Rosemary Tannock
2 de junho de 2014
© 2014 The Author. Journal of Child Psychology and Psychiatry. © 2014 Association for Child and Adolescent Mental Health.
The recent publication of the latest version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) unleashed a torrent of dissatisfaction with and criticism of the ongoing use of a symptom-based classification system, based on descriptive phenotypical features. Thus, the ‘multiple memory systems’ framework proposed in the accompanying Annual Research Review by Goodman et al. (Journal of Child Psychology and Psychiatry, 55, 2014, XX) as a potential explanatory mechanism underlying several psychiatric disorders, is innovative and consistent with an emergent focus on transdiagnostic core mechanisms. In this commentary, I discuss the ‘merit and mire’ of the proposed framework, focusing on its potential for delineating neurobehavioral mechanisms underlying psychiatric disorders, with particular emphasis on one of the neurodevelopmental disorders, ADHD.
The recent publication of the latest version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013) unleashed a torrent of dissatisfaction with and criticism of the ongoing use of a symptom-based classification system, based on descriptive phenotypical features (e.g., Casey et al., 2013; Greenberg, 2013). Thus, the ‘multiple memory systems’ framework proposed in the accompanying Annual Research Review by Goodman, Marsh, Peterson, 2013and Packard (2013) as a potential explanatory mechanism underlying several psychiatric disorders, is innovative and consistent with an emergent focus on transdiagnostic core mechanisms (e.g., Cuthbert & Insel, 2013; Goschke, 2014).
According to the multiple memory systems framework espoused by Goodman and colleagues, two memory systems not only are dissociable behaviorally and neuroanatomically, but also mature at different rates. Specifically, the authors argue that a dorsal-striatal-dependent memory system, which supports stimulus-response ‘habit’ learning and memory, matures earlier than a hippocampal-dependent memory system, which supports a ‘cognitive/declarative’ form of memory. The major premise of the framework is that the asynchronous maturation rate of these two memory systems establishes the ‘preconditions for emergence of developmental psychopathologies’ when one or the other system fails to develop properly or becomes otherwise dysfunctional' (Goodman et al., 2013, p. 584). In their Annual Research Review, the authors apply this framework and their hypotheses to account for habit-like symptoms in a diverse group of psychiatric disorders, namely: Tourette syndrome (TS), attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), obsessive-compulsive disorder (OCD), and eating disorders. In this commentary, I discuss the ‘merit and mire’ of the proposed framework, focusing on its potential for delineating neurobehavioral mechanisms underlying psychiatric disorders, with particular emphasis on one of the neurodevelopmental disorders, ADHD.
The major premise of the authors' theoretical framework is that it is the discrepant maturation of the two memory systems (striatal-habit memory matures earlier than hippocampal memory) that underlies the emergence of habit-like symptoms in all of the disorders (ADHD, TS, ASD, OCD, eating disorders). However, it is useful to disaggregate and review evidence for the underlying set of propositions (some implicit) that comprise this major premise: (i) two memory systems exist that are anatomically and functionally distinct (striatal-dependent habit memory; hippocampal-dependent cognitive or declarative memory);(ii) these two memory systems develop at different rates, with the striatal-habit memory system maturing before the hippocampal memory system; (iii) the differing maturation rates of the two memory systems contribute to the early emergence of habit-like symptoms in childhood and adolescence; (iv) habit-like symptoms are core features of many psychiatric disorders that begin in childhood of adolescence; (v) these disorders are associated with anatomical and functional abnormalities in striatal circuits and sometimes in hippocampal circuits; (vi) the different rates of maturation of the striatal-habit memory and hippocampal cognitive/declarative memory contribute to the habit-like features associated with these psychiatric disorders; and (vii) abnormal functioning of one or both of the two memory systems may also contribute to the manifestation and maintenance of habit-like symptoms in these psychiatric disorders.
Multiple memory systems framework: Are there two distinct memory systems?
I use the term ‘framework’ rather than ‘theory’ or ‘model’ since the framework and its hypotheses are somewhat general and not falsifiable (Popper, 1976). The multiple memory systems framework, which was first theorized in the 1970's (Tulving, 1972) and now widely accepted, posits that anatomically and functionally distinct mnemonic circuits support different types of learning and memory: namely, a dorsal-striatal system that supports ‘habit’ or ‘implicit’ memory, and a hippocampal circuit that supports ‘cognitive’ or ‘explicit, declarative’ memory.
Burgeoning evidence from animal lesion studies and amnesic patients over the past few decades indicate that memory is not a unitary process but rather consists of multiple processes that involve distinct neural substrates. One of the most popular theories within the multiple memory systems framework is the declarative-nondeclarative theory (e.g., Squire, 1992; Tulving, 1972). The declarative or explicit memory system consists of two dissociable types (episodic/events, semantic/facts), whereas the nondeclarative, implicit system comprises several types of memory processes (habit, priming, simple classical conditioning, nonassociative learning). These two relatively distinct memory systems are believed to engage in parallel and may either interact synergistically or compete with one another for control of behavior (Poldrack & Packard, 2003). However, these two systems interact and at times hippocampal-depending ‘cognitive’ memory may contribute to habit learning and memory; conversely, striatal-dependent memory system may contribute to declarative memory. That is, habit-tasks can engage both memory systems as can cognitive/declarative type tasks.
Other frameworks have been proposed, including the ‘processing modes framework’ which links explicit memory with conceptual processing and implicit with perceptual (e.g., Roediger, Welson, & Challis, 1989), and a ‘component process framework’, which posits numerous processing components that are recruited in various combinations by memory tasks and yield complex patterns of associations and dissociations (e.g., Cabeza & Moscovitch, 2013; Moscovitch, 1992). Nonetheless, the extant evidence for the two memory systems central to the major premise in question is adequate and permits review of the second premise: namely that the two memory systems mature at different rates, with the striatal-habit memory system maturing earlier than the hippocampal-declarative memory system.
Do the two memory systems develop asynchronously in humans?
As reviewed by Goodman and colleagues, both neurobiological and neurobehavioral evidence from rodents and nonhuman primates is fairly consistent and robust: the striatal-dependent habit memory system matures within the first month of life, whereas the hippocampal-dependent memory systems develops about 2–3 years later (also, see review by Bachevalier, 2013). By contrast, neurobiological findings of different maturation rates of these two memory systems are limited and inconsistent, so these authors were required to base their premise primarily on neurobehavioral evidence, which also yields equivocal evidence, in part because findings often appear to be task-dependent. Nonetheless, some evidence suggests that children show a gradual shift from the use of a striatal-dependent S-R response type learning and memory to a hippocampal-dependent declarative type process between 5 and 10 years of age, although findings are inconsistent (e.g., Bohbot et al., 2011).
What are ‘habits’?
The authors liken habits to the clinical manifestations of “impulsive and semi-compulsory thoughts and behavioral symptoms…that have escaped regulatory control” manifest in the specified psychiatric disorders, and characterize these symptoms as being ‘rigid and invariant’ (Goodman et al., 2013; pages 590 and 601, respectively). Historically, the term ‘habit learning’ was first coined in the 1970's (Hirsh, 1974) to describe a particular type of memory and learning system that was not hippocampally based and thus did not involve contextual encoding in memory (i.e., does not involve information about the spatial or temporal context). Later, Miskin and colleagues (Miskin, Malamut, & Bachevalier, 1984) defined habit learning as a relatively primitive and slow form of learning that was reliable but inflexible, and appears earlier in development as well as in evolution. In the animal literature, a habit is defined as a behavior that is elicited by environmental stimuli to which it has become strongly tied (i.e., under the control of S-R mechanisms), and that have become insensitive to both outcome value and response-outcome contingency (e.g., Dickinson, 1985). Automaticity, as another key characteristic of habits was emphasized by Graybiel (2008), who defined them as mannerisms, customs, or rituals that are largely learned, occur repeatedly over the course of days or years, become remarkably fixed, and performed almost automatically, allowing attention to be focused elsewhere. Furthermore, Graybiel (2008) proposed that habits and typically involving an ordered and structured action-sequence that is elicited by a particular context or stimulus, and which may comprise cognitive expressions of routine (i.e., habits of thought), as well as motoric expressions of routine (i.e., habitual behavior). However, it remains unclear whether a habit shares the same defining features with those of automatic behavior (Ashby, Turner, & Horvitz, 2010; Seger & Spiering, 2011). Nonetheless, whereas Goodman and colleagues emphasize rigidity, inflexibility, and lack of self-regulatory control, as defining characteristics of habits and habit-like behavior, others propose a broader array of characteristics, such as automaticity, insensitivity to devaluation of the reinforcer or in the response-outcome contingency, and having minimal or no demands on attention. Thus, next I consider the extent to which habit-like symptoms manifest in very young typically developing children and in children with ADHD or ASD.
Do individuals with ADHD manifest habit-like symptoms?
Assuming that the habit memory system does mature before the declarative memory system in humans, then presumably habit-like behavior must also be manifested in young typically developing children – an issue not addressed by Goodman and colleagues. Evidence is sparse, but interpretation of findings from a study of age-related performance of A-not-B tasks or Day-Night Stroop tasks in young children suggests that preschoolers manifest great difficulty inhibiting a well-learned ‘automatic’ or ‘habitual’ response to such tasks and in developing and remembering a higher-order representation of the task rules, whereas older children do not (Diamond, 2009; Marcovitch & Zelazo, 2009). However, it is unclear whether Goodman and colleagues are proposing that habit-like behavior in psychiatric disorders is attributable to abnormally large differences in rates of maturation of the two memory systems (i.e., even more protracted maturation of the hippocampal system), rather than the typical differential rate of maturation per se.
Goodman and colleagues propose that the impulsive and hyperactive symptoms of ADHD constitute habit-like behavior. The primary evidence cited in support of their claim comes from neuroimaging studies, which typically reveal reduced volumes of the basal ganglia nuclei, particularly the putamen in children, adolescents and adults with ADHD (for meta-analysis, see Frodl & Skokauskas, 2012). However, findings for compensatory increases in hippocampal volume are inconsistent (e.g., Perlov et al., 2008 vs. Onnink et al., 2014). Also, as acknowledged by Goodman and colleagues, neurobehavioral evidence of deficits in striatal-based learning and memory in ADHD is sparse and inconclusive. So, one key question is whether the impulsive or hyperactive symptoms manifest in ADHD constitutes habit-like behavior. Robust evidence exists for deficient inhibitory control in ADHD, which is believed to underlie impulsivity (e.g., Lipszyc & Schachar, 2010). If inhibitory deficits constitute habit-like behavior, then performance should be immune to the effects of reinforcement, as per definitions of habits. But recent evidence using a novel task that integrates measures of reinforcer anticipation within a go/no-go paradigm to measure inhibitory control, suggests that this is not the case (Demurie, Roeyers, Wiersema, & Sonuga-Barke, 2013). Positive reinforcement not only decreased inhibitory deficits in children with ADHD, as well as those with ASD, but also both groups of children showed sensitivity to different type of reinforcers: namely monetary ones improved go-response performance to a greater extent than did social reinforcers (e.g., Demurie et al., 2013). Furthermore, inhibitory performance of an ADHD+ASD group was normalized during the reinforced trials. Consistent with previous research (e.g., Kohls, Herpertz-Dahlmann, & Konrad, 2009; Marx, Höpcke, Berger, Wandschneider, & Herpertz, 2013), these findings do not support the proposition that inhibitory deficits observed in ADHD (at least as measured by go/no-go or stop-signal tasks) are rigid, fixed, and immune to changes in response-outcome contingencies – the characteristics of habit behavior.
Conclusions
The multiple memory systems framework account of habit-like behavior in psychiatric disorders is innovative and the authors have provided a comprehensive and critical review of the literature. Their primary postulate of discrepant maturation of two memory systems as setting the preconditions for an overreliance on the habit-learning/memory system is provocative. A somewhat different explanation of habit-like behavior in psychiatric disorders has been proposed: namely that inflexible habit-like behavior may derive from an imbalance or disrupted regulation of the balance between the two memory systems (Pittenger, 2013). This proposition endorses Goodman's notion of inadequate top-down regulatory control of striatal functions, resulting in an ‘overreliance on the habit learning system.” (Goodman et al.,2013, p.10).
Whereas this framework and its primary postulate garner support from research findings in addiction and obsessive-compulsive behavior, and even in Tourette syndrome, I find it less compelling as an explanation of impulsive/hyperactive symptoms in ADHD.
This Commentary was invited by the Editor-in-Chief of JCPP and has been subject to internal review; the author has declared that she has received research funding in kind from Pearson-Cogmed; honoraria from Shire for an unrestricted scientific talk; and travel and accommodation from Biomed Central to attend the annul Editors meeting. Also, Dr Tannock was a member of the DSM-5 workgroup for ADHD and externalizing disorders.
Artigo original: