Functional MRI of emotional memory in adolescent depression

Abstract

Introduction: Major Depressive Disorder (MDD) is a leading cause of disease burden worldwide. Mood-congruent biases in memory tasks are frequently reported in MDD patients, with facilitated memory for negative stimuli. Most functional MRI studies to date have examined the neural correlates of these biases in depressed adults, with fewer studies in adolescents with MDD. Investigation of MDD in adolescence may aid greater understanding of the aetiology and development of the disorder.

Methods: Cognitive biases were investigated in 56 MDD patients aged 11-17 years and a matched group of 30 healthy control participants with a self-referential memory task. Behavioural performance and BOLD fMRI data were collected during both encoding and retrieval stages.

Results: The neural response to encoding in adolescents with MDD was found to differ significantly from controls. Additionally, neural responses during encoding and retrieval showed differential relationships with age between patient and control groups, specifically in medial, temporal, and prefrontal regions.

Conclusions: These findings suggest that during adolescence neurophysiological activity associated with emotional memory differs in those with depression compared to controls and may be age sensitive.

Keywords: Adolescence; Depression; Emotion; Memory; fMRI.

Fig. 1

Schematic diagram of trial events in the encode/retrieve task. Boxes are labelled by stage, and box outline indicates duration: red = 30 s, black = 500 ms and blue is a jittered ITI with a mean of 3000 ms (range 1000–6000 ms), events depicted here, in order: encoding instructions, positive categorisation, negative categorisation, retrieval instructions, positive new word recognition (distractor: possible outcomes are correct rejection and false alarm), positive old word recognition (target: possible outcomes are hit and miss). Timeline indicates direction of progression, dashed line indicates where task continues but is not detailed in the diagram. In the retrieval stage, event types not shown are negative new word and negative old word. (For interpretation of the references to color in figure legend, the reader is referred to the web version of the article.)

Fig. 2

BOLD signal during successful encoding (hit > miss) of positive and negative words where greater activation was identified in controls compared to the depressed group (excluding patients taking antidepressant medications). Patients taking medications are included in the plot for illustrative purposes only. The 3D unthresholded statistical map can be found at http://neurovault.org/images/14547.

Fig. 3

BOLD signal during successful encoding of negative words (hit > miss) where a significant group by age interaction was identified (excluding patients taking antidepressant medications). Scatter plot illustrates association with age in controls and depressed patients.

Fig. 4

BOLD activation during encoding attempt (all words > baseline) where a significant group by age interaction was identified. (This group by age interaction was only evident when excluding the patients taking antidepressant medications.) The 3D unthresholded statistical map can be found at http://neurovault.org/images/14644.

Fig. 5

BOLD signal during retrieval attempt (hit > correct rejection) showing a group by age interaction (when excluding patients taking antidepressant medications). Scatter plot illustrates the group by age interaction in the two groups. The 3D unthresholded statistical map can be found at http://neurovault.org/images/14714.

Fig. 6

BOLD activation during retrieval attempt (all retrieval trails > baseline) where a significant group by age interaction was identified (excluding patients taking antidepressant medications). The 3D unthresholded statistical map can be found at http://neurovault.org/images/14798.

Fig. 7

BOLD activation during retrieval attempt (all retrieval trails > baseline) showing a significant correlation with SMFQ in the depression group (not taking antidepressant medications). The medicated group is included in the scatter plot for illustrative purposes only. The 3D unthresholded statistical map can be found at http://neurovault.org/images/14795.