br Materials and methods br Results

Materials and methods
Results
Discussion Behavioural changes in reaction times for emotional categorisation and emotional recognition of both positive and negative material increased with age across both groups, however no case–control differences were identified. This is inconsistent with some previous literature identifying case–control differences in the affective bias of memory processing (Dalgleish and Werner-Seidler, 2014; Bradley and Mathews, 1988). However the current findings may be consistent with reports of age related changes in memory performance (Neshat-Doost et al., 1998). In line with these findings, BOLD activity during encoding and retrieval stages were also found to be associated with age. This indicates a possible developmental aspect to the behavioural performance and neural activation during this task. Group differences were also identified during the encoding stage of the task, but were not seen in retrieval. Differential activation during successful encoding and encoding attempt was identified in regions including the occipital cortex and cingulate gyrus. This finding suggests that depressed youth may have a differing neural response to encoding of emotional state words. This finding is consistent with previous studies in adults (Hamilton and Gotlib, 2008; Arnold et al., 2011; van Tol et al., 2012) and adolescents (Roberson-Nay et al., 2006) with MDD which have identified case–control differences in comparable gingerol regions during memory encoding. Difference in the structure and function of the cingulate, in particular, has been consistently implicated in depression (Clark et al., 2009). Here we also report reduced activation in the supramarginal gyrus, lateral occipital and superior parietal cortex, which have been implicated in language and visual processing. Previous fMRI studies have focused on the encoding stage of this task rather than on retrieval, and hence further work is needed to determine if group differences are in fact characteristic during this aspect of memory in MDD. In the current study, MDD self-report symptom scores indexing episode severity were linked with altered neural activity during retrieval attempt. This is consistent with reports of dysfunction in the thalamus in previous reports of the neural effects of MDD (Graham et al., 2013). When considering the wider cohort (including patients taking SSRIs) there was also an association between depression severity and the neural response to affective bias for encoding identified in frontal regions. This finding is consistent with literature describing frontal lobe dysfunction in MDD (Clark et al., 2009; Roiser and Sahakian, 2013; Kerestes et al., 2014). These findings suggest an association between differing neural activation in these regions and severity of depressive symptomology. Further work is needed for a more detailed examination of the relationship between neural activation with symptom severity and other characteristics such as duration and the number of depressive episodes. Adolescence appears to be a crucial period in the emergence and aetiology of MDD. It has been proposed that aberrant developmental processes themselves may play a key role in the onset of disorder (Goodyer, 2008; Paus et al., 2008). Further work is required to specifically test this hypothesis. Investigation of grey matter volume in this cohort also revealed the presence of age-related differences in ACC and thalamus (Hagan et al., 2015), emphasising the need for further investigation of MDD across this period of neural development. Such investigations would benefit from including factors that may influence the typically developing brain, including measures of childhood adversity that may provide valuable explanatory power regarding the differences we have observed in adolescents (Dannlowski et al., 2012; Teicher and Samson, 2013). Indeed exposure to an adverse family environment in childhood carrying a moderate to severe risk of childhood maltreatment increases the liability for emerging adolescent depression by 14 years of age by 3–5 times (Dunn et al., 2011). Furthermore, proof of principle findings have suggested that childhood adversities may exert effects on neural systems once past and current mental state factors are taken into account (Walsh et al., 2014) and may indicate a mechanism for atypical neuro-cognitive development. Inclusion of such measures would enable testing of the neurodevelopmental model and determine the contribution of such experience to neural scars in adult life.