Emily E Bernstein1, Alexandre Heeren2, Richard J McNally3. 1. Department of Psychology, Harvard University, 33 Kirkland Street, Cambridge, MA, 02138, USA. Electronic address: ebernstein@g.harvard.edu. 2. Psychological Sciences Research Institute & Institute of Neuroscience, Université Catholique de Louvain, Place du Cardinal Mercier, 10, 1348, Louvain-la-Neuve, Belgium. Electronic address: alexandre.heeren@uclouvain.be. 3. Department of Psychology, Harvard University, 33 Kirkland Street, Cambridge, MA, 02138, USA. Electronic address: rjm@wjh.harvard.edu.
Abstract
BACKGROUND AND OBJECTIVES: Rumination is strongly associated with risk, maintenance, and worsening of depressive and related symptoms, and it predicts poor treatment response and relapse. More work is needed to clarify the nature and malleability of rumination. We propose reexamining trait rumination as a system of interacting components ("nodes"). METHODS: A regularized partial correlation network was first computed to estimate the functional relations among items from the Ruminative Responses Scale (RRS) (N = 403). We then tested whether items constitute multiple distinguishable sub-networks or communities, and if so, if particular items function as "bridges" connecting them. RESULTS: RRS items were not interchangeable, with network components varying widely in their centrality. We identified three communities of nodes and the nodes bridging these communities. LIMITATIONS: Data were derived from a heterogeneous community sample and include items from a single measure. Thus, results should not be interpreted as definitive, but instead as hypothesis-generating and highlighting the utility of rethinking the conceptualization and measurement of rumination. CONCLUSIONS: Of the larger set of cognitive patterns forming the rumination construct, the high centrality nodes were largely passive and self-critical processes. Community detection analyses identified a sub-network largely comprising items from the RRS that have traditionally been labeled reflective pondering and adaptive; however, strong bridge nodes were also from this community. This implies that in isolation or at low levels such processes may not be problematic, but that their persistence or intensification could be associated with the activation of more maladaptive processes.
BACKGROUND AND OBJECTIVES: Rumination is strongly associated with risk, maintenance, and worsening of depressive and related symptoms, and it predicts poor treatment response and relapse. More work is needed to clarify the nature and malleability of rumination. We propose reexamining trait rumination as a system of interacting components ("nodes"). METHODS: A regularized partial correlation network was first computed to estimate the functional relations among items from the Ruminative Responses Scale (RRS) (N = 403). We then tested whether items constitute multiple distinguishable sub-networks or communities, and if so, if particular items function as "bridges" connecting them. RESULTS: RRS items were not interchangeable, with network components varying widely in their centrality. We identified three communities of nodes and the nodes bridging these communities. LIMITATIONS: Data were derived from a heterogeneous community sample and include items from a single measure. Thus, results should not be interpreted as definitive, but instead as hypothesis-generating and highlighting the utility of rethinking the conceptualization and measurement of rumination. CONCLUSIONS: Of the larger set of cognitive patterns forming the rumination construct, the high centrality nodes were largely passive and self-critical processes. Community detection analyses identified a sub-network largely comprising items from the RRS that have traditionally been labeled reflective pondering and adaptive; however, strong bridge nodes were also from this community. This implies that in isolation or at low levels such processes may not be problematic, but that their persistence or intensification could be associated with the activation of more maladaptive processes.
Authors: Jens Allaert; Rudi De Raedt; Frederik M van der Veen; Chris Baeken; Marie-Anne Vanderhasselt Journal: Sci Rep Date: 2021-06-02 Impact factor: 4.379