Symposia
Cognitive Science/ Cognitive Processes
Amy Gillespie, Ph.D.
University of Oxford
Oxford, England, United Kingdom
Negative affective bias has been shown to precede and predict subsequent depression, and remediation of negative bias is an early marker of treatment response. Tasks such as the Facial Expression Recognition Task (FERT), in which participants must identify the expression of ambiguous positive or negative faces, can be used to assess affective bias. We therefore measured negative bias as an objective marker of depression vulnerability, in an online longitudinal study during the COVID-19 pandemic. Our aim was to identify protective or risk factors, to inform the identification of potential interventions. Between April 2020 and February 2021, we conducted an online study of a diverse UK sample (n=2043, aged 18+) at four timepoints (83.5% retention at 10 months). Participants completed self-report questionnaires and tasks assessing affective bias (including the FERT) and reward learning. Data were analysed controlling for demographics, current antidepressant use, current self-reported depression, exercise, alcohol use and psychiatric history. We confirmed that depression (current diagnosis, past history, and current symptom severity) is associated with increased negative bias, and found that negative bias (specifically, increased recognition of angry faces) predicts depression at 10 month follow up (β=3.61, p=0.027). High levels of behavioural activation were associated with increased recognition of happy faces (F=4.87, p=0.027), indicating a protective effect, whereas loneliness was associated with reduced recognition of happy faces indicating vulnerability (F=7.63, p=0.0058). I will discuss how this study highlights core factors associated with depression resilience and vulnerability during the COVID-19 pandemic, such as behavioral activation and loneliness. I will also discuss how measuring depression vulnerability using an objective emotional cognition biomarker helps circumvent many of the limitations of self-report measures and allows us to investigate the manifestation of vulnerability across time and how this may be mitigated.