Neural and Behavioral Correlates of Impaired Insight and Self-Awareness in Substance Use Disorder

Abstract

Purpose of Review:

Clinical insight is an emerging interest in substance use disorder research, but is difficult to study empirically. We reviewed recent research examining the behavioral and neural correlates of several psychological processes tapping into self-awareness that may in turn contribute to insight.

Recent Findings:

Individuals with substance use disorder exhibit deficits in self-monitoring (especially self-report / behavior dissociations), metacognition, alexithymia, readiness for behavior change, and interoception. Behavioral impairments have been further linked to neural abnormalities in a diverse network of brain regions encompassing cortical midline areas, insula, and frontal cortex, among others. Various treatment modalities may target these deficits, though corresponding effects on neural circuitry remain to be determined.

Summary:

Compromised insight in substance use disorder may be relevant to a constellation of behaviors that suggest a lack of behavioral awareness linked to drug use. Future work needs to refine and advance the measurements, continuing to investigate insight problems in addiction that may become important therapeutic targets.

Introduction

Individuals with substance use disorder (SUD) display a wide spectrum of behavioral deficits that suggest problems with insight and self-awareness linked to their addiction. They report having a high degree of personal control over their drug use and subsequent recovery (e.g., endorsing items such as, “I can control my drinking any time I want to”) (1–3), and rate themselves as having less emotional and cognitive dysfunction than do close informants (3–5). In national surveys with representative samples of the United States population, a plurality of individuals meeting diagnostic criteria for SUD do not perceive a need for drug treatment (6).

In the clinical setting, insight describes a continuous, multidimensional construct attributable to a patient’s clinical presentation and cognitive style (7, 8). Over the past century, the definition has evolved to include a patient’s awareness of having a disorder, awareness of having symptoms, recognition of symptoms as a manifestation of the disorder, recognition of the sequelae of symptoms, appreciation of the need for treatment, and treatment adherence (8–10). Insight, when viewed through this lens, further includes the impact of having an illness on one’s mood, cognitions, and behaviors – including the ability for monitoring and self-correcting (10) – and is influenced by a combination of pathophysiological causes, neurocognitive compromises, and cognitive styles (8). Although insight has been traditionally invoked in psychosis, particularly schizophrenia (10, 11), emerging theory and research have begun to underscore the importance of insight, and additional processes pertaining to self-awareness of addictive behavior, in understanding core pathophysiology of SUD (8, 12–14). This is because, transdiagnostically, insight is associated with quality of life, psychosocial functioning, symptom severity, and therapeutic outcomes (i.e., compliance, readmission, and relapse rates) (10).

For SUDs, insight may have important implications for psychiatric treatment planning and prognosis. According to a recent review, better insight in SUD was associated with abstinence and treatment adherence, and better identification of the negative consequences from substance use and readiness for change (14). For example, people with lower insight into their cocaine use and associated disinhibition problems (i.e., mood instability, impulsivity) exhibited poor treatment motivation maintenance (15) and increased cocaine spending (14, 16). Ultimately, though, a comprehensive understanding of insight in SUD is likely to require examination of the complex interplay between neurocognitive functions (e.g., executive functioning, memory, attention, etc.) and self-referential functions (e.g., self-monitoring, interoception, metacognition, etc.), and the neural mechanisms that underpin them. This inherent complexity presents notable challenges for scientific investigation, especially for laboratory-based studies incorporating sensitive behavioral and neuroimaging assays. One approach is to separate insight into constituent processes more amenable to empirical investigation.

Here, we describe and highlight emerging research in SUD, primarily which has been published in the last five years (2016–2021), that has sought to characterize the behavioral, emotional, and neural deficits that ostensibly relate to insight and self-awareness processes in SUD. These include: (A) self-monitoring: dissociation between participants’ self-reports and more objective measures of behavior ascertained by tasks, brain imaging, and/or collateral reports; (B) metacognition; (C) alexithymia; (D) readiness to change drug-use behavior; and (E) interoception. We do not intend to provide an exhaustive or comprehensive review of any particular domain, function, or research area; rather, we aim to present an overview of recent work in multiple basic functions that may contribute to problems of insight in SUD. Where applicable, we highlight studies examining the underlying brain circuits, including abnormalities seen in key brain areas such as the dorsolateral prefrontal cortex (DLPFC), anterior cingulate cortex (ACC) extending to the ventromedial PFC, and temporal areas (12, 13, 17). The insula, particularly its anterior portion which has been consistently linked to self-awareness, is also likely to be involved (18). We conclude with remaining challenges and treatment implications.

Self-Monitoring: Dissociations Between Self-Reports and Objective Behavior

In many cases, individuals with SUD rate themselves as having fewer impairments or clinical severity than objective evidence would indicate, which is highly relevant to insight. A recent study among participants with methamphetamine use disorder juxtaposed self-reports of emotional experience against psychophysiological metrics including skin conductance and startle response (19). After viewing videos meant to elicit anger, fear, amusement, and joy, respectively, the methamphetamine group had a higher startle response to the anger video than healthy controls, but the groups did not differ on self-reported anger (19). Thus, methamphetamine users had a greater bodily response to anger that they did not – perhaps were unable to – self-report. We reported a similar dissociation between craving and EEGs (20). In this study, we studied a cohort of individuals with cocaine use disorder (CUD) and binned them based on length of cocaine abstinence (2 days, 1 week, 1 month, 6 months, and 1 year). We found that, as abstinence increased in duration within this cross-sectional study, self-reported cocaine craving declined linearly. In contrast, the late positive potentials (LPPs), acquired in response to viewing cocaine-related images [and previously linked to craving (21)], showed a strikingly different pattern, following a quadratic (inverted U-shaped) function (20). Notably, this “incubation” of drug-related reactivity and craving coheres with preclinical studies (22), providing validity to the LPPs. The implication is that individuals with CUD could be unaware of the extent of their drug cue-reactivity, a potential vulnerability to relapse (20).

We have also examined differences between self-reports and a different objective outcome in SUD: choice behavior. During this task, research participants made choices to view drug-related images in comparison with standardized pleasant, unpleasant, and neutral images (23). At the conclusion of the task, participants self-reported their most chosen picture category, and we then examined for coherence (or discrepancy) between self-reports of choice and objective choice (16). Recently, we tested this paradigm in chronic pain patients who use opioids (24). Results showed that patients with impaired self-awareness of their opioid-choice behavior were more cue-reactive (i.e., more likely to choose the drug images for viewing) and had more severe opioid misuse (24).

A related method of measuring self-awareness impairments has been to examine self-other discrepancy scores, testing whether individuals with SUD underestimate their cognitive deficits as identified by close informants. In one study using this approach (3), individuals with CUD and their designated collaterals completed the Frontal Systems Behavior Scale (FrSBe), a questionnaire measure of executive function (25). Previously, these authors had reported that individuals with addiction underestimated their executive impairments, particularly when actively using drugs (less so once in recovery) (4). In the later study, the authors correlated the difference score between self- and collateral ratings with gray matter volume (GMV), ascertained using voxel-based morphometry (VBM). The more that addicted individuals underestimated their apathy and disinhibition deficits relative to informants (i.e., suggesting poorer self-awareness), the greater was the GMV in the left dorsal striatum and left orbitofrontal cortex (OFC), respectively. Furthermore, the DLPFC showed unique differential correlations with the self-other discrepancy score: in this region, underestimation of deficits was negatively correlated with GMV, whereas overestimation of deficits was positively correlated with GMV (3).

Metacognition

Metacognition broadly refers to the ability to monitor and evaluate one’s own cognition and behavior (26). Problems with metacognition may be linked to problems with insight in SUD, in that decrements in self-monitoring may culminate in unreflective, compulsive drug use that is lacking in self-awareness. Metacognition can be described as a special case of more general performance- and self-monitoring functions, which have been studied in users of stimulants (27), cannabis (28), and alcohol (29). Whereas the self-other discrepancy approach (above) tests whether self-reports are validated by external information, metacognition tests whether self-reports are appropriately synched with behavior within the same person and context. Thus, they are separable, albeit related, constructs (30). Metacognition has been investigated across multiple functional domains in the laboratory, most notably with respect to perception and memory. Although historically metacognition has been examined with self-report measures (31–33), more recent studies have advanced research in this area by using laboratory-based measurements. In such laboratory studies, typically conducted in healthy controls to date, metacognition has been investigated by asking participants trial-by-trial to render a second-order evaluation (e.g., confidence rating) of their ongoing task accuracy (e.g., during a simple perceptual judgment). Metacognition is then operationalized as the degree to which higher confidence correlates with better performance (34). Behaviorally-measured metacognition has been convincingly linked to functional and structural integrity of regions comprising the anterior PFC (35–41).

Recent studies are beginning to translate these interesting and precise metacognition paradigms to SUD. In our own work, actively-using CUD participants exhibited a weaker correlation between actual performance and performance confidence on a perceptual decision-making task, indicating compromised metacognition in this group compared with controls (and compared with abstinent CUD participants, who did not have metacognitive deficits). In this same study, we used VBM analysis to test for GMV integrity in the rostral (also known as the perigenual) ACC, using a priori coordinates taken from a prior study (42). The VBM analyses squared with the metacognition behavioral analyses, showing lower GMV in active CUD than in abstinent CUD and controls; and, importantly, behavior and rostral ACC GMV were positively correlated: the higher the GMV, the better the metacognition (43). Another study from a different laboratory has since replicated the behavioral metacognitive deficit in SUD: in this latter study (44), methadone-maintained individuals with opioid use disorder (OUD) exhibited metacognition impairments while performing a perceptual decision-making task quite similar to our own. Future studies will need to evaluate the direct neural correlates of metacognition impairment in SUD, for example using fMRI tasks where behavior and neuroimaging are collected simultaneously (36, 40).

Another line of work has shown disruptions of metamemory (i.e., metacognition applied to memory), particularly in alcohol use disorder (AUD). Especially during early abstinence, individuals with AUD can be unaware of their memory deficits and can overestimate their mnemonic capacities, perhaps stemming from alcohol-induced impairments to brain circuits encompassing the default mode network (DMN) and insula, among other regions (45). In a recent study, individuals with AUD and controls performed an episodic memory task engaging both retrospective (confidence judgments of accuracy) and prospective (future confidence of accuracy) monitoring processes. The individuals with AUD showed impairments in predicting their future memory performance, and these impairments were further correlated with lower insular volumes (46); and, in a subsequent study, such impairments were correlated with lower insula activation and altered insula-vmPFC connectivity (47). Future studies can test metacognition and metamemory within the same patient sample, to examine for domain-specific versus domain-general processes. In healthy controls, for example, domain-specific functions were linked to the lateral PFC, whereas domain-general processes were more prominently linked to the self-referential network (36). Another direction will be to test for metacognitive deficits in high-order capacities such as social cognition, again as previously studied in healthy samples (48, 49).

Alexithymia

Alexithymia is characterized by difficulty identifying feelings, difficulty describing feelings, restricted imagination, and an externally-oriented thinking style, widely assessed using the Toronto Alexithymia Scale (TAS-20) (50). Difficulties with understanding one’s own emotions reflect a type of impaired self-awareness that may contribute to problems with insight. A recent coordinate-based meta-analysis reported that higher levels of alexithymia were associated with smaller volumes of the insula, amygdala, OFC, and striatum (51). Multiple frontal cortical regions also appear to be involved (52, 53). Alexithymia has been suggested as a trait risk factor for problematic substance use, most convincingly demonstrated in AUD and drinking severity (54–59). One mechanism through which alexithymia may impact AUD dependence and severity is through a heightened experience of craving and other negative emotions (5, 60–62).

There is also emerging work on alexithymia in other SUDs. For example, positive correlations between alexithymia and craving were reported in cigarette smokers during nicotine withdrawal (63). In the same study, higher alexithymia in smokers predicted weakened resting-state functional connectivity between the right anterior insula and ventromedial PFC. Additional studies have reported above-control levels of alexithymia among users of heroin (64, 65) and stimulants (66). Notably, the latter study further also examined relationships between alexithymia and D2-type receptor availability using [¹⁸F]fallypride in the ACC and insula. Results revealed a positive correlation between alexithymia and D2-type receptor availability in these regions in the control group but lack of correlation in the methamphetamine group, for whom there may be an abnormal decoupling between dopamine signaling and emotional processing (66). In another study, this time pilot work in methadone-maintained individuals who met criteria for CUD, alexithymia was positively correlated with fMRI activation in multiple clusters, including the insula, midbrain, inferior frontal gyrus, middle frontal gyrus, and DLPFC, during the prospect of both rewards and losses when performing a monetary incentive delay task (67). There is also some early evidence showing that functional brain correlations with alexithymia may be modulated by gender (68).

Readiness to Change

A core component of insight is recognizing the need for treatment to reduce symptoms. In SUD, it is important for individuals to identify the need to curb problematic drug use and engage with drug treatment. Many patients are unprepared or unwilling to make such changes (69). For example, in a large cohort consisting of 1066 heroin users with OUD, a majority (~65%) of the participants were found to lack insight into their heroin-use behavior, characterized by the following: (1) denial about having problems with self-control related to heroin use, (2) inability to link behavioral changes with chronic use of heroin, and (3) belief that opioid medications are useless as a treatment strategy (70). Similar findings have emerged for other addictions. In a sample of 1008 individuals with AUD, the majority (~80%) did not seek treatment, and the most common reason for not doing so was a lack perceived need (71); and, lack of perceived need is often intractable over years (72). Even among those who enter treatment, those with low perceived need to change tend to exhibit disengagement, ambivalence, and poor outcomes (73). In contrast, greater recognition of the need to change behavior predicts better outcomes (74–76). Treatment itself may augment readiness to change from pretreatment to post-treatment (77), although this relationship may be nuanced (78, 79).

Several lines of research in this area have drawn upon psychosocial models such as the Transtheoretical Model of Behavior Change (80). According to the Transtheoretical Model, motivation to change behavior proceeds through a series of stages. It begins with no or low recognition of a problem or need for change (Precontemplation), followed by increased recognition and motivation to confront the problem (Contemplation), and finally by behaviors or thoughts meant to initiate or sustain positive change (Action, Maintenance). In some studies, stages of change have been related to imaging measures to examine for neural correlates. In an earlier study of AUD, those in the “precontemplation” phase (i.e., denying a need for change) or “contemplation” phase (i.e., readiness to change but without taking steps to do so) had lower GMV across multiple brain regions (including cerebellum, fusiform gyri, and multiple regions comprising the ventral anterior PFC) compared with healthy controls and compared with individuals with AUD were in the “action” phase (i.e., taking steps toward recovery) (81).

In a subsequent fMRI study conducted among individuals with methamphetamine use disorder, higher levels of precontemplation were positively correlated with weakened resting-state functional connectivity between the rostral ACC (seed region) and the amygdala, parahippocampal gyrus, medial and lateral OFC, bilateral precentral gyrus, temporal lobe, bilateral occipital cortex, and cerebellum (82). There were also negative correlations in this study between precontemplation score and resting-state functional connectivity between the precuneus (second seed region) and the midbrain, brainstem and cerebellum (82). Interestingly, this same study found that higher levels of precontemplation were correlated with lower cognition measured by a battery of neuropsychological tests (82) [a relationship which has since been replicated in AUD (83)]. Such findings support the view that low perceived need to change, relevant to impaired insight, may stem from quantifiable neurocognitive deficits rather than a deliberate self-presentation of denial (8).

Still, an interpretive limitation with such studies is that the measure of insight/readiness to change was completed outside the scanning environment. To more directly probe insight-related circuitry, we recently developed and piloted an “fMRI-insight task” to probe insight- and self-referential circuitry in CUD. In this proof-of-concept study (limited sample size), individuals with CUD and healthy controls were scanned with fMRI while responding to precontemplation and contemplation (reverse scored) statements adapted from the University of Rhode Island Change Assessment (URICA) (84), an extensively validated questionnaire also based on the Transtheoretical Model. Behavioral results showed that while individuals with CUD reported a greater perceived need to change drug use behavior than healthy controls, as expected given their DSM-5 diagnosis, the ratings of CUD participants fell short of “agreeing” with a need to alter drug use (85). The fMRI analyses showed a similar pattern of group difference in the medial OFC / ventromedial PFC, such that CUD participants had higher activations in this region than controls (85), which may reflect increased personal relevance of answering drug-change questions in this population compared with non-drug users (13). It will be important for future studies to confirm these relationships, while also expanding the paradigm to other addictions and uncovering novel longitudinal relationships between the behavior/circuitry and future drug use/treatment outcomes. In this regard, it is interesting to note that the ventromedial PFC has been implicated in “treatment resilience,” according to a recent meta-analysis (86).

Interoception

Interoception refers to the ability to perceive one’s internal body state, including visceral and emotional sensations for maintaining homeostasis (18). Problems with interoceptive awareness could lead to misinterpretation of drug-relevant internal states, such as craving, that in turn could contribute to individuals misevaluating their SUD severity. Of the regions that have been implicated in interoception, the insula has received particular attention. The insula processes incoming sensory information and integrates this information with emotional, cognitive, and motivational signals from interconnected brain regions (87). One of the most common interoception paradigms has been testing self-awareness of one’s own heartbeat. A meta-analysis, conducted in healthy controls, reported that heartbeat detection involved a network of brain regions including the posterior right and left insula (as expected), right claustrum, precentral gyrus, and medial frontal gyrus (88).

Emerging studies have begun characterizing interoceptive processes and abnormalities in SUD. In a recent transdiagnostic study, individuals with SUD were compared with individuals with anxiety/depression and healthy controls, testing the hypothesis that patients would have less variance in heartbeat monitoring during an experimental challenge that ought to enhance interoceptive awareness (i.e., holding one’s breath). Consistent with hypotheses, the healthy controls showed improved performance during the breathing challenge, whereas the patients did not, indicating impaired modulation of interoceptive sensitivity (89). In a recent fMRI study, stimulant users and opioid users completed a cued attention task in which they either focused attention on their heartbeat or stomach sensations (two interoceptive conditions), or else focused attention on an external task stimulus (exteroceptive condition) (90). The stimulant users groups had lower insula fMRI activation than controls when focusing attention on their heartbeat. Opioid users showed a similar pattern, though the difference from controls was not significant, perhaps due to a smaller sample size. Interestingly, both SUD groups differed from controls on attention to stomach sensations (90).

Treatment Considerations

These lines of research suggest the potential utility of remediating self-awareness and insight deficits in SUD; several existing treatment modalities may have such an effect. Mindfulness-based interventions promote insight and adaptive action in response to inner bodily signals, enhancing awareness of cognitions, emotions, and behaviors in the present moment (91, 92). Mindful Awareness in Body-oriented Therapy (MABT), for example, is designed to teach interoceptive awareness and related skills to fortify emotion regulation and improve interoceptive awareness (93, 94). Mindfulness-Oriented Recovery Enhancement (MORE) targets substance craving, substance use patterns, and relapse prevention, by addressing neurocognitive mechanisms involved in reward learning and executive functioning (95). MORE and related psychological interventions may reduce opioid misuse risk in patients with chronic pain and improve posttreatment abstinence from cigarette smoking (95, 96). Incorporating imaging measures into randomized trials could be a valuable future direction (97).

Insight may also be improved via motivational interviewing, which facilitates open, reflective communication to promote behavioral change. Collaborative and integrated models work particularly well for patients with long-term and case management needs, as they are associated with better health outcomes, treatment adherence, and lower outpatient costs (8). Psychotherapy that helps facilitate mentalization (e.g., Mentalization-Based Therapy) and increase reflective functioning may also be beneficial, with an emphasis on tolerance of affect and emotion regulation (98). Modulating one’s embodied experience with interventions like these may enhance insula reactivity and increase frontal control network, in turn possibly affecting drug-seeking behavior and urges to use substances (99).

Challenges and Future Directions

Self-awareness and insight in SUD is an emerging area of work, and moving the field forward requires resolving multiple challenges. First, the insight construct itself lacks sufficient operationalization in addiction research and clinical practice; there is not yet an agreed-upon set of questions or attributes on which to focus. We have identified several prospects in this review (see Table 1, for an overview), but there are undoubtedly others. There is also the notable challenge of fitting the constituent pieces together (Figure 1).

Table 1.

Constructs potentially involved in insight and their relevant behavioral and neural abnormalities, by drug type.

Constructs	Drugs Studied	Behavioral Effects	Possible Neural Correlates	Relevant Citations
Self-monitoring	Cocaine	Underestimation of executive impairments and clinical severity Lack of awareness of drug cue-reactivity	GMV in DLPFC, left dorsal striatum, and left OFC	3, 4, 20
Metacognition	Cocaine	Compromised ability to monitor and evaluate one's own cognition and behavior	GMV in rostral ACC	42, 43
	Alcohol	Awareness of memory deficits (i.e., metamemory) Overestimation of mnemonic capacities	Insular volumes and fMRI activation Insula-vmPFC functional connectivity	45, 46, 47
Alexithymia	Alcohol	Heightened experience of craving and other negative emotions	Structure / volume of insula, amygdala, OFC, and striatum GABA concentrations in rostral ACC	51, 52, 53, 54–59
	Nicotine	Increased craving during nicotine withdrawal	Right insula-vmPFC connectivity	63
	Methamphetamine	Deficits in emotion recognition and processing	D2-type dopamine receptor availability in ACC and insula	66
	Cocaine	Increased activation to reward Differential activation to stress by gender	fMRI activation in insula, midbrain, IFG, MFG, DLPFC, striatum, amygdala, and thalamus	67, 68
Readiness for change	Alcohol	Denying a need for change (i.e., “precontemplation” stage of change) or readiness to change but without taking steps to do so (“contemplation” stage of change)	GMV in cerebellum, fusiform gyri, and ventral anterior PFC	81
	Methamphetamine	Higher levels of Precontemplation associated with lower cognition	Connectivity between rostral ACC and multiple corticolimbic areas	82
	Cocaine	Disagreeing that there is a need to reduce drug use	Medial OFC / vmPFC fMRI activation	85
Interoception	Opioid	Altered processing of interoceptive signals	Insula fMRI activation	90
	Stimulant (cocaine and amphetamine)	Altered processing of interoceptive signals	Insula fMRI activation	90

Note. The Table provides an overview of the key constructs and outcomes from the review. It is not meant to be an exhaustive description of all studies, but rather to highlight representative studies that help clarify the behavioral and/or neural effects of each construct, separately for each substance of abuse. Abbreviations: GMV = gray matter volume; DLPFC = dorsolateral prefrontal cortex; OFC = orbitofrontal cortex; ACC = anterior cingulate cortex; vmPFC = ventromedial prefrontal cortex; GABA = gamma-aminobutyric acid; IFG = interior frontal gyrus; MFG = middle frontal gyrus.

Figure 1.

Simplified model characterizing the underpinnings and outcomes of insight in substance use disorder. Insight is expected to relate to treatment outcomes in a bidirectional manner: the processes and circuitry underlying insight are expected to facilitate treatment outcomes; in turn, successful treatment is expected to remediate the same processes and circuitry. Various behavioral and neural deficits related to self-awareness likely contribute both to insight and treatment success, with relationships between these variables again likely being bidirectional. Finally, a full understanding of the relationship between insight and treatment outcomes is likely to be incomplete without taking into account various personal characteristics and environmental contexts that may exert a modulatory effect.

Second, and relatedly, difficult methodological issues must be considered and overcome. Some instruments assessing insight have been developed and validated for use in other psychopathologies, for example referring to the experience of psychosis rather than drug-seeking. Many studies additionally rely on self-report measures, which may be inherently suspect among patients with compromised self-awareness. Future studies will need to continue to develop addiction-centric measures for ascertaining impaired insight in SUD. Where possible, objective measures of substance use and drug-seeking should be incorporated.

Third, prolonged substance use and its antecedents may exacerbate cognitive impairments that in turn worsen and otherwise change the trajectory of insight as the addiction progresses. Longitudinal studies are needed for this purpose, including tests of whether insight can predict treatment outcomes for SUDs. Another important developmental question, which would need to be studied in younger samples, concerns whether insight is a predictor or a consequence of SUD. In studying developmental trajectories, special consideration should be paid to the potential effects of comorbid trauma on the PFC leading to impulsive and self-destructive behavior (14, 15).

Finally, from a clinical standpoint, it is important to recognize that insight is not a panacea, as enhanced insight also poses challenges for symptomatic individuals. For instance, higher self-reflection, which is a component of cognitive insight, is associated with increased rumination and depressed mood (10), which notably are triggers for drug-seeking especially in those with comorbid dysphoria (100). Insight should be broached within the context of a supportive therapeutic environment.

Conclusion

Insight is an emerging construct of vital interest for research and clinical care in SUD. It is also very difficult to study scientifically, due to its complex and multidimensional nature. In this review, we highlighted several processes ostensibly linked to insight in SUD, but that are more amenable to scientific study. Our review indicates that individuals with SUD display a constellation of behavioral deficits encompassing impaired metacognition, abnormal interoception, reduced readiness for behavior change, problems with self-reports, and high alexithymia. Neuroimaging results, where available, suggest that these behavioral deficits may be underpinned by abnormal structure and functioning of cortical midline areas, lateral PFC regions, and insula-centric networks, potentially illuminating symptom-related mechanisms and providing novel targets for intervention. Our working hypothesis is that dysfunction of this neural circuitry may contribute to lack of clinical insight in patients with SUD, with impaired insight being a special case of more global abnormalities. Such findings challenge the long-held belief that patients with SUDs are “in denial” about their substance use and its effects. Instead, it may be a combination of neurocognitive deficits and circuit abnormalities that impair insight into illness for these patients (8).

Despite many remaining questions, insight has important implications for treatment development, access, delivery, and adherence; and for prognosis and overall quality of life. The scientific study of insight provides the opportunity to enhance knowledge into multiple interesting processes that relate to self-monitoring, emotional experience, and behavior change in SUD. These concepts and relationships can augment our understanding of drug addiction, as well as potentially inform the study of self-referential processing and self-knowledge in cognitive neuroscience.