BACKGROUND: Longitudinal studies examining the temporal association between mental health outcomes during the COVID-19 outbreak are needed. It is important to determine how relationships between key outcomes, specifically loneliness and depressive symptoms, manifest over a brief timeframe and in a pandemic context. METHOD: Data was gathered over 4 months (March - June 2020) using an online survey with three repeated measures at monthly intervals (N = 1958; 69.8% females; Age 18-87 years, M = 37.01, SD = 12.81). Associations between loneliness, depression symptoms, and emotion regulation difficulty were tested using Pearson's product moment correlations, and descriptive statistics were calculated for all study variables. Cross-lagged structural equation modelling was used to examine the temporal relationships between variables. RESULTS: The longitudinal association between loneliness and depressive symptoms was reciprocal. Loneliness predicted higher depressive symptoms one month later, and depressive symptoms predicted higher loneliness one month later. The relationship was not mediated by emotion regulation difficulties. Emotion regulation difficulties and depressive symptoms were also reciprocally related over time. LIMITATIONS: Limitations include the reliance on self-report data and the non-representative sample. There was no pre-pandemic assessment limiting the conclusions that can be drawn regarding the mental health impact of the COVID-19 crisis. CONCLUSIONS: Loneliness should be considered an important feature of case conceptualisation for depression during this time. Clinical efforts to improve mental health during the pandemic could focus on interventions that target either loneliness, depression, or both. Potential approaches include increasing physical activity or low-intensity cognitive therapies delivered remotely.