Intraindividual phenotyping of depression in high-risk youth: An application of a multilevel hidden Markov model

Abstract

Background: Traditionally, depression phenotypes have been defined based on interindividual differences that distinguish between subgroups of individuals expressing distinct depressive symptoms often from cross-sectional data. Alternatively, depression phenotypes can be defined based on intraindividual differences, differentiating between transitory states of distinct symptoms profiles that a person transitions into or out of over time. Such within-person phenotypic states are less examined, despite their potential significance for understanding and treating depression. Methods: The current study used intensive longitudinal data of youths (N = 120) at risk for depression. Clinical interviews (at baseline, 4, 10, 16, and 22 months) yielded 90 weekly assessments. We applied a multilevel hidden Markov model to identify intraindividual phenotypes of weekly depressive symptoms for at-risk youth. Results: Three intraindividual phenotypes emerged: a low-depression state, an elevated-depression state, and a cognitive-physical-symptom state. Youth had a high probability of remaining in the same state over time. Furthermore, probabilities of transitioning from one state to another did not differ by age or ethnoracial minority status; girls were more likely than boys to transition from a low-depression state to either the elevated-depression state or the cognitive-physical symptom state. Finally, these intraindividual phenotypes and their dynamics were associated with comorbid externalizing symptoms. Conclusion: Identifying these states as well as the transitions between them characterizes how symptoms of depression change over time and provide potential directions for intervention efforts

Machine learning in toxicological sciences: opportunities for assessing drug toxicity

“…For example, an HMM was recently used to predict schizophrenia (the hidden state) based on motor activity (the observed state) (Boeker et al, 2023). Also, a study evaluating different states of depression in youth over 90 weeks (time-series data is well-suited for this algorithm) used HMMs to report that some were more likely to transition from a low to a highly depressed state and the need for intervention (Liu et al, 2023). In evaluating opioid use disorder patients over 12 months, an HMM model could predict the effect of addiction consult services on the disposition of those individuals (King et al, 2021).…”