The aim of this systematic review was to assort the recent literature addressing the effect of cannabis use on impulsivity, disinhibition, and motor control. To the best of our knowledge, this is the first systematic review specifically disentangling the effect of cannabis on impulsivity or disinhibition.

Because impulsivity is a multifaceted construct (cf. [45]), comparisons between different study designs have to address the different operationalization used in that construct. Two facets of impulsivity often predominate research on impulsivity in the context of drug abuse [46], namely impulsive choice and impaired inhibition (see also [47]). In the current review, we included studies on cannabis and impulsivity measuring the following domains: response inhibition, motor control, interference/suppression of information, and executive functions of divided/sustained attention and decision making/risk taking. Neurophysiological definitions of impulsivity share all these aspects (see introduction). The focus of this review was to assess structural and functional imaging studies examining the effect of cannabis on impulsivity. We analyzed a broad cross-section of studies which each applied at least one impulsivity measure. These included acute and non-acute designs, different abstention periods and different consumption intensities.

On the behavioral level, the included studies did not find performance differences on inhibition in cannabis users, but trends towards faster reaction times and higher rates on commission errors were observed [34]. In contrast, behavioral studies of inhibitory and motor control found deficits of inhibitory control in cannabis users under acute exposure [48] [49, 50] [32, 51, 52] 53, and in regular users after abstention of use [36, 54] 55. Duration, time of onset and total lifetime amount of cannabis use, as well as hits per month have a detrimental impact on executive functions [56] [57] [58]. Earlier onset, longer duration, and heavier usage have been associated with altered structural frontal integrity and decreased activations in users compared to non-users in the superior and middle frontal gyrus [34, 37]. Behavioral studies showed impulsive decision-making in cannabis users under acute [32, 50, 59] and non-acute conditions [54] [60], while others did not find significant differences [61] [62] [53]. In recreational users a positive correlation between reduced volume in the medial prefrontal cortex and age of first marijuana use was found [39] while chronic users reported higher scores on the “non-planning” but lesser scores on the “motor impulsivity” subscale of the BIS-11 [38]. Put together, mere behavioral studies show detrimental effects of cannabis consumption on different domains of impulsivity even after prolonged abstention. While the behavioral data of neuroimaging studies show only trends, there are correlations between severity of consumption and structural and functional alterations in these studies.

Beyond behavioral findings, we also reviewed neuroimaging data with a focus on alterations in dorsal and ventral PFC regions including the dorsolateral (dl-), ventrolateral (vl-), ventromedial (vm-)PFC, subgenial, dorsal, or rostral ACC, inferior frontal gyrus (IFG), and orbital frontal cortex (OFC), because these brain regions have been proposed to be critically implicated in drug addiction [24].

PET studies of acute administration of cannabis or delta9-THC found pre- and post-treatment increases in rCBF and brain metabolism in the PFC (mesial, orbital, middle and medial frontal cortices) and ACC during task performance (O’Leary et al.,2003, Weinstein et. al., 2007), indicating an increased brain effort to perform equally (see also [63]). The only fMRI study we reviewed here found significant bold signal decreases after delta-9-THC and CBD administration among healthy non-smoking volunteers in the ACC and in the inferior frontal gyrus during response inhibition [14]. These functional brain imaging data suggest that delta-9-THC acutely attenuates the engagement of brain regions that mediate response inhibition. Independent from pre- to post-treatment increases in rCBF, between-group comparison showed significant less frontal activations in chronic versus occasional users [33]. These results support the iRISA model proposed by Goldstein and Volkow [24] models by showing that acute cannabis exposure reduced frontal activations in chronic users and that delta9-THC has a direct diminishing impact on brain areas relevant for executive control.

With respect to non-acute studies of cannabis administration in two fMRI studies, chronic and recreational cannabis users showed greater activation and a more diffuse and bilateral activation pattern of the dlPFC and middle frontal gyrus during inhibition compared to non-users [41, 42]. More diffuse and bilateral activations in users compared to more focal and lateralized cluster in non-using controls may reflect a compensatory brain effort mechanism [42] to achieve similar inhibitory control.

Attentional functions involved in error detection, interference and stop signaling are mediated mainly by the ACC. Reduced ACC activity has been linked with altered metabolisms in cingulate cortex (e.g. [79, 80]). There is evidence for reduced focal ACC activity during task performance in chronic cannabis users [35, 41, 43]. Chronic users also show a more diffuse activation pattern in the cingulum among different research groups with greater activations in the middle compared to the anterior cingulate cortex, and together with activations in parietal regions [34, 35, 41, 42]. This favors the hypothesis of additional recruitment of parietal cortices to maintain attentional competence in order to equally perform during task. Garavan et al. [79] proposed two interacting cortical systems of response inhibition based on fMRI and EEG data depending on high- or low-absent-mindedness: recruitment of right prefrontal and parietal regions on the one hand, and activation of cingulate regions on the other hand. Hence, different cingulum activations and additional parietal activation in cannabis users would represent difficulties or inefficacies to sustain sufficient attention when inhibiting responses.

Abnormal monitoring abilities may furthermore affect decision-making [34, 35, 41, 42] and this corresponds with reduced activations in right dlPFC and right lateral OFC [44]. Cannabis users also show diminished choice optimization and decreased functional activation of orbital prefrontal regions during performance of the IGT [81]. Regions of interest in fMRI research of the IGT are the OFC and dlPFC [82, 83]. The OFC is involved in reward attribution and altered in addiction [84, 85]. Poor decision-making is a hallmark of addiction, and the IGT can discriminate controls from patients also with ventral medial lesions [86, 87], substance-dependence, or pathological gamblers [44, 88, 89].

Frontal correlates of response inhibition are linked to the pre-supplementary motor cortex to simply suppress unwanted movements, while the dlPFC needs to be recruited during tasks involving higher working memory load [64, 65]. Inhibition-related functions have been found to be significantly heritable [66-68] [69]. It has been shown, by using a variety of imaging modalities, that the prefrontal cortex (most prominently the inferior prefrontal cortex and the supplementary motor area (SMA), as well as subcortical striato-thalamic projection targets), represents a neural correlate of response inhibition [70-78]. The neuroimaging findings reviewed here indicate both reduced task-specific and predefined region-specific (region of interests - ROI) activations in PFC, OFC and ACC in acute as well as non-acute studies, and greater activations in a more diffuse, bilateral pattern together with supplementary parietal cortex recruitments. On the behavioral level, neuroimaging data showed no significant differences, but tendencies in impairments towards task complexities as well as correlations with earlier onset and heavier consumption. Thus, cannabis under acute and non-acute condition attenuates the focal engagement of inhibitory networks, namely the right dorsolateral and inferior as well as the orbital prefrontal cortex and ACC. To perform equally as a compensatory mechanism, cannabis leads to bilateral and broader, more diffuse activations depending on the specific task. To maintain attentional competence, greater activation occurs in the middle, while a broader, more diffuse pattern of activity is found within the anterior cingulum. Supplementary parietal activations have also been found. When cognitive load in the task is higher, chronic users decompensate and have less PFC activation (hence prefrontal control) than occasional users. Impulsive decision-making is associated with reductions in dlPFC and OFC.

Out of the five structural imaging studies one found a reduction in the medial prefrontal cortex in recreational users [39]. Recently, findings from structural neuroimaging studies have found positive, negative or even inverse results of persisting alterations on brain morphology after long-term marijuana use that underlies deficits in attention, learning, memory, executive functions and emotional processes [5, 90, 91], for review see [92]. Abnormalities in neuropsychological testing and fMRI activation are consistent with imaging studies demonstrating morphological changes in cannabis users [5]. Some reports have described reduced grey matter in limbic areas [92, 93] and abnormal gyrification [94]. Recent reviews of structural neuroimaging studies revealed no significant differences on global measures of brain volume following cannabis use [5]. Region-specific brain volume alterations have been found in different areas, most consistently in hippocampal and parahippocampal regions, but were not consistent across studies [92, 95]. Inconsistencies might arrive from different imaging modalities, as well as from different analysis approaches such as region-of-interest (ROI) or whole brain voxel-based morphometry.

DTI of white matter microstructures has identified reduced fractional anisotropy (FA) in cannabis users [37, 41]. In accordance, increases of axonal diffusivity have also been found in fronto-temporal brain connections [96]. Furthermore Arnone et al. [97, 98] found a marijuana-related damage in the corpus callosum. Chronic marijuana use may result in reduced FA and increased diffusivity, which may be associated with increased impulsivity, and ultimately contribute to the initiation of marijuana use or the inability to discontinue use [37].

Morphological changes of neuronal structures that are rich in cannabinoid receptors have been reported most consistently in the hippocampus [99] [100, 101] and the cerebellum [102, 103]). However, Zalesky et al. [104] found no evidence for brain-wide cannabis-induced shortening of axonal fibers except in the right fimbria of the hippocampus (fornix), splenium of the corpus callosum and commissural fibers. Another study used (¹H-MRS-)markers to investigate neurogenesis, synaptogenesis (NAA), and synaptic plasticity (Cho) in recreational cannabis users and found reduced neuronal and axonal integrity in the dlPFC and ACC. Thus, there is growing evidence of structural disintegration after prolonged cannabis exposure not only in whole-brain analyses, but also specifically in brain regions that are crucial for impulse control [29]. Whether structural alterations remain persistent after prolonged discontinuation of cannabis use is still in debate. There are hazardous white matter connectivity impacts when the developing brain is exposed to long-term cannabis usage, but these alterations could be reversible after abstinence or functional adaptation [104].

In general, reviews show either the effects of cannabis on brain structures in psychiatric patients, individuals at a high risk of developing psychosis, or the interplay between psychosis and co-morbidities. These data may be distinct from non-psychiatric samples [5,91,95,105]. Nevertheless, cannabis use in clinical subjects is associated with decreases in some brain regions, particularly those rich in CB₁–receptors such as the cingulum, the dlPFC, the cerebellum, and striatum. According to the iRISA model, the cingulum, PFC and striatum are crucial regions of the inhibition network required to monitor, suppress and countermand hot drives and to instill cold, volitional control. Chang and Chronicle [106] reviewed the chronic effects in cannabis users and found equivocal evidence that it might result in structural brain changes, blood-oxygenation-level-dependent-fMRI studies consistently show alterations, or neuroadaptation, in the activation of brain networks responsible for higher cognitive functions, but whether changes are reversible with abstinence was unclear. Likewise Chang and Chronicle [106], Batalla and colleagues [107] reviewed neuroimaging data in chronic adolescent and adult cannabis users with no other psychiatric disorder. Probands had to have consumed cannabis at least several times per week for a minimum of two years. Morphological alterations were found in medial, temporal and frontal cortices, as well as in the cerebellum. They also found different patterns of resting in global and brain activity during different task performances, which may also have indicated compensatory effects in chronic exposure. In contrast to that review, we focused on the impact of cannabis use on impulsivity measures and included acute and non-acute studies of recreational drug users and drug-naïve individuals. Comparisons of studies are often limited because of the differing definitions of cannabis exposure. Only a few studies provide absolute numbers of consumption in order to use cannabis abuse as a continuous variable. This is important when taking into account the possibility of a threshold effect of cannabis on brain morphology. Furthermore, we reviewed the literature with respect to impulsivity, using the iRISA model as a framework, to provide a conclusive overview of neuroimaging data regarding the impact of cannabis on brain structures and the functionality of habitual control networks in humans.

Impulsivity is a proposed endophenotype for substance dependence as well as a possible consequence of prolonged drug use [108]. Drugs of addiction and subconscious craving are likely to increase impulsiveness due to the loss of frontal cortical inhibition of impulses and increased limbic drive (see [24] [109]). Neural substrates of impulsivity encompass circuitry involving both cortical (top-down volitional control) and subcortical mechanisms, particularly within the basal ganglia, and have mainly been studied on response inhibition in humans and experimental animals. Although there are insights into the interplay of compulsive stimulant taking and impulsivity, this remains unclear in the case of cannabis. However, there are behavioral and epidemiological findings, which emphasize such a relationship, as well as the fact that cannabis use may lead to a loss of attention, reflection and inhibitory control. Associated cerebral activations consist of “right lateralized fronto-parietal” attentional or response selection mechanisms, as well as “midline performance monitoring processes” (ACC). To stop already pre-processed activation inferior frontal region are involved to countermand responses (see [80]). Thus, giving the broad epidemiological “burden of society” cannabis leads to dose-related impairments of psychomotor performance with implications for car driving, and personal health risk [110].

Limitations of this review include the analysis of different user types (heavy vs. recreational), small study groups, and a failing to control for co-use between studies. Whole brain search strategies often fail to show significant results, but localized differences may be found in connectivity analyses in future structural and functional neuroimaging studies.