Abnormal Ventral and Dorsal Attention Network Activity during Single and Dual Target Detection in Schizophrenia

Abstract

Early visual perception and attention are impaired in schizophrenia, and these deficits can be observed on target detection tasks. These tasks activate distinct ventral and dorsal brain networks which support stimulus-driven and goal-directed attention, respectively. We used single and dual target rapid serial visual presentation (RSVP) tasks during fMRI with an ROI approach to examine regions within these networks associated with target detection and the attentional blink (AB) in 21 schizophrenia outpatients and 25 healthy controls. In both tasks, letters were targets and numbers were distractors. For the dual target task, the second target (T2) was presented at three different lags after the first target (T1) (lag1 = 100 ms, lag3 = 300 ms, lag7 = 700ms). For both single and dual target tasks, patients identified fewer targets than controls. For the dual target task, both groups showed the expected AB effect with poorer performance at lag 3 than at lags 1 or 7, and there was no group by lag interaction. During the single target task, patients showed abnormally increased deactivation of the temporo-parietal junction (TPJ), a key region of the ventral network. When attention demands were increased during the dual target task, patients showed overactivation of the posterior intraparietal cortex, a key dorsal network region, along with failure to deactivate TPJ. Results suggest inefficient and faulty suppression of salience-oriented processing regions, resulting in increased sensitivity to stimuli in general, and difficulty distinguishing targets from non-targets.

Keywords: RSVP; attentional blink; fMRI; schizophrenia; visual attention.

FIGURE 1

Single and dual target RSVP task illustration. Twenty stimuli (targets and distractors) per trial were presented for 85 ms with a 15 ms interstimulus interval. A 5000 ms response period prompted participants to indicate whether T1 was a vowel or a consonant (single and dual target tasks) and whether T2 was X or Y (dual target task). In the dual target task trial shown here, T2 is presented at lag 1.

FIGURE 2

Location of regions of interest (ROI) within the ventral (top) and dorsal (bottom) attention networks. Each ROI (9 mm sphere) centered on coordinates shown in Montreal Neurological Institute (MNI) standard space. ACC, anterior cingulate cortex; AI, anterior insula; aIPC anterior intraparietal cortex; LFC, lateral frontal cortex; pIPC, posterior intraparietal cortex; TPJ, temporo-parietal junction.

FIGURE 3

Behavioral performance data. (A) Single target task performance (mean accuracy) by group. (B) Dual target task performance by group as a function of lag. Assessed using the probability of correct T2 identification given correct T1 identification (conditional probability). ^∗∗p < 0.01.

FIGURE 4

Single target task ROI analysis by network (contrast: single target task > baseline). (A) Ventral attention network: patients showed significantly greater deactivation of TPJ than controls. (B) Dorsal attention network: no significant group differences. ^∗p < 0.05.

FIGURE 5

Dual target task ROI analysis by network (all lags combined; contrast: dual target task > baseline). (A) Ventral attention network: both groups showed activation of ACC and AI, but deactivation of TPJ. (B) Dorsal attention network: no significant group differences. ^∗p < 0.05.

FIGURE 6

Dual target task vs. single target task ROI analysis by lag. (A) Controls showed significantly greater deactivation of temporo-parietal junction (TPJ) than patients across lags. (B) Patients showed greater activation of posterior intraparietal cortex (pIPC) than controls across lags.