Abstract

Efficient interaction with the sensory environment requires the rapid reallocation of attentional resources between spatial locations, perceptual features, and objects. It is still a matter of debate whether one single domain-general network or multiple independent domain-specific networks mediate control during shifts of attention across features, locations, and objects. Here, we employed functional magnetic resonance imaging to directly compare the neural mechanisms controlling attention during voluntary and stimulus-driven shifts across objects and locations. Subjects either maintained or switched voluntarily and involuntarily their attention to objects located at the same or at a different visual location. Our data demonstrate shift-related activity in multiple frontoparietal, extrastriate visual, and default-mode network regions, several of which were commonly recruited by voluntary and stimulus-driven shifts between objects and locations. However, our results also revealed object- and location-selective activations, which, moreover, differed substantially between voluntary and stimulus-driven attention. These results suggest that voluntary and stimulus-driven shifts between objects and locations recruit partially overlapping, but also separable, cortical regions, implicating the parallel existence of domain-independent and domain-specific reconfiguration signals that initiate attention shifts in dependence of particular demands.

Introduction

Continuously faced with an overwhelming amount of information, our visual system has to select which sensory input should be preferentially processed at the expense of other information. A key mechanism to overcome capacity limitations is the ability to shift attention and select relevant information for more detailed analysis. The guidance of this attentional selection is not a unitary process, but instead involves endogenous (voluntary/goal-directed) as well as exogenous (involuntary/stimulus-driven) factors. The neural mechanisms underlying such voluntary and stimulus-driven attentional selection have been extensively investigated using neurophysiological recordings in primates and functional neuroimaging in humans (for recent reviews, see Maunsell and Treue 2006; Corbetta et al. 2008; Reynolds and Heeger 2009). Attention can operate on spatial locations (Posner 1980; Heinze et al. 1994), simple features (Treisman and Gelade 1980; Schoenfeld et al. 2007), or objects (Duncan 1984; Schoenfeld et al. 2003). However, the neural correlates of voluntary and stimulus-driven orienting have not been compared across different perceptual domains as, for example, between spatial and nonspatial shifts of attention. Thus, it is not clear whether attentional control during different types of voluntary and stimulus-driven shifts (e.g. between objects and locations) is mediated by one domain-general system (Yantis and Serences 2003; Corbetta et al. 2008) or separate domain-specific networks (Rushworth et al. 2001; Wager et al. 2004).

The domain-independent account is supported by findings that one common network of frontoparietal regions is activated by attentional transitions, regardless of the respective perceptual domain (Corbetta et al. 2008). As such, similar activation foci within the intraparietal sulcus (IPS), the superior parietal lobes (SPLs), and the frontal eye fields (FEFs) have been observed for attention shifts between locations (Hopfinger et al. 2000; Yantis et al. 2002), features (Liu et al. 2003), objects (Serences et al. 2004), sensory modalities (Macaluso et al. 2002; Shomstein and Yantis 2004), or even for switches between different task sets (Slagter et al. 2006). However, direct evidence for the existence of one domain-general system by direct within-subject comparison of different types of attention shifts [e.g. shifts between objects and locations (Shomstein and Behrmann 2006), features and locations (Slagter et al. 2007), or between voluntary and stimulus-driven orienting (Peelen et al. 2004)] is sparse. Other recent investigations, in contrast, also provided evidence in favor of the existence of several domain-specific or at least one unitary but compartmentalized network mediating attentional control (Rushworth et al. 2001).

So far, no study directly compared the mechanisms of voluntary and stimulus-driven attentional control across different domains, for example, between spatial and nonspatial shifts of attention in a within-subject design. Spatial location confound problems were eliminated by employing overlapping transparent surfaces of moving dots that could be separately attended to (He and Nakayama 1995; Schoenfeld et al. 2003). To this end, subjects were explicitly cued to (1) maintain their attention at a currently attended surface, (2) switch to another surface at the same location, or (3) switch to a surface located in the opposite visual field. In addition, the subjects’ attention could be involuntarily captured by target-like movements (4) of an unattended surface at the attended location or (5) of an unattended surface located in the opposite visual field. This design permitted us, for the first time, to directly compare the neural modulations between multiple shift types, namely voluntary or stimulus-driven shifts between objects or locations, in the absence of any sensory confound.

Materials and Methods

Subjects

Sixteen neurologically normal right-handed subjects (9 females), all with normal or corrected-to-normal vision, participated as paid volunteers in the study [mean age: 25.9 ± 0.8 (standard error of the mean, SEM) years]. All gave written informed consent before participation, and the local Ethics Committee approved the study. To ensure high performance, all subjects completed 3 practice sessions outside and 1 in the scanner before participating in the main experiment. This led to high accuracy rates for all participants (mean accuracy >95% and false responses <3%; see Results).

Stimuli and Experimental Design

Stimuli were presented against a dark background (0.5 cd/m2) within 2 square apertures (4.0° × 4.0°) centered 6.8° to the left and right of, and 4.0° above, a central fixation cross (Fig. 1). Each aperture contained 100 randomly distributed isoluminant white dots (brightness 200 cd/m2 and dot size 0.1°). Within both apertures, each half of the dots continuously moved coherently in the opposite direction (horizontal in the left aperture and vertical in the right aperture; velocity 8.7°/s). In this way, 2 transparent moving surfaces located in the same region of visual space were generated within each aperture.

Figure 1.

Schematic illustration of the experimental design. (A) Motion sequences presented during the experiment. Subjects were presented with 2 apertures located in the left and right visual fields, in which 2 overlapping transparent surfaces continuously moved into opposite horizontal (left aperture) and vertical (right aperture) directions. At the beginning of each run (and every 10th trial thereafter), a central cue indicated the surface to be attended (upper-left screenshot). Fast movements in the predominant motion direction of the surface (left-middle screenshot) served as targets (fast movement of the currently attended surface), to which the subjects were required to perform a button press response. However, fast movements could also occur within the unattended surfaces, which—due to their target similarity—involuntarily captured the subjects’ attention across surfaces (fast movement of an unattended object at the attended location) or across spatial locations (fast movement of an unattended object at the unattended location). These fast movements within the unattended surfaces, however, were completely task-irrelevant and subjects were instructed to ignore them. The cue sequences that prompted subjects to voluntarily redirect their attention consisted of 2 short displacements orthogonal to the predominant motion direction of the attended surface (right-middle and lower-right screenshot). See Materials and Methods for a detailed description of these cue sequences. (B) Schematic illustration of trial types during which attention was voluntarily controlled. The columns indicate which particular surface had to be attended at trial onset (left column) and which surface had to be attended after completion of the trial (right column). Red arrows (middle column) indicate the motion directions of the cue sequence that guided the subjects’ voluntary attention. (C) Schematic illustration of the targets and both types of capture trials. The columns indicate which particular surface had to be attended at trial onset (left column), the particular surface that executed a fast movement (indicated by the direction of the orange arrows; middle column), and which surface had to be attended after trial completion (right column). Target trials were defined as fast movements in the predominant motion direction of the attended surface, whereas fast movements of the unattended surface involuntarily captured the subjects’ attention across objects (fast movement of an unattended object at the attended location) or across spatial locations (fast movement of an unattended object at the unattended location); importantly, these latter trials represented nontargets and did not require a button press.

Figure 1.

Schematic illustration of the experimental design. (A) Motion sequences presented during the experiment. Subjects were presented with 2 apertures located in the left and right visual fields, in which 2 overlapping transparent surfaces continuously moved into opposite horizontal (left aperture) and vertical (right aperture) directions. At the beginning of each run (and every 10th trial thereafter), a central cue indicated the surface to be attended (upper-left screenshot). Fast movements in the predominant motion direction of the surface (left-middle screenshot) served as targets (fast movement of the currently attended surface), to which the subjects were required to perform a button press response. However, fast movements could also occur within the unattended surfaces, which—due to their target similarity—involuntarily captured the subjects’ attention across surfaces (fast movement of an unattended object at the attended location) or across spatial locations (fast movement of an unattended object at the unattended location). These fast movements within the unattended surfaces, however, were completely task-irrelevant and subjects were instructed to ignore them. The cue sequences that prompted subjects to voluntarily redirect their attention consisted of 2 short displacements orthogonal to the predominant motion direction of the attended surface (right-middle and lower-right screenshot). See Materials and Methods for a detailed description of these cue sequences. (B) Schematic illustration of trial types during which attention was voluntarily controlled. The columns indicate which particular surface had to be attended at trial onset (left column) and which surface had to be attended after completion of the trial (right column). Red arrows (middle column) indicate the motion directions of the cue sequence that guided the subjects’ voluntary attention. (C) Schematic illustration of the targets and both types of capture trials. The columns indicate which particular surface had to be attended at trial onset (left column), the particular surface that executed a fast movement (indicated by the direction of the orange arrows; middle column), and which surface had to be attended after trial completion (right column). Target trials were defined as fast movements in the predominant motion direction of the attended surface, whereas fast movements of the unattended surface involuntarily captured the subjects’ attention across objects (fast movement of an unattended object at the attended location) or across spatial locations (fast movement of an unattended object at the unattended location); importantly, these latter trials represented nontargets and did not require a button press.

At the beginning of each run, a central cue (a white double arrow pointing in 1 of the 4 standard movement directions of the transparent surfaces for 1500 ms) indicated which of the surfaces had to be attended initially by the subjects. Beside target stimuli (discussed subsequently), 1 of the 4 simple motion sequences (cue sequences) could occur within the attended surface (for illustration, see Fig. 1). Each of these cue sequences consisted of a combination of 2 short subsequent displacements orthogonal to the standard movement direction of the attended surface (each displacement lasted for 300 ms and was separated by an interval of 200 ms; velocity 21.2°/s). Intense training prior to the scanning session ensured that upon these motion sequences, subjects maintained their attention at the same surface, switched their attention to the other surface within the same aperture, or switched their attention to one specific surface located in the opposite visual field (a detailed description of the individual motion sequences that were used to guide the subjects’ attention voluntarily is given in Fig. 1B). Thus, the instructional cue sequences resulted in 3 attention conditions that prompted the subjects to voluntarily direct their attention to 1 of the presented surfaces (Hold Attention, Switch Objects, and Switch Locations).

The subjects’ task was to deploy attention according to the instructional cue sequences and to perform a button press response whenever they detected the occurrence of a fast coherent movement (21.2°/s) in the predominant motion direction of the currently attended surface. In addition, fast movements could also occur within 1 of the nonattended surfaces, thereby capturing the subjects’ attention in a stimulus-driven manner (a detailed description of the target and capture trials is given in Fig. 1C). These capture trials could occur either within the unattended surface at the attended location (attentional capture across objects) or within 1 of the surfaces located in the unattended visual field (attentional capture across space). Thus, the fast movements not only served as targets, but also resulted in 2 additional attention conditions, in which the subjects’ attention was captured in a stimulus-driven manner by target-like stimuli (Capture across Objects and Capture across Space). Importantly, these fast movements within the nonattended surfaces were completely task-irrelevant, and subjects were instructed to ignore them.

To make sure that participants could reengage in the correct task in case that they had lost track of which surface needed to be attended (which was rare), a central cue was presented every 10th trial (a white arrow pointing in the motion direction of the currently to-be-attended surface for 1500 ms). The prior training ensured, however, that the attended surface was only rarely lost. All experimental manipulations (arrow cues, targets, capture trials, and instructional cue sequences) were considered as trials of independent conditions. The interval between the trials randomly varied between 3 and 8 s (mean intertrial interval: 3.7 s), following a gamma function to allow for trial separation in an event-related analysis (Hinrichs et al. 2000). Subjects performed 7 scanning runs of 6.3 min, which consisted of 12 blocks (time between arrow presentations) of 9 trials each, resulting in 51–65 trials per condition.

Functional Magnetic Resonance Imaging (fMRI) Data Acquisition

MR data were acquired on a 3 T MR scanner (Siemens Magnetom Trio, Erlangen, Germany) using an 8-channel head coil. An LCD projector back-projected the stimuli on a screen positioned behind the head coil, which was viewed by the subjects via a mirror attached to the coil. Functional images were acquired with a T2*-weighted echo planar imaging (EPI) gradient echo sequence [32 AC-PC oriented slices, thickness = 3.5 mm, in-plane resolution 64 × 64 mm, field of view (FoV) 224 × 224 mm, no gap, resulting voxel size = 3.5 × 3.5 × 3.5 mm, repetition time (TR) = 2000 ms, echo time (TE) = 30 ms, flip angle = 80°] in an odd–even interleaved sequence. Each scanning session consisted of 190 volumes. In a structural session, whole-head T1-weighted images of each subject's entire brain were collected using a magnetization-prepared rapid gradient echo sequence (96 sagittal slices, thickness = 2 mm, FoV 256 × 256 mm, no gap, spatial resolution = 1 × 1 × 2 mm, TR = 1650 ms, TE = 5 ms, TI = 1100 ms).

fMRI Data Analysis

Preprocessing and statistical analysis of the fMRI data were performed using the SPM5 software package (Wellcome Department of Cognitive Neurology, University College London, UK) and MATLAB 7.0 (The Mathwork Inc.). The functional volumes were realigned to the first volume and spatially normalized to an EPI template in standard Montreal Neurological Institute (MNI) space. After resampling to a final voxel size of 3 × 3 × 3 mm, the spatially normalized images were smoothed with an isotropic 8 mm full-width at half-maximum Gaussian kernel and highpass-filtered (cut-off 128 s).

For statistical analysis, blood-oxygen level-dependent (BOLD) responses were modeled by delta functions at the time of stimulus onset. For each subject, the resultant event regressors (Hold Attention, Switch Objects, Switch Locations, Capture across Objects, Capture across Locations, and Targets) were entered into a general linear model and convolved with the standard hemodynamic response function as well as with time and dispersion derivatives implemented in SPM5. The movement parameters derived from the realignment procedure were included as covariates in the model (Friston et al. 1998).

To identify regions commonly or differentially activated by voluntary and stimulus-driven shifts of attention between objects and locations, contrasts between individual conditions and conjunctions between the resultant contrasts were computed as follows. Group analyses were performed by submitting the individual-subject contrast estimates to a second-level, random-effects analysis, treating intersubject variability as a random effect to account for interindividual variance. Individual maxima within contiguous activation clusters are reported if they are separated by more than 16 mm. Stereotactic coordinates for voxels with maximal T-values within significant activation clusters are reported in the MNI standard space using an auxiliary voxel-level threshold of P < 0.005 (uncorrected) with subsequent cluster-level correction for multiple testing at P < 0.05 (corrected). The resultant activation maps were visualized using the MRIcron software package (http://www.mccauslandcenter.sc.edu/mricro/mricron/).

  1. Regions differentially activated by voluntary versus stimulus-driven shifts of attention were identified by contrasting “Switch Objects + Switch Locations” versus “Capture across Objects + Capture across Locations” trials.

  2. The common basis to voluntarily controlled shifts between objects and between locations was identified by forming a conjunction map of the “Switch Objects > Hold” and “Switch Locations > Hold” contrasts, whereas activity specific to both shift processes was assessed by directly contrasting “Switch Objects” versus “Switch Locations” trials against each other.

  3. A conjunction analysis between the “Capture across Objects > Hold” and “Capture across Locations > Hold” contrasts yielded activations common to both kinds of capture trials, whereas activity specific to each type was identified by a direct comparison of the “Capture across Objects” versus “Capture across Locations” trials.

  4. Activity common to voluntary and stimulus-driven shifts across locations was assessed by a conjunction of the “Switch Locations > Hold” and “Capture across Locations > Hold” contrasts, whereas regions specifically involved in controlling voluntary and stimulus-driven spatial shifts were identified by a direct comparison of “Switch Locations” versus “Capture across Locations” trials.

  5. Finally, a conjunction between the “Switch Objects > Hold” and “Capture across Objects > Hold” contrasts revealed activity that was common to both voluntary and involuntary objects-based shifts of attention, whereas differential activations to both shift types were identified by contrasting “Switch Objects” versus “Capture across Objects” trials against each other.

Eye Tracking

Eye movements were monitored during data acquisition using a custom-built MR-compatible eye-tracking device (for a detailed description of the eye-tracking system, see Kanowski et al. 2007). Recordings were performed using a modified version of the “Pupiltracker” software package (HumanScan AG, Erlangen, Germany). Before each run, an elliptic part of the monitored eyes’ image was defined as the template for tracking. During data acquisition, the software stored the pupils’ position as X- and Y-coordinates by computing the best match of each actual image (each lasting for 20 ms) with the template image within an adjustable search area. Data of 4 subjects were excluded from the analysis due to technical problems during the eye tracking. For statistical evaluation, the data were subjected to a 6 × 2 (condition vs. side of stimulus presentation) within-subject repeated-measures analysis of variance (RANOVA). The significance threshold for the RANOVA was set to P < 0.05 following Greenhouse–Geisser correction for nonsphericity.

Results

Behavioral Results

Average target-detection performance across all subjects was high during the functional runs (mean ± SEM: 95.82 ± 0.87%), whereas false alarms were rare (mean ± SEM: 2.39 ± 0.58%). On average, the subjects’ reaction time was 848 ms (SEM: ±30 ms), ranging from 646 to 1024 ms. One-way RANOVA with the factor side of target presentation (left vs. right visual field) was separately performed on the reaction time data and on the subjects’ hit rate. These analyses revealed neither a significant main effect of the side of target presentation for the hit rate (F1,15 = 2.9, P> 0.1) nor for the reaction times (F1,15 = 0.7, P> 0.4) of the subjects.

In order to assess whether the subjects’ attention was in fact captured by the target-like stimuli, we compared the reaction times with targets that were either preceded by another target or by a capture stimulus. This analysis revealed a significantly increased reaction time [paired T-test; T(1,15) = 2.2; P < 0.05] to targets that were preceded by a capture stimulus (876.1 ± 25.4 ms) in comparison to those that were preceded by another target (841.3 ± 31.9 ms). These reaction time costs to targets subsequently following target-like stimuli suggest that attentional capture took place.

Eye-Movement Data

Analysis of the eye-movement data revealed a very low occurrence of saccades (deviation of the pupils’ position from fixation >1%), with an average percentage of saccades across all conditions of 2.1%, ranging from 1.4% to 3.2%. Analysis of these data by a 2-way RANOVA with the factors attention condition (Targets, Hold Attention, Switch Objects, Switch Locations, Capture across Objects, and Capture across Locations) and side of stimulus presentation (left vs. right visual field) neither showed significant main effects of the experimental condition (F5,55 = 1.0, P> 0.3) or of the side of stimulus presentation (F1,11 = 2.2, P> 0.1) nor a significant interaction between both factors (F5,55 = 0.3, P> 0.7). Thus, the percentage of eye movements was very low and did not differ between conditions.

fMRI Results

Voluntary Versus Stimulus-Driven Orienting

First, we aimed to identify regions that were differentially involved during voluntary versus stimulus-driven orienting in general (collapsed across object- and location-based shifts of attention). Higher activity during voluntary than during stimulus-driven orienting occurred in regions classically described to belong to the frontoparietal attention network, including IPS, SPL, FEF, and supplementary motor area (SMA) (depicted in red/yellow in Fig. 2A). In addition, left dorsolateral prefrontal cortex (DLPFC) and striatum, as well as the inferior parietal lobe (IPL), precuneus, and frontal operculum (FO) of the left hemisphere, were also more active during voluntary orienting (see upper rows of Table 1 for the respective activation maxima).

Table 1

Peak activation foci for the comparison of voluntary and stimulus-driven orienting

Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 “Switch Objects + Switch Locations” > “Capture across Objects + Capture across Locations” 
FEF −27 −9 −60 9.79 <0.001 1437 
27 −12 60 7.62 
SMA −3 51 7.66 
54 9.21 
Striatum −15 12 −9 7.05 <0.001 489 
18 15 −6 7.61 <0.001 268 
FO −54 18 6.19 0.005 109 
DLPFC −33 33 30 6.33 0.002 140 
Anterior IPS/SPL −12 −63 57 9.02 <0.001 1231 
15 −66 54 6.59 
IPL −39 −42 42 7.68 
Precuneus −21 −63 27 6.45 
 “Capture across Objects + Capture across Locations” > “Switch Objects + Switch Locations” 
SMG/TPJ −57 −63 36 6.56 0.04 84 
51 −57 33 11.91 <0.001 410 
Anterior MTG −63 −27 −12 9.16 <0.001 823 
60 −15 −27 8.76 <0.001 451 
IFG −45 39 −12 6.56 <0.001 211 
48 45 −15 7.27 <0.001 458 
MFG 48 12 51 7.27 <0.001 284 
DMPFC/pre-SMA – 27 60 10.75 <0.001 2345 
42 39 8.47 
Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 “Switch Objects + Switch Locations” > “Capture across Objects + Capture across Locations” 
FEF −27 −9 −60 9.79 <0.001 1437 
27 −12 60 7.62 
SMA −3 51 7.66 
54 9.21 
Striatum −15 12 −9 7.05 <0.001 489 
18 15 −6 7.61 <0.001 268 
FO −54 18 6.19 0.005 109 
DLPFC −33 33 30 6.33 0.002 140 
Anterior IPS/SPL −12 −63 57 9.02 <0.001 1231 
15 −66 54 6.59 
IPL −39 −42 42 7.68 
Precuneus −21 −63 27 6.45 
 “Capture across Objects + Capture across Locations” > “Switch Objects + Switch Locations” 
SMG/TPJ −57 −63 36 6.56 0.04 84 
51 −57 33 11.91 <0.001 410 
Anterior MTG −63 −27 −12 9.16 <0.001 823 
60 −15 −27 8.76 <0.001 451 
IFG −45 39 −12 6.56 <0.001 211 
48 45 −15 7.27 <0.001 458 
MFG 48 12 51 7.27 <0.001 284 
DMPFC/pre-SMA – 27 60 10.75 <0.001 2345 
42 39 8.47 

Note: L, left; R, right.

Figure 2.

(A) Group activation maps for the comparison of goal-directed and stimulus-driven allocation of attention. Warm colors (red/yellow) depict regions that are more active during voluntary than during stimulus-driven orienting, whereas the inverse contrast (stimulus-driven > voluntary shifts of attention) is shown in blue/cyan. For both comparisons, data were collapsed to include trials in which attention was shifted across objects as well as across locations. (B) Group activation maps for the comparison of voluntary shifts between objects and locations. Activity common to both voluntary shift types (conjunction of the “Switch Objects > Hold” and “Switch Locations > Hold” contrasts) is shown in red/yellow, whereas activations specific to both voluntary shift processes are illustrated in blue/cyan (“Switch Locations > Switch Objects”) and green (“Switch Objects > Switch Locations”). (C) Group activation maps for the comparison of stimulus-driven shifts between objects and locations. Activations common to both kinds of capture trials (conjunction of the “Capture across Objects > Hold” and “Capture across Locations > Hold” contrasts) are depicted in red/yellow. Brain areas that were more strongly activated during stimulus-driven orienting across locations than across objects (“Capture across Locations > Capture across Objects”) are shown in blue/cyan, while we did not observe any significant activation for the inverse comparison.

Figure 2.

(A) Group activation maps for the comparison of goal-directed and stimulus-driven allocation of attention. Warm colors (red/yellow) depict regions that are more active during voluntary than during stimulus-driven orienting, whereas the inverse contrast (stimulus-driven > voluntary shifts of attention) is shown in blue/cyan. For both comparisons, data were collapsed to include trials in which attention was shifted across objects as well as across locations. (B) Group activation maps for the comparison of voluntary shifts between objects and locations. Activity common to both voluntary shift types (conjunction of the “Switch Objects > Hold” and “Switch Locations > Hold” contrasts) is shown in red/yellow, whereas activations specific to both voluntary shift processes are illustrated in blue/cyan (“Switch Locations > Switch Objects”) and green (“Switch Objects > Switch Locations”). (C) Group activation maps for the comparison of stimulus-driven shifts between objects and locations. Activations common to both kinds of capture trials (conjunction of the “Capture across Objects > Hold” and “Capture across Locations > Hold” contrasts) are depicted in red/yellow. Brain areas that were more strongly activated during stimulus-driven orienting across locations than across objects (“Capture across Locations > Capture across Objects”) are shown in blue/cyan, while we did not observe any significant activation for the inverse comparison.

The inverse comparison (stimulus-driven > voluntary attention) revealed clusters of significant activation to stimulus-driven reorienting bilaterally within supramarginal gyrus [SMG; extending into the temporo-parietal junction (TPJ)], anterior portions of the middle temporal gyrus (MTG), and inferior frontal gyrus (IFG), but also in right hemispheric medial frontal gyrus (MFG) and midline dorsomedial prefrontal cortex (DMPFC)/pre-SMA (illustrated in blue/cyan in Fig. 2A; local activation maxima are shown in the bottom half of Table 1).

Voluntary Orienting Between Objects and Locations

With our first analysis, we identified networks that were differentially activated during voluntary versus stimulus-driven orienting (collapsed over location- and object-based shifts). These analyses, however, did not allow disentangling the networks that mediate attention shifts between locations and/or objects during either voluntary or stimulus-driven attentional orienting. Thus, to determine the regions specifically involved in the control of voluntary shifts between objects and locations, we directly contrasted “Switch Locations” versus “Switch Objects” trials and compared the respective activations with those identified by a conjunction analysis of both shift types (conjunction of the “Switch Locations > Hold” and “Switch Objects > Hold” contrasts).

The conjunction analysis revealed clusters of significant activation to both voluntary shift types within classical frontoparietal attentional control regions (FEF, SMA, and anterior IPS/SPL), but also in left posterior MTG and in the right SMG/TPJ (illustrated in red/yellow in Fig. 2B; local activation maxima are shown in the upper part of Table 2). The activations from the conjunction analysis within the anterior IPS/SPL were located in more lateral subregions of the parietal cortex than those for the comparison of spatial > object-based voluntary shifts (shown in blue/cyan in Fig. 2B; local activation maxima are shown in the middle part of Table 2). The activations for this comparison (“Switch Locations > Switch Objects” trials) were observed primarily within medial aspects of posterior parietal and occipital cortices [precuneus, posterior cingulate cortex (PCC), and SPL], except for 2 activation maxima located in the bilateral superior occipital gyrus. The only region that was significantly activated by the opposite comparison (“Switch Objects > Switch Locations”) was located in the dorsal anterior cingulate cortex (dACC) (depicted in green in Fig. 2B; corresponding coordinates, Tmax, and cluster size are shown in the bottom part of Table 2).

Table 2

Peak activations for the comparison of voluntary shifts between objects and locations

Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Switch Locations > Hold” and “Switch Objects > Hold” 
Anterior IPS/SPL −18 −69 57 6.52 <0.001 1626 
24 −63 57 5.53 
FEF −24 −9 60 5.79 0.001 396 
27 −23 57 6.19 0.002 327 
SMA 12 −6 66 4.23 
Posterior MTG −33 −75 18 5.16 0.02 226 
SMG/TPJ 57 −45 21 5.01 0.05 153 
 “Switch Locations > Switch Objects” 
SPL – −3 −54 57 6.75 <0.001 3937 
Precuneus/PCC −12 −66 27 6.24 
12 −51 18 5.79 
−12 −78 45 5.97 
15 −78 42 4.16 
Superior occipital gyrus −42 −81 27 6.04 
42 −84 27 4.59 
 “Switch Objects > Switch Locations” 
dACC – 27 42 5.36 0.02 152 
Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Switch Locations > Hold” and “Switch Objects > Hold” 
Anterior IPS/SPL −18 −69 57 6.52 <0.001 1626 
24 −63 57 5.53 
FEF −24 −9 60 5.79 0.001 396 
27 −23 57 6.19 0.002 327 
SMA 12 −6 66 4.23 
Posterior MTG −33 −75 18 5.16 0.02 226 
SMG/TPJ 57 −45 21 5.01 0.05 153 
 “Switch Locations > Switch Objects” 
SPL – −3 −54 57 6.75 <0.001 3937 
Precuneus/PCC −12 −66 27 6.24 
12 −51 18 5.79 
−12 −78 45 5.97 
15 −78 42 4.16 
Superior occipital gyrus −42 −81 27 6.04 
42 −84 27 4.59 
 “Switch Objects > Switch Locations” 
dACC – 27 42 5.36 0.02 152 

Note: L, left; R, right.

Stimulus-Driven Orienting Across Objects and Locations

In the same way as for the comparison of voluntary shifts between objects and locations (as described earlier), we compared activations with exogenously provoked shifts of attention across objects and locations. For this purpose, we contrasted “Capture across Locations” versus “Capture across Objects” trials and compared the particular activations with those detected by the conjunction analysis of both capture types (conjunction of “Capture across Locations > Hold” and “Capture across Objects > Hold”).

Within occipito-parietal cortex, activations common to both capture types were present only within closely confined regions of the right SMG/TPJ (shown in red/yellow in Fig. 2C; local activation maxima are shown in the upper part of Table 3), whereas in the frontal lobe, significantly activated regions covered widespread regions of right MFG (extending into caudal aspects of midline DMPFC/pre-SMA) and right IFG (adjacent to 1 additional activation cluster observed in the anterior MTG). Higher activations to involuntary location- versus object-based orienting (shown in blue/cyan in Fig. 2C) were observed across several distinct clusters located in the rostral DMPFC and ventromedial PFC (VMPFC), anterior portions of MTG and superior temporal gyrus, left SMG/TPJ, and right lingual gyrus (local activation maxima for all clusters are shown in the middle rows of Table 3). Moreover, an additional cluster covered large parts of the precuneus and PCC and extended into dorsal parts of the anterior SPL. For the inverse contrast (“Capture across Objects > Capture across Locations”), we observed no activations at the cluster correction threshold of P < 0.05.

Table 3

Peak activations for the comparison of stimulus-driven shifts between objects and locations

Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Capture across Locations > Hold” and “Capture across Objects > Hold” 
SMG/TPJ 51 −51 27 7.54 <0.001 654 
DLPFC/MFG 51 15 48 7.12 <0.001 1052 
DMPFC/pre-SMA – 12 33 48 7.43 
IFG 45 45 −12 5.92 <0.001 735 
Anterior MTG 60 −15 −21 5.51 0.03 205 
 “Capture across Locations>Capture across Objects” 
Precuneus/PCC – −6 −45 51 7.17 <0.001 3324 
Anterior SPL −30 −27 57 6.92 
 30 −39 63 6.35 
PCC – −51 18 6.40 
Rostral DMPFC/VMPFC – −6 63 15 6.77 <0.001 1186 
  36 −21 6.03 
Anterior MTG −30 −27 −24 6.26 <0.001 860 
 48 −36 6.70 <0.001 698 
SMG/TPJ −45 −75 30 5.75 0.001 119 
STG −36 −30 5.01 <0.001 151 
 33 −27 12 4.43 0.002 106 
LG 24 −69 −9 4.98 0.002 100 
 “Capture across Objects >Capture across Locations” 
– – 
Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Capture across Locations > Hold” and “Capture across Objects > Hold” 
SMG/TPJ 51 −51 27 7.54 <0.001 654 
DLPFC/MFG 51 15 48 7.12 <0.001 1052 
DMPFC/pre-SMA – 12 33 48 7.43 
IFG 45 45 −12 5.92 <0.001 735 
Anterior MTG 60 −15 −21 5.51 0.03 205 
 “Capture across Locations>Capture across Objects” 
Precuneus/PCC – −6 −45 51 7.17 <0.001 3324 
Anterior SPL −30 −27 57 6.92 
 30 −39 63 6.35 
PCC – −51 18 6.40 
Rostral DMPFC/VMPFC – −6 63 15 6.77 <0.001 1186 
  36 −21 6.03 
Anterior MTG −30 −27 −24 6.26 <0.001 860 
 48 −36 6.70 <0.001 698 
SMG/TPJ −45 −75 30 5.75 0.001 119 
STG −36 −30 5.01 <0.001 151 
 33 −27 12 4.43 0.002 106 
LG 24 −69 −9 4.98 0.002 100 
 “Capture across Objects >Capture across Locations” 
– – 

Note: L, left; R, right; STG, superior temporal gyrus.

Voluntary Versus Stimulus-Driven Orienting Between Locations

Networks mediating voluntary and stimulus-driven shifts between locations were revealed by contrasting “Switch Locations” versus “Capture across Locations” trials and comparing the activations with those common to both location-based shift types (conjunction of the “Switch Locations > Hold” and “Capture across Locations > Hold” contrasts). Higher activity to voluntary than to involuntary spatial shifts (shown in blue/cyan in Fig. 3A) was observed within the classical frontoparietal attention network (FEF, SPL, and SMA), but also in SMG/IPL, striatum, and the FO of the left hemisphere (local activation maxima for all clusters are shown in Table 4). The inverse comparison (“Capture across Locations > Switch Locations”) revealed activations in bilateral anterior MTG, IFG, and in the DLPFC, DMPFC, and VMPFC (illustrated in green in Fig. 3A; local activation maxima are shown in the bottom part of Table 4). Within posterior brain regions, activity was observed in bilateral SMG/TPJ and the precuneus/PCC extending into the left anterior SPL. Finally, activity common to both voluntary and stimulus-driven shifts across locations was identified by the conjunction analysis that revealed significant clusters of activity in posterior brain regions (illustrated in red/yellow in Fig. 3A; local activation maxima are shown in the upper part of Table 4), including precuneus/PCC, SPL, and bilateral middle occipital gyrus (extending into caudo-ventral portions of the bilateral angular gyrus/TPJ). Note that the bilateral activations in middle occipital and angular gyri were located more ventral than those observed in the “Capture across Locations > Switch Locations” contrast (compare red/yellow and green activation maps in Fig. 3A).

Table 4

Peak activations for the comparison of voluntary and stimulus-driven shifts between locations

Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Switch Locations > Hold” and “Capture across Locations > Hold” 
Precuneus/PCC −6 −45 48 8.33 <0.001 848 
 −48 48 8.55 
SPL 18 −57 66 4.31 
Angular gyrus/TPJ middle occipital gyrus −45 −75 24 6.11 <0.001 237 
  −27 −81 18 4.99 
 48 −69 24 6.01 <0.001 294 
  27 −81 21 4.84 
 “Switch Locations > Capture across Locations” 
FEF −27 −9 42 7.81 <0.001 840 
 18 −3 48 7.37 
SMA −3 −3 57 5.83 
 −3 63 5.76 
FO −54 18 5.61 <0.05 92 
Striatum −24 5.41 <0.005 155 
SPL −12 −66 60 7.14 <0.001 1032 
 12 −63 57 5.20 
SMG/IPL −45 −39 33 5.83 
 “Capture across Locations >Switch Locations” 
Anterior MTG −60 −21 −15 12.53 <0.001 1724 
IFG  −45 33 −9 6.49 
DMPFC/pre-SMA 12 24 60 10.78 <0.001 4240 
Anterior MTG  66 −36 −9 9.16 
DLPFC/MFG  48 12 54 6.83 
VMPFC – 39 −9 6.23 
IFG 42 33 −9 5.58 
SMG/TPJ 57 −60 33 10.36 <0.001 299 
Precuneus/PCC −15 −48 33 5.41 <0.001 725 
 −51 36 7.39 
Anterior SPL −36 −30 66 6.56 
SMG/TPJ −57 −63 36 6.54 <0.01 137 
Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 Conjunction of “Switch Locations > Hold” and “Capture across Locations > Hold” 
Precuneus/PCC −6 −45 48 8.33 <0.001 848 
 −48 48 8.55 
SPL 18 −57 66 4.31 
Angular gyrus/TPJ middle occipital gyrus −45 −75 24 6.11 <0.001 237 
  −27 −81 18 4.99 
 48 −69 24 6.01 <0.001 294 
  27 −81 21 4.84 
 “Switch Locations > Capture across Locations” 
FEF −27 −9 42 7.81 <0.001 840 
 18 −3 48 7.37 
SMA −3 −3 57 5.83 
 −3 63 5.76 
FO −54 18 5.61 <0.05 92 
Striatum −24 5.41 <0.005 155 
SPL −12 −66 60 7.14 <0.001 1032 
 12 −63 57 5.20 
SMG/IPL −45 −39 33 5.83 
 “Capture across Locations >Switch Locations” 
Anterior MTG −60 −21 −15 12.53 <0.001 1724 
IFG  −45 33 −9 6.49 
DMPFC/pre-SMA 12 24 60 10.78 <0.001 4240 
Anterior MTG  66 −36 −9 9.16 
DLPFC/MFG  48 12 54 6.83 
VMPFC – 39 −9 6.23 
IFG 42 33 −9 5.58 
SMG/TPJ 57 −60 33 10.36 <0.001 299 
Precuneus/PCC −15 −48 33 5.41 <0.001 725 
 −51 36 7.39 
Anterior SPL −36 −30 66 6.56 
SMG/TPJ −57 −63 36 6.54 <0.01 137 

Note: L, left; R, right.

Figure 3.

(A) Group activation maps for the comparison of voluntary versus stimulus-driven shifts between locations. Activations common to both location-based shift types (conjunction of the “Switch Locations > Hold” and “Capture across Locations > Hold” contrasts) are illustrated in red/yellow, whereas activations specific to voluntary or stimulus-driven spatial shifts are illustrated in blue/cyan (“Switch Locations > Capture across Locations”) and green (“Capture across Locations > Switch Locations”). (B) Group activation maps for the comparison of voluntary versus stimulus-driven shifts between objects. Activity common to both voluntary and stimulus-driven shifts across objects (conjunction of the “Switch Objects > Hold” and “Capture across Objects > Hold” contrasts) is depicted in red/yellow. Brain areas showing higher activity to voluntary compared with stimulus-driven shifts across objects are shown in blue/cyan (“Switch Objects > Capture across Objects”), whereas significant activations for the opposite comparison are shown in green (“Capture across Objects > Switch Objects”).

Figure 3.

(A) Group activation maps for the comparison of voluntary versus stimulus-driven shifts between locations. Activations common to both location-based shift types (conjunction of the “Switch Locations > Hold” and “Capture across Locations > Hold” contrasts) are illustrated in red/yellow, whereas activations specific to voluntary or stimulus-driven spatial shifts are illustrated in blue/cyan (“Switch Locations > Capture across Locations”) and green (“Capture across Locations > Switch Locations”). (B) Group activation maps for the comparison of voluntary versus stimulus-driven shifts between objects. Activity common to both voluntary and stimulus-driven shifts across objects (conjunction of the “Switch Objects > Hold” and “Capture across Objects > Hold” contrasts) is depicted in red/yellow. Brain areas showing higher activity to voluntary compared with stimulus-driven shifts across objects are shown in blue/cyan (“Switch Objects > Capture across Objects”), whereas significant activations for the opposite comparison are shown in green (“Capture across Objects > Switch Objects”).

Voluntary Versus Stimulus-Driven Orienting Between Objects

In analogy to the previous comparisons, we compared activations between endogenously and exogenously initiated attention shifts across objects. “Shift Objects” were contrasted against “Capture across Objects” trials, and the activation maps were compared to those revealed by the conjunction analysis of both object-based shift types (conjunction of “Shift Objects > Hold” and “Capture across Objects > Hold”). Higher activity to voluntary than to stimulus-driven shifts between objects was observed in the classical frontoparietal attention network (IPS, SPL, FEF, and SMA). Moreover, significant activity was observed in the bilateral striatum and in the left MFG (depicted in blue/cyan in Fig. 3B; local activation maxima are shown in Table 5). The inverse contrast identified activations in 4 narrowly circumscribed regions located in the DMPFC/pre-SMA, left DLPFC/MFG, as well as in the right IFG and right SMG/TPJ (illustrated in green in Fig. 3B; local activation maxima are shown in the bottom part of Table 5). Finally, the conjunction analysis of both object-based shift types revealed only 1 activation cluster in the right SMG/TPJ, which was located rostro-ventral to the SMG/TPJ activation observed in the “Capture across Objects > Shift Objects” contrast.

Table 5

Peak activations for the comparison of voluntary and stimulus-driven shifts between objects

Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 “Conjunction of Switch Objects > Hold” and “Capture across Objects > Hold” 
SMG/TPJ 57 −48 24 5.65 <0.05 63 
 “Switch Objects > Capture across Objects” 
FEF −24 −9 57 9.23 <0.001 1313 
24 −12 57 6.39 
SMA −3 −3 60 6.30 
−9 69 6.39 
Anterior IPS −36 −42 45 9.03 <0.001 1127 
30 −51 48 5.82 
SPL −15 −69 60 8.71 
15 −66 57 5.45 
Striatum 18 15 −3 8.42 <0.001 318 
−18 12 8.39 <0.001 592 
MFG −30 36 27 5.96 <0.01 112 
 “Capture across Objects > Switch Objects” 
DLPFC/MFG 45 15 36 8.75 <0.001 314 
DMPFC/pre-SMA – 27 60 8.71 <0.001 379 
SMG/TPJ 48 −60 33 8.00 <0.001 376 
IFG 45 42 −9 6.82 <0.001 255 
Anatomical structure Hemisphere MNI coordinates (x, y, z)
 
Max. T-value Corrected cluster P-value Cluster size 
 “Conjunction of Switch Objects > Hold” and “Capture across Objects > Hold” 
SMG/TPJ 57 −48 24 5.65 <0.05 63 
 “Switch Objects > Capture across Objects” 
FEF −24 −9 57 9.23 <0.001 1313 
24 −12 57 6.39 
SMA −3 −3 60 6.30 
−9 69 6.39 
Anterior IPS −36 −42 45 9.03 <0.001 1127 
30 −51 48 5.82 
SPL −15 −69 60 8.71 
15 −66 57 5.45 
Striatum 18 15 −3 8.42 <0.001 318 
−18 12 8.39 <0.001 592 
MFG −30 36 27 5.96 <0.01 112 
 “Capture across Objects > Switch Objects” 
DLPFC/MFG 45 15 36 8.75 <0.001 314 
DMPFC/pre-SMA – 27 60 8.71 <0.001 379 
SMG/TPJ 48 −60 33 8.00 <0.001 376 
IFG 45 42 −9 6.82 <0.001 255 

Note: L, left; R, right.

Discussion

The involvement of frontoparietal regions in the control of voluntary (Yantis et al. 2002; Liu et al. 2003; Serences et al. 2004, 2005) and stimulus-driven shifts of attention (Kincade et al. 2005; Serences et al. 2005; Indovina and Macaluso 2007; Shulman et al. 2009) has been subject of intense research. However, no study directly compared the neural correlates of voluntary and stimulus-driven attentional control across different domains, for example, between spatial and nonspatial shifts of attention. The current study employed fMRI to directly compare the mechanisms of voluntary and stimulus-driven attentional control during object- and location-based attention shifts in a within-subject design. Consistent with previous studies, we observed activations in a widespread network of frontoparietal, extrastriate visual, and default-mode network regions. Within this network, multiple areas were commonly recruited by either voluntary or stimulus-driven shift of attention between objects and locations. Therein, activity common to both voluntary shift types was observed within classical frontoparietal attentional control regions (FEF, SMA, and anterior IPS/SPL) and in the right SMG/TPJ, whereas stimulus-driven reorienting between objects and locations commonly activated midline DMPFC/pre-SMA and ventral attention network regions (right hemispheric IFG, MFG, and SMG/TPJ).

More importantly, however, we also observed object- and location-selective activations, which moreover differed substantially between voluntary and stimulus-driven re-/orienting. Higher activity to spatial than object-based voluntary shifts activated medial parts of posterior parietal and occipital cortices (precuneus, PCC, and SPL) and 2 additional areas located in the bilateral superior occipital gyrus, whereas the opposite comparison (“Switch Objects > Switch Locations”) only revealed 1 cluster located in the dACC. During stimulus-driven reorienting, higher activity for space- than for object-based shifts was mainly observed within rostral DMPFC and VMPFC and across large parts of the occipito-parietal cortex, including precuneus, PCC, and dorsal portions of the anterior SPL. The inverse contrast (“Capture across Objects > Capture across Locations”) did not reveal any clusters of significant activation.

Finally, we investigated the commonalities and differences between voluntary and stimulus-driven attentional control when attention was shifted either between objects or between locations. Higher activity for voluntary than for stimulus-driven orienting for both object- and location-based attention shifts was observed in classical dorsal attention network regions, whereas the inverse comparisons both showed activations in the ventral frontoparietal cortex. Activity common to voluntary and stimulus-driven shifts across locations was confined to the same medial parts of posterior parietal and occipital cortices (precuneus, PCC, and SPL) as observed in the comparison of spatial > object-based attention shifts, whereas activity common to voluntary and stimulus-driven shifts across objects was only evident in 1 cluster located in the right SMG/TPJ.

These results show that voluntary and stimulus-driven shifts between objects and locations recruit partially overlapping, but also separable, cortical regions, indicative of the parallel existence of domain-independent and domain-specific reconfiguration signals in dependence of the particular attentional demands.

Neural Networks for Voluntary Versus Stimulus-Driven Orienting

The comparison of voluntary versus stimulus-driven attentional control (collapsed across object- and location-based shifts) revealed higher activity to voluntary shifts within the classical dorsal frontoparietal attention network and in the bilateral striatum (Fig. 2A). Stimulus-driven reorienting, in contrast, elicited activity across several ventral frontoparietal and default-mode network regions. This pattern is in line with current theories on attentional control, postulating 2 interacting systems that mediate the allocation of resources to environmental events (Hopfinger and Mangun 1998; Hopfinger and West 2006; Corbetta et al. 2008). The activations in the dorsal frontoparietal system to voluntary shifts are in accordance with its well-known role in the generation of endogenous signals that bias the processing of particular features, objects, or spatial locations, according to expectations and current goals (Kastner et al. 1999; Corbetta et al. 2000; Hopfinger et al. 2000). The ventral frontoparietal network, in contrast, is not activated by expectations or task preparation, but is recruited when attention is involuntarily oriented toward unexpected or behaviorally relevant events (Kincade et al. 2005; Serences et al. 2005; Indovina and Macaluso 2007; Shulman et al. 2009), as observed in response to our target-like attention-capturing stimuli.

Besides these ventral frontoparietal activations, stimulus-driven shifts also recruited distinct default-mode network areas. The functional connectivity of this network is known to correlate negatively with the activation state of the dorsal frontoparietal network (He et al. 2007) in terms of a push–pull relationship between the 2 systems (Sridharan et al. 2008). Therefore, it appears likely that the clusters observed in the default-mode network were found due to a lack of, or at least less pronounced, deactivation in response to stimulus-driven than goal-directed shifts of attention. Such a lack of deactivation during stimulus-driven reorienting might in turn simply reflect a failure to suppress distractive information and thus to deploy attentional resources to the current task (Rule et al. 2002).

Cortical Networks Common or Specific for Voluntary Shifts Between Objects and Locations

As noted earlier, our current task design not only permitted us to compare the neural mechanisms of voluntary and stimulus-driven orienting, but also allowed us to investigate the commonalities and differences between object- and location-based shifts during voluntary and stimulus-driven attentional control, respectively. A direct comparison of voluntarily initiated shifts between objects and locations revealed activity common to both processes within classical dorsal frontoparietal regions (FEF, SMA, and SPL/IPS) and in the right SMG/TPJ. This pattern corroborates earlier findings of a transient reconfiguration signal in the dorsal frontoparietal cortex, which has been observed for attention shifts between locations (Hopfinger et al. 2000; Yantis et al. 2002; Slagter et al. 2007), features (Liu et al. 2003; Slagter et al. 2007; Greenberg et al. 2010), objects (Serences et al. 2004), sensory modalities (Macaluso et al. 2002; Shomstein and Yantis 2004), or even for switches between different task sets (Slagter et al. 2006).

More importantly, however, we also observed domain-specific activations for both object- and location-based voluntary orienting. Spatial shifts recruited distinct portions of the SPL and adjacent regions of precuneus, PCC, and superior occipital gyrus. These location shift-related activations covered more medial/posterior parts of the parietal cortex than those common to both object- and location-based shifts (Fig. 2B). This pattern is in line with previous studies showing increased activations in mediodorsal parts of the posterior parietal cortex to spatial in comparison to nonspatial attentional reconfigurations (Vandenberghe et al. 2001; Giesbrecht et al. 2003; Molenberghs et al. 2007; Slagter et al. 2007; Shulman et al. 2009), but also with classical neuropsychological findings of selectively impaired spatial orienting in patients with lesions to these regions (Mesulam 1981, 1990; Posner et al. 1984; Cammalleri et al. 1996; Katayama et al. 1999).

On a theoretical account, shifts between objects and locations differ with respect to the computational demands crucial for accurate shift execution: the receptive fields of neurons, which process the particular object–features before and after shift execution, either cover the same (shifts across objects) or different parts of visual space (shifts across locations). Thus, spatial shifts first require the computation of their target coordinates before the correct subset of neurons can be selected for modulation. These computations would be expected to recruit additional resources, resulting in a larger cortical net reconfiguration signal, which matches the pattern we observed in mediodorsal parts of the posterior parietal cortex and PCC. The posterior parietal cortex has been shown to be retinotopically organized (Sereno et al. 2001; Saygin and Sereno 2008) and, together with the PCC and precuneus, has been implicated in the generation of spatial reference frames that are used to guide reaching, grasping, mental navigation, as well as gaze and attention shifts (Andersen et al. 1993; Duhamel et al. 1997; Ghaem et al. 1997; Maguire 1997; Andersen and Buneo 2002; Shulman et al. 2009). Likewise, transcranial magnetic stimulation of posterior but not anterior parts of the IPS has been demonstrated to impair spatial updating of remembered target locations across saccades (Morris et al. 2007). Thus, the observed activations in mediodorsal parietal regions might in fact reflect the processing specific to the spatial updating processes underlying shifts between locations.

The inverse comparison (higher activity to voluntary object- than location-based shifts), in contrast, revealed only 1 narrowly circumscribed cluster located in the dACC. This region has recently been characterized to be part of the so-called “saliency network” (Dosenbach et al. 2007; Sridharan et al. 2008), which is recruited by perceptually or cognitively demanding tasks or in response to salient events. It has been suggested to mediate states of increased alertness and readiness for action, enabling the redistribution of processing resources toward events that pose an actual or potential challenge to an organism (Menon and Uddin 2010; Sterzer and Kleinschmidt 2010). In this context, voluntary shifts between objects might in fact be more challenging than spatial ones, insofar as the instructional cue sequence occurs in the particular object from which attention needs to be released after completion of the respective trial. This object, however, is perceptually primed by the motion onset of the cue sequence, imposing the need for additional processing resources to overcome this perceptual facilitation to enable the shift toward the nonprimed object at the same spatial location.

Cortical Networks Common or Specific for Stimulus-Driven Orienting Across Objects and Locations

In the same way as for voluntarily controlled attention shifts, we also investigated the commonalities and differences between object- and location-based attentional control during stimulus-driven reorienting. The conjunction analysis revealed activity common to both involuntary shifts types in the right SMG/TPJ and more widespread activations in frontal/prefrontal areas, including DMPFC/pre-SMA as well as right IFG and MFG. Some of these regions (SMG/TPJ, IFG, and MFG) constitute the ventral frontoparietal attention network as described in previous studies on stimulus-driven orienting (Kincade et al. 2005; Serences et al. 2005; Indovina and Macaluso 2007; Shulman et al. 2009). Moreover, the pre-SMA has previously also been implicated in stimulus-driven orienting, for example, toward target-colored distractor stimuli (Serences et al. 2005), whereas the large activation cluster in the DMPFC/pre-SMA is likely to result from processes related to conflict monitoring and inhibitory control due to the fact that our attention-capturing stimuli had to be identified as nontargets upon which the subjects had to withhold a manual response (Ridderinkhof et al. 2004; Li et al. 2006; Sharp et al. 2010).

Beyond this network common to both involuntary shift types, we also observed domain-specific activations for stimulus-driven shifts between objects and locations. Higher activity to involuntary spatial compared with object-based shifts was observed within anterior portions of the bilateral SPL and adjacent PCC/precuneus, and an additional frontal activation cluster covered large parts of rostral DMPFC and VMPFC (BA 10). The parietal activations to involuntary spatial shifts were located rostrally to those observed for voluntary attentional control (compare Fig. 2B,C). Activity within the same anterior portions of the SPL and PCC/precuneus has very recently been described in a study investigating the effects of voluntary and stimulus-driven attention on episodic memory formation (Uncapher et al. 2011). In this study, activity within this region was positively associated with encoding success for items presented in the focus of attention (i.e. enhanced activity for events that were later remembered vs. forgotten), whereas the opposite effect was observed for items presented at unattended locations (i.e. enhanced activity for events that were later forgotten vs. remembered). Based on these findings, the authors suggested that activity within the anterior SPL constitutes a biasing signal promoting memory formation for stimuli presented within the focus of attention, at the expense of information appearing at unattended locations. In concert with our current results, these data imply that the anterior SPL might be recruited to overcome involuntary shifts toward irrelevant locations, thereby promoting more efficient allocation of processing resources toward relevant (voluntarily to-be-attended) parts of visual space. The concurrent activation of rostral DMPFC/VMPFC and PCC/precuneus is also well in line with this notion, in that these regions have been suggested to prevent distractibility by external events (e.g. by inhibition of dorsal attention network regions), thereby strengthening the goal-directed orienting of attentional resources in accordance with the currently relevant attentional set (Small et al. 2003).

Cortical Networks Common or Specific for Voluntary and Stimulus-Driven Shifts Between Locations or Between Objects

Finally, our task design allowed for a direct comparison not only between voluntary and stimulus-driven orienting in general, but also separately for both perceptual domains, that is, during object- or location-based orienting. We observed higher activity for voluntary than for stimulus-driven orienting within overlapping regions of the dorsal frontoparietal cortex across both perceptual domains (FEF, IPS, SPL, and SMA). Similarly, both inverse comparisons (stimulus-driven > voluntary orienting across objects or across locations) also showed highly overlapping activation patterns, which were confined to the ventral attention network (IFG, MFG, and SMG/TPJ). These data suggest that the control of both object- and location-based orienting involves very similar networks located in dorsal and ventral frontoparietal cortices (compare Fig. 3A,B and Tables 4 and 5).

However, besides these highly overlapping networks, the analyses also revealed domain-specific activations. Both voluntary and stimulus-driven shifts across locations commonly activated medial posterior parietal and occipital cortices (precuneus, PCC, and SPL), which emphasize the role of these regions in the spatial updating processes underlying spatial shifts. Moreover, spatial shifts commonly activated caudo-ventral portions of the bilateral angular gyrus/TPJ, whereas activity common to object-based orienting was confined to a single cluster in the right rostro-ventral SMG/TPJ.

Previous studies have shown that activity in the TPJ is invoked not only by spatial but also by nonspatial information (McCarthy et al. 1997; Linden et al. 1999; Marois et al. 2000), which led to the suggestion that the TPJ might mediate the disengagement of attention (Posner et al. 1984; Friedrich et al. 1998). In this view, the disengagement initiates the transitions between the ventral and dorsal attention networks that control orientation responses to different environmental events (Corbetta et al. 2008; Doricchi et al. 2010). Our data support this idea as we observed TPJ activations to voluntary and stimulus-driven orienting not only between spatial locations, but also between different objects. Importantly, these activations were located within spatially distinct parts of the angular gyrus/SMG/ TPJ, suggesting that spatial and nonspatial orienting recruit different parts of the TPJ complex.

Conclusions

Together, the present findings reveal the parallel existence of both common and separable neural substrates mediating attentional control during shifts of attention between objects and locations. Importantly, these domain-general and domain-specific networks differed substantially between situations in which attention was controlled either in a goal-directed or in a stimulus-driven manner. These results suggest that voluntary and stimulus-driven shifts between objects and locations recruit partially overlapping, but also separable, cortical regions, which convey the specific reconfiguration signals in dependence of the particular attentional demands. These findings provide a useful framework for future investigation of the mechanisms underlying attentional control across different perceptual domains.

Funding

This work was supported by the following grants: Scho 1217/1-2 and SFB 779-A1 from the Deutsche Forschungsgemeinschaft (DFG) awarded to M.A.S.

Notes

The authors thank Dr Michael Scholz for technical advice. Conflict of Interest: None declared.

References

Andersen
RA
Buneo
CA
Intentional maps in posterior parietal cortex
Annu Rev Neurosci
 , 
2002
, vol. 
25
 (pg. 
189
-
220
)
Andersen
RA
Snyder
LH
Li
CS
Stricanne
B
Coordinate transformations in the representation of spatial information
Curr Opin Neurobiol
 , 
1993
, vol. 
3
 (pg. 
171
-
176
)
Cammalleri
R
Gangitano
M
D'Amelio
M
Raieli
V
Raimondo
D
Camarda
R
Transient topographical amnesia and cingulate cortex damage: a case report
Neuropsychologia
 , 
1996
, vol. 
34
 (pg. 
321
-
326
)
Corbetta
M
Kincade
JM
Ollinger
JM
McAvoy
MP
Shulman
GL
Voluntary orienting is dissociated from target detection in human posterior parietal cortex
Nat Neurosci
 , 
2000
, vol. 
3
 (pg. 
292
-
297
)
Corbetta
M
Patel
G
Shulman
GL
The reorienting system of the human brain: from environment to theory of mind
Neuron
 , 
2008
, vol. 
58
 (pg. 
306
-
324
)
Doricchi
F
Macci
E
Silvetti
M
Macaluso
E
Neural correlates of the spatial and expectancy components of endogenous and stimulus-driven orienting of attention in the Posner task
Cereb Cortex
 , 
2010
, vol. 
20
 (pg. 
1574
-
1585
)
Dosenbach
NU
Fair
DA
Miezin
FM
Cohen
AL
Wenger
KK
Dosenbach
RA
Fox
MD
Snyder
AZ
Vincent
JL
Raichle
ME
Distinct brain networks for adaptive and stable task control in humans
Proc Natl Acad Sci USA
 , 
2007
, vol. 
104
 (pg. 
11073
-
11078
)
Duhamel
JR
Bremmer
F
BenHamed
S
Graf
W
Spatial invariance of visual receptive fields in parietal cortex neurons
Nature
 , 
1997
, vol. 
389
 (pg. 
845
-
848
)
Duncan
J
Selective attention and the organization of visual information
J Exp Psychol Gen
 , 
1984
, vol. 
113
 (pg. 
501
-
517
)
Friedrich
FJ
Egly
R
Rafal
RD
Beck
D
Spatial attention deficits in humans: a comparison of superior parietal and temporal–parietal junction lesions
Neuropsychology
 , 
1998
, vol. 
12
 (pg. 
193
-
207
)
Friston
KJ
Fletcher
P
Josephs
O
Holmes
A
Rugg
MD
Turner
R
Event-related fMRI: characterizing differential responses
NeuroImage
 , 
1998
, vol. 
7
 (pg. 
30
-
40
)
Ghaem
O
Mellet
E
Crivello
F
Tzourio
N
Mazoyer
B
Berthoz
A
Denis
M
Mental navigation along memorized routes activates the hippocampus, precuneus, and insula
Neuroreport
 , 
1997
, vol. 
8
 (pg. 
739
-
744
)
Giesbrecht
B
Woldorff
MG
Song
AW
Mangun
GR
Neural mechanisms of top-down control during spatial and feature attention
NeuroImage
 , 
2003
, vol. 
19
 (pg. 
496
-
512
)
Greenberg
AS
Esterman
M
Wilson
D
Serences
JT
Yantis
S
Control of spatial and feature-based attention in frontoparietal cortex
J Neurosci
 , 
2010
, vol. 
30
 (pg. 
14330
-
14339
)
He
BJ
Snyder
AZ
Vincent
JL
Epstein
A
Shulman
GL
Corbetta
M
Breakdown of functional connectivity in frontoparietal networks underlies behavioral deficits in spatial neglect
Neuron
 , 
2007
, vol. 
53
 (pg. 
905
-
918
)
He
ZJ
Nakayama
K
Visual attention to surfaces in three-dimensional space
Proc Natl Acad Sci USA
 , 
1995
, vol. 
92
 (pg. 
11155
-
11159
)
Heinze
HJ
Mangun
GR
Burchert
W
Hinrichs
H
Scholz
M
Münte
TF
Gös
A
Scherg
M
Johannes
S
Hundeshagen
H
Combined spatial and temporal imaging of brain activity during visual selective attention in humans
Nature
 , 
1994
, vol. 
372
 (pg. 
543
-
546
)
Hinrichs
H
Scholz
M
Tempelmann
C
Woldorff
MG
Dale
AM
Heinze
HJ
Deconvolution of event-related fMRI responses in fast-rate experimental designs: tracking amplitude variations
J Cogn Neurosci
 , 
2000
, vol. 
12
 
Suppl 2
(pg. 
76
-
89
)
Hopfinger
JB
Buonocore
MH
Mangun
GR
The neural mechanisms of top-down attentional control
Nat Neurosci
 , 
2000
, vol. 
3
 (pg. 
284
-
291
)
Hopfinger
JB
Mangun
GR
Reflexive attention modulates processing of visual stimuli in human extrastriate cortex
Psychol Sci
 , 
1998
, vol. 
9
 (pg. 
441
-
447
)
Hopfinger
JB
West
VM
Interactions between endogenous and exogenous attention on cortical visual processing
NeuroImage
 , 
2006
, vol. 
31
 (pg. 
774
-
789
)
Indovina
I
Macaluso
E
Dissociation of stimulus relevance and saliency factors during shifts of visuospatial attention
Cereb Cortex
 , 
2007
, vol. 
17
 (pg. 
1701
-
1711
)
Kanowski
M
Rieger
JW
Noesselt
T
Tempelmann
C
Hinrichs
H
Endoscopic eye tracking system for fMRI
J Neurosci Methods
 , 
2007
, vol. 
160
 (pg. 
10
-
15
)
Kastner
S
Pinsk
MA
De Weerd
P
Desimone
R
Ungerleider
LG
Increased activity in human visual cortex during directed attention in the absence of visual stimulation
Neuron
 , 
1999
, vol. 
22
 (pg. 
751
-
761
)
Katayama
K
Takahashi
N
Ogawara
K
Hattori
T
Pure topographical disorientation due to right posterior cingulate lesion
Cortex
 , 
1999
, vol. 
35
 (pg. 
279
-
282
)
Kincade
JM
Abrams
RA
Astafiev
SV
Shulman
GL
Corbetta
M
An event-related functional magnetic resonance imaging study of voluntary and stimulus-driven orienting of attention
J Neurosci
 , 
2005
, vol. 
25
 (pg. 
4593
-
4604
)
Li
CS
Huang
C
Constable
RT
Sinha
R
Imaging response inhibition in a stop-signal task: neural correlates independent of signal monitoring and post-response processing
J Neurosci
 , 
2006
, vol. 
26
 (pg. 
186
-
192
)
Linden
DE
Prvulovic
D
Formisano
E
Vollinger
M
Zanella
FE
Goebel
R
Dierks
T
The functional neuroanatomy of target detection: an fMRI study of visual and auditory oddball tasks
Cereb Cortex
 , 
1999
, vol. 
9
 (pg. 
815
-
823
)
Liu
T
Slotnick
SD
Serences
JT
Yantis
S
Cortical mechanisms of feature-based attentional control
Cereb Cortex
 , 
2003
, vol. 
13
 (pg. 
1334
-
1343
)
Macaluso
E
Frith
CD
Driver
J
Supramodal effects of covert spatial orienting triggered by visual or tactile events
J Cogn Neurosci
 , 
2002
, vol. 
14
 (pg. 
389
-
401
)
Maguire
EA
Hippocampal involvement in human topographical memory: evidence from functional imaging
Phil Trans R Soc Lond B Biol Sci
 , 
1997
, vol. 
352
 (pg. 
1475
-
1480
)
Marois
R
Leung
HC
Gore
JC
A stimulus-driven approach to object identity and location processing in the human brain
Neuron
 , 
2000
, vol. 
25
 (pg. 
717
-
728
)
Maunsell
JH
Treue
S
Feature-based attention in visual cortex
Trends Neurosci
 , 
2006
, vol. 
29
 (pg. 
317
-
322
)
McCarthy
G
Luby
M
Gore
J
Goldman-Rakic
P
Infrequent events transiently activate human prefrontal and parietal cortex as measured by functional MRI
J Neurophysiol
 , 
1997
, vol. 
77
 (pg. 
1630
-
1634
)
Menon
V
Uddin
LQ
Saliency, switching, attention and control: a network model of insula function
Brain Struct Funct
 , 
2010
, vol. 
214
 (pg. 
655
-
667
)
Mesulam
MM
A cortical network for directed attention and unilateral neglect
Ann Neurol
 , 
1981
, vol. 
10
 (pg. 
309
-
325
)
Mesulam
MM
Large-scale neurocognitive networks and distributed processing for attention, language, and memory
Ann Neurol
 , 
1990
, vol. 
28
 (pg. 
597
-
613
)
Molenberghs
P
Mesulam
MM
Peeters
R
Vandenberghe
RR
Remapping attentional priorities: differential contribution of superior parietal lobule and intraparietal sulcus
Cereb Cortex
 , 
2007
, vol. 
17
 (pg. 
2703
-
2712
)
Morris
AP
Chambers
CD
Mattingley
JB
Parietal stimulation destabilizes spatial updating across saccadic eye movements
Proc Natl Acad Sci USA
 , 
2007
, vol. 
104
 (pg. 
9069
-
9074
)
Peelen
MV
Heslenfeld
DJ
Theeuwes
J
Endogenous and exogenous attention shifts are mediated by the same large-scale neural network
NeuroImage
 , 
2004
, vol. 
22
 (pg. 
822
-
830
)
Posner
MI
Orienting of attention
Quart J Exp Psychol
 , 
1980
, vol. 
32
 (pg. 
3
-
25
)
Posner
MI
Walker
JA
Friedrich
FJ
Rafal
RD
Effects of parietal injury on covert orienting of attention
J Neurosci
 , 
1984
, vol. 
4
 (pg. 
1863
-
1874
)
Reynolds
JH
Heeger
DJ
The normalization model of attention
Neuron
 , 
2009
, vol. 
61
 (pg. 
168
-
185
)
Ridderinkhof
KR
Ullsperger
M
Crone
EA
Nieuwenhuis
S
The role of the medial frontal cortex in cognitive control
Science
 , 
2004
, vol. 
306
 (pg. 
443
-
447
)
Rule
RR
Shimamura
AP
Knight
RT
Orbitofrontal cortex and dynamic filtering of emotional stimuli
Cogn Affect Behav Neurosci
 , 
2002
, vol. 
2
 (pg. 
264
-
270
)
Rushworth
MF
Paus
T
Sipila
PK
Attention systems and the organization of the human parietal cortex
J Neurosci
 , 
2001
, vol. 
21
 (pg. 
5262
-
5271
)
Saygin
AP
Sereno
MI
Retinotopy and attention in human occipital, temporal, parietal, and frontal cortex
Cereb Cortex
 , 
2008
, vol. 
18
 (pg. 
2158
-
2168
)
Schoenfeld
MA
Hopf
JM
Martinez
A
Mai
HM
Sattler
C
Gasde
A
Heinze
HJ
Hillyard
SA
Spatio-temporal analysis of feature-based attention
Cereb Cortex
 , 
2007
, vol. 
17
 (pg. 
2468
-
2477
)
Schoenfeld
MA
Tempelmann
C
Martinez
A
Hopf
JM
Sattler
C
Heinze
HJ
Hillyard
SA
Dynamics of feature binding during object-selective attention
Proc Natl Acad Sci USA
 , 
2003
, vol. 
100
 (pg. 
11806
-
11811
)
Serences
JT
Schwarzbach
J
Courtney
SM
Golay
X
Yantis
S
Control of object-based attention in human cortex
Cereb Cortex
 , 
2004
, vol. 
14
 (pg. 
1346
-
1357
)
Serences
JT
Shomstein
S
Leber
AB
Golay
X
Egeth
HE
Yantis
S
Coordination of voluntary and stimulus-driven attentional control in human cortex
Psychol Sci
 , 
2005
, vol. 
16
 (pg. 
114
-
122
)
Sereno
MI
Pitzalis
S
Martinez
A
Mapping of contralateral space in retinotopic coordinates by a parietal cortical area in humans
Science
 , 
2001
, vol. 
294
 (pg. 
1350
-
1354
)
Sharp
DJ
Bonnelle
V
De Boissezon
X
Beckmann
CF
James
SG
Patel
MC
Mehta
MA
Distinct frontal systems for response inhibition, attentional capture, and error processing
Proc Natl Acad Sci USA
 , 
2010
, vol. 
107
 (pg. 
6106
-
6111
)
Shomstein
S
Behrmann
M
Cortical systems mediating visual attention to both objects and spatial locations
Proc Natl Acad Sci USA
 , 
2006
, vol. 
103
 (pg. 
11387
-
11392
)
Shomstein
S
Yantis
S
Control of attention shifts between vision and audition in human cortex
J Neurosci
 , 
2004
, vol. 
24
 (pg. 
10702
-
10706
)
Shulman
GL
Astafiev
SV
Franke
D
Pope
DL
Snyder
AZ
McAvoy
MP
Corbetta
M
Interaction of stimulus-driven reorienting and expectation in ventral and dorsal frontoparietal and basal ganglia-cortical networks
J Neurosci
 , 
2009
, vol. 
29
 (pg. 
4392
-
4407
)
Slagter
HA
Giesbrecht
B
Kok
A
Weissman
DH
Kenemans
JL
Woldorff
MG
Mangun
GR
fMRI evidence for both generalized and specialized components of attentional control
Brain Res
 , 
2007
, vol. 
1177
 (pg. 
90
-
102
)
Slagter
HA
Weissman
DH
Giesbrecht
B
Kenemans
JL
Mangun
GR
Kok
A
Woldorff
MG
Brain regions activated by endogenous preparatory set shifting as revealed by fMRI
Cogn Affect Behav Neurosci
 , 
2006
, vol. 
6
 (pg. 
175
-
189
)
Small
DM
Gitelman
DR
Gregory
MD
Nobre
AC
Parrish
TB
Mesulam
MM
The posterior cingulate and medial prefrontal cortex mediate the anticipatory allocation of spatial attention
NeuroImage
 , 
2003
, vol. 
18
 (pg. 
633
-
641
)
Sridharan
D
Levitin
DJ
Menon
V
A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks
Proc Natl Acad Sci USA
 , 
2008
, vol. 
105
 (pg. 
12569
-
12574
)
Sterzer
P
Kleinschmidt
A
Anterior insula activations in perceptual paradigms: often observed but barely understood
Brain Struct Funct
 , 
2010
, vol. 
214
 (pg. 
611
-
622
)
Treisman
AM
Gelade
G
A feature-integration theory of attention
Cogn Psychol
 , 
1980
, vol. 
12
 (pg. 
97
-
136
)
Uncapher
MR
Hutchinson
JB
Wagner
AD
Dissociable effects of top-down and bottom-up attention during episodic encoding
J Neurosci
 , 
2011
, vol. 
31
 (pg. 
12613
-
12628
)
Vandenberghe
R
Gitelman
DR
Parrish
TB
Mesulam
MM
Location- or feature-based targeting of peripheral attention
NeuroImage
 , 
2001
, vol. 
14
 (pg. 
37
-
47
)
Wager
TD
Jonides
J
Reading
S
Neuroimaging studies of shifting attention: a meta-analysis
NeuroImage
 , 
2004
, vol. 
22
 (pg. 
1679
-
1693
)
Yantis
S
Schwarzbach
J
Serences
JT
Carlson
RL
Steinmetz
MA
Pekar
JJ
Courtney
SM
Transient neural activity in human parietal cortex during spatial attention shifts
Nat Neurosci
 , 
2002
, vol. 
5
 (pg. 
995
-
1002
)
Yantis
S
Serences
JT
Cortical mechanisms of space-based and object-based attentional control
Curr Opin Neurobiol
 , 
2003
, vol. 
13
 (pg. 
187
-
193
)