Self-Dislike and Sadness are Central Symptoms of Depression in College Students: A Network Analysis

Abstract

Depression in college students is an important public health issue. There is growing recognition that some depression symptoms may have a more central role in the disorder than others. We use network analyses to identify the most central symptoms of depression in college students. In a cross-sectional network the BDI-SF was collected at a single time point (N = 10,005) and in a symptom change network the BDI-SF was collected at two time points separated on average by approximately two months (N = 700). In both samples, the most central depression symptoms were self-dislike and sadness. Centrality scores for the cross-sectional network were highly correlated (r = .86) with the centrality scores for the symptom change network. Self-dislike and sadness may have important roles in college student depression. Additional work experimentally targeting and changing these central symptoms is needed to determine whether they have a causal role in the disorder.

Comorbid anxiety and depression psychopathology in university students: a network approach

“…For instance, research shows that central depression symptoms predict whether participants will experience a major depressive episode in the following 6 years better than non-central symptoms (Boschloo et al, 2016). Our results are consistent with previous network studies with university students ( self-dislike ; Mullarkey et al, 2018), general population adolescents ( self-dislike , pessimism , and loneliness ; Mullarkey et al, 2019), and clinical population adults ( anhedonia ; Bringmann et al, 2015).…”

“…The findings of this study have implications for depression treatment and assessment in university students. First, if we are to accept central symptoms as good predictors of psychopathology outcome (course, prognosis, and aetiology; Borsboom, 2017), then the current approach of using aggregate depression measure scores falls short when it comes to accounting for the differential importance of individual symptoms (Cramer et al, 2016; Mullarkey et al, 2018). Using weighted sum scores based on symptom centrality, as recommended in the network literature, may be a better option for operationalizing the phenomenology of depression (Boschloo et al, 2016; Bringmann & Eronen, 2018; Fried et al, 2016).…”

“…This new psychopathology paradigm resulted in an analytic method called psychometric network models, capable of assessing network structures of between-groups and repeated-measure data. Network models examine the individual relationship between each symptom pair while at the same time holding all the other symptoms in the network constant and bringing spurious associations to zero (Epskamp & Fried, 2018; Mullarkey et al, 2018).…”

“…Although a number of studies have to date used network analysis to explore depression as a possible system of causal interactions between symptoms (e.g., Bringmann et al, 2015; Mullarkey et al, 2018, 2019; van Borkulo et al, 2015), network studies concerning the relations between anxiety and depression symptoms (among university students and other populations) have generally been limited. To be sure, only two network studies have to date explored the association between anxiety and depression: one with older adults (An et al, 2019) and another with a psychiatric sample (Beard et al, 2016).…”

Understanding the Unique Relationships Between Self-Compassion, Mindfulness, and Individual Adolescent Depressive Symptoms: A Network Analysis

“…To use an example from physical health, we would expect headaches and painful, dark skin pustules to occur at the same time in the presence of the smallpox virus, but these two symptoms would be unrelated in the disease's absence (Fried, 2015). However, the associations between individual depressive symptoms are not actually stronger among people diagnosed with major depressive disorder compared to people below the cutoff for major depressive disorder (Mullarkey, Stewart, Wells, & Beevers, 2018;Santos, Fried, Asafu-Adjei, & Ruiz, 2017). This implies that a latent disease model, where symptoms more strongly co-occur in the presence of the disease than in its absence, is not fully capturing the phenomenology of depression.…”