As nicely exhibited by Redish et al., we are converging on an understanding that addiction is a disease impacting specific brain systems that control and adapt motivated behavior. Evidence reported here, and continually mounting (Belin & Everitt Reference Belin and Everitt2008), paints an increasingly clear picture that addiction is not merely a vague notion of so-called psychological dependence or some unchecked need to feel good all the time. Rather, it is a disease of neurological character, involving brain systems and phenomenological themes reminiscent of Parkinson's or Huntington's disease. Like these diseases, addiction involves progressive changes to the basal ganglia, the primary neural system involved in the ordering and procedural memory of behavioral programming (Everitt & Robbins Reference Everitt and Robbins2005; Haber Reference Haber2003; Volkow et al. Reference Volkow, Wang, Telang, Fowler, Logan, Childress, Jayne, Ma and Wong2006). Perhaps it has taken us this long to appreciate addiction as a type of progressive cortical-striatal disease because it uniquely seems to involve functions of “free will” that many view as unapproachable as a biomedical topic of exploration. After all, Parkinson's and Huntington's are so obviously motoric in nature, and involuntary, whereas any action choice (i.e., to use a drug) appears voluntary, whether in the addict or in the drug-naive.

But now a new picture is emerging: While dorsal cortical-striatal systems govern the relatively inflexible execution of procedural motor programs (i.e., behavior), ventral cortical-striatal systems flexibly and hierarchically govern the ordering, prioritization, and selection of these dorsally represented programs (i.e., motivation) (Gerdeman et al. Reference Gerdeman, Partridge, Lupica and Lovinger2003; Haber et al. Reference Haber, Fudge and McFarland2000; Kelley Reference Kelley2004b; Yin & Knowlton Reference Yin and Knowlton2006). As facilitated by ever-changing environmental contingencies that provoke rapid changes in striatal dopamine transmission (Bardo et al. Reference Bardo, Donohew and Harrington1996; Finlay & Zigmond Reference Finlay and Zigmond1997; Spanagel & Weiss Reference Spanagel and Weiss1999), neuroplastic alterations spanning these striatal regions change the way ventral and dorsal compartments represent information and communicate with one another (Chambers et al. Reference Chambers, Bickel and Potenza2007; Graybiel Reference Graybiel1998; Hyman & Malenka Reference Hyman and Malenka2001). Thus, behavioral repertoires are dynamically evolving, highly complex structures, or maps, in which specific motor programs are like destinations interconnected by motivational routes (Chambers et al. Reference Chambers, Bickel and Potenza2007). Depending on developmental age, environmental conditions, or inherent individual attributes, cortical-striatal circuits that generate these motivational-behavioral repertoires are themselves physically changing in an unending attempt to provide the most adaptive mapping of behavioral organization of the individual to the external world (Chambers et al. Reference Chambers, Bickel and Potenza2007).

Equipped with this understanding, there is no need to equate a biomedical understanding of addiction with a hubristic claim that we can absolutely define all the physical mechanisms of “free will.” Who knows if we shall ever fully grasp the extreme complexity of human behavioral repertoires? But we can say that if this complexity is synonymous with “free will,” then addiction is a disease of neural systems that govern free will, in which the adaptive complexity of free will is reduced and compromised. Understanding this process mechanistically may actually be within our grasp.

Redish et al. rightly centralize a general concept of impaired decision-making in the pathophysiology of addiction. Of all contexts in which addiction vulnerability has been identified, two stand out most prominently: adolescent neurodevelopment and mental illness (Chambers et al. Reference Chambers, Taylor and Potenza2003; Kessler Reference Kessler2004). Notably, these contexts entail features that are associated with a general theme of impulsivity, whether it be couched in terms of “novelty-seeking,” “high risk-taking,” “poor judgment” or “decision-making,” or “behavioral inhibition deficits.” To suggest a succinct, yet inclusive formulation of impulsivity: it is an operational condition in which ventral cortical-striatal circuits consistently fail to produce adaptive mapping of behavioral organization onto the external world, despite normative intelligence. Of course, in normal adolescence, such impulsivity is actually quite adaptive, because developing healthy brains have the need and capacity for learning from their mistakes (Chambers & Potenza Reference Chambers and Potenza2003). But in mental illness, such capacities are episodically or chronically compromised into adulthood, and the mapping of behavioral organization is maladaptively reduced in complexity and/or relatively inflexible to change based on environmental demands.

Across the broad span of mental illnesses in which addiction comorbidity is robustly present (including schizophrenia, bipolar disorder, antisocial and borderline personality, post-traumatic stress disorder [PTSD], various impulse control disorders, and many others) (Kessler Reference Kessler2004), one or more of the following three brain regions are pathologically altered: the prefrontal cortex, the amygdala, and the hippocampus (Charney et al. Reference Charney, Nestler and Bunney1999). All of these areas normally and robustly send direct glutamatergic projections into the ventral striatum (Groenewegen et al. Reference Groenewegen, Wright, Beijer and Voorn1999; Haber Reference Haber2003; Swanson Reference Swanson2000). There, these projections not only cooperatively help generate and modulate a rich diversity of firing pattern representations spanning ventral striatal networks (representing motivational information), but they act in concert with, or are acted upon by dopamine, resulting in altered neural connectivity (Goto & Grace Reference Goto and Grace2005a; Hyman & Malenka Reference Hyman and Malenka2001; O'Donnell et al. Reference O'Donnell, Greene, Pabello, Lewis and Grace1999; Vanderschuren & Kalivas Reference Vanderschuren and Kalivas2000). Such plasticity likely instantiates the acquisition of new motivational representations contributing to a more complex, highly nuanced, and adaptive motivational repertoire that optimally directs behavioral programming. But, if one or more among the prefrontal cortex, the amygdala, or the hippocampus is compromised due to a pre-existing neurodevelopmental condition (i.e., mental illness), then the complexity and/or adaptive flexibility of the motivational-behavioral repertoire, as generated by striatal networks, is reduced (Chambers Reference Chambers2007). In other words, since limbic inputs to the ventral striatum are relatively impoverished, then the catalogue of motivational representations that may be generated by it are also impoverished in number, complexity, or changeability.

Although some dopamine-mediated neuroplasticity (and adaptive behavioral flexibility) persists that can produce real change in the motivational repertoire, such change may be abnormally limited to conditions or stimuli that produce particularly strong and/or prolonged dopamine (DA) signals (e.g., pharmacological actions of addictive drugs). Without, or before, addictive drug exposure, this situation is clinically perceived as impulsivity, poor decision-making, or other motivational disturbances of mental illness. Upon addictive drug exposure we have a massive epidemic of dual diagnosis (substance disorder comorbidity in mental illness) on our hands to the tune of greater than 50% of all mentally ill or drug-addicted patients seeking treatment (Dixon Reference Dixon1999; RachBeisel et al. Reference RachBeisel, Scott and Dixon1999). In the de-institutionalization era, no wonder we face such tremendous medical morbidity and mortality from addictions, and homelessness and criminal incarceration of the mentally ill (Lasser et al. Reference Lasser, Boyd, Woolhandler, Heimmerlstein, McCormick and Bor2000; Rosen et al. Reference Rosen, Rosenheck, Shaner, Eckman, Gamache and Krebs2002; Schmetzer Reference Schmetzer2006).

Consistent with these translational perspectives, animal modeling of dual diagnosis has demonstrated that if certain neurodevelopmental lesion models of mental illness are combined with addictive drug exposure, addiction vulnerability phenotypes are accentuated. For instance, neonatal ventral hippocampal lesions (a comprehensive animal model of schizophrenia) produce a host of biological changes involving prefrontal cortical and ventral striatal circuits (Goto & O'Donnell Reference Goto and O'Donnell2002; Lipska et al. Reference Lipska, Lerman, Khaing and Weinberger2003; Tseng et al. Reference Tseng, Lewis, Lipska and O'Donnell2007). These aspects correspond to increases in acquisition of cocaine self-administration, resistance to extinction, and drug-induced relapse of drug-seeking (Chambers & Self Reference Chambers and Self2002). The lesion also produces elevations of an impulsive approach to natural reward before drug exposure, and synergistic worsening of this trait after cocaine exposure in a manner not seen in control animals (Chambers et al. Reference Chambers, Jones, Brown and Taylor2005).

As suggested by Redish et al., identifying differential styles or mechanisms of impulsivity as predictive markers of addiction vulnerability, illness trajectory, or treatment options, is surely a next step for the field. Studying differential forms and patterns of dual diagnosis in both animal models and human populations should represent a major avenue of the exploration. This work will help us elucidate the extent to which the 10 decision-making vulnerabilities suggested by Redish et al. represent truly independent facets of addiction vulnerability or redundant manifestations of the same underlying principle. By determining which of these vulnerabilities carry the most weight of addiction liability in most people, we may arrive at a more parsimonious and “winning” theory of addiction.