Co-morbid obsessive–compulsive disorder and depression: a Bayesian network approach

McNally, Richard J.; Mair, Patrick; Mugno, Beth L.; Riemann, Brad C

doi:10.1017/s0033291716003287

Cited by 188 publications

(183 citation statements)

References 43 publications

Supporting

Mentioning

164

Contrasting

Unclassified

Order By: Relevance

“…However, replicability of specific edges was notably worse in the split-half pairs, with a median of a quarter to over a third of the edges failing to replicate. This finding is in contrast to McNally et al’s (2017) suggestion that this method “depicts only those edges nearly certain to be genuine” (p. 1207) and highlights the sensitivity of DAGs to small differences in the relationships among symptoms in the network. Overall, the lack of stability and generalizability in the DAGs is to be expected once we understand the assumptions, discussed below, that underlie the estimation and interpretability of the models.…”

Section: Discussioncontrasting

confidence: 93%

See 1 more Smart Citation

Evidence that psychopathology symptom networks have limited replicability.

Forbes¹,

Wright²,

Markon³

et al. 2017

Journal of Abnormal Psychology

274

215

View full text Add to dashboard Cite

Network analysis is quickly gaining popularity in psychopathology research as a method that aims to reveal causal relationships among individual symptoms. To date, four main types of psychopathology networks have been proposed: (1) association networks, (2) regularized concentration networks, (3) relative importance networks, and (4) directed acyclic graphs. We examined the replicability of these analyses based on symptoms of major depression and generalized anxiety between and within two highly similar epidemiological samples (i.e., the National Comorbidity Survey – Replication [n = 9282] and the National Survey of Mental Health and Wellbeing [n = 8841]). While association networks were stable, the three other types of network analysis (i.e., the conditional independence networks) had poor replicability between and within methods and samples. The detailed aspects of the models—such as the estimation of specific edges and the centrality of individual nodes—were particularly unstable. For example, 44% of the symptoms were estimated as the “most influential” on at least one centrality index across the six conditional independence networks in the full samples, and only 13–21% of the edges were consistently estimated across these networks. One of the likely reasons for the instability of the networks is the predominance of measurement error in the assessment of individual symptoms. We discuss the implications of these findings for the growing field of psychopathology network research, and conclude that novel results originating from psychopathology networks should be held to higher standards of evidence before they are ready for dissemination or implementation in the field.

show abstract

Section: Discussioncontrasting

confidence: 93%

“…DAGs were computed based on Bayesian network analyses (i.e., the hill-climbing algorithm from the R package bnlearn ; Scutari, 2010), as described in McNally, Mair, Mungo, and Riemann (2017). The hill-climbing algorithm adds, removes, and reverses edges until a target Bayesian Information Criterion score is reached.…”

Section: Methodsmentioning

confidence: 99%

Evidence that psychopathology symptom networks have limited replicability.

Forbes¹,

Wright²,

Markon³

et al. 2017

Journal of Abnormal Psychology

274

215

View full text Add to dashboard Cite

show abstract

“…Finally, studying comorbid depression and obsessive-compulsive disorder (OCD), McNally et al [74] found that both clusters of symptoms were connected through “sadness,” but sleep and appetite symptoms were not connected to either depression or OCD. These types of results illustrate that not all symptoms play the same role when developing comorbid disorders.…”

Section: Resultsmentioning

confidence: 99%

The Study of Psychopathology from the Network Analysis Perspective: A Systematic Review

Contreras

Nieto

Valiente

et al. 2019

Psychother Psychosom

264

217

View full text Add to dashboard Cite

Background: Network analysis (NA) is an analytical tool that allows one to explore the map of connections and eventual dynamic influences among symptoms and other elements of mental disorders. In recent years, the use of NA in psychopathology has rapidly grown, which calls for a systematic and critical analysis of its clinical utility. Methods: Following PRISMA guidelines, a systematic review of published empirical studies applying NA in psychopathology, between 2010 and 2017, was conducted. We included the literature published in PubMed and PsycINFO using as keywords any combination of “network analysis” with the terms “anxiety,” “affective disorders,” “depression,” “schizophrenia,” “psychosis,” “personality disorders,” “substance abuse” and “psychopathology.” Results: The review showed that NA has been applied in a plethora of mental disorders in adults (i.e., 13 studies on anxiety disorders; 19 on mood disorders; 7 on psychosis; 1 on substance abuse; 1 on borderline personality disorder; 18 on the association of symptoms between disorders), and 6 on childhood and adolescence. Conclusions: A critical examination of the results of each study suggests that NA helps to identify, in an innovative way, important aspects of psychopathology like the centrality of the symptoms in a given disorder as well as the mutual dynamics among symptoms. Yet, despite these promising results, the clinical utility of NA is still uncertain as there are important limitations on the analytic procedures (e.g., reliability of indices), the type of data included (e.g., typically restricted to secondary analysis of already published data), and ultimately, the psychometric and clinical validity of the results.

show abstract

“…To ensure the stability of the DAG, we bootstrapped 10,000 samples, computing a network for each sample (McNally et al, 2017). We then averaged them to obtain the final, resultant network.…”

Section: Methodsmentioning

confidence: 99%

“…First, we computed an undirected, regularized partial correlation graph – the most popular method for modelling psychopathology (Epskamp & Fried, 2016). Next, we used Bayesian methods to compute a directed, acyclic graph – a relatively new approach to modelling psychopathology (McNally, Mair, Mugno, & Riemann, 2017). Both approaches test for links between symptoms after adjusting for the influence of other symptoms, but their strengths and weaknesses are mirror images of one another.…”

mentioning

confidence: 99%

A Bayesian network analysis of posttraumatic stress disorder symptoms in adults reporting childhood sexual abuse

McNally

Heeren

Robinaugh

2017

European Journal of Psychotraumatology

161

214

View full text Add to dashboard Cite

Background: The network approach to mental disorders offers a novel framework for conceptualizing posttraumatic stress disorder (PTSD) as a causal system of interacting symptoms. Objective: In this study, we extended this work by estimating the structure of relations among PTSD symptoms in adults reporting personal histories of childhood sexual abuse (CSA; N = 179). Method: We employed two complementary methods. First, using the graphical LASSO, we computed a sparse, regularized partial correlation network revealing associations (edges) between pairs of PTSD symptoms (nodes). Next, using a Bayesian approach, we computed a directed acyclic graph (DAG) to estimate a directed, potentially causal model of the relations among symptoms. Results: For the first network, we found that physiological reactivity to reminders of trauma, dreams about the trauma, and lost of interest in previously enjoyed activities were highly central nodes. However, stability analyses suggest that these findings were unstable across subsets of our sample. The DAG suggests that becoming physiologically reactive and upset in response to reminders of the trauma may be key drivers of other symptoms in adult survivors of CSA. Conclusions: Our study illustrates the strengths and limitations of these network analytic approaches to PTSD.

show abstract

Co-morbid obsessive–compulsive disorder and depression: a Bayesian network approach

Abstract: Bayesian analysis can expand the repertoire of network analytic approaches to psychopathology. We discuss clinical implications and limitations of our findings.

Cited by 188 publications

References 43 publications

Evidence that psychopathology symptom networks have limited replicability.

Evidence that psychopathology symptom networks have limited replicability.

The Study of Psychopathology from the Network Analysis Perspective: A Systematic Review

A Bayesian network analysis of posttraumatic stress disorder symptoms in adults reporting childhood sexual abuse

Contact Info

Product

Resources

About