A tutorial on regularized partial correlation networks.

Published on Mar 29, 2018in Psychological Methods
· DOI :10.1037/MET0000167
Sacha Epskamp33
Estimated H-index: 33
(UvA: University of Amsterdam),
Eiko I. Fried33
Estimated H-index: 33
(UvA: University of Amsterdam)
Recent years have seen an emergence of network modeling applied to moods, attitudes, and problems in the realm of psychology. In this framework, psychological variables are understood to directly affect each other rather than being caused by an unobserved latent entity. In this tutorial, we introduce the reader to estimating the most popular network model for psychological data: the partial correlation network. We describe how regularization techniques can be used to efficiently estimate a parsimonious and interpretable network structure in psychological data. We show how to perform these analyses in R and demonstrate the method in an empirical example on posttraumatic stress disorder data. In addition, we discuss the effect of the hyperparameter that needs to be manually set by the researcher, how to handle non-normal data, how to determine the required sample size for a network analysis, and provide a checklist with potential solutions for problems that can arise when estimating regularized partial correlation networks. (PsycINFO Database Record (c) 2018 APA, all rights reserved).
📖 Papers frequently viewed together
302 Citations
668 Citations
240 Citations
#1René MõttusH-Index: 30
#2Michael AllerhandH-Index: 21
14 Citations
#1Sacha Epskamp (UvA: University of Amsterdam)H-Index: 33
#2Denny Borsboom (UvA: University of Amsterdam)H-Index: 76
Last. Eiko I. Fried (UvA: University of Amsterdam)H-Index: 33
view all 3 authors...
The usage of psychological networks that conceptualize behavior as a complex interplay of psychological and other components has gained increasing popularity in various research fields. While prior publications have tackled the topics of estimating and interpreting such networks, little work has been conducted to check how accurate (i.e., prone to sampling variation) networks are estimated, and how stable (i.e., interpretation remains similar with less observations) inferences from the network s...
743 CitationsSource
#1Eiko I. Fried (UvA: University of Amsterdam)H-Index: 33
#2Marloes B. EidhofH-Index: 3
Last. Karen-Inge KarstoftH-Index: 19
view all 10 authors...
The growing literature conceptualizing mental disorders like posttraumatic stress disorder (PTSD) as networks of interacting symptoms faces three key challenges. Prior studies predominantly used (a) small samples with low power for precise estimation, (b) nonclinical samples, and (c) single samples. This renders network structures in clinical data, and the extent to which networks replicate across data sets, unknown. To overcome these limitations, the present cross-cultural multisite study estim...
174 CitationsSource
#1Junhao Pan (SYSU: Sun Yat-sen University)H-Index: 9
#2Edward H. Ip (Wake Forest University)H-Index: 33
Last. Laurette Dubé (Desautels Faculty of Management)H-Index: 57
view all 3 authors...
: As a commonly used tool for operationalizing measurement models, confirmatory factor analysis (CFA) requires strong assumptions that can lead to a poor fit of the model to real data. The post hoc modification model approach attempts to improve CFA fit through the use of modification indexes for identifying significant correlated residual error terms. We analyzed a 28-item emotion measure collected for n = 175 participants. The post hoc modification approach indicated that 90 item-pair errors w...
24 CitationsSource
#1Marie K. Deserno (UvA: University of Amsterdam)H-Index: 8
#2Denny Borsboom (UvA: University of Amsterdam)H-Index: 76
Last. Hilde M. Geurts (UvA: University of Amsterdam)H-Index: 42
view all 4 authors...
Given the heterogeneity of autism spectrum disorder, an important limitation of much autism spectrum disorder research is that outcome measures are statistically modeled as separate dependent variables. Often, their multivariate structure is either ignored or treated as a nuisance. This study aims to lift this limitation by applying network analysis to explicate the multivariate pattern of risk and success factors for subjective well-being in autism spectrum disorder. We estimated a network stru...
36 CitationsSource
#1Aaron J. Fisher (University of California, Berkeley)H-Index: 21
#2Jonathan W. Reeves (University of California, Berkeley)H-Index: 5
Last. Julian A. Rubel (University of Trier)H-Index: 23
view all 5 authors...
: Individual variation is increasingly recognized as important to psychopathology research. Concurrently, new methods of analysis based on network models are bringing new perspectives on mental (dys)function. This current work analyzed idiographic multivariate time series data using a novel network methodology that incorporates contemporaneous and lagged associations in mood and anxiety symptomatology. Data were taken from 40 individuals with generalized anxiety disorder (GAD), major depressive ...
110 CitationsSource
#1Miriam K. Forbes (UMN: University of Minnesota)H-Index: 21
#2Aidan G. C. Wright (University of Pittsburgh)H-Index: 56
Last. Robert F. Krueger (UMN: University of Minnesota)
view all 4 authors...
: Network analysis is quickly gaining popularity in psychopathology research as a method that aims to reveal causal relationships among individual symptoms. To date, 4 main types of psychopathology networks have been proposed: (a) association networks, (b) regularized concentration networks, (c) relative importance networks, and (d) directed acyclic graphs. The authors examined the replicability of these analyses based on symptoms of major depression and generalized anxiety between and within 2 ...
148 CitationsSource
#1Eiko I. Fried (UvA: University of Amsterdam)H-Index: 33
#2Angélique O. J. Cramer (UvA: University of Amsterdam)H-Index: 27
Since the introduction of mental disorders as networks of causally interacting symptoms, this novel framework has received considerable attention. The past years have resulted in over 40 scientific publications and numerous conference symposia and workshops. Now is an excellent moment to take stock of the network approach: What are its most fundamental challenges, and what are potential ways forward in addressing them? After a brief conceptual introduction, we first discuss challenges to network...
253 CitationsSource
#1Sacha Epskamp (UvA: University of Amsterdam)H-Index: 33
#2Joost Kruis (UvA: University of Amsterdam)H-Index: 4
Last. Maarten Marsman (UvA: University of Amsterdam)H-Index: 25
view all 3 authors...
Network models, in which psychopathological disorders are conceptualized as a complex interplay of psychological and biological components, have become increasingly popular in the recent psychopathological literature (Borsboom, et. al., 2011). These network models often contain significant numbers of unknown parameters, yet the sample sizes available in psychological research are limited. As such, general assumptions about the true network are introduced to reduce the number of free parameters. ...
83 CitationsSource
#1Giulio Costantini (University of Milan)H-Index: 17
#2Juliette Richetin (University of Milan)H-Index: 17
Last. Marco Perugini (University of Milan)H-Index: 58
view all 6 authors...
Networks have been recently proposed for modeling dynamics in several kinds of psychological phenomena, such as personality and psychopathology. In this work, we introduce techniques that allow disentangling between-subject networks, which encode dynamics that involve stable individual differences, from within-subject networks, which encode dynamics that involve momentary levels of certain individual characteristics. Furthermore, we show how networks can be simultaneously estimated in separate g...
89 CitationsSource
Cited By400
#2María Montero-López Lena (UNAM: National Autonomous University of Mexico)H-Index: 5
In the study of collective action, there are multiple explanatory models, however, there is no known model in the context of collective action against violence against women in Mexico. Therefore, the purpose of this study was to explore how feminist identity, gender, self-efficacy, expectations of results, collective efficacy, perception of injustice, indignation, subjective norms, and collective actions are related amongst 488 women from the Metropolitan Area of Mexico City, obtained in October...
#1Guangzhe Frank Yuan (UM: University of Macau)
#2Wei Shi (Sichuan University)
Last. Brian J. Hall (New York University Shanghai)H-Index: 36
view all 7 authors...
Research has demonstrated that posttraumatic stress disorder (PTSD) is associated with internet-related problematic behaviors. However, studies have not explored the linkage between PTSD symptoms and internet gaming disorder (IGD) symptoms. The current study aimed to investigate the relationship between posttraumatic stress symptoms (PTSS) and IGD symptoms via network analysis. We conducted a cross-sectional study with 341 Chinese young adults directly exposed to a typhoon and examined the netwo...
#1Timo Slotta (University of Cologne)
#2Michael Witthöft (University of Mainz)H-Index: 29
Last. Anna Pohl (University of Cologne)H-Index: 6
view all 4 authors...
Abstract null null Interoception is the sense of the body's physiological state. Both the ability to accurately perceive bodily signals, and the self-reported sensibility for the body were assumed to be reciprocally associated with negative affect. Up to now, the specific interplay between dimensions of interoception and negative affect remains largely unknown. We investigated the interplay of cardioceptive accuracy assessed by a heartbeat tracking task, facets of interoceptive sensibility as me...
#1Tessa F. BlankenH-Index: 12
#2Joe Bathelt (RHUL: Royal Holloway, University of London)H-Index: 1
Last. Linda Douw (VU: VU University Amsterdam)H-Index: 32
view all 6 authors...
Abstract null null In recent years, there has been an increase in applications of network science in many different fields. In clinical neuroscience and psychopathology, the developments and applications of network science have occurred mostly simultaneously, but without much collaboration between the two fields. The promise of integrating these network applications lies in a united framework to tackle one of the fundamental questions of our time: how to understand the link between brain and beh...
#1Christina Ralph-Nearman (University of Louisville)
#2Brenna M. Williams (University of Louisville)H-Index: 4
Last. Cheri A. Levinson (University of Louisville)H-Index: 23
view all 5 authors...
Abstract null null Background null Sleep, anxiety, and worry are strongly related to psychiatric illness and in particular to eating disorder (ED) symptoms. However, it is unclear how these specific sleep and anxiety symptoms are interrelated with anorexia nervosa (AN) pathology. null null null Methods null We utilized network analysis to test our theoretically-based conceptual model, by identifying core features and illness (i.e., bridge) pathways among sleep disturbance, anxiety, worry, and ED...
1 CitationsSource
#2Marian D. IlieH-Index: 1
Last. Daniel E. IancuH-Index: 1
view all 4 authors...
Over time, the academics’ approaches to teaching (i.e., content- or learning-focused approach) were intensively studied. Traditionally, studies estimated the shared variance between the items that describe a behavioral pattern (i.e., the psychometric approach), defined as a learning- or content-focused approach to teaching. In this study, we used a different perspective (i.e., network analysis) to investigate academics’ approaches to teaching. We aimed to bring in new insights regarding the inte...
This study investigated the effect of the psychological network on the willingness to communicate in English among Japanese people. Previous studies have shown that psychological factors affect the willingness to communicate in English for Japanese people. However, the network structure of psychological factors and their effects have not been revealed yet. The present study conducted a network analysis with 644 Japanese people. Consequently, the edge between perceived communication competence an...
#1Sameer Ashaie (Rehabilitation Institute of Chicago)H-Index: 3
#2Nichol Castro (UB: University at Buffalo)H-Index: 6
Purpose Aphasia is a complex, neurogenic language disorder, with different aphasia syndromes hallmarked by impairment in fluency, auditory comprehension, naming, and/or repetition. Broad, standardized assessments of language domains and specific language and cognitive assessments provide a holistic impairment profile of a person with aphasia. While many recognize the correlations between assessments, there remains a need to continue understanding the complexity of relationships between assessmen...
#1Samantha Furman (University of Pittsburgh)H-Index: 1
#2Andrew M. Stern (University of Pittsburgh)H-Index: 22
Last. Chakra Chennubhotla (University of Pittsburgh)H-Index: 18
view all 6 authors...
Summary null null Tumors are dynamic ecosystems comprising localized niches (microdomains), possessing distinct compositions and spatial configurations of cancer and non-cancer cell populations. Microdomain-specific network signaling coevolves with a continuum of cell states and functional plasticity associated with disease progression and therapeutic responses. We present LEAPH, an unsupervised machine learning algorithm for identifying cell phenotypes, which applies recursive steps of probabil...
#1Natasha Benfer (AU: Auburn University)H-Index: 6
#2Joseph R. Bardeen (AU: Auburn University)H-Index: 24
Last. Travis A. Rogers (AU: Auburn University)H-Index: 3
view all 4 authors...
Abstract null null Two prominent conceptual models of posttraumatic stress disorder (PTSD) are the cognitive model, associated with cognitive processing therapy (CPT; Resick & Schnicke, 1992), and the functional contextualist model, underlying acceptance and commitment therapy (ACT; Hayes et al., 1999). Network analysis was used to examine dynamic interactions among cognitive (relating to CPT) and functional contextualistic (relating to ACT) variables and PTSD symptoms in a sample of 722 trauma-...