Historically, a key question pervading both psychiatry and clinical psychology has been the identification of the elusive boundary between normal and abnormal behavior (e.g., Wakefield 1997; Widiger and Clark 2000). One specific form that this question has taken pertains to establishing meaningful boundaries between different “types” of abnormal behavior. In this respect, two opposing views have been popular over the years: (1) the “splitter” approach, which posits that the best way to understand dysfunction is by identifying an increasingly larger number of diagnoses, and (2) the “lumper” approach, which emphasizes functional similarities across pathologies and proposes dimensional models of psychopathology (e.g., Barlow et al. 2014; Cuthbert 2005; Mennin et al. 2008).

Recently, a middle ground between these two positions emerged with the advent of the transdiagnostic approach (e.g., Harvey et al. 2004; Kring and Sloan 2009). Its primary goal has been to identify which pathological processes cut across different disorders (i.e., transdiagnostic) and which are confined to specific forms of psychopathology (i.e., disorder-specific). For example, elevated negative affect is present across all mood and anxiety disorders (e.g., Watson 2009) and, as such, can be conceptualized as a transdiagnostic feature of these conditions. On the other hand, reduced positive affect characterizes only depression and social anxiety disorder (e.g., Brown 2007). Thus, while it is transdiagnostic in relation to those two conditions, it is disorder-specific when taking into account the larger landscape of mood and anxiety pathology. In this way, the transdiagnostic approach retains traditional, symptom-based categories (i.e., the “splitter” approach) while seeking to identify underlying dimensions of pathological functioning (i.e., the “lumper” approach). Given this integrative nature, it is not surprising that it has become an extremely valuable framework for identifying factors leading to the development and/or maintenance of disorders and for identifying patterns of comorbidity (e.g., Aldao 2012; Gruber et al. 2008; Hofmann et al. 2012; Nolen-Hoeksema and Watkins 2011; Wade et al. 2006) as well as for treating comorbid conditions (e.g., Barlow et al. 2004; Ehrenreich-May et al. 2012; Fairburn et al. 2008; Hayes et al. 1999; Norton et al. 2004).

However, in the past few years, the growth of this transdiagnostic approach appears to have stalled as the field has once again gravitated towards the two extremes of “splitting” and “lumping”. On the one hand, the latest edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM; American Psychiatric Association 2013) has adopted a “splitter” approach, as exemplified, for example, by the creation of new disorder categories (e.g., in DSM-IV-TR obsessive compulsive disorder was in the “anxiety disorders” category whereas in DSM-5 it is the anchor disorder of its own diagnostic category). On the other hand, the National Institute of Mental Health (NIMH) has developed research guidelines that reject traditionally defined symptom-based diagnoses and emphasize the identification of underlying mechanisms. Specifically, the Research Domain Criteria (RDoC; Insel et al. 2010) embodies the essence of the “lumper” approach, as it proposes to study pathological processes (e.g., negative valence systems, systems for social processes) dimensionally and across multiple units of analyses (e.g., self-reports, behavior, physiology, circuitry, genes).

We can think about the DSM and RDoC approaches as analogous to selecting ice cream flavors. Let us imagine that all the ice cream parlors in a town undergo a major overhaul in their menus: some parlors adopt a DSM-inspired menu and others choose an RDoC-inspired one. In the parlors with DSM-inspired menus, we would find that each of the flavors had been broken down into many more flavors by taking into account combinations with toppings. For example, chocolate with sprinkles on top would become its own flavor, as would chocolate with a cherry on top. The menu would continue to grow to the point that there would be no space on the walls to display more menu boards. Customers would have a difficult time finding and remembering what they want. On the other hand, in RDoC-inspired parlors, the menus would only list ingredients. Customers would then have to figure out exactly how much milk, cream, eggs, cocoa, vanilla, etc., to request in order to get the flavor they want. Although the occasional ice cream connoisseur might have a lot of fun with this process, most customers would be very confused. Thus, both, the DSM- and RDoC-inspired ice cream parlors would generate confusion and frustration in customers and, their day-to-day operations would be quite inefficient.

One solution would be to go back to a classic menu consisting of a handful of flavors, such as chocolate, strawberry, and vanilla. From there, the parlor owners could experiment with a few variations in the menu: some menus would list flavors whereas others would list ingredients. In a similar way, a fruitful approach to maximizing the benefits from the DSM and RDoC approaches involves relying on symptom-based categories while systematically identifying dimensions of dysfunction that cut across such categories. In other words, it might be useful to go back to the “classic” transdiagnostic approach. Below I describe two ways in which this can be done.

Conducting Studies That Include More Than One Diagnostic Group

Close inspection of the clinical science literature reveals that many studies still include only one diagnostic group and one control group. Doing so is problematic because it does not allow investigators to conclude whether the observed group differences are a function of psychopathology in general or of a specific condition. By systematically including more than one diagnostic group, however, studies can separate which processes pertain to multiple forms of psychopathology versus which have a certain amount of specificity to a given diagnostic group. Thus, in order to advance our understanding of dysfunction in psychopathology, it is essential that we more systematically include multiple diagnostic groups in our studies.

Analyzing Symptoms Dimensionally

In addition to utilizing diagnosed samples, it is important to analyze symptoms dimensionally, both across clinical and healthy groups and also within clinical groups. This practice, although uncommon, is crucial in order to better understand functional overlap at the symptom level. In addition, it can help elucidate the dose-dependent relationship between symptoms and other pathological processes (e.g., do emotion regulation deficits become more pronounced as the severity of a disorder increases?). A dimensional approach can help us elucidate whether such relationships are linear or quadratic, and/or whether they are moderated by the presence of symptoms of additional conditions. Of particular importance, the adoption of a developmental psychopathololgy approach (Cicchetti et al. 1995) would allow one to identify risk and protective factors as well as sequential comorbidity (e.g., Beauchaine and McNulty 2013; Nolen-Hoeksema and Watkins 2011).

Goal of This Special Issue

The main goal of this Special Issue is to advance psychopathology research by assembling a collection of studies based on the classic transdiagnostic approach. To that end, I included 12 studies that either by themselves or, in combination with others, shed light onto whether pathological processes could be considered transdiagnostic or disorder-specific. In all, studies covered a wide range of conditions, including generalized anxiety disorder (GAD), social anxiety disorder (SAD), major depressive disorder (MDD), bipolar disorder (BPAD), borderline personality disorder (BPD), anorexia nervosa (AN), bulimia nervosa (BN), substance use disorder (SUD), and oppositional defiant disorder (ODD).

In terms of the pathological process of interest, this Special Issue focused on emotion regulation. Three reasons guided this choice. First, the study of emotion regulation has grown exponentially in the past 15 years (Gross 2013) and it has received much attention within the clinical science literature (as reviewed in Aldao et al. 2010, 2015; Tracy et al. 2014; Sheppes et al. 2015). In fact, the term “clinical affective science” has begun to be utilized to denote this fast growing field. As such, it is an important framework to understand dysfunction. Second, there is growing evidence suggesting that emotion regulation is a transdiagnostic process that spans a wide range of conditions (as reviewed by Aldao et al. 2010; Harvey et al. 2004; Hofmann et al. 2012; Kring and Sloan 2009; Nolen-Hoeksema and Watkins 2011). Third, the process of emotion regulation is multi-faceted, and thus cuts across several units of analyses, such as subjective feelings, facial expressivity, physiological arousal, and neural activation (e.g., Aldao 2013; Mauss et al. 2005; Gross 2015). As such, it is an ideal target for the study of dimensional processes that span multiple systems (Insel et al. 2010). Below, I provide a brief overview of the scope of these articles. A more in-depth discussion can be found in the commentary written by James Gross and his colleagues.

Conducting Studies That Include More Than One Diagnostic Group

A number of studies compared emotion regulation in relation to more than one form of psychopathology. In this respect, Van Meter and Youngstrom (2015) contrasted BPAD to MDD, MacNamara et al. (2015), Seeley et al. (2015) contrasted GAD with MDD, and Naumann et al. (2015) compared AN to BN.

Analyzing Symptoms Dimensionally

Van Meter and Youngstrom (2015), MacNamara et al. (2015), and Naumann et al. (2015) examined comorbid types of psychopathology (e.g., bipolar disorder, depression, anxiety, eating disorders) both dimensionally and categorically. Also, Rosenthal et al. (2016), Shapero et al. (2015), and Hofmann et al. (2016) analyzed symptoms of comorbid conditions (e.g., anxiety and depression, BPD ad PSTD).

Other studies sought to model the complex interplay between different types of symptoms and emotion regulation. In this respect, Farris et al. (2015) examined whether the link between emotion regulation difficulties and cigarette smoking relapse would vary as a function of past year psychopathology. Poon et al. (2015) tested whether trait emotion regulation difficulties with interact with cortisol reactivity to a social stressor to predict substance abuse and symptoms of MDD and ODD. Similarly, Dixon-Gordon et al. (2015) examined the link between adolescents’ negative affect and maternal problem solving and validation and symptoms of BPD. Crucially, Heleniak et al. (2015) adopted a developmental psychopathology approach and tested whether emotion regulation would mediate the longitudinal association between child maltreatment and internalizing and externalizing symptoms. Lastly, Blalock et al. (2015) examined the daily use of regulation strategies in participants diagnosed with social anxiety disorder and healthy controls. Although they did not examine comorbid symptoms, they tested the temporal link between social anxiety and mood fluctuations. As such, this work illustrates how to elegantly assess emotion regulation outside of laboratory settings (e.g., Aldao 2013).

Concluding Remarks

In recent years, there has been much debate as to whether the study of psychopathology should adopt DSM- or RDoC-based approaches. The main goal of this Special Issue is to reconcile these perspectives by advocating for a “classic” transdiagnostic perspective that retains symptom-based categories while systematically identifying which processes are transdiagnostic versus disorder-specific. I hope that these superb studies as well as the commentary (James Gross and his colleagues) stimulate future research in clinical affective science that can help us identify the role of affective dysfunction across various forms of psychopathy as well as develop a better understanding of functional relationships among comorbid conditions.