A comprehensive understanding of the underlying neurobiology of alcohol and other drug use will likely prove essential in developing more effective prevention, early intervention, and treatment strategies. Ethical limitations prevent the use of many of the techniques developed to study brains in animal models of addiction; however, technical advancements in neuroimaging are increasingly allowing for more in-depth brain assessments in humans. This chapter reviews findings on brain structure and function in individuals with alcohol and other substance use disorders.
These insights into brain alterations associated with addictions are often limited by questions of causality. Most studies report cross-sectional comparisons between addicted individuals and healthy controls, and thus obscuring whether differences are due to substance use, preexisting differences that may have contributed to risk, or some combination of these two factors. To help address this, relevant studies examining at-risk individuals prior to the onset of substance use disorders are also discussed in this chapter.
Differences in brain volume can be measured noninvasively through high-resolution structural magnetic resonance imaging (MRI) scans. Studies investigating differences in gray or white matter volume may manually trace regions of interest in high-resolution MRI scans and calculate volumes; however, automated analysis techniques like brain voxel-based morphometry are more common, as they are less time consuming and easier for comparing multiple brain regions than manual tracing methods. Despite these advantages, automated methods can have problems with accuracy in regions with anatomical variations between subjects, such as folds in the cortex.
Gray Matter Volume
Alcohol dependence is associated with reduced gray matter in the prefrontal cortex, insula, ventral striatum, amygdala, and hippocampus, and gray matter deficits appear more pronounced in female than male patients, suggesting that female patients may be more sensitive to the neurotoxic effects of alcohol. Similarly, a recent voxel-based meta-analysis of gray matter volume comparisons in heavy users of cocaine, methamphetamine, and amphetamine and healthy controls found that the most consistently reported findings were gray matter volume reductions in the insula, thalamus, middle and inferior frontal gyrus, and anterior cingulate. Reduced insula, thalamus, and anterior cingulate volumes have also been correlated with greater methamphetamine cravings in users. Reduced insula and striatal volumes have also been observed in chronic heroin users. A recent meta-analysis on gray matter deficits in marijuana users reported somewhat mixed findings, although reduced hippocampal volumes do appear to be a consistent finding. Finally, tobacco smokers show decreased insula, orbitofrontal, prefrontal, and cingulate cortex volumes. Collectively these findings suggest that gray matter volume deficits are present across different classes of substance use disorders, with insula, prefrontal, and cingulate cortexes commonly affected.
It is unclear to what extent gray matter deficits associated with substance use disorders are a direct consequence of substance use. Prefrontal gray matter deficits were found to worsen over a 5-year period in people with chronic alcoholism, and there is evidence that gray matter deficits are attenuated with extended alcohol abstinence. Similarly, daily morphine administration in patients with chronic pain reduces volumes of orbitofrontal cortex and amygdala providing further evidence that at least some gray matter deficits may be a consequence of chronic use. However, nonabusing individuals who have family histories of alcohol or other substance use disorders (FH+) have reduced amygdala volumes, and although findings in other brain regions are somewhat mixed, this does suggest that at least some gray matter deficits may predate excessive substance use.
White Matter Volume
Alcohol dependent patients have decreased frontal white matter volumes, and these deficits are more pronounced in patients with Wernicke encephalopathy. Similar to sex differences in gray matter deficits, female patients with alcoholism show more pronounced white matter deficits than male patients. There is mixed evidence for greater white matter volume deficits in alcoholic patients with comorbid cocaine dependence and tentative evidence for frontal white matter deficits in cocaine-dependent patients without comorbid alcohol dependence. Findings on white matter volumes in marijuana users are mixed, but frontal white matter volume deficits have been observed in individuals who abused marijuana along with cocaine and or heroin but not alcohol. Similarly, reduced white matter volumes were found in opioid-dependent patients and long-term tobacco smokers. Collectively these findings suggest that frontal white matter volume deficits are also broadly associated with different classes of substance use disorders, although they may be a marker for more extensive substance use than gray matter reductions.
White matter volume deficits appear to be a consequence of problem substance use, at least in the case of alcohol use disorders. White matter volume improvements during abstinence have been documented in alcohol-dependent patients, although it is not clear that full recovery is possible. Unlike gray matter volume deficits, white matter volume deficits have not been reported in nonabusing FH+ individuals, suggesting that these deficits are not present before the onset of problem substance use.
White Matter Microstructure
White matter microstructure can be assessed using diffusion tensor imaging to index fractional anisotropy (FA) of water diffusion in white matter tracts. FA describes the directional selectivity of the random diffusion of water molecules. Higher FA values (maximum theoretical value is 1.0) are observed along heavily myelinated white matter tracts. The structure of the axonal cell membranes and myelin sheath hinders the diffusion of water molecules in all directions except along the fiber tract, thereby producing highly anisotropic water diffusion. Thus decreased FA values in white matter tracts may indicate decreased myelin levels or damage to axons due to processes such as inflammation.
Decreased frontal white matter FA values have been found in nearly all classes of substance use disorders. For instance, alcohol-dependent patients had reduced frontal white matter FA that was associated with higher self-reported impulsivity, longer duration of drinking, and slower processing speeds on cognitive measures. Frontal white matter FA deficits were also present in adolescent binge drinkers, suggesting that relatively limited exposure to large alcohol doses may impair white matter integrity in youths, while frontal white matter FA increases in alcohol-dependent patients with extended abstinence. Similarly, frontal white matter FA deficits have also been reported in cocaine-dependent patients, and FA deficits in cocaine-dependent patients were also associated with greater impulsivity and was more pronounced in individuals who abused additional substances. Similarly, frontal white matter FA deficits have also been observed in individuals who abuse methamphetamine, opioids, marijuana, and tobacco. Collectively, these findings suggest that white matter FA deficits may be a common phenotype across multiple classes of substance use disorders.
Although there is evidence of improvement in frontal white matter FA during abstinence, there is also evidence that FA deficits are at least partially premorbid. Nonabusing FH+ adolescents and young adults have been found to have decreased frontal white matter FA values, and these deficits have been shown to be proportional to the number of biological parents and grandparents with substance use disorders, suggesting a possible hereditary influence. Of interest, similar FA deficits were not observed in FH+ youths with low levels of externalizing behaviors and other risk-related characteristics, suggesting that frontal white matter FA deficits may be closely linked to the behavioral phenotype underlying risk in FH+, typically referred to as behavioral undercontrol or neurobehavioral inhibition.
Collectively, these findings indicate that white matter microstructure deficits are present in individuals with substance use disorders and appear to both partially predate substance use and worsen as a consequence of heavy substance use. Additional studies are needed to determine how frontal white matter FA deficits may contribute to risk of substance use disorder, as well as what specific underlying white matter dysfunctions the decreased FA values reflect.
Summary of Structural Brain Findings
Collectively, these studies indicate that, across different classes, substance use disorders are associated with structural brain alterations in both gray and white matter tissue, commonly involving the prefrontal and frontal brain regions. In several instances these alterations were related to increased impulsivity and other problem behaviors, suggesting a possible neural substrate for behavioral dysfunctions commonly associated with substance use disorders. Brain alterations found in substance users are often assumed to be a consequences of substance use, and there is evidence that substance use is directly associated with structural brain changes. However, given that tobacco use is both directly linked to frontal gray and white matter alterations and not controlled for in most studies, it is difficult to rule out a direct contribution of smoking to the findings described earlier. Finally there is evidence that some gray matter and white matter alterations may predate problem substance use.
Dopamine and Other Neurotransmitter Systems
It is possible to study neurotransmitter systems in humans using techniques such as positron emission tomography (PET), which involves administering radiolabeled ligands and thus exposing participants to low levels of radiation. Although these studies are generally quite safe, since radiation is involved they are generally restricted to adult populations and not children and adolescents. Thus, PET imaging studies are more readily used in adult populations with existing substance use disorders rather than in youth populations prior to the onset of substance use disorders or with limited substance use histories.
Dopamine (DA) has received extensive attention in the addiction field, due in no small part to the finding that nearly all drugs of abuse induce increased dopamine levels in the striatum, and the DA system has been studied extensively in addicted populations using PET imaging. With PET imaging it is possible to study DA-receptor availability and distribution and stimulant-induced DA release using radiolabeled ligands for specific DA receptors (D2, D3, or D4) or DA transporters, and to monitor DA synthesis using radiolabeled levodopa ( l -DOPA).
Decreased striatal D2 receptor availability has been observed across most drug addictions included alcohol, cocaine, methamphetamine, opiates, and tobacco (see Trifilieff et al. for review), and these deficits appear to persist even with extended abstinence (e.g., Volkow et al. ). It appears that decreased striatal D2 receptor availability may be at least partially a consequence of heavy substance use, as prolonged exposure to drugs such as cocaine or methamphetamine decreases D2 receptor density in animal models. Furthermore, striatal D2 receptor availability in FH+ adults with no substance use disorders has been reported to be increased relative to FH− controls or no different from FH− controls. However, other studies have also associated decreased D2 receptor availability with substance use disorder risk factors such as increased impulsivity in both individuals with substance use disorders and healthy controls as well as in animal models (see Trifilieff et al. ). Thus D2 receptor availability deficits also may at least partially predate substance use disorders and contribute to substance use disorder risk.
PET studies have also been used to index stimulant-induced striatal DA release by comparing DA receptor binding before and after administering drugs like amphetamine. Similar to findings on DA receptor availability, individuals with alcohol, cocaine, methamphetamine, opiates, and tobacco all show blunted DA release in this assay, and again similar findings are observed in animal models with prolonged exposure to drugs of abuse (see Trifilieff et al. for review). Findings in FH+ adults are mixed, with one study reporting blunted DA release in this assay, and another reporting no differences in DA release from FH− controls. Blunted stimulant-induced striatal DA release has also been associated with increased impulsivity both in healthy normal adults and individuals with substance use disorders (see Trifilieff et al. ). Thus as with D2 receptor availability deficits, decreased stimulant-induced DA release also may partially predate substance use disorders and contribute to substance use disorder risk.
In contrast with findings on D2 receptor availability and stimulant-induced DA release, findings on DA transporter levels in individuals with substance use disorders appear mixed, with some studies reporting decreases relative to controls and other studies reporting no differences (see Hou et al. for reviews). Findings on presynaptic DA synthesis in individuals with substance use disorders are also divergent, with some studies reporting increases, decreases, and no differences, respectively (see Hou et al. for reviews). These findings suggest that DA transporter levels and presynaptic DA release are less consistently affected in individuals with substance use disorder, although additional studies with larger sample sizes may be necessary to clarify this.
Other Neurotransmitters Systems
PET imaging has been used to a lesser extent to study other neurotransmitter systems in addicted populations. During abstinence, μ-opioid receptor binding increases in cortical and striatal regions in heavy users of alcohol, cocaine, and opioids, and these increases appear to be related to cravings (see Jones et al. and Volkow et al. for reviews). Some studies report both increases and decreases in serotonin transporters in the mid- and hind-brains of heavy users of alcohol and other drugs. Reduced γ-aminobutyric acid A (GABA A ) receptor binding in cortical and striatal regions has been observed in heavy users of alcohol and opiates. Reduced metabotropic glutamate receptor binding in the prefrontal cortex has been observed in smokers but not cocaine users, suggesting that this may be specific to nicotine dependence. Finally, there is reduction in cannabinoid-type 1 receptor binding in both heavy users of marijuana and alcohol, with both groups showing deficits in cortical regions and only alcohol users showing deficits in subcortical regions.
Summary of Neurotransmitter Findings
Collectively these results indicate that substance use disorders are associated with widespread disruptions in the DA system along with other neurotransmitter systems. Decreased striatal D2 receptor availability and decreased striatal DA release appear to be clear addiction biomarkers and provide insight into how reward circuitry is altered in addictions. However, additional studies are needed to better understand how these DA and other neurotransmitter systems are synergistically affected in addicted individuals and how these systems affect chronic drug use. Restrictions on the use of PET in children and adolescents make studying these systems in nonabusing FH+ and other at risk populations prior to the onset of substance use more difficult, thereby making it harder to determine to what extent alterations may predate substance use.
Functional Imaging Findings
Functional brain activity is frequently measured using functional MRI (fMRI) to monitor changes in blood oxygenation level–dependent (BOLD) signaling. Increasing activity in a brain region is followed a few seconds later by an increase in BOLD signaling, reflecting an increase in oxygen carrying blood to that area. Thus BOLD levels are not a direct marker of neural activity, but rather a delayed indicator with a temporal resolution (changes over seconds) on a vastly different scale than the activity of neurons (changes over milliseconds). Despite these limitations, the combinations of spatial and temporal resolution possible with fMRI make it an extremely popular choice for studying human brain activity. This section reviews findings in addicted populations, focusing on brain activations associated with alcohol and other drug cues, nondrug rewards, delay discounting, response inhibition, and, finally, resting state connectivity.
Alcohol and Other Drug-Cue Reactivity
Exposure to alcohol and other drug-related cues consistently induced activations in reward circuitry in addicted populations. Alcohol and other drug cues (stimuli that have been associated with substance use through repeated pairings) can induce strong attentional biases and cravings in heavy users. Imaging studies have most often compared activations to alcohol, cocaine, tobacco, or other drug cues relative to neutral cues in heavy users of those substances. The drug cues used typically involve viewing substance use–related words, images, and videos, hearing scripts about substance use; handling substance use paraphernalia; and/or smelling or tasting alcohol or drugs. Cues that involve more than one sensory modality appear to invoke neural activations with more robust relationships with craving and other clinical indicators of addiction severity. Recent meta-analyses have found that drug cue–induced activations appear largely consistent across drug classes. For instance, heavy drinkers and patients with alcohol use disorders exposed to alcohol cues had increased activations in regions including the striatum, amygdala, anterior cingulate, middle frontal cortex, and orbitofrontal cortex. Similarly, heavy cocaine users exposed to cocaine cues had increased activations in the striatum, amygdala, anterior and posterior cingulate, and insula, and heavy tobacco users exposed to tobacco cues had increased activity in the striatum, amygdala, and anterior cingulate cortex.
Collectivity, these studies indicate that stimuli associated with substance use produce a generally common pattern of robust activations in reward-related circuitry across different classes of substance use disorders. These neural activations show significant promise as biomarkers for predicting treatment response and measuring treatment outcomes.
Nondrug Reward Reactivity
Many studies have found that patients with substance use disorders have altered functioning of reward circuitry in response to nondrug rewards like money; however, the direction of these effects are varied. For example, one of the most commonly used paradigms involving nondrug rewards has been the monetary incentive delay task, which examines neural responses to cues predicting monetary gains and losses as well as gain and loss outcomes. Alcohol-dependent patients performing this task have been reported to have striatal activity that is reduced, increased, or not different from healthy controls. Similarly, striatal activity in cocaine-dependent individuals performing this task has been reported to be reduced, increased, or similar to controls. Likewise, marijuana users performing these tasks have also been reported to have striatal activity that is reduced, increased, or similar to controls. A recent meta-analysis of studies of individuals with substance disorders performing fMRI tasks involving reward anticipation and delivery (including monetary incentive delay tasks) reported overall that individuals with substance use disorders have relatively reduced striatal activity during the anticipation of rewards and relatively increased striatal activity during the delivery of rewards, which the authors note to be difficult to interpret in the context of current neurobiological theories of addiction.
Delay discounting has received extensive attention in the addiction field because discounting or devaluing of delayed rewards has been observed in nearly all classes of substance use disorders as well in nonabusing FH+ individuals. However, relatively few functional imaging studies on delay discounting in addicted populations have been published, and findings are mixed. For instance, heavy alcohol users have been reported to have greater activations in prefrontal, frontal, and occipital cortical regions. In contrast, heavy methamphetamine users performing delay-discounting tasks have been reported to have decreased activity in frontoparietal regions, and smokers have been reported to have decreased activity in parietal and occipital cortical regions.
The mixed results observed may be partially due to things like variations in tasks designs and analytical strategies, so caution is warranted in interpreting differences in delay-discounting associated neural activity across different substance use disorders. Furthermore, the limited published studies on delay discounting in addicted populations likely reflect inherent difficulties in adapting complex paradigms like delay discounting for imaging to compare different populations. Thus, despite clear behavioral evidence of increased delay discounting in addicted populations, it is less clear how neural activity underlying delay discounting may be altered in these individuals.
Response inhibition, or the ability to inhibit proponent responding, has also been found to be consistently impaired in individuals with substance use disorders as well as in nonabusing FH+ individuals, typically measured with stop signal or go/no go tasks. These tasks activate a predominantly right-sided network of structures including the inferior frontal gyrus, anterior cingulate cortex, presupplementary motor area, and dorsolateral prefrontal cortex. Findings in individuals with substance use disorders are generally mixed, although it does appear somewhat more common for addicted individuals to have relative decreases in activity in these structures. Thus similar to delay discounting, it is not entirely clear how neural activity is altered in addicted populations to account for their impairments in response inhibition.
Resting State Functional Connectivity
Resting state functional connectivity (rsFC) is emerging as a potentially powerful measure for studying psychiatric disorders. Unlike fMRI tasks, where participants are shown images or required to perform tasks, resting state scanning procedures typically involve imaging subjects instructed to let their minds wander and not fixate on any particular topic. At rest, brain circuits such as the default mode network show correlated patterns of neural activity, and alterations in rsFC have been observed in psychiatric disorders such as depression, autism, and schizophrenia.
Individuals with substance use disorders have been reported to have altered rsFC patterns; however, the directions of findings are mixed. For instance, alcohol-dependent patients have been reported to have decreased rsFC between the ventral striatum and ventromedial prefrontal cortex, insula, and putamen, with poorer connectivity predicting relapse in short-term abstinent alcoholics. In contrast, although cocaine-dependent patients have been reported to show poorer rsFC between the putamen and insula and postcentral gyrus, compulsive cocaine use has been associated with increased rsFC between the striatal and anterior prefrontal/orbital prefrontal cortex, and relapse in cocaine-dependent patients has been predicted by increased rsFC between the posterior hippocampus and posterior cingulate. Likewise, opioid-dependent patients have been reported to have decreased rsFC between both the ventral striatum and insula and caudate and dorsolateral prefrontal cortex, but have also been reported to have increased rsFC, both between the ventral striatum and prefrontal and orbitofrontal cortexes and between the hippocampus and the caudate, putamen, insula, and posterior cingulate. Increases and decreases in rsFC have also been observed for regular marijuana users and tobacco smokers. Collectively, these findings suggest that substance users have altered rsFC, although the direction of the findings may depend on both the drug class and specific circuits examined.
Functional Imaging Summary
The most reliable functional imaging findings in addicted populations appear to be increased reactivity in reward-related circuitry in response to alcohol and other drug cues, and this holds promise as a biomarker for predicting treatment outcomes and indexing therapeutic effects. Findings in other areas appear much more varied; however, several caveats are important to keep in mind. Functional imaging studies place greater demands on participants than structural imaging studies, and in some paradigms it can be difficult to measure how fully participants are complying with instructions and engaging with procedures. Many studies, particularly older ones, used fairly small samples, which may not only have resulted in some false-positive findings but also considerable heterogeneity across study samples in potential confounding variables such as comorbid psychiatric conditions and other substance use. Thus studies with enough participants to find significant group differences in a particular study may not have been large enough to find activation differences reliably associated with substance use disorders across studies. Finally, differences in analytical procedures may also have contributed to significant variability across studies.