMedics, Salt Lake City, UT). Participants completed an incremental exercise test on a treadmill following the Bruce protocol until volitional fatigue (for full details, see Sullivan et al., Reference Sullivan, Wallace, Wade, Swartz and Lisdahl2020; Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019). Criteria for VO2 max were based on Howley, Bassett, and Welch (Reference Howley, Bassett and Welch1995).
MRI acquisition
Structural MRI (sMRI) scans were acquired on a 3T Signa LX MRI scanner (GE Healthcare, Waukesha, WI) using a 32-channel quadrature transmit/receive head coil. Anatomical images were acquired using a T1-weighted spoiled gradient-recalled at steady-state pulse sequence (TR = 8.2 ms, TE = 3.4 s, TI = 450, and flip angle of 12°). The in-plane resolution of the anatomical images was 256 × 256 with a square field of view of 256 mm. One hundred fifty slices were acquired at 1-mm thickness. This resulted in a 1 mm × 1 mm × 1 mm voxel resolution.
Processing pipeline
Participant structural scans were processed in a standard processing pipeline within FreeSurfer version 5.3 (explained further in Supplementary Materials).
Statistical Analysis
Differences in demographic variables were examined using ANOVAs and Chi-square tests in R (R Development Core Team, 2010). A series of multivariate regressions were run on whole-brain regional gray matter volume with CAN group, sex, VO2 max levels, and their interactions (CAN × Sex and CAN × VO2 max) as the independent variables of interestFootnote 1; covariates included past-year alcohol and cotinine level at the time of aerobic fitness testing (see Supplementary Materials for results of a power analysis). Analyses were completed separately between each hemisphere (right and left) and smoothed with a global Gaussian blur at full width at half maximum (FWHM) of 15. Corrections for multiple comparisons were made using Monte Carlo simulations at a vertex-wise/cluster-forming threshold of p < .05 (i.e., 1.3) and cluster-wise probability (cwp) of p = .05, while correcting across both hemispheric spaces; no minimum number of voxels required to achieve significant cluster results were set (Greve & Fischl, Reference Greve and Fischl2018). Regional effect sizes (ESs) were computed through dividing the residual error standard deviation by the contrast ES for significant effects within the analyses.
A series of linear regressions were run in R which examined subcortical volume ROIs (hippocampus, amygdala, cerebellum, caudate, and putamen) with CAN group, sex, VO2 max levels, and their interactions (CAN × Sex and CAN × VO2 max) as the independent variables of interestFootnote 2; covariates included past-year alcohol use and cotinine level at the time of aerobic fitness testing. Corrections for multiple comparisons using false discovery rate (FDR) were computed for the series of subcortical volume ROI linear regressions (Benjamini & Hochberg, Reference Benjamini and Hochberg1995), both raw p-values and FDR-corrected p-values are reported below.
Follow-up exploratory analyses examined correlations between corrected significant clusters or subcortical ROIs and neuropsychological performance on aforementioned neuropsychological tests. Correlation matrices were computed using Pearson’s r correlations. Decisions on reporting were made at p < .05. Correlations were run separately between users and nonusers for CAN × VO2 max interactions, and between users and nonusers by sex for CAN × Sex interactions to interpret specific effects. Correlations for each group were compared using two-tailed Fisher’s z computation (Ramseyer, Reference Ramseyer2015).
RESULTS
Demographic Data
There were no significant differences between CAN and nonuser groups in regard to age (p = .27), sex distribution (p = .23), ethnicity (p = .26), race (p = .44), educational attainment (p = .78), VO2 max (p = .30), and body fat percentage (p = .27). As expected, there were significant differences in lifetime (p < .001) and past-year CAN use (p < .001), past-year tobacco use (p = .008), cotinine levels at VO2 max testing (p = .003), and alcohol consumed within the past year (p < .001); cotinine levels and alcohol use were included as a covariate in all analyses. Within the CAN users, there was no difference between sexes for past year (p = .20) or lifetime (p = .19) CAN use, days of CAN abstinence prior to sMRI (p = .27), or age of first regular CAN use onset (p = .55).
Primary Analyses
CAN findings
There were no significant CAN group findings observed in whole-brain or subcortical volume outcomes.
CAN × Sex findings
Whole-brain volume. Interactions were observed between CAN group and sex in the left lateral orbitofrontal [t(58) = −3.99, ES = −.29, cwp = .019], left inferior temporal [t(58) = −2.73, ES = −.28, cwp = .017], left precuneus [t(58) = −2.67, ES = −.29, cwp = .034], left caudal middle frontal [t(58) = −2.40, ES = −.27, cwp = .0003], right superior frontal [t(57) = 4.42, ES = .30, cwp = .001], and the right paracentral [t(57) = 3.19, ES = .29, cwp = .005] regions (see Figure 1). CAN-using females demonstrated greater volume in left lateral orbitofrontal, left precuneus, left caudal middle frontal, and right paracentral regions compared to nonusing females, whereas CAN-using males had reduced volume in these regions compared to nonusing males. However, in the left inferior temporal and right superior frontal, CAN-using females demonstrated less volume compared to nonusing females, similar to the relationship in males, yet CAN-using males demonstrated the most robust decrease in these regions compared to nonusing males (see Supplementary Table 1 and Supplementary Figure 2). Subcortical volume. Significant interactions between CAN group and sex were observed in right amygdala [t(63) = −2.41, p = .019, FDR = 0.13] and right caudate [t(63) = −2.04, p = .046, FDR = 0.23] regions. CAN-using males demonstrated smaller right amygdala volume compared to nonusing males, whereas CAN-using females exhibited larger volume in the right amygdala compared to nonusing females. Both male and female CAN users demonstrated less right caudate volume compared to male and female nonusing controls, respectively; yet, this was more robust for females compared to males. However, neither region survived correction for multiple comparisons.
Cannabis × Sex findings. Lateral view of CAN group and sex interactions observed in (a) left lateral orbitofrontal, left inferior temporal, left precuneus (not pictured), left caudal middle frontal, right superior frontal, and (b) right paracentral volumes. CAN-using males exhibited less volume compared to nonusing males. Contrarily, CAN-using females demonstrated more volume in aforementioned regions compared to nonusing females, except for left inferior temporal and right superior frontal volume where less volume was observed in CAN-using females compared to nonusing females.
VO2 findings
Whole-brain volume. A significant relationship between increased VO2 max and larger volume was observed in two separate areas of the left inferior parietal [t(58) = 5.21, ES = .37, cwp = .0001; t(58) = 4.31, ES = .32, cwp = .0001], left rostral middle frontal [t(58) = 2.89, ES = .27, cwp = .039], right inferior parietal [t(57) = 3.40, ES = .30, cwp = .035], right fusiform [t(57) = 3.11, ES = .29, cwp = .001], and right precuneus [t(57) = 3.06, ES = .29, cwp = .02] regions (see Figure 2) (see Supplementary Table 1). Subcortical volume. There was a significant relationships between increased VO2 max and larger volume in left [t(63) = 3.04, p = .003, FDR = 0.03] and right [t(63) = 2.84, p = .006, FDR = 0.046] caudate, and in left [t(63) = 3.01, p = .004, FDR = 0.03] and right [t(63) = 2.48, p = .016, FDR = 0.11] cerebellum, though the finding in the right cerebellum did not survive corrections.
VO 2 findings. Lateral view of VO2 findings observed in left inferior parietal, left rostral middle frontal, right inferior parietal, right fusiform, and right precuneus (not pictured) volumes. Increased VO2 was positively associated with more volume in these regions.
CAN × VO2 findings
Whole-brain volume. A significant interaction was observed between VO2 max and CAN group in the left superior temporal region [t(58) = −3.58, ES = −.30, cwp = .0001] (see Supplementary Table 1). Nonusing controls demonstrated a positive relationship between increased VO2 max and more volume, whereas no trend was observed for the CAN group (see Supplementary Figure 1). Subcortical volume. There were no VO2 max-by-CAN group interactions observed for subcortical volume.
Exploratory Brain–Behavior Correlations
Correlations in VO2 associated regions are located in Supplementary Materials. See Table 2 for correlation coefficients between brain volume and cognitive tasks in regions that differed according to CAN × VO2 or CAN × Sex interactions. Fisher’s z scores were calculated to determine whether correlation coefficients significantly differed by CAN group status for the males and females in the sample.
Correlations between volume and neuropsychological performance in regions that differed in primary analyses
Notes. Fisher’s Z comparisons were run to determine whether correlation coefficients significantly differed between CAN and controls. CAN: Cannabis users; PASAT: Paced Auditory Serial Addition Task.
a Region identified from Cannabis × VO2 analysis; controls demonstrated a more robust positive relationship between VO2 max and volume.
b Regions identified from Cannabis × Sex analyses where CAN-using males had smaller volume compared to nonusing males.
c Regions identified from Cannabis × Sex analyses where CAN-using females had smaller volume compared to non-using females.
d Regions identified from Cannabis × Sex analyses where CAN-using females had greater volume compared to nonusing females.
*p < .05.
DISCUSSION
Given ongoing policy debates (Carliner, Brown, Sarvet, & Hasin, Reference Carliner, Brown, Sarvet and Hasin2017) and prevalence of use in adolescents and young adults (Johnston et al., Reference Johnston, Miech, O’Malley, Bachman, Schulenberg and Patrick2020), further characterizing brain structure as it relates to regular CAN use in this population is of continued importance. Yet, findings from the current literature are largely heterogeneous and there is a call to examine potential influencers (Lorenzetti et al., Reference Lorenzetti, Chye, Silva, Solowij and Roberts2019). The current study sought to further elucidate this relationship by investigating two potentially moderating factors—sex and aerobic fitness—on the associations between CAN group and brain volume in a group of healthy adolescents and young adults who underwent 3 weeks of monitored abstinence. There were no main effects of CAN group on volume after accounting for sex, aerobic fitness, past-year alcohol use, and current nicotine use. However, CAN-by-sex interactions were observed in frontal, temporal, paracentral, and precuneus volumes. Males demonstrated smaller volumes, whereas female users generally had larger volumes compared to their nonusing same-sex counterparts. Exploratory and preliminary brain–behavior analyses largely demonstrated that the pattern of volume findings in both male and female CAN users was linked with disadvantageous neuropsychological performance. Whole-sample findings with aerobic fitness were diffusely observed with increased cortical volume; and furthermore, a CAN-by-aerobic fitness interaction was demonstrated in left superior temporal volume, with nonusers showing a positive association, whereas no relationship was observed for CAN users. Overall, aerobic fitness was linked with greater brain volume and was in turn associated with superior neuropsychological performance.
Notably, we did not find any main effects of CAN on volume in the present study. This is inconsistent with prior studies, which have demonstrated differences between CAN users and nonusers in several regions (Ashtari et al., Reference Ashtari, Avants, Cyckowski, Cervellione, Roofeh, Cook and Kumra2011; Lisdahl et al., Reference Lisdahl, Tamm, Epstein, Jernigan, Molina, Hinshaw and Group2016; Maple et al., Reference Maple, Thomas, Kangiser and Lisdahl2019; Matochik, Eldreth, Cadet, & Bolla, Reference Matochik, Eldreth, Cadet and Bolla2005; Medina et al., Reference Medina, Nagel and Tapert2010; Medina, Schweinsburg, Cohen-Zion, Nagel, & Tapert, Reference Medina, Schweinsburg, Cohen-Zion, Nagel and Tapert2007; Price et al., Reference Price, McQueeny, Shollenbarger, Browning, Wieser and Lisdahl2015). The present null main-effect findings could be due to novel sampling of balanced aerobically fit and unfit CAN users, varying frequency of use to determine inclusion criteria, or due to the longer-than-average length of abstinence (3 weeks) our sample maintained (Batalla et al., Reference Batalla, Bhattacharyya, Yucel, Fusar-Poli, Crippa, Nogue and Martin-Santos2013; Lisdahl et al., Reference Lisdahl, Gilbart, Wright and Shollenbarger2013; Scott et al., Reference Scott, Slomiak, Jones, Rosen, Moore and Gur2018), but nonetheless, lend further evidence to the overall heterogeneous aberrant volumetric findings in CAN users and potentially influential effect of moderators (Lorenzetti et al., Reference Lorenzetti, Chye, Silva, Solowij and Roberts2019).
To that end, the present analyses revealed several CAN-by-sex interactions in left frontal, temporal, and precuneus volumes, and right frontal volume. It may be that null main-effect findings could be due to accounting for these significant interactions that are demonstrating opposing effects in several regions. For example, CAN-using males demonstrated smaller volume compared to nonusing males and CAN-using females generally showcased larger volume compared to nonusing females in left lateral orbitofrontal, left precuneus, left caudal middle frontal, and right paracentral regions. Yet, in the left inferior temporal and right superior frontal volume, both CAN-using males and females exhibited smaller volume compared to their nonusing same-sex counterparts; however, this difference was starker in males. This general trend in findings aligns with previous literature showcasing larger amygdala volume in CAN-using females (McQueeny et al., Reference McQueeny, Padula, Price, Medina, Logan and Tapert2011) and CAN-by-sex interactions in prefrontal volume (Medina et al., Reference Medina, McQueeny, Nagel, Hanson, Yang and Tapert2009) in a somewhat younger cohort who also underwent 30 days of monitored abstinence.
In follow-up preliminary analyses to understand brain–behavior relationships in the male and female CAN users and nonusers, we found an overall pattern in the male CAN users that linked smaller volume in right paracentral and left caudal middle frontal regions with poorer sustained attention. Controls also demonstrated significantly more robust correlations between smaller volumes and poorer neuropsychological performance compared to male CAN users in left inferior temporal and left precuneus regions. Thus, consistent with prior studies, we again found that smaller volume in conjunction with CAN use is disadvantageously linked to neuropsychological function in males (Medina et al., Reference Medina, McQueeny, Nagel, Hanson, Yang and Tapert2009; Price et al., Reference Price, McQueeny, Shollenbarger, Browning, Wieser and Lisdahl2015). Among female CAN users, larger volumes in the left precuneus and smaller volumes in the right superior frontal region were correlated with worse sequencing and processing speed performance, whereas smaller volume in the left inferior temporal lobe was linked with poorer sustained attention. This pattern generally suggested that abnormal volumes observed in female CAN users compared to female nonusing controls were disadvantageous. Interestingly, female CAN users had more robust correlations between volume and neuropsychological performance in left inferior temporal and right superior frontal regions. However, it is notable that the smallest sample size was of female CAN users; thus, these findings need replication in larger samples as the magnitude of brain–behavior relationships is potentially smaller than previously recognized (Palmer et al., Reference Palmer, Zhao, Loughnan, Zou, Fan, Thompson and Jernigan2020). Importantly, previous research has demonstrated neuropsychological differences prior to CAN initiation, which represent a risk for use (Cheetham et al., Reference Cheetham, Allen, Whittle, Simmons, Yucel and Lubman2012; Jackson et al., Reference Jackson, Isen, Khoddam, Irons, Tuvblad, Iacono and Baker2016; Tervo-Clemmens, Quach, Calabro, Foran, & Luna, Reference Tervo-Clemmens, Quach, Calabro, Foran and Luna2020), and thus the present findings—particularly the preliminary brain–behavior relationships—are noted as associations rather than causal relationships. Even so, present findings suggest negative links with cognition associated with aberrant brain volume morphology between CAN -using and nonusing groups, which require further replication in large-scale studies.
More broadly, CAN-by-sex findings may be due to several factors. Sex-specific pruning patterns may be impacted by introducing exogenous CAN exposure into staggered developmental trajectories (Medina et al., Reference Medina, McQueeny, Nagel, Hanson, Yang and Tapert2009; Rubino & Parolaro, Reference Rubino and Parolaro2015), which temporally differ between the sexes (Giedd et al., Reference Giedd, Blumenthal, Jeffries, Castellanos, Liu, Zijdenbos and Rapoport1999; Lenroot et al., Reference Lenroot, Gogtay, Greenstein, Wells, Wallace, Clasen and Giedd2007). Furthermore, increased CB1 receptor density in males compared to females has also been observed in preclinical models (Burston et al., Reference Burston, Wiley, Craig, Selley and Sim-Selley2010; Rubino et al., Reference Rubino, Vigano, Realini, Guidali, Braida, Capurro and Parolaro2008). In addition, male CAN users tend to use more frequently, severely, and with higher potency products (Cuttler et al., Reference Cuttler, Mischley and Sexton2016), potentially contributing to a more consistent picture of reduced brain volume and cognitive deficits (Lisdahl & Price, Reference Lisdahl and Price2012).
Investigating the associations between aerobic fitness and brain morphometry revealed robust positive associations between superior aerobic fitness and larger volume in bilateral inferior parietal, left rostral middle frontal, right fusiform, and right precuneus regions—regardless of CAN group status. In our Supplementary Material, we demonstrated a pattern of significant preliminary positive correlations between brain volume and cognitive functioning. Chiefly, these findings are supported by previous literature examining the relationship between increased aerobic fitness levels and brain morphometry (Herting & Chu, Reference Herting and Chu2017; Herting & Keenan, Reference Herting, Keenan and Watson2017; Wittfeld et al., Reference Wittfeld, Jochem, Dorr, Schminke, Glaser, Bahls and Grabe2020) and cognitive function, particularly on sustained attention and psychomotor speed tasks in young adults (Hwang, Castelli, & Gonzalez-Lima, Reference Hwang, Castelli and Gonzalez-Lima2017; Lee et al., Reference Lee, Wong, Lau, Lee, Yau and So2014; Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019). Moreover, associations were observed between superior aerobic fitness and larger left cerebellar and bilateral caudate volumes, which are consistent with previous findings demonstrating a positive link between physical exercise and subcortical volume in children (Ortega et al., Reference Ortega, Campos, Cadenas-Sanchez, Altmae, Martinez-Zaldivar, Martin-Matillas and Campoy2019) and adults (Wittfeld et al., Reference Wittfeld, Jochem, Dorr, Schminke, Glaser, Bahls and Grabe2020). An interesting finding was an interaction between aerobic fitness and CAN group in left superior temporal volume, where nonusing controls exhibited a robust positive association and no trend was observed for CAN users. Left superior temporal volume was also positively related to sustained attention in nonusers, but not CAN users. Intriguingly, this region has previously been identified as a benefactor to increased aerobic fitness in healthy adults (Wittfeld et al., Reference Wittfeld, Jochem, Dorr, Schminke, Glaser, Bahls and Grabe2020); although present findings indicate this relationship may not be as evident for young adult CAN users, suggesting that CAN use may disrupt this benefit. Further, this is consistent with our prior study findings that CAN users did not have as robust of a relationship between fitness and cortical surface structure in cuneus and occipital regions (Sullivan et al., Reference Sullivan, Wallace, Wade, Swartz and Lisdahl2020). Taken together, this may suggest CAN users demonstrate gains in neurocognitive indices following aerobic activity; yet, these gains may not be as apparent in some brain regions; however, this needs to be confirmed in a clinical trial design.
Still, it is notable that regional links between aerobic fitness and brain volume exist across participants—regardless of CAN group membership—while accounting for sex, alcohol use, and cotinine level in a physically healthy cohort of adolescents and young adults. This represents a novel finding in the aerobic fitness literature. One possible mechanism underlying these findings is that recent studies have revealed that aerobic exercise releases endocannabinoids (Heyman et al., Reference Heyman, Gamelin, Goekint, Piscitelli, Roelands, Leclair and Meeusen2012; Hillard, Reference Hillard2018; Meyer, Crombie, Cook, Hillard, & Koltyn, Reference Meyer, Crombie, Cook, Hillard and Koltyn2019; Watkins, Reference Watkins2018). This may lessen the negative impact of repeated and regular exogenous CAN exposure in youth. Another proposed explanation for aerobic fitness main effects on brain structure in CAN users is that physical activity may metabolize exogenous cannabinoids out of the body at a faster rate, which has been previously examined experimentally (Westin, Mjones, Burchardt, Fuskevag, & Slordal, Reference Westin, Mjones, Burchardt, Fuskevag and Slordal2014; Wong et al., Reference Wong, Montebello, Norberg, Rooney, Lintzeris, Bruno and McGregor2013); hence, reducing the time cannabinoids cycle through the body and perhaps diminishing their overall impact on brain morphometry. As aforementioned, engaging in aerobic exercise has been additionally linked with increased BDNF release (Huang et al., Reference Huang, Larsen, Ried-Larsen, Moller and Andersen2014), VGF (Fleenor et al., Reference Fleenor, Marshall, Durrant, Lesniewski and Seals2010), and neurogenesis (Nokia et al., Reference Nokia, Lensu, Ahtiainen, Johansson, Koch, Britton and Kainulainen2016). Moreover, these structural findings add to our previous research in our lab that found superior performance on visual memory, psychomotor speed, and sequencing ability in more aerobically fit CAN users compared to nonfit users (Wade et al., Reference Wade, Wallace, Swartz and Lisdahl2019). Taken together, these findings suggest that aerobic fitness may be a moderating factor between CAN exposure and neurocognitive health, and this could be harnessed in prevention and treatment efforts. Future studies are needed to help elucidate potential underlying mechanisms explaining the relationship between aerobic fitness, brain structure, and neurocognition in CAN-using youth. Furthermore, understanding which types of physical activity (e.g., muscle strength, balance, and resistance training) influences fitness and, potentially, substance use and brain–behavior relationships may be an important future direction.
It is worth noting potential limitations of the present study. Causality cannot be determined from the present sample due to CAN use initiation occurring prior to study protocols. Moreover, although attempts were made to balance the sample according to active and sedentary individuals, the VO2 max of our sample was lower than average age-based norms (Pescatello, Reference Pescatello2014). Assessing a more representative sample of the population (i.e., a more aerobically fit group) may demonstrate stronger ameliorative associations between aerobic fitness and brain structure in CAN users. Although we did find sex differences, the smallest cell in the present analysis was the CAN-using females (n = 13), which limits our power; we expect that a larger sample size altogether could reveal more robust findings across and within sexes. In addition, there are other influential factors on the relationship between CAN use and brain morphometry, including, genetics (Filbey, Schacht, Myers, Chavez, & Hutchison, Reference Filbey, Schacht, Myers, Chavez and Hutchison2010; Shollenbarger, Price, Wieser, & Lisdahl, Reference Shollenbarger, Price, Wieser and Lisdahl2015; Verweij et al., Reference Verweij, Zietsch, Lynskey, Medland, Neale, Martin and Vink2010; Zinkstok et al., Reference Zinkstok, Schmitz, van Amelsvoort, de Win, van den Brink, Baas and Linszen2006) and psychopathological comorbidities (Crippa et al., Reference Crippa, Zuardi, Martin-Santos, Bhattacharyya, Atakan, McGuire and Fusar-Poli2009; Lev-Ran et al., Reference Lev-Ran, Roerecke, Le Foll, George, McKenzie and Rehm2014). The present study did not have the capacity to account for potentially influential effects of genetics and excluded for major Axis I disorder. Furthermore, prenatal substance use was measured through parental self-report, which may minimize reporting of use. Future investigations should prioritize specific characteristics of CAN use, including, but not limited to, age of first regular onset, severity of use, or CAN potency. In addition, despite CAN metabolites (i.e., THCCOOH) cycling out within a 3- to 4-week period (Goodwin et al., Reference Goodwin, Darwin, Chiang, Shih, Li and Huestis2008), future studies are needed to determine whether subtle abnormalities would recover with longer periods of sustained abstinence.
The current analysis found that after 3 weeks of monitored abstinence, sex moderated the relationship between CAN use and brain volume. In CAN-using males, smaller volumes were observed in lateral orbitofrontal, superior frontal, caudal middle frontal, inferior temporal, precuneus, and paracentral volumes compared to nonusing males. CAN-using females generally exhibited larger volume in these areas compared to nonusing females, except for in the left inferior temporal and right superior frontal, where they also demonstrated smaller volumes. Preliminary brain–behavior correlations generally indicate that abnormal volumes were not advantageous in either the male or female CAN users. We also found robust associations between aerobic fitness and greater inferior parietal, rostral middle frontal, inferior parietal, fusiform, precuneus, cerebellum, and caudate volumes in both CAN users and nonusers. Greater volume in these regions was linked with superior neuropsychological performance. These findings, coupled with existing literature, suggest that aerobic interventions may be a potential low-cost ameliorative tool in the recovery of chronic and repeated CAN use. Taken together, we found that sex and aerobic fitness may be factors that help explain heterogeneity in findings and future studies examining the impact of CAN use on brain volume need to consider these significant factors. Additional prospective and longitudinal studies, such as the ABCD Study® (Lisdahl et al., Reference Lisdahl, Sher, Conway, Gonzalez, Feldstein Ewing, Nixon and Heitzeg2018), are needed to confirm causality and replicate findings.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S135561772100062X
ACKNOWLEDGMENTS
This work was supported by the National Institute on Drug Abuse to K.M.L. (R01 DA030354, U01 DA041025, R21 DA049109) and the National Institute on Alcohol Abuse and Alcoholism to N.E.W. (T32AA013525; P.I.: Riley/Tapert to NEW). Requests for data are welcomed and can be submitted to the corresponding author.
CONFLICTS OF INTEREST
There are no conflicts of interests to report.
