At baseline, respondents were also asked to self-identify whether they had previously experienced depression and, if so, whether they had seen a clinician or counsellor for treatment. Our first analysis, therefore, contrasts those participants with those without a history of depression, and the second analysis further differentiated amongst those with a history of depression who had or had not sought professional treatment. We also estimated current depression based on the Goldberg Depression Scale. In order to evaluate potential reverse-causality, our models evaluated whether reported history of depression prior to the baseline assessment was associated with the consumption of prudent and western dietary patterns at the baseline assessment, after controlling for current depression (estimate of caseness). In common with our previous analysis, we took into account potentially confounding variables including age, gender, measures of socioeconomic circumstances, cardiovascular risk factors and other health behaviours (Jacka et al., 2014).