Attending to remember and remembering to attend in depression

Introduction. Numerous investigations on cognitive risk factors in depression indicate that emotional biases in attention and memory are closely related to one’s emotional state. However, current insights into interactions between such biases in cognition are limited. Here we will present experiments testing how emotional attention may modulate what is remembered and how emotional long-term memory (LTM) may in turn guide attention. Method. Experiments were conducted in mixed samples of healthy and subclinically depressed individuals. In examining effects of attention biases at encoding and retrieval on memory, eye movements were recorded during encoding (e.g., scrambled sentences test) and retrieval (e.g., remember/know recognition decisions) tasks. Gaze behavior (e.g., fixation time/frequency) was inspected in relation to recollected material. In examining attentional guidance by LTM, stimuli from an affective learning phase were presented as task-irrelevant features in a subsequent visual search task. Performance on this task was inspected in relation to memory for acquired associations. Results. At encoding, sustained attention bias directly predicted recollection of negative vs. positive material, whereas biased selective orienting was indirectly related to emotional memory via interpretation. At retrieval, preliminary data suggests that selective orienting and sustained attention biases are associated with recollection of negative vs. positive material, but only during controlled and not automatic retrieval. Memory-guided attention was evident when positive –not negative– task-irrelevant material interfered with target detection during visual search, which was predicted by more accurate memory for positive associations. Discussion/Conclusion. This research contributes to our understanding of each individual cognitive bias and depression-linked distortions in cognitive functioning as a whole. Namely, the finding that emotional biases in attention may control and be controlled by emotional LTM points to highly interactive cognitive biases that cannot be regarded isolated, independent phenomena. This has important implications for the development and implementation of cognitive training procedures which will be discussed