After a period of decline, use of the recreational “club drug” 3, 4-methylenedioxymethamphetamine (MDMA or “Ecstasy”) has been rising in the United States since 2005 and recent data suggests that over 12 million Americans have used MDMA (Johnston et al. 2008
;Substance Abuse and Mental Health Services Administration 2003). Ecstasy remains highly popular throughout the world, especially in North America, Western Europe and Oceania (United Nations Office on Drugs and Crimes 2008
Numerous animal studies support MDMA-induced long term serotonergic (5-HT) alterations through degeneration of presynaptic axon terminals and depletion of brain 5-HT upon exposure to MDMA of adequate magnitude and chronicity (Gibb et al. 1990
;Green et al. 2003
;Ricaurte et al. 2000
). This has led to significant concerns regarding the neurotoxic potential of MDMA. However, loss of brain 5-HT markers can occur in the absence of axonal loss (Fantegrossi et al. 2004
;Wang et al. 2004
;Wang et al. 2005
), suggesting that MDMA may produce functional consequences for 5-HT neurotransmission even in the absence of axon degeneration. Animal data have also shown an association between MDMA use and long term behavior change, which may be mediated by chronic alterations in the 5-HT system (Easton et al. 2006
). These data raise significant concern about the possibility of chronic neurological effects of MDMA use in humans.
Many, but not all (Back-Madruga et al. 2003
;Gouzoulis-Mayfrank et al. 2005
), studies have found evidence of impairment across multiple neurocognitive domains including verbal working memory (Jacobsen et al. 2004
), episodic memory (Morgan 2000
) and visual memory (Back-Madruga et al. 2003
) in subjects with a history of MDMA exposure. Recent meta-analyses found effect sizes ranging from small to substantial for impaired verbal memory and for short and long term memory impairment in MDMA users compared to non-MDMA using control subjects (Kalechstein et al. 2007
;Laws et al. 2007
). A recent prospective cohort study compared subjects who started using MDMA with non-MDMA polysubstance users, and found significantly reduced immediate and delayed verbal recognition on the Ray Auditory-Verbal Learning Test (RAVLT) in the MDMA users (Schilt et al. 2007
Several associative studies have shown a relationship between neurocognitive deficits in MDMA users and underlying neural alterations. Bolla et al. studied subjects abstinent from MDMA for a median of 4 weeks and found that verbal and visual memory deficits correlated with a reduction in levels of the 5-HT metabolite 5-HIAA (Bolla et al. 1998
). Using magnetic resonance spectroscopy (MRS), Reneman et al. found significant deficits in delayed word recall in MDMA users compared to control subjects that were strongly associated with a reduced NAA (N-acetylaspartate)/Cr (creatine) ratio in the prefrontal cortex (Reneman et al. 2001d
). McCann et al. (McCann et al. 1998
) found that the relationship between 5-HTT binding and cognition seen in non-MDMA users was disrupted in MDMA users, suggesting that MDMA disrupts the relationship between 5-HT function and cognition. However a causal link between neural alterations and neurocognitive dysfunction remains to be established.
Imaging studies have also supported the association between MDMA exposure and functional brain alterations. Reneman et al. found increased regional cortical blood flow and increased apparent diffusion coefficient of water in the globus pallidi of MDMA users who had been abstinent for at least 3 weeks (Reneman et al. 2001c
). This is consistent with subchronic vasodilation in recently abstinent MDMA users. Using FDG PET, Obrocki et al. found significantly reduced left hippocampal metabolism in MDMA users compared to oncology controls (Obrocki et al. 1999
). 2 subsequent reports using similar techniques found reduced resting metabolism in bilateral caudate/putamen and in the left amygdala in MDMA users compared to oncology controls (Buchert et al. 2001
;Obrocki et al. 2002
). Using an N-back fMRI paradigm activating working memory, Daumann et al. found significantly greater activation in the right superior parietal lobe and significantly reduced activation in left posterior cingulate cortex, bilateral inferior temporal gyri and bilateral angular gyri in recently abstinent relatively pure MDMA users compared to controls (Daumann et al. 2003
). Using a prospective technique in a similar study, Daumann et al. found increased activation in parietal cortex compared to baseline assessment activation in subjects who continued to use MDMA or amphetamines. The activation was positively correlated with lifetime MDMA use (Daumann, Jr. et al. 2004
). These studies support reduced resting brain metabolism in selected brain regions and alterations in regional brain activation that are associated with MDMA exposure.
Cowan et al. (Cowan et al. 2003
) used voxel-based morphometry (VBM) to study neocortical gray matter volume in MDMA polysubstance users and non-MDMA polysubstance users and found decreased gray matter concentrations in bilateral Brodmann area (BA) 18, left BA 21 and left BA 45. Interestingly, a recent positron emission tomography (PET) study (Lee et al. 2002
) demonstrated that similar regions (BA 9, 18, 21/22 and 45) were involved in semantic memory. Based on the findings of gray matter loss in the Cowan et al. report, we hypothesized that BA 18, 21/22 (we included BA 22 because our fMRI task confluently activated both BA 21/22) and 45 would show altered activation during performance of a verbal memory task. We adapted the semantic memory task described by Lee et al. (Lee et al. 2002
) with two goals in mind: 1) to produce activation in BA 9, 18, 21/22 and 45 to specifically probe the function of these brain regions and 2) to probe aspects of verbal memory that are potentially impaired in MDMA users. Our paradigm consisted of an encoding phase followed by a recognition phase after a very brief delay; as such, this paradigm tapped elements of semantic and working memory.
Our primary hypothesis was that increased lifetime episodes of MDMA use would be associated with altered percent BOLD signal change in BA 18, 21/22 and 45 but not BA 9 during semantic recognition. Our secondary hypothesis was that increased lifetime episodes of MDMA use would be associated with poorer performance response times and increased errors on the semantic recognition task.