Role of the dorsal attention network in distracter suppression based on features

Figures & data

Figure 1. Feature-based variant of the sustained attention to response task. The task-relevant stream(s) was/were indicated by a colored circle. Participants were instructed to press a button on a response box when a target (digit ≠ ‘3ʹ) appeared in the task-relevant stream(s) (= go trials), but had to inhibit the response when a non-target (digit = ‘3ʹ) appeared in a task-relevant stream (= no-go trials). (a) In experimental blocks where only one stream of digits was task-relevant, two types of go trials occurred: target (T) (task-relevant digit ≠ ‘3ʹ) with high (task-irrelevant digit = ‘3ʹ) or low (task-irrelevant digit ≠ ‘3ʹ) priority distracter (D). (b) In experimental blocks where both streams of digits were task-relevant, only one type of go trial occurred (T + T) (both digits ≠ ‘3ʹ).

Table 1. Experimental conditions.

Experimental block	Experimental condition	Number of trials per run	Relevant stream(s)	Irrelevant stream	Trial type
One Relevant Stream	T + Low Priority D	168 (7/9)	1 Target (≠ 3)	1 Low Priority Distracter (≠ 3)	Go
	T + High Priority D	24 (1/9)	1 Target (≠ 3)	1 High Priority Distracter (= 3)	Go
	Non-Target	24 (1/9)	1 Non-Target (= 3)	1 Low Priority Distracter (≠ 3)	No-Go
Two Relevant Streams	T + T	96 (8/9)	2 Targets (≠ 3)	N.A.	Go
	Non-Target	12 (1/9)	1 Non-Target (= 3) & 1 Target (≠ 3)	N.A.	No-Go

For experimental blocks where only one stream of digits was task-relevant, three experimental conditions occurred: Target (T) + Low Priority Distracter (D), Target (T) + High Priority Distracter (D), and Non-Target. For experimental blocks where both streams of digits were task-relevant, two experimental conditions occurred: Target (T) + Target (T), and Non-Target.

Figure 2. Behavioral results. (a) Proportion misses, and (b) Reaction times on go trials. * = p < .05, ** = p < .01, *** = p < .001, n.s. = not significant (p > .05). Abbreviations: ‘T’ = Target, ‘D’ = Distracter.

Table 2. Significant clusters in (a) Contrast 1: effect of increasing distracter priority, (b) Contrast 2: effect of selecting one (low priority distracter) versus attending two stream(s) of stimuli, and (c) Contrast 3: effect of selecting one (high priority distracter) versus attendingtwo stream(s) of stimuli.

Contrast	Region	x, y, z (MNI)	Number of voxels (27 mm^3)	Peak Z-score	FWE-corr. p
a	Left IPS + SPL	−27, −52, 44	546	5.01	<.001
	Left FEF + MFG	−27, 2, 50	404	4.70	<.001
	Right IPS + SPL	12, −58, 38	465	4.53	<.001
	Right FEF	30, 8, 56	112	4.09	<.001
	Right MFG	48, 29, 26	70	3.91	.002
	Left MT+	−39, −61, −13	39	3.73	.032
b	Right SPL	15, −61, 59	69	4.18	.004
c	Left IPS + SPL	−27, −52, 44	687	5.08	<.001
	Left FEF + MFG	−27, 2, 50	522	4.89	<.001
	Right IPS + SPL	9, −70, 59	646	4.88	<.001
	Right FEF	30, 8, 56	183	4.59	<.001
	Right MFG	48, 29, 23	117	4.52	<.001

For each significantly activated cluster, the anatomical region, stereotactic MNI (Montreal Neurological Institute) coordinates, number of voxels, peak Z-score and FWE (familywise error)-corrected p-value is reported. Here, the FWE-corrected p-value was calculated based on the number of voxels in each cluster. Cluster-extent based thresholding was used with primary threshold uncorrected p < .001 and FWE-corrected p < .05. Abbreviations: IPS = intraparietal sulcus, SPL = superior parietal lobe, FEF = frontal eye fields, MFG = middle frontal gyrus, MT+ = middle temporal.

Figure 3. fMRI results. (a) Contrast 1: effect of increasing distracter priority, (b) Contrast 2: effect of selecting one (low priority distracter) versus attending two stream(s) of stimuli, and (c) Contrast 3: effect of selecting one (high priority distracter) versus attending two stream(s) of stimuli. Results of correct responses on go trials presented as statistical t-maps in MNI (Montreal Neurological Institute) space. Cluster-extent based thresholding was used with voxel-based threshold uncorrected p < .001 and cluster-based threshold FWE (familywise error)-corrected p < .05.