Cannabis is one of the most commonly used substances in the United States with approximately 24.6% of adolescents (grades 8, 10, 12) and 42.0% of young adults (aged 19–28) reporting past year use (1, 2). Due to increased prevalence rates in this age range, the scientific community is increasingly interested in understanding the impact of repeated and regular cannabis use on adolescent and young adult neurodevelopment.

Exogenous cannabinoids primarily consist of cannabidiol and delta-9-tetrahydrocannabinol (THC). THC use can impact the brain by interacting with the endogenous cannabinoid system (3), which includes cannabinoid receptor 1 (CB1) (4) and is primarily distributed throughout the central nervous system (5). Receptors for this system are principally involved in neuromodulation (6) and are at peak density throughout adolescence, particularly in prefrontal and limbic regions (7–9). Repeated and regular use of THC can affect CB1 binding (10) and in turn, affect downstream functional and structural changes to the brain (11). To that end, frequency of cannabis use has been related to significant reductions in the density of CB1 receptors throughout the cortex, with recovery reportedly occurring after one month of abstinence (12). Moreover, chronic exposure to cannabis has been related to alterations in cognitive control, inhibitory control, affective processing, and reward functioning (i.e., continuing beyond approximately 1-month of abstinence) (11, 13–15). Albeit the link between chronic cannabis use and affective reward processing following sustained abstinence in adolescents and young adult has yet to be fully characterized. Measuring these affective properties after abstinence is important to distinguish from short-term, relatively acute effects of cannabis, vs. longer-lasting impacts that may continue to disrupt treatment outcomes or results in impairments even after abstinence is achieved. Furthermore, this may reveal specific neuronal networks that are sensitive to longer-term impacts of chronic cannabis exposure during the adolescent and young adult years.

Among typically developing adolescents and young adults, reward circuitry centers are broadly characterized by strong signaling as they attune to evaluating and appraising rewarding stimuli (16, 17). Important regions for higher-order processes of inhibitory control (i.e., one's prepotent ability to withhold immediate responding) (18)—including the prefrontal cortex—remain structurally and functionally underdeveloped relative to subcortical areas (19, 20). During this dyssynchronous development, it is suggested that reward signaling may be less downregulated by cortical control systems resulting in increased impulsive or reward-driven outward behavior during adolescence (21, 22). Somerville et al. (23) examined this phenomenon in adolescents and young adults with an emotional Go/No-go paradigm examining positive (i.e., “happy”) faces, where teens (aged 13–17) displayed stronger activation in the ventral striatum—a subcortical region characterized by its role in the dopaminergic system related to reward processing—compared to children and adults (23). Throughout adolescence, the ventral striatum—along with other reward-related regions—experiences an uptick in dopaminergic receptors (24, 25) and are recruited differentially depending on the context of reward processing. The striatum has been dually associated with an increase in functional activation when responding to a reward (26, 27) and a decrease in functional activation when anticipating a reward (26, 28, 29). With the onset of substance use generally occurring during adolescence (1), elucidating differences in reward processing and inhibition as it pertains to substance use risk and consequences has garnered greater attention in the broader field of addiction neuroscience (30).

It has been demonstrated both across the animal and human literature that introduction of substance use alters and reprograms key reward appraisal regions by devaluing standard rewards and thus, rewards of more salient value are required to recruit reward functioning that mirrors similar levels of activation compared to non-using counterparts (31–34). In preclinical studies, THC has been shown to decrease sensitivity to rewards and result in aberrant functioning of the reward systems (i.e., the mesolimbic dopamine system, including the ventral striatum) (35–39). Within humans, functional connectivity at rest between these reward regions is disrupted in cannabis-using adolescents and young adults, with non-dependent individuals demonstrating increased connectivity in prefrontal regions and dependent individuals exhibiting more subcortical region connectivity (15, 40, 41), with broad functional connectivity differences that are partially recovered after one month of abstinence in young adult cannabis-using males (42). Yet, less is known on whether functional differences—while task-engaged—may persist following abstinence and could theoretically interfere with treatment outcomes and interpersonal functioning.

Functional responding in these reward regions are typically assessed through the Monetary Incentive Delay (MID) task (26, 27, 43), with pronounced striatal hypoactivation seen when anticipating reward (44–46) and striatal hyperactivation during reward appraisal (45, 46) [see, Luijten et al. (30) for review]. This research indicates that cannabis use interplays with reward systems through exaggerated responding when engaged in reward processing, however, these findings are largely exclusive to the MID task and there is a paucity of research examining differences in functional reward responding to other pleasurable cues outside of monetary appraisal—such as through an emotional faces go/no-go task (23). The latter may be more ecologically valid in the context of social reward processing and interpersonal relationship functioning, concepts important for both healthy adolescent/young adult development and treatment outcomes (47, 48). Yet, minimal studies have investigated emotional go/no-go in the context of adolescent substance use, despite the focus in other prevalent psychopathologies in adolescence (e.g., depression, anxiety) (49–51). Moreover, considering the mentioned context-dependent functional connectivity differences in cannabis-using samples on fMRI task data (40, 41), continued investigation of functional co-activation while engaged in eliciting paradigms is needed. Thus, examination of reward responding activation and subsequent co-activation with reward-related areas—identified in control samples (i.e., ventral striatum) (23)—may elucidate potential risks for or effects of cannabis use on brain function.

Within investigations of cannabis use and emotional faces paradigms, studies have typically focused on negative affective stimuli processing, with cannabis-using groups showing amygdala reactivity and aberrant brain activation particularly in frontal and cingulate regions (52–55). Comparatively, only a few small-sample studies have examined functional neuroimaging differences between cannabis-using and non-using when viewing positive affective stimuli. Gruber et al. (52) examined functional brain response to positive face viewing in a sample of cannabis-using adults, where more activation was observed in cingulate region and less activation was observed in the left temporal lobe and amygdala when compared to non-using adults. Conversely in other studies, cannabis-using adults displayed no differences in brain activation when viewing positive situations compared to non-using adults (56); and, no associations were observed between functional brain activation to positive faces and cannabis use disorder scores in adolescents, albeit, this treatment-seeking sample comprised of increased psychopathological comorbidity prevalence (57). As these studies comprise of passive stimuli viewing, no studies have examined positive face processing while also engaging in inhibitory control contexts (i.e., Go/No-go), despite broad aberrant inhibitory findings (13, 55, 58–61). Investigations on the interplay between positive face processing and inhibitory control may shed light on problematic substance use development into adulthood, as this interplay is a key aspect of withdrawal stages within working models of addiction (33, 47). In addition, given the neurodevelopmental timing of our sample (i.e., adolescents to young adults) (23), investigating positive face processing—in the context of abstinent cannabis use—has important implications for potential downstream effects on reward-related regions (25) which have varying functional activation dependent on context of rewarding stimuli (26).

Therefore, the aims of the study were to: (1) Examine differences in whole-brain blood oxygenation level dependent (BOLD) imaging to positive faces between abstinent cannabis-using and non-using adolescents and young adults during an emotional go/no-go paradigm in both inhibitory and approach contexts (13, 23). (2) Conduct context-dependent functional analyses to examine ventral striatum activation and connectivity differences between cannabis-using and non-using groups. Whole-brain approaches were employed due to uncertainty in which specific regions may differ between groups during affective reward processing and inhibitory control while processing positive affect; we hypothesize that cannabis-using groups would demonstrate increased BOLD response in ventral striatal regions during passive viewing (i.e., Happy Go) and inhibitory control (Happy No-Go) conditions, due to aforementioned literature on reward in cannabis use. These aims were investigated in a group of nontreatment-seeking cannabis-using adolescent and young adults with no psychopathological comorbidities who underwent at least two-weeks of monitored abstinence, to mitigate effects of acute cannabis ingestion, leading up to functional imaging acquisition.

Cannabis is one of the more commonly used substances in the United States (1, 2), thus, understanding the impact of cannabis use on neurodevelopment amongst adolescents and young adults is of increasing importance. Prior research has shown aberrant associations between cannabis use and reward circuitry activity (30, 40, 41, 44, 45); yet minimal studies have investigated response to positive stimuli during inhibitory control conditions in abstinent cannabis-using adolescents and young adults. In the present study, we found that cannabis-using participants demonstrated increased BOLD response in left middle cingulum and decreased BOLD response in left supplementary motor area while engaging in positive viewing (i.e., happy Go), despite similar behavioral performance. In positive inhibition (i.e., happy No-go), cannabis-using participants displayed decreased activation in a cluster within the left superior frontal region, relative to controls. In addition, an ROI activation analysis revealed abstinent cannabis-using participants demonstrated hyperactivity during engagement of positive stimuli and hypoactivity during inhibition of positive stimuli within the ventral striatum compared to controls; although, when conducting a context-dependent connectivity analysis, no connectivity differences were observed.

Cannabis-using adolescents and young adults displayed increased activation in the left middle cingulum and ventral striatum, and decreased activation in left supplemental motor area relative to control participants while freely responding to positive faces (i.e., happy Go). The cluster identified as the left middle cingulum is positioned primarily within anterior cingulum white matter tract, which corresponds with increased cingulate activity when viewing positive faces (52), though, also shares proximity with the inferior parietal region. Notably, the inferior parietal region has been implicated in cannabis fMRI investigations, demonstrating increased BOLD responding in Go conditions among adolescents (61), decreased BOLD responding in positive word conditions in young adults (83), and aberrant inferior parietal-cerebellar connectivity on a Go/No-go task (84, 85). As this cluster additionally sits within the anterior cingulum bundle, it is worth discussing the relevance of this primarily white-matter activity finding as well given that white matter exhibits activation and functional BOLD differences are worthy of documentation (i.e., instead of the approach to censor for solely grey matter activation) (86). It is possible that activation within these white matter may represent broader recruitment of cingulum and surrounding regions for task performance, and their corresponding communicating fibers, however, this should be further investigated using network-based analyses of task-related BOLD data to comprehensively examine gray and white matter contributions to these group differences. Notwithstanding, research has identified impacted cingulum development due to cannabis use across young adults (87) with reduced structural volumes observed (88, 89). Notably, the anterior cingulum plays a role in both executive functioning and emotion processing (90) and has been implicated in Go/No-go task performance within community samples (91). Follow-up analyses should continue to investigation structural/functional network-related characteristics of cannabis use disorder in adolescence.

Analyses additionally observed increased ventral striatum activity during positive response conditions, which corresponds with hyperactivation seen during the reward outcome trials on the MID task among cannabis-using groups (30, 45, 46). This may possibly represent shared activation patterns for positive facial stimuli and monetary reward stimuli. Hyperactive regions observed within our cannabis group may be indicative of increased recruitment in order to complete the task similar to their non-using peers (i.e., maintain rule set to initiate motor response to positive faces); alternatively, increased activation in cingulum and ventral striatum regions could demonstrate sensitized responses to reward processing within cannabis-using groups, despite similar behavioral performance and abstinence length. Lastly, decreased left supplemental motor activation observed is consistent with findings specifically on tasks requiring motor response (92, 93). As all three regions (left middle cingulum, left supplemental motor, and ventral striatum) are rich in CB1 receptors (94, 95), aberrant activation in these areas may be suggestive of compensatory mechanisms by which cannabis-using participants are more heavily recruiting middle cingulum and ventral striatal regions due to positive nature of task condition, while non-using controls more readily engage supplementary motor areas to respond to positive stimuli.

Decreased left superior frontal and ventral striatum activations were observed in positive response inhibition (i.e., happy No-go) for cannabis-using participants relative to controls. Studies of BOLD activation elicited by inhibitory tasks have shown that cannabis-using groups generally display aberrant BOLD response in frontal regions (59, 61, 84, 96). In a similar sample, Wallace et al. (13) found increased activation in left frontal gyrus when examining response inhibition to calm (i.e., neutral) faces. Further, given the mixed findings on BOLD response elicited by positive stimuli (52, 56, 57), we demonstrate a unique pattern of left superior frontal activation in inhibitory trials with added component of inhibition to positive stimuli. While we do see in non-using teens that more frontal activation is recruited for successful inhibitory response to positive faces (21, 23), it is hypothesized that cannabis-using participants may not be actively recruiting frontal regions; further evidenced by connectivity studies of reward paradigms, where frontal connectivity network differences are observed even following a period of monitored abstinence (40, 41). Hypoactivation was observed in positive inhibition trials which additionally aligns with previous MID studies of reward anticipation (30, 44–46), however, as this task did not involve “anticipation” per se, these findings may instead represent discrepant BOLD responses elicited by positive inhibition. Findings could be attributed to more effortful inhibition of the ventral striatum when cannabis-using individuals are asked to inhibit response to positive (i.e., rewarding) faces in order to complete the task (relative to the non-using literature 97). Overall, the positive inhibition findings may represent that when cognitive control components are added it could disrupt typical reward processing. Future studies with tasks engaging both inhibition and reward processing are needed as these could present more ecologically valid day-to-day interactions with rewarding stimuli.

Interestingly, we did not observe connectivity differences with bilateral ventral striatal activity and other cortical regions between cannabis-using and non-using participants, despite observing group differences on ventral striatal activity—which was dependent on task condition. Null findings align with ventral striatal connectivity differences that are largely recovered after one month of cannabis abstinence in young adult males (42). Indeed, our own sample—on average—had 31 days of abstinence at the time of MRI. Thus, it is possible that fronto-striatal connectivity differences are resolved after at least two-weeks of monitored abstinence, and differences would be more evident with shorter durations of abstinence (15, 40, 41). Yet, the present analyses did observe varying ventral striatal activity dependent on task condition as described above. Correspondingly, research on the ventral striatum shows differential activation dependent on emotional valence (greater deactivation for inhibition to neutral compared to positive lexical stimuli) (97) and hyperactivation for positive faces in Go trials compared to No-go trials in teens (23). These findings lend evidence to a larger theory of dysregulated ventral striatum activity within cannabis-using samples across fMRI tasks (30), positing the ventral striatum in either the risk for or consequence of cannabis use in this age range (98, 99).

Taken together, we observed discrepant BOLD response elicited by positive face responding in cannabis-using participants compared to controls both in various cortical regions necessary for task completion and specifically in reward-related structures implicated in positive task conditions (i.e., ventral striatum) (21, 23). These findings are evident following a monitored period of abstinence and thus may represent a consequence of prolonged cannabis use during these developmental periods and could ameliorate after longer sustained abstinence (12). More broadly, aberrant functional activation seen in cannabis-using individuals relative to controls across fMRI tasks has been implicated in the literature previously, particularly in tasks requiring increased attention, inhibition, and/or set shifting (11, 100, 101). This further posits the aforementioned mechanism of compensation—through dysregulation of CB1 receptors (12, 61, 102, 103)—to exhibit similar behavioral performance relative to their non-using peers, particularly in these high-density CB1 regions (94, 95). Specifically, significant regions in the present analysis are implicated in affective processing paradigms (21, 23, 90–92) and facilitate other task completion processes (e.g., motor component in Go trials), possibly lending evidence towards compensation. Additionally, a novel and interesting interplay between reward processing and inhibitory control is achieved with this task and more studies are needed to ascertain the contributions of each process in this relationship. Importantly, the culmination of these findings could have important implications for downstream impacts to treatment outcomes or interpersonal functioning, specifically in this neurodevelopmental period.

Notably, we did also observe unique covariate findings of interest. Past year alcohol use was associated with increased BOLD activation in right postcentral and bilateral cingulum regions in positive responding, relative to neutral responding, conditions. Most research on BOLD response to rewarding cues in alcohol-dependent individuals relates to alcohol-related cues; the present findings add-on to the literature demonstrating aberrant frontal activation as it relates to positive cues (104–106). Additionally, higher levels of cotinine at the time of scan were associated with BOLD activation decreases in left medial orbitofrontal and increases in right superior frontal regions. Minimal research has been conducted on nicotine in positive face responding, but reward-related findings have implicated blunted striatal reward (107) and generally altered amygdala function (108). Thus, these frontal cotinine findings observed demonstrate continued aberrant findings related to nicotine levels in frontolimbic regions. These provide rationale for the need of cannabis analyses to control for patterns related to polysubstance use. Further, these outcomes bolster general findings on aberrant brain activity in affective contexts seen across substance use more broadly (109, 110).

It is worth noting limitations of the present analysis. Notably, causality cannot be determined from the sample due to cannabis use initiation occurring prior to the study protocols. Longitudinal studies can assess functional activation elicited by affective processing tasks in substance-naïve youth and determine whether this represents a risk factor for cannabis use or is affected by use. Further, these studies can also investigate developmentally-dependent (i.e., age-related) changes in functional activation across adolescence to young adulthood in a larger longitudinal sample. Second, the sample was balanced for aerobic fitness and excluded individuals who could not undergo acute aerobic fitness measurements; this decreases generalizability to adolescent and young adult populations who may be primarily sedentary. Third, the study excluded for those with mood-related comorbidities, which allowed for more conclusive implications on cannabis use's association with positive responding and inhibition but may be more varied in those with comorbid depressive disorders (111, 112). In line with the above, a sample willing to abstain from cannabis was recruited in order to examine cannabis effects that remain after withdrawal, this limits generalizability to individuals who may be unwilling or unable to sustain abstinence, such as those with more severe cannabis use disorder. Fifth, it is possible that gPPI connectivity analyses were underpowered (113) to detect small effect sizes and might explain null connectivity differences between cannabis-using and non-using participants. Finally, cannabis metabolites cycle out within a three-to-four week period (114); thus, future studies are needed to determine whether differences exist at the acute stage (i.e., no abstinence period) and if subtle differences would recover with longer periods of sustained abstinence.

The current study found that after at least two-weeks of monitored abstinence, cannabis-using adolescents and young adults displayed increased BOLD response in left middle cingulum and ventral striatum, and decreased BOLD response in left supplemental motor areas in positive Go conditions. In addition, decreased BOLD response in a left superior frontal cluster and ventral striatum in positive No-go conditions was observed in cannabis-using participants. These aberrant BOLD activations align with previous findings and may further implicate compensatory theories within cannabis-using individuals who are recruiting other functional regions to assist in responding to positive stimuli on a Go/No-go task or, may represent increased activation when responding to positive cues and disrupted functional response when an inhibitory control component is added. Ventral striatum outcomes correspond with monetary-incentive investigations (30), while also representing a novel result in cannabis research relative to the neurodevelopmental adolescent literature on positive affective processing (23). Further, we did not observe significant context-dependent connectivity differences between the ventral striatum and other cortical regions (15). Overall, these findings, coupled with the existing literature, suggest that BOLD activity elicited by either engagement or inhibition to positive faces—even following at least two-weeks of monitored abstinence—differs between cannabis-using and non-using adolescents and young adults. Jointly investigating response inhibition and reward processing in cannabis-using individuals may present more ecologically valid methods of examining contributions to escalation or continuation of use into adulthood, as this interplay is described within withdrawal stages of addiction models (33, 47). Future prospective, longitudinal studies, are needed to further elucidate the causal relationship between escalating cannabis use and functional activation elicited by positive stimuli, specifically examining the contributions of the ventral striatum in this relationship.