A team of researchers have developed a predictive model to recognize patterns of persistent negative thinking, or rumination, using machine learning. Researchers hypothesized that the variance of dynamic connectivity between certain brain regions, such as the dorsal medial prefrontal cortex (dmPFC), could be associated with rumination.
The objective of this study is to use machine learning algorithms to analyze patterns of rumination from the RDoC perspective. The goal is to identify which variables predict high levels of maladaptive rumination in a diverse group of individuals.
We collected data from a sample of 200 consecutive outpatient referrals who had been diagnosed with various mental health conditions, including schizophrenia, schizoaffective disorder, bipolar disorder, depression, anxiety disorders, obsessive-compulsive disorder, and post-traumatic stress disorder. Our machine learning algorithms took into account a range of factors such as sociodemographics, levels of immune markers (IL-6, IL-1β, IL-10, TNF-α, and CCL11) and BDNF in the serum, psychiatric symptoms and disorders, history of suicide and hospitalizations, functionality, medication use, and comorbidities.
After applying recursive feature elimination, our best model included the following variables: socioeconomic status, illness severity, worry, generalized anxiety and depressive symptoms, and current diagnosis of panic disorder. Through linear support vector machine learning, we were able to differentiate individuals with high levels of rumination from those with low levels, achieving an AUC of 0.83, sensitivity of 75, and specificity of 71.
Rumination is known to have a negative impact on mental health prognosis. Our study indicates that rumination is a maladaptive coping style associated not only with worry, distress, and illness severity but also with socioeconomic status. Additionally, we found a specific association between rumination and panic disorder.
Paper Source : ScienceDirect
