Abstract

There is an ongoing debate whether visual object representations can be formed outside the focus of voluntary attention. Recently, implicit behavioral measures suggested that grouping processes can occur for task-irrelevant visual stimuli, thus supporting theories of preattentive object formation (e.g., Lamy, D., Segal, H., & Ruderman, L. Grouping does not require attention. Perception and Psychophysics, 68, 17–31, 2006; Russell, C., & Driver, J. New indirect measures of “inattentive” visual grouping in a change-detection task. Perception and Psychophysics, 67, 606–623, 2005). We developed an ERP paradigm that allows testing for visual grouping when neither the objects nor its constituents are related to the participant's task. Our paradigm is based on the visual mismatch negativity ERP component, which is elicited by stimuli deviating from a regular stimulus sequence even when the stimuli are ignored. Our stimuli consisted of four pairs of colored discs that served as objects. These objects were presented isochronously while participants were engaged in a task related to the continuously presented fixation cross. Occasionally, two color deviances occurred simultaneously either within the same object or across two different objects. We found significant ERP differences for same- versus different-object deviances, supporting the notion that forming visual object representations by grouping can occur outside the focus of voluntary attention. Also our behavioral experiment, in which participants responded to color deviances—thus, this time the discs but, again, not the objects were task relevant—showed that the object status matters. Our results stress the importance of early grouping processes for structuring the perceptual world.

INTRODUCTION

Visual Object Formation—The Attention Issue

Our visual system has to integrate the mixture of information emanating from the environment into coherent perceptual units or object representations. There is a long-standing debate whether the integration process is confined to attended stimuli or whether it also occurs for unattended stimuli. According to Treisman's feature integration theory (see e.g., Treisman & Schmidt, 1982; Treisman & Gelade, 1980), the binding of different features (e.g., color, shape, orientation) belonging to a given object requires that attention is focused on the respective stimulus location. In contrast, object-based theories of attention suggest that attentional processes mainly operate on preattentively formed perceptual units or object representations (e.g., Driver, Davis, Russell, Turatto, & Freeman, 2001; Scholl, 2001; Duncan, 1984; for detailed discussions of both approaches, see, e.g., Yantis & Serences, 2003; Egeth & Yantis, 1997). Using an event-related brain potential method, we designed the current study to test whether focused attention is a necessary prerequisite of the formation of visual object representations.

Empirical evidence for object-based theories of attention mostly comes from studies in which participants judge different stimulus features of the same versus different objects. For example, in Duncan's (1984) experiment, participants looked at two superimposed objects (a tilted line and a rectangle with a small gap in its contour) and had to make judgments regarding two different stimulus features. Participants responded more accurately and/or faster when the task-relevant features belonged to the same object (e.g., the size of the rectangle and the position of the gap in the contour of the rectangle) compared with when the target features belonged to two different objects (e.g., the size of the rectangle and the style of the line). This so-called same-object advantage has been observed in numerous behavioral studies using a large variety of stimulus material (e.g., Palmer & Beck, 2007; Beck & Palmer, 2002; Awh, Dhaliwal, Christensen, & Matsukura, 2001; Holcombe & Cavanagh, 2001; Valdes-Sosa, Cobo, & Pinilla, 2000; Watson & Kramer, 1999; Chen, 1998; Lavie & Driver, 1996). Electrophysiological (Rodriguez & Valdes-Sosa, 2006; Khoe, Mitchell, Reynolds, & Hillyard, 2005; Lopez, Rodriguez, & Valdes-Sosa, 2004; Pinilla, Cobo, Torres, & Valdes-Sosa, 2001; Valdes-Sosa, Bobes, Rodriguez, & Pinilla, 1998) and fMRI evidence (O'Craven, Downing, & Kanwisher, 1999) provided further support for the notion of object-based attentional modulations.

A common feature of these studies was that the perceptual objects tested, or at least parts of them, were task relevant and thus had to be attended due to instruction. Because of this, the same-object advantage has been seen as an indicator of object-based modulations of attention rather than as an indicator of preattentive object formation. Formation of object representations outside the focus of voluntary attention can only be investigated by studying same- versus different-object effects for task-irrelevant objects. There are some studies indicating grouping (or object formation) outside the focus of attention by using implicit behavioral measures. In these experiments, participants were instructed to judge the relative length of two target lines, which were presented on a task-irrelevant background consisting of a matrix of dots painted with two different colors. Dots could be grouped based on color similarity into arrowheads, which flanked the task-relevant lines in such a way that if grouping occurred, they would induce the Müller–Lyer illusion. The participants' judgments of the line lengths varied in accordance with the illusion brought about by the organization of the background dot patterns although the background was task irrelevant, and participants were unable to describe the background pattern after the trials (Lamy, Segal, & Ruderman, 2006; Moore & Egeth, 1997). Furthermore, Driver et al. (2001) showed that the accuracy in a change detection task performed on matrices presented successively at the center of the display was influenced by the organization of the task-irrelevant and unattended background (Russell & Driver, 2005). For example, participants performed the task more accurately, when both the target matrix and the organization of the task-irrelevant background pattern remained constant or when they changed at the same time. However, even with these types of implicit behavioral measures cited above, the task-irrelevant background forming the objects was in some respect related to the task-relevant stimulus. According to some theories, for example, the theory of event coding (Hommel, Müsseler, Aschersleben, & Prinz, 2001), such relations can possibly be exploited by the system. That is, the object status is not completely irrelevant for solving the task. To avoid this trap, we used an ERP approach, in which the object status is not related to the task.

An ERP Measure for Investigating Object Formation outside the Focus of Attention

ERP components accompanying the detection of deviance can be used to study object formation because they may show modulation according to how the stimuli are grouped (e.g., Atienza et al., 2003; Winkler et al., 2003; Yabe et al., 2001; Sussman, Ritter, & Vaughan, 1999). The visual mismatch negativity (vMMN; for reviews see, Czigler, 2007; Pazo-Alvarez, Cadaveira, & Amenedo, 2003) is elicited by task-irrelevant to be ignored stimuli deviating from a stimulus sequence containing some regularity defined by stimulus features, such as color, spatial frequency, shape, luminance, size or movement direction, and speed. vMMN responses can be elicited irrespective of whether the deviant stimuli are ignored or attended (Winkler, Czigler, Sussman, Horvath, & Balazs, 2005), and the vMMN amplitude is unaffected by variations of the difficulty of the primary task (Pazo-Alvarez, Amenedo, & Cadaveira, 2004; Heslenfeld, 2003). Winkler et al. (2005) used the vMMN as an indicator of preattentive binding of visual features. They demonstrated that deviant background patterns defined by an irregular conjunction of two features (color and orientation) elicited vMMN irrespective of whether participants attended the test patterns or performed a demanding detection task of changes in the central fixation cross. These results support the view that the integration of different visual features may occur even for stimuli outside the focus of attention.

The Present Study

In the present study, we used the vMMN measure to investigate the possibility of preattentive formation of visual object representations. Visual objects were created by connecting two colored discs with a bracket line. Four of these objects were evenly placed around the fixation cross. Similarly to previous studies (e.g., Winkler et al., 2005), participants performed a detection task, continuously monitoring the fixation cross. All other stimuli (including the objects) were task irrelevant and (unlike in the reported behavioral studies above) also not related to the task. In the frequently presented standard displays, all of the discs had the same color. Occasionally, two adjacent discs took on a deviant color. This could occur either on the same object or on two different objects. Spatial relations between the two color irregularities were identical in the same- and different-object deviation case. Based on previous results, we expected that violations of the color regularity would be detected outside the focus of voluntary attention, eliciting the vMMN: We expected that if vMMN worked on preattentively formed object representations, vMMNs elicited by deviations occurring on the same object versus on two different objects would differ from each other. This is because no other feature of the current design distinguished between the pairs of deviations occurring on same versus on different objects. We had no a priori expectation regarding the nature of the difference between the two cases. On the one hand, two deviations cooccurring on the same object might be more salient than one of the two deviations alone, which, according to previous results, produces higher vMMN amplitudes (Czigler, Balazs, & Winkler, 2002) and/or shorter vMMN latencies (Maekawa et al., 2005). On the other hand, two concurrent although lower amplitude vMMNs elicited by deviations occurring on different objects may sum to produce a higher overall response than a single higher amplitude vMMN.

In a second experiment, we tested the effects of same- versus different-object deviances on participants' performance in a task in which deviances were task relevant but objects were again task irrelevant. We expected to find a same-object advantage; that is, same-object deviances would be processed more efficiently than different-object deviances.

EXPERIMENT 1

Methods

Participants

Fifteen healthy volunteers (1 man; aged 20–46 years, mean age = 28.7 years) participated in the experiment for either course credit or payment. Written informed consent was obtained from the participants after the experimental procedure was explained to them. All of them reported normal or corrected-to-normal vision.

Stimuli

Stimuli were created and presented using Cogent2000v1.25. They appeared on a light gray background of a 19-in. CRT monitor (Iiyama HM 903DT, refreshing rate 100 Hz; Tokyo, Japan) placed 1 m in front of the participant. Eight colored discs (0.92° radius each) surrounded by a thin black outline (0.11° width) were arranged at equidistant locations on a circle surrounding the central fixation cross (Figure 1). The distance between the center of the display and the center of each colored disc was 3.84°. Each disc was connected to exactly one of its neighbors either by an elliptic or by a rectangular black bracket (0.34° line width, 5.61° maximal distance from the fixation cross), thus creating four pairs of discs, the test objects (Figure 1). In different displays, the position of the brackets varied in such a way that the two discs forming an object fell either within the same visual quadrant (noncardinal configuration) or across adjacent visual quadrants (cardinal configuration). Noncardinal and cardinal object configurations occurred equiprobably within each block. To prevent the formation of larger scale objects, each display included two elliptic and two rectangular bracket objects whose arrangement varied randomly.

Figure 1. 

Schematic display sequence presented in Experiments 1 and 2. In Experiment 1, participants were instructed to detect infrequent increases of the size of the fixation cross, whereas the objects formed by the small discs and the brackets were task irrelevant. Deviant-colored discs were related either to the same object (A and B) or to different objects (C and D). In Experiment 2, participants were instructed to indicate whether deviant-colored discs were located in the same (A and C) or in different visual quadrants (B and D); that is, color changes were task-relevant but objects remained task-irrelevant. We give examples of deviance displays in the lower panel.

Figure 1. 

Schematic display sequence presented in Experiments 1 and 2. In Experiment 1, participants were instructed to detect infrequent increases of the size of the fixation cross, whereas the objects formed by the small discs and the brackets were task irrelevant. Deviant-colored discs were related either to the same object (A and B) or to different objects (C and D). In Experiment 2, participants were instructed to indicate whether deviant-colored discs were located in the same (A and C) or in different visual quadrants (B and D); that is, color changes were task-relevant but objects remained task-irrelevant. We give examples of deviance displays in the lower panel.

In the frequent standard displays (p = 90%), all discs had the same color (green or red, in different blocks). In the infrequent deviant displays (p = 10%), two adjacent discs differed from the color of the other six discs (red when the standard color was green and green when the standard color was red). The two color-deviant discs belonged either to the same object or to adjacent different objects, with both cases occurring equiprobably. As a consequence of this arrangement, the color-deviant discs occurring on the same object versus on different objects were separated by the same spatial distance. Noncardinal and cardinal object configurations occurred equiprobably for both types of deviant displays within each block (see also Figure 1); that is, possible differences between the processing of same- and different-object deviances cannot be attributed to different object configurations. The pairs of color-deviant discs were presented equiprobably at each of the eight possible positions on the display.

Each test-stimulus display was shown for 120 msec and was followed by an interstimulus interval of 600 msec. Standard and deviant displays were presented in randomized order with the restriction that deviant displays were always followed by at least one standard display. A black fixation cross (0.69° × 0.69°) within a white disc (0.95° radius) was constantly displayed at the center of the monitor throughout the experimental blocks.

Procedure

Participants sat in a dimly lit, electrically shielded, and sound-attenuated chamber. They were instructed to keep their eyes on the fixation cross and to respond as quickly as possible to small extensions of the fixation cross (from 0.69° to 0.74°), which were displayed for 120 msec duration. These extended fixation crosses occurred unpredictably with a mean frequency of 3.125 per minute (i.e., 12 times per block). Responses had to be given by pressing the left button of a computer mouse with the right index finger. Participants were instructed to ignore the peripheral stimuli, which were introduced as task irrelevant.

Stimuli were delivered in 12 blocks of 320 trials each. In half of the blocks, the standard color of discs was red; in the other half, it was green. Blocks with red or green standard color were presented in randomized order.

EEG Recordings and Data Analysis

The EEG was recorded continuously with an ActiveTwo amplifier system (BioSemi, Amsterdam, The Netherlands), with 128 active electrodes from standard locations according to the ABC electrode system (http://www.biosemi.com/headcap.htm). Horizontal and vertical eye movements were monitored by electrodes placed at the outer canthi of both eyes and above and below the left eye, respectively. EEG and EOG activities were sampled at 512 Hz.

Off-line, EEG activity was re-referenced to the activity recorded from an electrode placed at the tip of the nose, and EEG and EOG activity was filtered (1–35 Hz band-pass digital finite impulse response filter with a length of 1025 points). EEG and EOG activities were averaged for epochs beginning 100 msec before and extending until 600 msec after the onset of the test displays. The first 100 msec of each epoch served as the baseline interval. Epochs containing amplitude changes exceeding 100 μV at any EEG or EEG channel were excluded from analysis. Epochs during which a target (i.e., an extended fixation cross) occurred or the participant made a button press were also excluded from analysis, as were epochs of standard displays immediately following a deviant display.

Epochs were averaged separately for the standard and for the two types of deviant displays (factor Object; same vs. different). On average, 148 ± 19 epochs for same-object deviances and 149 ± 19 epochs for different-object deviances were available for each participant. To analyze responses to deviances, we calculated difference waves by subtracting ERPs elicited by the standard displays from those elicited by the different deviant displays. By visual inspection, we selected three ROIs (for their location, see the top panel of Figure 2), where the deviant-minus-standard difference exhibited prominent negative deflections. To improve the signal-to-noise ratio, we averaged EEG signals over six neighboring electrode leads at both hemispheres, separately for each ROI (termed ROI responses). Statistical analyses were conducted for individual mean ROI response amplitudes measured from the 30-msec time windows centered on the peak latencies of the two prominent negative deflections in the grand average ROI difference potentials. We tested the elicitation of the differential responses with one-sample, one-tailed Student's t tests comparing the ROI difference amplitudes against zero. Responses were compared across different types of deviances by repeated measures ANOVA with the factors of Object (same vs. different) × Time Window (205–235 vs. 250–280 msec) × ROI (fronto-central [fc] vs. posterio-temporal [pt] vs. occipital [o]). Follow-up analyses comparing vMMN amplitudes elicited by same- and different-object deviances were carried out by one-sample, two-tailed Student's t tests. The alpha level criterion for all statistical analyses was set to .05. Greenhouse–Geisser corrections of the degrees of freedom were applied when appropriate to correct for violations of the assumption of sphericity.

Figure 2. 

Top: Electrode locations signals were averaged over for fronto-central (left column), posterio-temporal (middle column), and occipital (right column) ROIs. Middle: Group-averaged (N = 15) ERPs elicited by standards (dotted lines), same-object deviances (dashed lines), and different-object deviances (solid lines). Bottom: Deviant-minus-standard difference waveforms are displayed below the corresponding ERPs. Gray shading indicates the time windows used to determine the mean amplitudes for the deviance-related negativities. For latency comparisons, the peak latencies of the P2 and N2 components are shown by thin dotted vertical lines.

Figure 2. 

Top: Electrode locations signals were averaged over for fronto-central (left column), posterio-temporal (middle column), and occipital (right column) ROIs. Middle: Group-averaged (N = 15) ERPs elicited by standards (dotted lines), same-object deviances (dashed lines), and different-object deviances (solid lines). Bottom: Deviant-minus-standard difference waveforms are displayed below the corresponding ERPs. Gray shading indicates the time windows used to determine the mean amplitudes for the deviance-related negativities. For latency comparisons, the peak latencies of the P2 and N2 components are shown by thin dotted vertical lines.

We applied a scalp current density (SCD) analysis to display the distributions of the deviance-related negativities. SCDs were calculated by computing the second spatial derivative of the surface scalp potentials, which were interpolated by the spherical spline method (Perrin, Pernier, Bertrand, & Echallier, 1989). This method highlights localized cortical sources of EEG activity while attenuating the contribution of deep and distributed ones (Srinivasan, 2005, p. 168). The maximum degree of the Legendre polynomials was set to be 50 and the order of splines to be 4. The calculation and the plotting of SCDs were carried out using the sphspline plug-in for EEGlab (Delorme & Makeig, 2004) written by Andreas Widmann, University of Leipzig (2006).

Results

Performance in the Continuous Detection Task

On average, participants detected 90.4% of the extensions of the fixation cross (SD = 8.0%). The false alarm rate was lower than 0.1%. The mean RT was 465 msec (SD = 39 msec).

Event-related Potentials

Figure 2 displays the grand average ERPs, superimposing responses elicited by standard, same-object, and different-object deviant stimuli, together with the respective deviant-standard difference waveforms. Signals were collapsed separately for the three ROIs: fronto-central (fc), posterio-temporal (pt), and occipital (o). In the latency range of the P1 (ca. 90 msec) and the N1 peak (ca. 165 msec), ERPs for both standard and deviant stimuli show nearly identical prominent deflections at posterior ROIs. In contrast, in the P2–N2 latency range (peaks at 230 and 285 msec, respectively), deviant ERPs generally show a more negative response than standards. The difference waveforms show that this deviance-related negativity exhibits two peaks at 220 and 265 msec, respectively (for amplitude measurements, see Table 1). Difference waveforms for deviances related to same and different objects are rather similar at the earlier (220 msec) peak. In contrast, at the later (265 msec) peak, same-object deviances elicited an enhanced negativity compared with different-object deviances (Table 1).

Table 1. 

Mean Deviant-Minus-Standard Difference Amplitudes (μV) Elicited by Deviances Occurring on the Same Object versus on Different Objects

Windows (msec)
ROI
Deviances Related to The Same Object
Deviances Related to Different Objects
Amplitude (μV)
T(df = 14)
p
Amplitude (μV)
T(df = 14)
p
Early 
205–235 o −0.86 (0.35) 2.48 .014 −1.17 (0.39) 3.05 .005 
205–235 pt −0.68 (0.26) 2.59 .011 −0.91 (0.27) 3.35 .003 
253–283 fc −0.63 (0.34) 1.88 .041 −0.56 (0.21) 2.72 .009 
 
Late 
250–280 o −1.42 (0.38) 3.72 .001 −0.74 (0.32) 2.34 .018 
250–280 pt −1.26 (0.37) 3.45 .002 −0.77 (0.29) 2.26 .009 
302–332 fc −0.66 (0.17) 3.35 .001 −0.30 (0.25) 1.19 >.1 
Windows (msec)
ROI
Deviances Related to The Same Object
Deviances Related to Different Objects
Amplitude (μV)
T(df = 14)
p
Amplitude (μV)
T(df = 14)
p
Early 
205–235 o −0.86 (0.35) 2.48 .014 −1.17 (0.39) 3.05 .005 
205–235 pt −0.68 (0.26) 2.59 .011 −0.91 (0.27) 3.35 .003 
253–283 fc −0.63 (0.34) 1.88 .041 −0.56 (0.21) 2.72 .009 
 
Late 
250–280 o −1.42 (0.38) 3.72 .001 −0.74 (0.32) 2.34 .018 
250–280 pt −1.26 (0.37) 3.45 .002 −0.77 (0.29) 2.26 .009 
302–332 fc −0.66 (0.17) 3.35 .001 −0.30 (0.25) 1.19 >.1 

Responses are displayed separately for the early and the late time window and for three ROIs: occipital (o) ROI, posterio-temporal (pt) ROI, and fronto-central (fc) ROI. SEM are given in parentheses. Bold numbers highlight mean amplitudes differing significantly (p < .05) from zero. Student's t tests and corresponding one-tailed p values are given in italics.

We analyzed response amplitudes by a repeated measures ANOVA including the factors Object (same vs. different) × Time Window (205–235 vs. 250–280 msec) × ROI (fronto-central vs. posterio-temporal vs. occipital). The ANOVA showed that amplitudes of deviance-related negativities were higher at posterior as compared with more anterior ROIs, main effect ROI, F(2,28) = 8.24, p < .006, ɛ = .66, η2 = .37, with a significant contrast between ROIs (o) and (pt) against ROI (fc), F(1,14) = 10.21, p < .006, η2 = .42. There were no main effects of the factors Object and Time Window (both F < 1.1, p > .3, η2 < .07). However, a significant interaction of the Object and the Time Window factors, F(1,14) = 9.24, p < .009, η2 = .4, reflected differences in the amplitude ratios between early and late negativities elicited by same- versus different-object deviances. In the 205- to 235-msec time window, deviance-related negativities of similar amplitudes were observed for same- and different-object deviances, t(1,14) = 0.68, p > .5. In contrast, in the 250- to 280-msec time window, the amplitude of the deviance-related negativity was higher for same-object relative to different-object deviances, t(1,14) = −2.92, p < .02. No other interactions reached significance.

SCD maps depicted in Figure 3 reveal an occipital/occipito-temporal distribution of deviance-related negativities elicited by both same- and different-object deviances. The distribution is characterized by bilateral broad occipital/occipito-temporal sinks and a source in the central–parietal region.

Figure 3. 

Topographic distribution (SCD) of the deviance-related negativities elicited by deviances occurring on different objects (left column) or on the same object (middle column). SCDs of the contrast between the two conditions are shown in the right column. SCDs are displayed separately for deviance-related negativities elicited in the 205- to 235-msec (upper panel) and the 250- to 280-msec time window (lower panel). Distances between isocontour lines are 0.05 mA/m3. A smoothing parameter of lambda = 10−5 was applied.

Figure 3. 

Topographic distribution (SCD) of the deviance-related negativities elicited by deviances occurring on different objects (left column) or on the same object (middle column). SCDs of the contrast between the two conditions are shown in the right column. SCDs are displayed separately for deviance-related negativities elicited in the 205- to 235-msec (upper panel) and the 250- to 280-msec time window (lower panel). Distances between isocontour lines are 0.05 mA/m3. A smoothing parameter of lambda = 10−5 was applied.

EXPERIMENT 2

In Experiment 2, using the same stimuli as in Experiment 1, we tested whether target features appearing on the same versus on different objects affect judgments about the spatial location of these features. We instructed participants to indicate whether the two simultaneously delivered color deviances occurred within the same visual quadrant or across two quadrants. Thus, although participants were (unlike in Experiment 1) required to attend to parts of the objects (i.e., the discs), the objects themselves were task irrelevant (just as in Experiment 1). Note that the spatial distance between the two color-deviant discs was identical, irrespective of whether they appeared within the same quadrant or across different quadrants as well as whether they were carried by the same object or by two different objects.

Methods

Participants

Fourteen right-handed healthy volunteers (6 men; aged 20–32 years, mean age = 24.3 years) participated in the experiment for either course credits or payment. Participants were naive to the purpose of the experiment, and none of them had participated in Experiment 1. Written informed consent was obtained from the participants after the experimental procedure was explained to them. All of them reported normal or corrected-to-normal vision. Due to a hit rate of more than two standard deviations below the mean hit rate of all the other participants, one additional participant's data was excluded from the analysis.

Stimuli and Procedure

We presented the same stimulus material as in Experiment 1. However, to shorten the experimental session, deviant displays occurred with a probability of 20% instead of 10%. As in Experiment 1, the pairs of small discs belonging to the same object were either located within the same or across different visual quadrants, and the two discs taking on the deviant color belonged either to the same or to two different objects. The combination of these two factors resulted in four types of deviant displays (same/different object × same/different quadrant), each type occurring equiprobably.

In Experiment 2, participants were instructed to indicate whether the two simultaneously presented color-deviant discs appeared within the same or across two different visual quadrants.1 Thus, the location of the deviances was task relevant, whereas the relation between deviances and objects was task irrelevant. This is in contrast to Experiment 1, in which participants were instructed to ignore the discs. Responses were to be given by pressing the left/right mouse button with the left/right index finger. For each participant, the assignment of the two response buttons to the two types of targets (deviances located within the same vs. across two different visual quadrants) was reversed after half of the stimulus blocks. At the beginning of each stimulus block, the participant was prompted with examples of the two types of target displays and the current response-button assignment. In these examples, discs with the target color were presented without the connecting brackets. Unlike in Experiment 1, the size of the fixation cross remained constant throughout the stimulus blocks. Stimuli were delivered in eight blocks of 160 trials each.

Data Analysis

Responses were accepted within the range of 200–1200 msec from the onset of the deviant displays. We conducted a two-way repeated measures ANOVA with the factors Quadrant (same vs. different quadrants) × Object (same vs. different objects) for reactions times and hit rates, separately.

Results

RTs and hit rates (mean ± SEM) are summarized in Table 2.

Table 2. 

Mean RTs and Hit Rates in Experiment 2


Deviances Occurring in The Same Quadrant
Deviances Occurring in Different Quadrants
RT (msec)
Hit Rate (%)
RT (msec)
Hit Rate (%)
Deviances related to the same object 631 (18) 94.6 (1.5) 664 (21) 92.7 (1.1) 
Deviances related to different objects 643 (19) 94.0 (0.9) 666 (19) 93.4 (1.2) 

Deviances Occurring in The Same Quadrant
Deviances Occurring in Different Quadrants
RT (msec)
Hit Rate (%)
RT (msec)
Hit Rate (%)
Deviances related to the same object 631 (18) 94.6 (1.5) 664 (21) 92.7 (1.1) 
Deviances related to different objects 643 (19) 94.0 (0.9) 666 (19) 93.4 (1.2) 

Participants were instructed to respond according to whether the color-deviant discs related to either the same object or to different objects appeared in the same versus different visual quadrants. SEM are given in parentheses.

Participants responded significantly faster to color-deviant pairs of discs located in the same as compared with different quadrants, F(1,13) = 19.74, p < .001, η2 = .60. More important regarding the main purpose of this experiment, participants responded faster to color-deviant pairs of discs appearing on the same as compared with different objects, F(1,13) = 5.29, p < .039, η2 = .29. There was no significant interaction between the two factors Quadrant and Object, F(1,13) < 1, p > .5, η2 < .05.

On average, participants responded correctly in 93.7 ± 0.9% of all target trials. Hit rates were not significantly affected by either factor, F(1,13) = 2.72, p < .12, η2 = .17 and F(1,13) < .1, p > .9, η2 < .1, for the Quadrant and Object factors, respectively. There was no significant interaction between the factors, F(1,13) < 0.5, p > .5, η2 < .05.

Discussion

In the present study, we analyzed ERP effects of object formation when objects were ignored (Experiment 1) and behavioral effects of object formation when parts of the objects were task relevant and therefore attended, but objects per se were task irrelevant (Experiment 2). Thus, we investigated whether voluntary focused attention is a necessary prerequisite of the formation of visual object representations.

In Experiment 1, we compared the ERP responses elicited by simultaneously presented color-deviant pairs of discs when they appeared on the same object with when they appeared on two different objects. No part of the test objects was task relevant because participants performed a detection task that required focusing their attention continuously on the fixation cross at the center of the visual display, away from the task-irrelevant stimuli presented peripherically.

Results showed that same- and different-object deviances elicited two significant negative waveforms in the time range of 200–300 msec (peaking at 220 and 265 msec from stimulus onset). The amplitude of the earlier peak was of similar size whether deviations occurred on the same object or on two different objects. In contrast, the later peak had higher amplitude when the two deviations appeared on the same object as compared with different objects, although the deviant discs were separated by the same spatial distance in the two cases. Thus, the configuration of objects affected deviance detection for task-irrelevant stimuli. Therefore, we conclude that the mechanisms responsible for the detection of task-irrelevant regularity violations operate based on object representations generated outside the focus of voluntary attention. This conclusion is further supported by the finding of a same-object advantage (shorter RTs) for spatial judgments based on the same deviations, when the object configuration was task irrelevant (Experiment 2). On the one hand, our results are in line with studies using implicit behavioral measures to confirm that elements of a display can be grouped outside the focus of attention (Lamy et al., 2006; Russell & Driver, 2005). Furthermore, our results corroborate recent behavioral (Mordkoff & Halterman, 2008) and electrophysiological studies (Winkler et al., 2005) showing that visual features (i.e., color/shape and color/orientation, respectively) can be integrated without directing attention toward the respective stimuli. On the other hand, our results (as well as the studies cited above) appear to contradict traditional theories of visual perception according to which focal attention is a necessary prerequisite for feature binding and object formation (Wolfe, 1994; Treisman & Gelade, 1980). This seeming contradiction might be due to specifics of the respective paradigms: In studies reporting object formation or feature integration outside the focus of attention, the corresponding stimuli were task irrelevant; that is, there was no need to map the relevant stimuli to a specific response. In contrast, in experiments supporting the notion that attention is required to form objects or feature conjunctions, stimuli were task relevant. More specifically, in these experiments, predefined objects or feature conjunctions must be related to an arbitrary, instruction-based response; that is, they must undergo a stimulus–response (S-R) mapping. Thus, one can speculate that S-R translation is dependent on attention, but the formation of visual feature conjunctions itself is not. Indeed, recently, Mordkoff and Halterman (2008) found strong evidence for this line of argumentation by showing that unattended stimuli comprising specific feature conjunctions that were implicitly associated with certain S-R translations (so-called correlated flankers) influenced the processing of central target stimuli.

Our participants were engaged in a continuous detection task during the whole experiment, and they reached a hit rate of more than 90% on average. Furthermore, task-relevant and task-irrelevant stimuli appeared at different positions of the display, and the presentation of targets and task-irrelevant stimuli was temporally uncorrelated. Thus, we minimized the possibility that attention leakage contributed to the processing of the task-irrelevant stimuli (Lachter, Forster, & Ruthruff, 2004). Nevertheless, one could argue that the salient color deviances could attract attention reflexively (e.g., Hopfinger & Ries, 2005; Theeuwes, 2004; Hopfinger & Mangun, 2001; but see also, Eimer & Kiss, 2008; Folk, Remington, & Johnston, 1992). Thus, it would be interesting to know whether an equivalent same–different-object distinction could also be made by the system when no concurrent color deviance comes into play. In other words, our experimental design cannot rule out the object-specific processing is contingent on the processing of a physical (here, color) deviance. However, as the object status, the color and even the stimuli themselves and the location they were presented at were task irrelevant (and as we did not find any ERP indication [e.g., N2b, P3a] expected to be associated with an attentive processing of deviant stimuli); the establishment of the differential processing of same- and different-object deviances must have occurred outside the focus of voluntary attention.

We did not observe latency differences between the vMMNs elicited by same- versus different-object deviances. However, previous studies reported that the vMMN peak latency decreases with increasing salience of the deviation (Maekawa et al., 2005). Therefore, the differences we found between processing deviances within one object or across two different objects cannot be conceptualized as salience effects.

It is possible that the same-object advantage reported in behavioral studies is based on more efficient preattentive processing of same-object as compared with different-object deviances. In Experiment 2, we showed a behavioral same-object advantage for processing the location of deviances appearing on the same object as compared with two different objects, although the task did not require the processing of object information. This same-object advantage could be associated with the higher late vMMN amplitudes found for same-object as compared with different-object deviances. However, we found no scalp distribution differences related to whether the deviations occurred either on the same versus on different objects or in the different time windows. Therefore, assuming different underlying processes might be plausible but still speculative. It remains to be investigated whether the two peaks of deviance-related negativities actually reflect different processes. For this purpose, further studies should compare the processing of simple feature-regularity violations and more complex object-related regularity violations to disentangle their contributions to visual deviance detection.

It should be noted that the SCD maps showed an occipital/occipito-temporal distribution of our deviance-related negativities. This is partly compatible with the results from a recent fMRI study that localized neural activity associated with the processing of task-irrelevant color- and orientation-deviant grating patterns in occipital-fusiform, posterior parietal, and, in addition, prefrontal regions (Yucel, McCarthy, & Belger, 2007). However, our results do not provide support for a prefrontal deviance-related activation, which is possibly due to the small size of our stimuli that resulted in relatively small deviance-related ERP effects.

Conclusion

In summary, differences in the vMMN amplitudes elicited by two simultaneously appearing deviations as a function of whether they appeared on the same or on two different task-irrelevant objects provide evidence for the formation of visual object representations outside the focus of voluntary attention. Importantly, in our new paradigm, neither the objects nor its constituents were related to the participant's task. Previous behavioral studies cannot exclude the possible contribution of such relations for object formation. Same-object deviances elicited higher late vMMN amplitudes than different-object deviances, which may be related to the same-object advantage found in behavioral tests. On a more general level, the establishment of visual object representations outside the focus of voluntary attention can be related to the auditory domain, for which object formation has been shown to occur at preattentive stages of information processing (e.g., Sussman, Horvath, Winkler, & Orr, 2007; Ritter, De Sanctis, Molholm, Javitt, & Foxe, 2006; Sussman, 2005). The attempt to extract both visual and auditory objects rather early (from a temporal and structural perspective) stresses the importance of organizing our perceptual world into objects.

Acknowledgments

This work was supported by the German Research Foundation (No. Schr 375/16), the Hungarian National Research Fund (OTKA T048383 and 71600), and the DAAD-MÖB grant (2006-2007/12). The experiment was realized using Cogent 2000 developed by the Cogent 2000 team at the FIL and the ICN. The authors thank Sabine Grimm and Andreas Widmann for numerous fruitful discussions, Alexandra Bendixen and Annekathrin Weise for proofreading the manuscript, and Nicole Koburger for assistance in data acquisition.

Reprint requests should be sent to Dagmar Müller, Institut für Psychologie I, Universität Leipzig, Seeburgstraße 14-20, D-04103 Leipzig, Germany, or via e-mail: dagmar_mueller@uni-leipzig.de.

Note

1. 

There were two additional experimental conditions, one presented before and the other after the one described above. (1) In the nonobject condition, only the eight small discs were presented on each display and participants were instructed to perform the quadrant task as described above. (2) In the object-task condition, stimulation was identical to the one described in the text and participants were instructed to indicate whether the two simultaneously presented deviances appeared on the same or across two different objects. Because these conditions did not provide results relevant for the aims of the current study, for conciseness, here we focus on results obtained in the condition described in the main text.

REFERENCES

REFERENCES
Atienza
,
M.
,
Cantero
,
J. L.
,
Grau
,
C.
,
Gomez
,
C.
,
Dominguez-Marin
,
E.
, &
Escera
,
C.
(
2003
).
Effects of temporal encoding on auditory object formation: A mismatch negativity study.
Cognitive Brain Research
,
16
,
359
371
.
Awh
,
E.
,
Dhaliwal
,
H.
,
Christensen
,
S.
, &
Matsukura
,
M.
(
2001
).
Evidence for two components of object-based selection.
Psychological Science
,
12
,
329
334
.
Beck
,
D. M.
, &
Palmer
,
S. E.
(
2002
).
Top–down influences on perceptual grouping.
Journal of Experimental Psychology: Human Perception and Performance
,
28
,
1071
1084
.
Chen
,
Z.
(
1998
).
Switching attention within and between objects: The role of subjective organization.
Canadian Journal of Experimental Psychology
,
52
,
7
16
.
Czigler
,
I.
(
2007
).
Visual mismatch negativity: Violation of nonattented environmental regularities.
Journal of Psychophysiology
,
21
,
224
230
.
Czigler
,
I.
,
Balazs
,
L.
, &
Winkler
,
I.
(
2002
).
Memory-based detection of task-irrelevant visual changes.
Psychophysiology
,
39
,
869
873
.
Delorme
,
A.
, &
Makeig
,
S.
(
2004
).
EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis.
Journal of Neuroscience Methods
,
134
,
9
21
.
Driver
,
J.
,
Davis
,
G.
,
Russell
,
C.
,
Turatto
,
M.
, &
Freeman
,
E.
(
2001
).
Segmentation, attention and phenomenal visual objects.
Cognition
,
80
,
61
95
.
Duncan
,
J.
(
1984
).
Selective attention and the organization of visual information.
Journal of Experimental Psychology: General
,
113
,
501
517
.
Egeth
,
H. E.
, &
Yantis
,
S.
(
1997
).
Visual attention: Control, representation, and time course.
Annual Review of Psychology
,
48
,
269
297
.
Eimer
,
M.
, &
Kiss
,
M.
(
2008
).
Involuntary attentional capture is determined by task set: Evidence from event-related brain potentials.
Journal of Cognitive Neuroscience
,
20
,
1423
1433
.
Folk
,
C. L.
,
Remington
,
R. W.
, &
Johnston
,
J. C.
(
1992
).
Involuntary covert orienting is contingent on attentional control settings.
Journal of Experimental Psychology: Human Perception and Performance
,
18
,
1030
1044
.
Heslenfeld
,
D. J.
(
2003
).
Visual mismatch negativity.
In J. Polich (Ed.),
Detection of change: Event-related potential and fMRI findings
(pp.
41
59
).
Boston
:
Kluver Academic Publishers
.
Holcombe
,
A. O.
, &
Cavanagh
,
P.
(
2001
).
Early binding of feature pairs for visual perception.
Nature Neuroscience
,
4
,
127
128
.
Hommel
,
B.
,
Müsseler
,
J.
,
Aschersleben
,
G.
, &
Prinz
,
W.
(
2001
).
The Theory of Event Coding (TEC): A framework for perception and action planning.
Behavioral and Brain Sciences
,
24
,
849
878
.
Hopfinger
,
J. B.
, &
Mangun
,
G. R.
(
2001
).
Tracking the influence of reflexive attention on sensory and cognitive processing.
Cognitive, Affective & Behavioral Neuroscience
,
1
,
56
65
.
Hopfinger
,
J. B.
, &
Ries
,
A. J.
(
2005
).
Automatic versus contingent mechanisms of sensory-driven neural biasing and reflexive attention.
Journal of Cognitive Neuroscience
,
17
,
1341
1352
.
Khoe
,
W.
,
Mitchell
,
J. F.
,
Reynolds
,
J. H.
, &
Hillyard
,
S. A.
(
2005
).
Exogenous attentional selection of transparent superimposed surfaces modulates early event-related potentials.
Vision Research
,
45
,
3004
3014
.
Lachter
,
J.
,
Forster
,
K. I.
, &
Ruthruff
,
E.
(
2004
).
Forty-five years after Broadbent (1958): Still no identification without attention.
Psychological Review
,
111
,
880
913
.
Lamy
,
D.
,
Segal
,
H.
, &
Ruderman
,
L.
(
2006
).
Grouping does not require attention.
Perception and Psychophysics
,
68
,
17
31
.
Lavie
,
N.
, &
Driver
,
J.
(
1996
).
On the spatial extent of attention in object-based visual selection.
Perception and Psychophysics
,
58
,
1238
1251
.
Lopez
,
M.
,
Rodriguez
,
V.
, &
Valdes-Sosa
,
M.
(
2004
).
Two-object attentional interference depends on attentional set.
International Journal of Psychophysiology
,
53
,
127
134
.
Maekawa
,
T.
,
Goto
,
Y.
,
Kinukawa
,
N.
,
Taniwaki
,
T.
,
Kanba
,
S.
, &
Tobimatsu
,
S.
(
2005
).
Functional characterization of mismatch negativity to a visual stimulus.
Clinical Neurophysiology
,
116
,
2392
2402
.
Moore
,
C. M.
, &
Egeth
,
H.
(
1997
).
Perception without attention: Evidence of grouping under conditions of inattention.
Journal of Experimental Psychology: Human Perception and Performance
,
23
,
339
352
.
Mordkoff
,
J. T.
, &
Halterman
,
R.
(
2008
).
Feature integration without visual attention: Evidence from the correlated flankers task.
Psychonomic Bulletin & Review
,
15
,
385
389
.
O'Craven
,
K. M.
,
Downing
,
P. E.
, &
Kanwisher
,
N.
(
1999
).
fMRI evidence for objects as the units of attentional selection.
Nature
,
401
,
584
587
.
Palmer
,
S. E.
, &
Beck
,
D. M.
(
2007
).
The repetition discrimination task: An objective method for studying perceptual grouping.
Perception and Psychophysics
,
69
,
68
78
.
Pazo-Alvarez
,
P.
,
Amenedo
,
E.
, &
Cadaveira
,
F.
(
2004
).
Automatic detection of motion direction changes in the human brain.
European Journal of Neuroscience
,
19
,
1978
1986
.
Pazo-Alvarez
,
P.
,
Cadaveira
,
F.
, &
Amenedo
,
E.
(
2003
).
MMN in the visual modality: A review.
Biological Psychology
,
63
,
199
236
.
Perrin
,
F.
,
Pernier
,
J.
,
Bertrand
,
O.
, &
Echallier
,
J. F.
(
1989
).
Spherical splines for scalp potential and current density mapping.
Electroencephalography and Clinical Neurophysiology
,
72
,
184
187
.
Pinilla
,
T.
,
Cobo
,
A.
,
Torres
,
K.
, &
Valdes-Sosa
,
M.
(
2001
).
Attentional shifts between surfaces: Effects on detection and early brain potentials.
Vision Research
,
41
,
1619
1630
.
Ritter
,
W.
,
De Sanctis
,
P.
,
Molholm
,
S.
,
Javitt
,
D. C.
, &
Foxe
,
J. J.
(
2006
).
Preattentively grouped tones do not elicit MMN with respect to each other.
Psychophysiology
,
43
,
423
430
.
Rodriguez
,
V.
, &
Valdes-Sosa
,
M.
(
2006
).
Sensory suppression during shifts of attention between surfaces in transparent motion.
Brain Research
,
1072
,
110
118
.
Russell
,
C.
, &
Driver
,
J.
(
2005
).
New indirect measures of “inattentive” visual grouping in a change-detection task.
Perception and Psychophysics
,
67
,
606
623
.
Scholl
,
B. J.
(
2001
).
Objects and attention: The state of the art.
Cognition
,
80
,
1
46
.
Srinivasan
,
R.
(
2005
).
High resolution EEG: Theory and practice.
In T. C. Handy (Ed.),
Event-related potentials: A methods handbook
(pp.
167
188
).
Cambridge, MA
:
MIT Press
.
Sussman
,
E.
,
Ritter
,
W.
,
Vaughan
,
H. G.
, &
,
(
1999
).
An investigation of the auditory streaming effect using event-related brain potentials.
Psychophysiology
,
36
,
22
34
.
Sussman
,
E. S.
(
2005
).
Integration and segregation in auditory scene analysis.
Journal of the Acoustical Society of America
,
117
,
1285
1298
.
Sussman
,
E. S.
,
Horvath
,
J.
,
Winkler
,
I.
, &
Orr
,
M.
(
2007
).
The role of attention in the formation of auditory streams.
Perception & Psychophysics
,
69
,
136
152
.
Theeuwes
,
J.
(
2004
).
Top–down search strategies cannot override attentional capture.
Psychonomic Bulletin & Review
,
11
,
65
70
.
Treisman
,
A.
, &
Schmidt
,
H.
(
1982
).
Illusory conjunctions in the perception of objects.
Cognitive Psychology
,
14
,
107
141
.
Treisman
,
A. M.
, &
Gelade
,
G.
(
1980
).
A feature-integration theory of attention.
Cognitive Psychology
,
12
,
97
136
.
Valdes-Sosa
,
M.
,
Bobes
,
M. A.
,
Rodriguez
,
V.
, &
Pinilla
,
T.
(
1998
).
Switching attention without shifting the spotlight object-based attentional modulation of brain potentials.
Journal of Cognitive Neuroscience
,
10
,
137
151
.
Valdes-Sosa
,
M.
,
Cobo
,
A.
, &
Pinilla
,
T.
(
2000
).
Attention to object files defined by transparent motion.
Journal of Experimental Psychology: Human Perception and Performance
,
26
,
488
505
.
Watson
,
S. E.
, &
Kramer
,
A. F.
(
1999
).
Object-based visual selective attention and perceptual organization.
Perception and Psychophysics
,
61
,
31
49
.
Winkler
,
I.
,
Czigler
,
I.
,
Sussman
,
E.
,
Horvath
,
J.
, &
Balazs
,
L.
(
2005
).
Preattentive binding of auditory and visual stimulus features.
Journal of Cognitive Neuroscience
,
17
,
320
339
.
Winkler
,
I.
,
Kushnerenko
,
E.
,
Horvath
,
J.
,
Ceponiene
,
R.
,
Fellman
,
V.
,
Huotilainen
,
M.
,
et al
(
2003
).
Newborn infants can organize the auditory world.
Proceedings of the National Academy of Sciences, U.S.A.
,
100
,
11812
11815
.
Wolfe
,
J. M.
(
1994
).
Guided search 2.0: A revised model of visual search.
Psychonomic Bulletin and Review
,
1
,
202
238
.
Yabe
,
H.
,
Winkler
,
I.
,
Czigler
,
I.
,
Koyama
,
S.
,
Kakigi
,
R.
,
Sutoh
,
T.
,
et al
(
2001
).
Organizing sound sequences in the human brain: The interplay of auditory streaming and temporal integration.
Brain Research
,
897
,
222
227
.
Yantis
,
S.
, &
Serences
,
J. T.
(
2003
).
Cortical mechanisms of space-based and object-based attentional control.
Current Opinion in Neurobiology
,
13
,
187
193
.
Yucel
,
G.
,
McCarthy
,
G.
, &
Belger
,
A.
(
2007
).
fMRI reveals that involuntary visual deviance processing is resource limited.
Neuroimage
,
34
,
1245
1252
.