Abnormal neural filtering of irrelevant visual information in depression

Abstract

The pathophysiology of major depressive disorder (MDD) includes both affective and cognitive dysfunctions. We aimed to clarify how regions regulating affective processing interact with those involved in attention, and how such interaction impacts perceptual processing within sensory cortices. Based on previous work showing that top-down influences from attention can determine the processing of external inputs within early sensory cortices, we tested with functional magnetic resonance imaging (fMRI) whether MDD alters attentional ("top-down") effects on the neural filtering of irrelevant, nonemotional visual stimuli. The present fMRI study was conducted in 14 nonmedicated patients with a first episode of unipolar MDD and 14 matched controls. During scanning, subjects performed two tasks imposing two different levels of attentional load at fixation (easy or difficult), while irrelevant colored stimuli were presented in the periphery. Analyses of fMRI data revealed that MDD patients show (1) an abnormal filtering of irrelevant information in visual cortex, (2) an altered functional connectivity between frontoparietal networks and visual cortices, and (3) a hyperactivity in subgenual cingulate/medial orbitofrontal cortex that was modulated by attentional load. These results demonstrate that biological abnormalities contribute to the cognitive deficits seen in major depression, and clarify how neural networks implicated in mood regulation influence executive control and perceptual processes. These findings not only improve our understanding of the pathophysiological mechanisms underlying cognitive dysfunctions in MDD, but also shed new light on the interaction between cognition and mood regulation.

Figure 1.

Stimuli and design in the attentional load experiment. A, A rapid continuous stream of colored “T” shapes appeared at central fixation during all blocks (500 ms duration each, 250 ms interval). B, Irrelevant peripheral stimulation included 20 s blocks with bilateral colored, “Mondrian-like” stimuli, alternating with blocks without any peripheral stimulation. C, Two main attentional load conditions required the subjects to either detect any red T regardless of its orientation (low load), or to detect all upright yellow T shapes and upside-down blue T shapes. Additional blocks of central fixation (baseline) were also included. The central visual stream remained identical in each task condition, only the task instructions differed.

Figure 2.

Main effect of high attentional load. A, Statistical maps from the group conjunction analysis showing increased activation in lateral occipital (x, y, z: −44, −68, −10; visual word form area) (Vinckier et al., 2007), superior parietal and inferior frontal regions in both MDD patients and controls. Maps are overlaid on average T1 structural scan, thresholded at p < 0.001 uncorrected. Color bar on the right indicates t values. B, Parameter estimates extracted from the statistical peaks demonstrate selective activation during the high-load task across all conditions of peripheral stimulation and in both populations. B, Baseline fixation; L, low attentional load; H, high attentional load.

Figure 3.

Main effects of low attentional load. A, Increased response during the low- minus the high-load condition in V4 for controls relative to MDD patients. B, Parameter estimates extracted from V4 peak show decreased activity during low versus high load in controls. Note that decreases from baseline to low load in response to peripheral colored stimuli was significant in patients (t = 2.36, p = 0.013) but not in controls (t = 0.98, p = 0.16). C, Decreased subgenual cingulate response during the high-load condition in MDD patients compared with controls. D, Parameter estimates showed decreased fMRI signal during high attentional load in patients but not in controls. For visualization purposes, statistical parametric maps are overlaid on average T1 structural scan and thresholded at p < 0.005 uncorrected. Color bars on the left indicate t values. B, Baseline fixation; L, low attentional load; H, high attentional load.