Cannabis is the most widely used ‘illicit’ drug in the United States, with nearly 20 million people reporting past month use (Substance Abuse and Mental Health Services Administration, 2011). Acute, detrimental effects of Δ9-tetrahydrocannabinol (THC), the primary psychoactive compound in cannabis, on performance of cognitive and psychomotor functioning are well documented. In double-blind, placebo-controlled studies, oral administration of 40–300 μg/kg THC caused acute, dose-dependent impairment in performance on memory, divided and sustained attention, reaction time, visual tracking and motor function tasks (Hall and Solowij, 1998; Leweke et al., 1998; Ameri, 1999; Hampson and Deadwyler, 1999; Curran et al., 2002; Lichtman et al., 2002; D’Souza et al., 2004; Ramaekers et al., 2004). Many of these tasks rely on the prefrontal cortex (PFC), which is heavily involved in processes such as executive control, working memory, novelty processing, attention and awareness, and integration of multimodal sensory information (Knight and Stuss, 2002). Cannabinoid 1 (CB1) receptors are located throughout the PFC, as well as in the striatum, amygdala, and cerebellum (Gardner, 2005). Therefore, studying the effects of THC on blood flow in the PFC during cognitive tasks testing any of the above processes is essential to increasing our understanding of how cannabis impacts cognitive performance.

Most brain imaging studies of acute cannabis exposure have demonstrated global cortical activity increases during administration of smoked cannabis or infused THC (Mathew et al., 1997, 1992). Mathew et al. (1992, 1997), using Positron Emission Tomography (PET), have consistently reported increased bilateral frontal lobe activation following administration of smoked cannabis or infused THC. In these studies, ratings of intoxication correlated with right hemispheric cerebral blood flow (CBF) increases (Mathew et al., 1997). Increases in blood flow to the PFC after a THC infusion are particularly pronounced in chronic marijuana users (Volkow et al., 1996).

In the current study, we examined changes in the PFC associated with THC administration during a working memory task using functional near-infrared spectroscopy (fNIRS) in 13 regular cannabis users. Functional NIRS is a portable, non-invasive, and inexpensive imaging modality that detects brain activity through local hemodynamics, similar to the BOLD activation measured by fMRI (Strangman et al., 2002; Keles et al., 2016). fNIRS has several advantages over other imaging modalities such as PET and functional magnetic resonance imaging (fMRI), in that fNIRS is economical, generally well-tolerated (e.g., participants can remain comfortably seated), and does not expose participants to radiation like in PET. Its main disadvantage is that near-infrared light pulsed into the forehead can only refracted from the tissues of the cortex up to depths of 3 cm, so that this method cannot be used to investigate subcortical regions. Because of this limitation, fNIRS is not as widely used as fMRI or PET, though fNIRS has been used to study acute effects of both medication [methylphenidate in children with ADHD; (Monden et al., 2012)] and addictive substances [heroin and methadone compared to saline in opiate-dependent and healthy subjects; (Stohler et al., 1999)]. To our knowledge, the current study is the first to use fNIRS to investigate functional changes in the human brain associated with THC administration. The goal of this preliminary study was to use fNIRS to detect potential changes in the human PFC associated with THC administration during a working memory task. Based on PET studies of cannabis intoxication (e.g., Mathew and Wilson, 1993; Volkow et al., 1996; Mathew et al., 1997, 1999) we hypothesized that, using fNIRS, we would observe significantly increased hemodynamic response during a working memory task after THC administration.

Study procedures were approved by the Partners Human Subjects Committee (IRB approval #2015P001516). All participants completed an informed consent process and signed a written informed consent form prior to initiation of study procedures. Participants were compensated $100, which was mailed to their residence in the form of a check after completion of the study.

Eighteen participants completed consent, met eligibility criteria and were enrolled. Five participants were excluded from the final analysis; three received a very low dose of dronabinol (5 mg) because their baseline blood pressure was >140/90; two subjects completed study procedures but their data were excluded from the analysis due to heavy motion artifacts in the fNIRS signal. The final cohort consisted of 13 cannabis users (10 males, 3 females, mean age 24 ± 6.75 years) who completed the fNIRS sessions. Please see Table 1 for participants’ baseline demographic and clinical information. Quantitative urine analysis of THC-COOH was implemented after the first three participants, and therefore, quantitative measures of circulating THC are only available for 10 participants.

To our knowledge, this is the first report of fNIRS to examine the effect of THC administration on prefrontal hemodynamic changes during a WM task. In this preliminary study, we observed a significant increase in HbO concentration after THC administration in several channels on the right PFC during 2-back WM condition. Though behaviorally we did not observe a performance decrement (likely due to the small sample size of 12 participants), increase in HbO may indicate that maintaining task performance was more difficult after THC administration.

An increase in HbO is consistent with the majority of brain imaging studies of smoked or orally administered cannabis. Previous studies have found that many effects of THC, including subjective levels of intoxication (Mathew and Wilson, 1992, 1993; Mathew et al., 1992), dissociative experiences, measures of depersonalization (Mathew and Wilson, 1993), measures of confusion (Mathew and Wilson, 1993), and changes in heart rate (Mathew and Wilson, 1992), correlate with increased global CBF after smoking marijuana cigarettes. Additionally, increased global CBF has been correlated with plasma THC levels (Mathew and Wilson, 1992). Task-based fMRI studies have also reported an increase in rCBF while participants were performing a cognitive task after acute THC (Borgwardt et al., 2008; Fusar-Poli et al., 2009).

Though the current study only assessed PFC changes, previous studies have indicated regional variation in the effects of THC throughout the brain. O’Leary et al. (2002), using PET, reported increased rCBF in the orbital and mesial frontal lobes (as well as in the insula, temporal poles, anterior cingulate, and cerebellum), but did not find mean global (whole-brain) CBF changes after THC administration. In fact, rCBF decreased after THC administration in auditory regions, visual cortex, parietal lobe, posterior frontal lobe, and thalamus, and rCBF was unchanged in the nucleus accumbens, basal ganglia, and hippocampus (O’Leary et al., 2002). The authors suggest that regional differences provide evidence that increased blood flow in the PFC specifically may in part underlie changes in executive functioning and/or mood that occur after THC exposure. O’Leary et al. (2007) also demonstrated that PFC increases were independent of the cognitive task being performed, indicating that changes in blood flow were likely the direct effect of the neurophysiological response of the brain and neurovascular system to THC, rather than task-related. Resting-state studies also support this conclusion (Mathew et al., 1992, 1997). In the current study, we generally observed increased signal during both the 0-back and 2-back portions of the task, though significance was only shown during the 2-back portion. This indicates that THC likely increases blood flow to the PFC irrespective of task difficulty, but this increased blood flow may be more pronounced when participants expend more effort.

Though we observed increased signal throughout the PFC, only channels in the right lateral orbitofrontal cortex (OFC) showed significantly increased blood flow after corrections for multiple comparisons. Other studies have demonstrated that cannabis users show altered structure (Cheetham et al., 2012) and function (Bolla et al., 2005) of the OFC specifically, raising the hypothesis that this region may by particularly susceptible to the effects of THC. The OFC is part of a neural network that underlies many processes in addiction and reward (Grant et al., 1996; Wang et al., 1999; Bartzokis et al., 2000; Bonson et al., 2002), and previous studies have found that alterations in this region have been associated with faulty decision-making (Bechara et al., 1994; Bolla et al., 2005). Though speculative, it is possible that chronic changes in OFC function are related to direct effects of drugs on this region during periods of acute intoxication. Other recreational drugs, such as methamphetamine (Volkow et al., 2001), cocaine (Volkow et al., 2005), and alcohol (Volkow et al., 2007), are also associated with altered function in the OFC, suggesting that several classes of psychoactive drugs may have mechanisms that contribute to deficits in prefrontal function that are commonly seen in addiction (Goldstein and Volkow, 2011).

There are many limitations to this study, and results should be interpreted cautiously. First, it is difficult to precisely measure vascular effects of THC on cerebral blood vessels (Ohlsson et al., 1980; Hollister et al., 1981), as THC does have a dilative effect on conjunctival and muscle blood vessels (Weiss et al., 1972; Ohlsson et al., 1980; Hollister et al., 1981). Previous fNIRS studies report that signal measured through fNIRS may in part arise from systemic changes that are not directly related to brain activity (Tachtsidis et al., 2008; Kirilina et al., 2012). However, in a PET study, smoked marijuana had no effect on forehead skin perfusion (Mathew et al., 1992), yet had effects on CBF. Furthermore, THC-induced vascular changes are likely to be global, while THC-induced changes in HbO in our study showed significant regional variations. Orbitofrontal channels showed significant changes, while medial ventral PFC generally did not. Furthermore, blood pressure did not change throughout the study, which is important to note given that blood pressure fluctuations can exert confounding effects on brain NIRS (Minati et al., 2011). However, we cannot rule out the possibility of drug-induced vascular changes contributing to the HbO response. In future studies, we plan to investigate the impact of systemic changes through the use of portable devices to monitor respiration, heart rate, or other physiological factors that could influence the fNIRS signal.

Second, this study was open-label, and a more rigorous exploration of the pharmacological effects of THC on cortical hemodynamics requires a double-blind placebo-controlled design. Factors such as expectation could have played a role in differences detected. Moreover, dose determination was partially based on self-report of cannabis use, and we were only able to grossly estimate the potency (THC content) of MJ used, which may influence participants’ tolerance and response to dronabinol.

Third, the current experimental design compared performance and fNIRS measurements before and after THC administration, and therefore, order and learning effects associated with n-back task could have influenced the results. However, this would bias the results against, rather than for, finding any differences between pre and post-THC. For separate sessions of the same task, several studies have shown greater magnitude of activation for the first session than for the second session, for both inhibitory control tasks (Langenecker and Nielson, 2003) and for language tasks (Schecklmann et al., 2008). Effects of habituation (Loubinoux et al., 2001; Fischer et al., 2003; Kiehl and Liddle, 2003) would also result in greater activation in the first than in the second task. It can therefore be expected that the HbO amplitude of the second measurement would be reduced, whereas in our study, we found that THC administration resulted in increased HbO amplitude in the second session. Furthermore, performance would be expected to improve with practice from the first to the second session, though we saw performance worsen somewhat (though not statistically significantly) after THC administration. A further methodological limitation is the use of different block lengths for the 2-back and 0-back conditions (30s vs. 20s). We chose to design the study in this way to control for the number of target trials, and because we were primarily interested in comparing each condition from post-THC to pre-THC, but the direct comparison between the 2-back and 0-back conditions may be affected by the different block lengths.

This study is also limited by the small sample size, which prevents exploration of effects such as dose, patterns of cannabis use, or interactions with factors such as gender or age. Our exploratory analysis of subjective drug response, which utilized a median split to show that participants who reported greater “high” from dronabinol exhibited greater signal increases from THC, may be especially prone to a false positive, and should be replicated in larger sample sizes with more robust statistical techniques. Given that this study represents the first report of fNIRS detecting changes in hemodynamic activity during cannabis intoxication, we chose to included this data in order to generate hypothesis that can be tested further in placebo-controlled studies with greater numbers of participants. Finally, though for ethical reasons, we could not administer dronabinol to cannabis-naïve individuals, the absence of a control group limits this study’s generalizability to non-regular cannabis users.

This study demonstrates that THC administration causes increased HbO concentration changes in the PFC, supporting rCBF findings in the PET literature. Though preliminary, the current study provides the first experimental demonstration that fNIRS may be used to investigate changes in hemodynamic activity during cannabis intoxication. This easily applied, economical, and well-tolerated method presents significant advantages over PET, which is expensive, labor-intensive, and exposes participants to radiation. Because fNIRS can be used repeatedly throughout the day on a single participant, future studies can use this technique to investigate the time-course of acute intoxication, and recovery from intoxication, which could greatly improve our understanding of the brain’s response to an intoxicating substance.

Conceived and designed the experiments: JMG, AEE, and HOK. Performed the experiments: HOK and MR. Oversaw clinical aspects: GP and AEE. Analyzed the data: HOK and JMG. Wrote the paper: JMG, HOK, and MR. All authors have approved the final article.

AEE received research grant funding and/or study supplies to her institution from Forum Pharmaceuticals, GSK, and Pfizer, and has performed consulting work for Reckitt Benckiser and Pfizer. JMG and AEE hold a provisional patent on use of fNIRS to determine impairment due to cannabis intoxication. The other authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.