Decreased Response to Positive Facial Affect in a Depressed Cohort in the Dorsal Striatum During a Working Memory Task-A Preliminary fMRI Study

Abstract

People with depression have shown alterations in processing emotional information and working memory functionality. There is some evidence that emotional content may interact with working memory update processes, however neurological correlates are current unknown. In this preliminary study we utilized a novel version of the emotional variant of the n-back working memory task in fMRI. We examined BOLD response of 14 healthy controls and 13 depressed participants in response to happy, sad, and neutral displays of facial affect. No accuracy or reaction time differences were found between the two groups. The depressed group showed significantly decreased BOLD response to happy faces compared to the control group areas of the dorsal striatum and anterior cingulate. Significant, moderate, positive associations were found between right caudate activation with anxiety score and anterior cingulate activation with depression score in those with depression. Our novel task was able to elicit group level differences in emotional processing during working memory update. These results suggest those with depression fail to differentiate between positive emotional stimuli and stimuli with no emotional content.

Keywords: depression; dorsal striatum; emotional processing; fMRI; n-back; working memory.

Figure 1

Stimuli presentation sequence for the 3-back task showing neutral faces, rest, and horizontal grating conditions.

Figure 2

Top: Median value for accuracy (Left) and reaction time (Right) for happy, sad, neutral, horizontal, vertical, and checked conditions. Bars indicate 95% confidence interval bootstrapped 10,000 times, points represent individual participant data. Bottom: Permuted null distribution with observed (red) and threshold (black) values marked as vertical bars.

Figure 3

Significantly activated areas for happy, sad, and neutral faces compared to the patterned baseline.

Figure 4

Significantly increased areas of activation in the control group compared to depression during presentation of happy faces (Top) and mean cluster estimates (Bottom).

Figure 5

(A) Depressed correlation matrix of Spearman's ρ shows unthresholded correlations in the top triangle and thresholded (p < 0.05 two tailed) correlations on the bottom. (B) Scatterplots of controls (left) and depressed (right) for DASS anxiety scale score and right caudate estimates (top) and DASS Depression scale score and right rostral anterior cingulate estimates (bottom). LOESS line (red; fraction = 0.5) indicates the direction of the monotonic relationship.