Neural and behavioral effects of interference resolution in depression and rumination

Abstract

Individuals diagnosed with major depressive disorder (MDD) often ruminate about their depression and their life situations, impairing their concentration and performance on daily tasks. We examined whether rumination might be due to a deficit in the ability to expel negative information from short-term memory (STM), and fMRI was used to examine the neural structures involved in this ability. MDD and healthy control (HC) participants were tested using a directed-forgetting procedure in a short-term item recognition task. As predicted, MDD participants had more difficulty than did HCs in expelling negative, but not positive, words from STM. Overall, the neural networks involved in directed forgetting were similar for both groups, but the MDDs exhibited more spatial variability in activation in the left inferior frontal gyrus (a region critical for inhibiting irrelevant information), which may contribute to their relative inability to inhibit negative information.

Fig. 1

Schematic diagram of the valenced directed forgetting task. Each run consisted of 16 trials that were balanced for the different trial-type combinations, lasted 7 min 38 s, and included 16 s of fixation at the beginning and end of each run

Fig. 2

Mean correct RTs for the lure – control contrast for the most negatively (ratings 1 and 2) and positively (ratings 6 and 7) rated words for both individuals with major depressive disorder (MDDs) and controls. Here, valence was determined by each individual participant. MDDs exhibit more difficulty in removing negative information from STM than positive. Controls do not show this pattern

Fig. 3

The difference in the lure – control contrast for the most negatively and positively valenced words as rated by individual participants, plotted against the Rumination score (RRS) of each participant. The linear equation is shown in the upper left of the figure

Fig. 4

Results of the conjunction analysis performed on the lure neg – control neg contrast for individuals with major depressive disorder and controls. This figure shows the overlap in LiFG and RiFG (the leftmost axial slice, z = 4, and the coronal slice, y = 28), the dorsal ACC and precuneus (the sagittal slice, x = −5), and the left and right parietal cortex with an additional overlap in bilateral middle frontal gyrus (rightmost axial slice, z = 51

Fig. 5

Three-dimensional renderings of mean activations in LiFG, for MDDs on the left and HCs on the right, for the lure neg – control neg contrast. Deactivations are in blue, and activations are in yellow/red. From the figure, one can see that MDDs activate LiFG more diffusely, while HCs activate this region more focally, with activation clusters centering in lower to middle portions of the LiFG. Mean contrast values range from −1.5 (blue) to +1.5 (red)

Fig. 6

Chromosome-style plot of LiFG for the lure neg – control neg contrast. The y-axis shows the contrast score for each voxel in LiFG for each participant. The x-axis shows the different slices through z, from 1 to 22 (ventral to dorsal). MDDs are displayed in the top panel, and HCs are displayed in the bottom panel. The data are vectorized over x and y, but each x, y pair is represented separately, and each participant is represented by a different color, as shown in the legend. HCs' contrast values, for the most part, stay between +4 and ‒4, while MDD's contrast values also range above and below +4 and – 4. From the figure, it is apparent that MDDs show greater variance at nearly every slice through z, indicating that the variance effect appears continuous throughout the region of interest