Editorial: Categories and dimensions: reflections on obsessive-compulsive disorder (OCD)


  • Daniel S. Pine

While the life of an editor is filled with joys and challenges, a few tasks uniquely combine the two experiences. One such task involves playing the role of “communicator”, to convey a sense about where the field has been and where it is moving. Many duties provide opportunities to communicate this sense, but none is greater than preparing an editorial. The current issue is ripe with such opportunity. However, Huyser et al. caught my eye, appearing sandwiched among a commentary and two additional original articles, all four of which focus on obsessive compulsive disorder (OCD). Together, these four articles provide a chance to reflect on OCD and broader issues raised when considering the nature of the diagnosis.

During the past decade, research has adopted dimensional perspectives with increasing frequency, which acknowledges the failure to find “points of rarity” related to risk. This also applies to research on OCD specifically. For example, sub-clinical expressions typically are found more frequently than frank cases in children of parents with various forms of mental illness. Similarly, in OCD, specific, relatively-narrow symptom dimensions, such as a tendency to hoard, tend to aggregate within families more strongly than the categorical diagnosis of OCD. Finally, sub-clinical expressions represent significant risks for later full-blown instances of psychopathology. Longitudinal studies for many conditions find smooth or continuous distributions connecting levels of symptoms, manifesting at one point in time, to levels of risk for frank disorders at later points in time. This raises questions on the value of categorical perspectives. Are we to abandon this approach completely? I certainly hope not.

Perhaps the most convincing evidence of value in the categorical perspective emerges when considering the experiences of the clinician. Even if risk ultimately is shown only to manifest continuously, for the clinician, decisions on treatment are made categorically. Because resources are limited and because treatments usually carry risk, clinicians regularly must answer a series of categorical questions, even if they are posed against a dimensional landscape. Is the level of suffering or impairment in this particular child sufficiently extreme to warrant treatment? And if so, does the clinician begin treatment with psychotherapy? Or should treatment begin with with medication? In other words, do the alternative risks of psychotherapy or psychopharmacology outweigh the distinct potential benefits for the two treatments? Research on therapeutics in OCD forces the clinician to confront this series of categorical questions with particular alacrity. This is because both cognitive behavioral therapy (CBT) and selective-serotonin-reuptake inhibitor medications have been shown to reduce symptoms of OCD. Moreover, there is some evidence that the combination of the two might provide more benefit than either treatment alone. Nevertheless, these two treatments carry complementary advantages and disadvantages, which require categorical, serial evaluation of each modality.

Beyond the trenchant nature of categorical approaches to clinical experience, other data raise questions on the appropriateness of categorical as opposed to dimensional perspectives on OCD. Clearly, as they do for major depressive disorder (MDD) and attention deficit hyperactivity disorder (ADHD), data from family and outcome studies in pediatric OCD provide stronger support for the dimensional than the categorical perspective. However, aspects of the clinical picture in OCD, at least when fulminate, can suggest a categorical departure from typical emotional experiences of the child. One can question the degree to which a child with severe OCD, that prevents the child from leaving the house due to their fears of contamination and need to recurrently wash everything surrounding them, suffers from a condition wholly distinct from children who exhibit mild contamination fears.

Questions on the appropriateness of the dimensional or categorical perspective might best be answered through review of the data on external validity. Indeed, the smooth distribution in the risk for many conditions in family-based and longitudinal work typically supports continuous perspectives. Pharmacological and biological studies provide other particularly-important opportunities to acquire data on external validity. While pharmacological data support dimensional perspectives on ADHD, relative to pharmacology, advances in brain imaging enable broader applications of clinical neuroscience. As such, brain imaging provides an unusual chance to evaluate categorical and dimensional perspectives. In the current issue of the Journal, it is this unique opportunity ultimately that drew me to Huyser and colleagues.

Ideally, future longitudinal, clinically-focused brain-imaging studies should be able to assess brain function dimensionally and then chart over time changes in brain function, as it relates to changes in symptoms of OCD. If these data were to reveal a smooth distribution between variations in brain function and variation in OCD symptoms, this would further support increasingly-popular dimensional perspectives. Nevertheless, in some areas of medicine, discrete, discontinuous associations are seen between biological markers and clinical status. For example, biological indicators of at least some cancers and forms of myocardial injury exhibit discontinuous relationships with clinical endpoints, justifying a categorical perspective. Indeed, for OCD, the sudden, fulminate appearance of symptoms following streptococcal infection similarly suggests that discontinuities could manifest in this condition. Much as serial stress-induced cardiograms can generate insights on cardiac pathology, longitudinal data on brain function and OCD might reveal discontinuities in brain-behavior relationships. Huyser et al. furthers the process of generating such data.

Huyser and colleagues chart symptoms, information-processing, and brain function in 25 patients with OCD and 25 matched comparisons over a few months, during which time patients with OCD received CBT. The article reveals complex associations among brain function, age, clinical status, and time. Such complexity has become the norm in research on brain function in children, followed over time. The truly unique contribution from Huyser and colleagues’ work is to go beyond most previous imaging studies not only by creatively using a cognitive-imaging approach but also by acquiring data in patients and healthy peers over time. This unique feature allows the study to generate insights on state- and trait-related correlates of OCD.

The primary contribution in the article is to parse brain regions based on the degree to which they show state- or trait-related associations with OCD. Two regions of frontal cortex, in the insula and anterior cingulate gyrus, are shown to manifest perturbed function in at least some sub-set of OCD patients. Before considering the broader implications of these findings, it is important to note the specific way in which Huyser et al. extend the work of others. Thus, they utilize a particular imaging approach, paired with a well-known interference task, to extend considerable neuropsychological work on perturbed error processing in OCD. Moreover, they focus on regions of fronto-striatal circuitry linked to OCD through prior imaging and basic science work. Finally, they examine changes in brain function after CBT, one of two established OCD treatments.

If these findings are replicated in more definitive, larger studies, they might encourage future research efforts. These efforts might pursue the explicit goal of weighing dimensional and categorical perspective on OCD, as it relates to underlying brain function. Such studies might show that frontal function exhibits a continuous relationship with concurrent and future manifestations of OCD. For instance, the risk for future OCD in offspring of OCD patients followed longitudinally might be shown to increase incrementally, in tandem with incremental changes in frontal function in these children, studied at an early, pre-symptomatic point in time. Alternatively, such studies might find that frontal function exhibits a discontinuous relationship with risk. Namely, the risk for future OCD might increase in a quantum fashion at some point along the distribution of frontal function, as charted in pre-symptomatic children. This would support categorical perspectives. While results from such studies are unlikely to emerge for many years if not decades, even the opportunity to consider using imaging data in such clinically-focused work illustrates just how far we have come as a field.

In closing, Huyser and colleagues, like so many other articles in the Journal, provide yet one more opportunity to weigh progress in research on developmental psychopathology, where it has been and where it is going. Nevertheless, in other ways, the process of distilling the message from Huyser and colleagues is unique. No doubt this partly reflects unique features of OCD, a condition with two established treatments and with a presentation that can be viewed alternatively as an extreme end of a continuum or as a categorical departure from normal development. Finally, and perhaps most importantly, the Huyser et al. article provides a rare opportunity to illuminate the increasing clinical relevance of brain imaging research, as it has advanced steadily this decade. Thus, we can now begin to use neuroscience in the service of clinical questions, including weighing the advantages and disadvantages of dimensional and categorical perspectives.



The opinions and assertions contained in this Editorial are the personal views of the author and are not construed as official or as reflecting the views of the NIMH or the US Department of Health and Human Services. D.S.P. has declared that he has no potential or competing interests.