c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
Available online at www.sciencedirect.com
ScienceDirect Journal homepage: www.elsevier.com/locate/cortex
Note
Time course of pseudoneglect in scene viewing Antje Nuthmann a,b,*,1 and Ellen Matthias c,1 a
University of Edinburgh, School of Philosophy, Psychology and Language Sciences, Psychology Department, Edinburgh, UK Center for Interdisciplinary Research (ZiF), Bielefeld University, Germany c University of Ulm, Department of Psychology and Pedagogy, Ulm, Germany b
article info
abstract
Article history:
When we view the visual world, our eyes move from one location to another about three
Received 1 August 2013
times each second. When looking at pictures of natural scenes, neurologically intact in-
Reviewed 18 September 2013
dividuals show a leftward bias in the direction of their first eye movement. The present
Revised 2 November 2013
study investigates the time course of this pseudoneglect and how it depends on task-
Accepted 18 November 2013
related control. Eye movements were recorded from 72 participants, each viewing 135
Action editor Paolo Bartolomeo
scenes under three different viewing instructions (memorization, esthetic preference
Published online 26 November 2013
judgment, object-in-scene search). In the memorization and preference tasks, pseudoneglect had a maximum extent of about 1 and lasted for about 1500 msec, or 5 fixations. The
Keywords:
effect was somewhat reduced in the preference task, which gave subjects free reign to
Scene perception
fixate anywhere they wanted to. During scene search, a task that is guided primarily by
Eye movements
top-down control, observers also showed a distinct pseudoneglect. Strikingly, a leftward
Attention
bias was present even when the search object was located in the right hemispace. Search
Pseudoneglect
performance was not affected by the observed spatial asymmetries. The effects likely arise
Hemispheric asymmetry
from a right-hemisphere dominance for visuo-spatial attention. ª 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
Humans do not attend equally to the left and the right sides of their visual world. Various systematic spatial asymmetries have been observed in healthy populations and neurological patients. Most prominently, a group of neurological patients shows a strong preference for objects located in the right side of space, known as visuo-spatial neglect (Bartolomeo & Chokron, 2002; Heilman, Watson, & Valenstein, 2003; Vallar, 1998). For example, in simple visual search tasks these patients neglect to fixate items on the left side of the display
(Husain et al., 2001; Lanyon & Denham, 2010). However, within the right hemispace they make leftward saccades as frequently as rightward ones (Niemeier & Karnath, 2000). According to the biased competition model of attention (Desimone & Duncan, 1995), the rightward spatial bias in neglect patients reflects lower “attentional weighting” for information presented in the left hemispace, which leads to a processing advantage for stimuli presented in the overattended (right) hemispace. Neurologically typical individuals also show a spatial processing bias, which is typically referred to as “pseudoneglect” (Bowers & Heilman, 1980). Notably, pseudoneglect is smaller
* Corresponding author. University of Edinburgh, Psychology Department, 7 George Square, Edinburgh EH8 9JZ, UK. E-mail address: [email protected] (A. Nuthmann). 1 Both authors contributed equally to this article. 0010-9452/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cortex.2013.11.007
114
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
in magnitude and is in the opposite direction: In contrast to neglect patients who show a pronounced rightward spatial bias, healthy individuals tend to exhibit a slight leftward spatial bias. A common tool for assessing asymmetries in visuo-spatial attention is the line bisection task, where participants are asked to select the middle point of a horizontal line (Jewell & McCourt, 2000; McCourt & Jewell, 1999; McCourt & Olafson, 1997). In this type of task, neurologically unimpaired individuals typically show a small but reliable error to the left (effect size between 0.37 and 0.44, Jewell & McCourt, 2000). A leftward bias has also been observed for visuo-spatial representations held in short-term memory (Della Sala, Darling, & Logie, 2010) and long-term memory (McGeorge, Beschin, Colnaghi, Rusconi, & Della Sala, 2007) such that participants reported more items from the left-hand side of remembered space. A few studies have reported a leftward bias in initial visual exploration through eye movements. In such experiments, observers initially fixate at the center of the screen, from which they could, in theory, move their eyes anywhere. However, participants undertaking a simple search task often start with a saccade to the left side of the display (Zelinsky, 1996). Similarly, when viewing pictures of real-world scenes, the very first saccade more often goes to the left than to the right side of the image (Dickinson & Intraub, 2009; Foulsham, Gray, Nasiopoulos, & Kingstone, 2013). Acknowledging that scene exploration is an active and dynamic process extending beyond the first saccade, the present study investigated the time course of pseudoneglect in scene perception and search: Is it a short-lived transient phenomenon or a longer lasting persistent one? In addition, we investigated how pseudoneglect depends on cognitive topdown influences, operationalized as probing effects of three different viewing tasks (e.g., Mills, Hollingworth, Van der Stigchel, Hoffman, & Dodd, 2011; Yarbus, 1967). First, a scene memorization task was included. Participants had to encode the scene to prepare for an old/new recognition test administered at the end of the experiment. In the memory test, half of the old and new scenes were mirrored horizontally. This served as a control condition ensuring that observed biases could not be attributed to left-right differences in the composition of the scenes (cf., Dickinson & Intraub, 2009). Second, an esthetic preference judgment was included in which participants rated how much they liked the scene. Third, a visual search task was employed in which participants had to look for a pre-specified object in the scene. Search objects were located on either the left or right side of the image. This allowed testing whether the bias toward the left side of space is also present when the search object is located in the right hemispace. In sum, the search task was experimenter-directed such that the target object was specified in the instructions. In contrast, the memorization and preference tasks were participant-directed as neither task specified a particular target object to which gaze should be directed. The preference task gave subjects free reign to fixate anywhere they choose to. The memorization task was less constrained than the search task but more constrained than the preference task, as it required observers to encode sufficient information in preparation for the memory test.
2.
Methods
2.1.
Participants, apparatus, and materials
Analyses were based on a large corpus of eye movements during scene viewing (Nuthmann & Henderson, 2010; Pajak & Nuthmann, 2013). Seventy-two participants (mean age ¼ 22.6 years, 34 males) each viewed 135 unique full-color photographs of real-world scenes from a variety of categories (indoor and outdoor). The 92 indoor scenes came from different sub-categories, ranging from common rooms in one’s house (e.g., living room, kitchen) to images from shops, garages etc. Scenes subtended 25.78 horizontally 19.34 vertically at a viewing distance of 90 cm. A chin rest with a head support was used to minimize head movement. The chin rest was placed such that the participant’s midpoint between the centers of the eyes was close to the vertical midline of the screen. In addition, table height was adjusted such that participants’ straight-ahead view was at the midpoint of the screen. Eye movements were recorded using an SR Research EyeLink 1000/2K system. Data from the right eye were analyzed.
2.2.
Design and procedure
Participants viewed each of the 135 scenes once, 45 scenes in each of the three viewing tasks (memorization, preference judgment, search). The order of tasks and the scenes used in each task were rotated across 9 subject groups (with 8 subjects each) using a dual Latin-square design. This ensured that every order of task and combination of scene with task was represented at least once across the 9 subject groups. All scenes were presented for 8 sec. Responses were collected using a controller with two side triggers and four front buttons. In the search task, participants indicated that they found the target by pressing either the left or right trigger button of the centrally located controller. In the preference task, participants responded by using the four buttons on the controller’s face (AeD for 1 dislike to 4 like). At the beginning of each trial a fixation point was presented at the center of the screen and acted as a fixation check. In the search task, prior to the fixation check a text label describing the target (e.g., basket) was presented for 800 msec. For details on the selection of search targets and their properties see Nuthmann and Henderson (2010). To keep viewing time constant across tasks, the scene remained on the screen until the 8 sec were over. However, analyses of spatial biases only considered fixations made until the button press terminating the search. After participants completed the three viewing tasks, the memory test was administered. It included 44 scenes from the memorization task, but half of them were mirrored horizontally. A further 22 new scenes were presented while half of these had also been flipped horizontally. Each scene was presented for 3 sec. Participants’ task was to identify whether a scene was (1) old, (2) altered (old scenes which have been mirrored horizontally), or (3) new. Participants responded by using three front buttons of the controller.
115
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
2.3.
Data analysis
Gaze raw data were converted into a fixation sequence matrix using SR Research Data Viewer. Data were further processed and analyzed using the R system for statistical computing (version 2.15.2; R Development Core Team, 2012) under the GNU General Public License (Version 2, June 1991). For analyses of the search data, scenes with search objects that covered the vertical midline of the image were excluded (n ¼ 6). Two of the three sets of scenes had more search objects on the left side of the image than on the right side (set 2: 24 vs 20, set 3: 25 vs 18). To ensure that the target appeared with equal probability in either the left or right visual field, randomly drawn scenes with left-side targets (set 2: 4, set 3: 7) were excluded from analyses.
3.
Results and discussion
Analogous to the line bisection task, during which the bisection error is measured as deviation from the middle of the line, we quantified pseudoneglect by calculating horizontal fixation deviation (FixDevx, in degrees of visual angle). To this end, horizontal fixation locations were analyzed as deviations from the vertical midline of the image. Fixations on the left/right side of the scene therefore have negative/positive horizontal deviations. If, at a given point in time, the deployment of attention and fixation was symmetric, we should observe a mean deviation of zero. Negative deviations are indicative of a
a
leftward bias in the distribution of attention. For time-course analysis, each fixation was further characterized by its position in the fixation sequence (ordinal number of fixation) as well as by the time passed between scene onset and the end of the current fixation (viewing time). In scene viewing, there is considerable variation in fixation durations (Nuthmann, Smith, Engbert, & Henderson, 2010), with individual differences explaining some of this variability. Observers who make comparatively long fixations make fewer fixations within a given trial. Other participants make many short fixations, and only these participants contribute data points to the higher fixation numbers. Therefore, the main analyses considered viewing time as a dependent variable. A generalized additive model (GAM) was fitted to each set of FixDevx raw data, using a penalized cubic regression spline. The implementation of the gam function from the mgcv package (Wood, 2011) supplied in R was used, called through the smoothing parameter of the ggplot function from the ggplot2 package (Wickham, 2009). The first analysis considered the data from the memorization and preference tasks, where observers made on average 27 (memorization) and 30 (preference) fixations during the 8sec viewing period. The GAM fit to the data describes the change of horizontal fixation deviation over the course of viewing (Fig. 1). For both tasks, there was an initial leftward bias, which can be characterized by its amplitude (maximal extent) and duration. The duration of the leftward bias is given by the point in time (Fig. 1a) or number of fixation (Fig. 1b) at which the regression line intersects with the point of zero deviation. First fixation denotes eye position following the first
b 8000
Task Memorization Preference
7000
Ordinal Number of Fixation
Viewing Time (ms)
6000
5000
4000
3000
2000
1000
0 −1.5
−1
−.5
0
.5
1
1.5
Horizontal Fixation Deviation (°)
19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 −1.5
−1
−.5
0
.5
1
1.5
Horizontal Fixation Deviation (°)
Fig. 1 e Pseudoneglect in the memorization and preference tasks. Spatial asymmetries are described as a change of horizontal fixation deviation ( ) over time. Time was measured in terms of viewing time (panel a), or ordinal number of fixation (panel b). Presented is the GAM fit to the data from the memorization (blue solid line) and preference (red broken line) tasks, with the shaded area depicting the 95% confidence interval. Negative deviations are indicative of a leftward bias in the distribution of attention and fixation. Similarly, positive deviations denote a rightward bias. See text for details.
116
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
a Mirrored during
b Not Mirrored during
Memory Test
Memory Test
3000
Viewing Time (ms)
2500
2000 Task Memory Encoding
1500
Memory Test 1000
500
0 −2
−1
0
1
2 −2
−1
0
1
2
Horizontal Fixation Deviation (°) Fig. 2 e Pseudoneglect in the memory-encoding and memory-test tasks. During the memory test, the scenes from the memory-encoding (memorization) block were repeated, with half of them mirror reversed. Panel a depicts how horizontal fixation deviation ( ) changes over time for the mirror-reversed scenes during the memory test (red broken line) and their original counterparts during memory encoding (blue solid line). For comparison, panel b shows the corresponding data for the scenes that were not mirrored during the memory test. During the memory test, scenes were presented for 3 sec. For correspondence, analysis of the memory-encoding data considered the initial 3 sec of the 8 sec viewing period.
saccade. The data displayed in Fig. 1b show that pseudoneglect was not restricted to the first eye movement. It lasted for approximately 1500 msec (Fig. 1a) or 5 fixations (Fig. 1b). The amplitude of the effect was around 1 . The leftward bias was larger and longer lasting in the memorization than the preference task and was followed by a rightward bias. Could the observed bias be attributed to lefteright differences in the composition of the scenes? Various image statistics were derived for our scenes, including contrast and edge density. The composed scenes tended to show the common bias toward having more visual features and objects2 in their center (e.g., Tatler, 2007), with no apparent lefteright asymmetries. The data from the memory test allowed us to assess more directly whether effects were due to idiosyncrasies of the scenes. During the memory test, the scenes from the memory-encoding (memorization) block were repeated, with half of them mirror reversed. For the old (not mirrored) and altered (mirrored) scenes, mean accuracies were 84.3% and 67%, respectively. The control analysis, including both correct and incorrect trials, compared the eye-movement data for the mirror-reversed scenes in the memory test with the eye-movement data for their original counterparts from memory encoding. If pseudoneglect was due to lefteright differences in the composition of the scenes, it should 2
This refers to all objects in the scenes.
disappear for the mirror-reversed scenes. This was not the case, as shown in Fig. 2a. There was, however, an effect of task such that the amplitude and duration of pseudoneglect were smaller during the memory test than during memory encoding. This effect was also present for scenes that were not mirrored during the memory test (Fig. 2b). The final set of analyses considered the data from the experimenter-directed scene search task. The average likelihood of correctly locating the target was 88.7%; the average search time for correct trials was 2856 msec. Analysis of spatial biases included only correct trials. In trials during which search time exceeded 3 sec, only fixations made during the initial 3 sec were analyzed. Pooling data from all scenes revealed a distinct pseudoneglect that lasted for about 1750 msec (Fig. 3a). The shortest distance between object center and the vertical midline was, on average, the same for left-side and right-side targets, t(116) ¼ 0.88, p ¼ .38. Towards the end of the search process, the mean deviation was therefore zero (Fig. 3a). A further analysis contrasted scenes in which the target was located on the left versus right side of the image (Fig. 3b). Not surprisingly, fixations on scenes with left-side targets showed a sustained bias toward the left side of space (solid line). Strikingly, this initial leftward bias was also present when the search object was located in the right hemispace (broken line). The effect was attenuated, as is evident from its reduced amplitude
117
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
b 3000
3000
2500
2500
2000
2000
Viewing Time (ms)
Viewing Time (ms)
a
1500
1000
500
1500
1000
Target location
500
left right 0
0 −2
−1
0
1
2
Horizontal Fixation Deviation (°)
−2
−1
0
1
2
Horizontal Fixation Deviation (°)
Fig. 3 e Pseudoneglect in the scene search task. Panel a depicts how horizontal fixation deviation ( ) changes over time for data from 118 search scenes in which the target appeared with equal probability on either side of the scene. In (panel b) the data are depicted separately for scenes in which the target was located on the left (blue solid line) or right (red broken line) side of the image.
(0.5 ). Notably, it still lasted for about 1 s. The search data allowed testing whether the observed spatial asymmetries affected performance. Does the initial leftward bias lead to longer search times for targets on the right side of the scene? Results from a linear mixed model specifying target location (left vs right) as fixed effect and subjects and scenes as (varying intercept) random effects suggested that this was not the case (fixed effect target location: t ¼ 0.36, p ¼ .72); the analysis was conducted using the lme4 package (Bates, Maechler, & Bolker, 2012) in R.
4.
General discussion
Previous research has reported a bias to start visual exploration of a natural scene with a saccade to the left side of the image (Dickinson & Intraub, 2009; Foulsham et al., 2013). The present study extends this research in two important ways. First, we provide a detailed time-course investigation of pseudoneglect during scene perception. Second, we show how pseudoneglect depends on cognitive top-down influences, operationalized by different viewing tasks. Fitting GAMs to the fixation raw data, a novel approach, allowed quantifying the amplitude and duration of pseudoneglect for a given task. In all tasks (memorization, preference judgment, scene search), participants’ gaze data were indicative of a pseudoneglect, which extended beyond the first eye movement. For
example, in the memorization task pseudoneglect had a maximum extent of about 1 and lasted for about 1500 msec, or 5 fixations. In the memorization and preference tasks, the initial leftward bias was followed by a rightward bias, possibly representing a compensatory mechanism to further visual exploration of previously unattended regions. During later scene viewing, the deployment of attention and fixation was symmetric, sometimes showing slight oscillations. The most striking aspect of the search data was the presence of a leftward bias even for right-sided targets. Notably, the amplitude and duration of pseudoneglect differed across tasks. In terms of viewing instructions, the preference task was the least constrained, followed by the memorization task. The search task was experimenterdirected and guided primarily by top-down control. Pseudoneglect was larger and longer lasting in the memorization than the preference task, and there was a distinct pseudoneglect in the search task. Collectively, the present data tentatively suggest that pseudoneglect may be more pronounced for tasks that evoke more top-down control. Pseudoneglect was smaller during the memory test than during memory encoding, which could be due to participants’ familiarity with the queried scenes or the shorter presentation time. What are the mechanisms that give rise to pseudoneglect in scene viewing? We argue that pseudoneglect is an attentional phenomenon rather than an oculomotor bias. We will then discuss the involvement of lateralized attention networks in the brain.
118
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
A number of studies have investigated the relative direction of saccades (relative to the previous fixation) in scene viewing and visual search. Averaged across the whole trial, rightward saccades are made as frequently as leftward ones (cf., Foulsham et al., 2013; Niemeier & Karnath, 2000),3 suggesting that pseudoneglect in scene viewing and search is not an oculomotor bias favoring the programming of left-directed saccades. Eye movements are an overt behavioral manifestation of the allocation of attention. A large body of literature has established that attentional orienting either precedes or coincides with saccade programming and that there is a mandatory link between the two (e.g., Deubel & Schneider, 1996). Interestingly, the general pattern of results reported heredthat is, an initial leftward bias followed by a rightward biasdcoincides well with results from a paradigm requiring covert attention shifts (Matthias et al., 2010) rather than overt attention shifts through eye movements. Most accounts of pseudoneglect focus on the advantage of the right hemisphere in directing spatial attention (Corbetta & Shulman, 2011; Foxe, McCourt, & Javitt, 2003; Mesulam, 1999; Mort et al., 2003). Considering the general principle of contralaterality in the brain, a right hemispheric specialization causes a superiority of stimuli presented in the left visual hemifield (Siman-Tov et al., 2007). Thus, pseudoneglect in tasks involving covert and/or overt attention shifts is likely to reflect cortical asymmetries with respect to the distribution of spatial attention. Attentional selection can occur as the result of top-down (cognitive) or bottom-up (stimulus-driven) processes. Accordingly, the model of selective attention by Corbetta and Shulman (2002) postulates two attentional networks: The bilateral IPs-FEF network is involved in top-down control, whereas the right-lateralized TPJ-VFC network is involved in bottom-up control.4 Crucially, the two networks interact such that the IPs provides the TPJ with task relevant information, either directly, or indirectly through top-down modulation of the visual cortex. Applied to the present scenario, viewing tasks that provoke more top-down control should lead to a stronger engagement of the IPs-FEF network and a stronger functional interaction between the IPs and TPJ. This may lead to enhanced activation of the right-lateralized bottom-up TPJVFC network, which in turn leads to an even stronger “overt orienting” of visuo-spatial attention towards the left hemispace. Investigation of this account within functional imaging studies will be instructive.
Acknowledgments We thank John M. Henderson for making the eye-tracking corpus data available and Tim J. Smith for his help with data pre-processing. The present article reports partial analyses of 3
This analysis ignores whether the fixation following the saccade is placed on the left or right half of the stimulus. The exception is the very first saccade, which is launched from screen center. 4 IPs ¼ intraparietal sulcus; FEF ¼ frontal eye field; TPJ ¼ temporoparietal junction, VFC ¼ ventral frontal cortex.
data collected under a grant from the Economic and Social Research Council (ESRC) of the UK to JMH (RES-062-23-1092).
references
Bartolomeo, P., & Chokron, S. (2002). Orienting of attention in left unilateral neglect. Neuroscience and Biobehavioral Reviews, 26(2), 217e234. http://dx.doi.org/10.1016/s0149-7634(01)00065-3. Bates, D. M., Maechler, M., & Bolker, B. (2012). lme4: Linear mixedeffects models using S4 classes (R package version 0.999999-0). http://CRAN.R-project.org/package¼lme4. Bowers, D., & Heilman, K. M. (1980). Pseudoneglect: effects of hemispace on a tactile line bisection task. Neuropsychologia, 18(4e5), 491e498. http://dx.doi.org/10.1016/0028-3932(80) 90151-7. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3(3), 201e215. http://dx.doi.org/10.1038/nrn755. Corbetta, M., & Shulman, G. L. (2011). Spatial neglect and attention networks. Annual Review of Neuroscience, 34, 569e599. http://dx.doi.org/10.1146/annurev-neuro-061010-113731. Della Sala, S., Darling, S., & Logie, R. H. (2010). Items on the left are better remembered. Quarterly Journal of Experimental Psychology, 63(5), 848e855. http://dx.doi.org/10.1080/17470211003690672. Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193e222. http://dx.doi.org/10.1146/annurev.neuro.18.1.193. Deubel, H., & Schneider, W. X. (1996). Saccade target selection and object recognition: evidence for a common attentional mechanism. Vision Research, 36(12), 1827e1837. http:// dx.doi.org/10.1016/0042-6989(95)00294-4. Dickinson, C. A., & Intraub, H. (2009). Spatial asymmetries in viewing and remembering scenes: consequences of an attentional bias? Attention Perception & Psychophysics, 71(6), 1251e1262. http://dx.doi.org/10.3758/app.71.6.1251. Foulsham, T., Gray, A., Nasiopoulos, E., & Kingstone, A. (2013). Leftward biases in picture scanning and line bisection: a gazecontingent window study. Vision Research, 78, 14e25. http:// dx.doi.org/10.1016/j.visres.2012.12.001. Foxe, J. J., McCourt, M. E., & Javitt, D. C. (2003). Right hemisphere control of visuospatial attention: line-bisection judgments evaluated with high-density electrical mapping and source analysis. NeuroImage, 19(3), 710e726. http://dx.doi.org/10.1016/ s1053-8119(03)00057-0. Heilman, K. M., Watson, R. T., & Valenstein, E. (2003). Neglect and related disorders. In K. M. Heilman, & E. Valenstein (Eds.), Clinical neuropsychology (pp. 296e346). Oxford: Oxford University Press. Husain, M., Mannan, S., Hodgson, T., Wojciulik, E., Driver, J., & Kennard, C. (2001). Impaired spatial working memory across saccades contributes to abnormal search in parietal neglect. Brain, 124(5), 941e952. http://dx.doi.org/10.1093/brain/ 124.5.941. Jewell, G., & McCourt, M. E. (2000). Pseudoneglect: a review and meta-analysis of performance factors in line bisection tasks. Neuropsychologia, 38(1), 93e110. http://dx.doi.org/10.1016/ s0028-3932(99)00045-7. Lanyon, L. J., & Denham, S. L. (2010). Modelling visual neglect: computational insights into conscious perception. PLoS ONE, 5(6), e11128. http://dx.doi.org/10.1371/journal.pone.0011128. Matthias, E., Bublak, P., Mu¨ller, H. J., Schneider, W. X., Krummenacher, J., & Finke, K. (2010). The influence of alertness on spatial and nonspatial components of visual attention. Journal of Experimental Psychology: Human Perception and Performance, 36(1), 38e56. http://dx.doi.org/10.1037/a0017602.
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
McCourt, M. E., & Jewell, G. (1999). Visuospatial attention in line bisection: stimulus modulation of pseudoneglect. Neuropsychologia, 37(7), 843e855. http://dx.doi.org/10.1016/ s0028-3932(98)00140-7. McCourt, M. E., & Olafson, C. (1997). Cognitive and perceptual influences on visual line bisection: psychophysical and chronometric analyses of pseudoneglect. Neuropsychologia, 35(3), 369e380. http://dx.doi.org/10.1016/s0028-3932(96)00143-1. McGeorge, P., Beschin, N., Colnaghi, A., Rusconi, M. L., & Della Sala, S. (2007). A lateralized bias in mental imagery: evidence for representational pseudoneglect. Neuroscience Letters, 421(3), 259e263. http://dx.doi.org/10.1016/j.neulet.2007.05.050. Mesulam, M. M. (1999). Spatial attention and neglect: parietal, frontal and cingulate contributions to the mental representation and attentional targeting of salient extrapersonal events. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 354(1387), 1325e1346. http://dx.doi.org/10.1098/rstb.1999.0482. Mills, M., Hollingworth, A., Van der Stigchel, S., Hoffman, L., & Dodd, M. D. (2011). Examining the influence of task set on eye movements and fixations. Journal of Vision, 11(8), 17, 1e15. http://dx.doi.org/10.1167/11.8.17. Mort, D. J., Malhotra, P., Mannan, S. K., Rorden, C., Pambakian, A., Kennard, C., et al. (2003). The anatomy of visual neglect. Brain, 126(9), 1986e1997. http://dx.doi.org/10.1093/brain/awg200. Niemeier, M., & Karnath, H. O. (2000). Exploratory saccades show no direction-specific deficit in neglect. Neurology, 54(2), 515e518. Nuthmann, A., & Henderson, J. M. (2010). Object-based attentional selection in scene viewing. Journal of Vision, 10(8), 20, 1e19. http://dx.doi.org/10.1167/10.8.20.
119
Nuthmann, A., Smith, T. J., Engbert, R., & Henderson, J. M. (2010). CRISP: a computational model of fixation durations in scene viewing. Psychological Review, 117(2), 382e405. http:// dx.doi.org/10.1037/a0018924. Pajak, M., & Nuthmann, A. (2013). Object-based saccadic selection during scene perception: evidence from viewing position effects. Journal of Vision, 13(5), 2, 1e21. http://dx.doi.org/10. 1167/13.5.2. Siman-Tov, T., Mendelsohn, A., Schonberg, T., Avidan, G., Podlipsky, I., Pessoa, L., et al. (2007). Bihemispheric leftward bias in a visuospatial attention-related network. Journal of Neuroscience, 27(42), 11271e11278. http://dx.doi.org/10.1523/ jneurosci.0599-07.2007. Tatler, B. W. (2007). The central fixation bias in scene viewing: selecting an optimal viewing position independently of motor biases and image feature distributions. Journal of Vision, 7(14), 4, 1e17. http://dx.doi.org/10.1167/7.14.4. Vallar, G. (1998). Spatial hemineglect in humans. Trends in Cognitive Sciences, 2(3), 87e97. http://dx.doi.org/10.1016/s13646613(98)01145-0. Wickham, H. (2009). ggplot2: Elegant graphics for data analysis. New York: Springer. Wood, S. N. (2011). Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society Series BStatistical Methodology, 73(1), 3e36. http://dx.doi.org/10.1111/ j.1467-9868.2010.00749.x. Yarbus, A. L. (1967). Eye movements and vision. New York: Plenum. Zelinsky, G. J. (1996). Using eye saccades to assess the selectivity of search movements. Vision Research, 36(14), 2177e2187. http://dx.doi.org/10.1016/0042-6989(95)00300-2.
Available online at www.sciencedirect.com
ScienceDirect Journal homepage: www.elsevier.com/locate/cortex
Note
Time course of pseudoneglect in scene viewing Antje Nuthmann a,b,*,1 and Ellen Matthias c,1 a
University of Edinburgh, School of Philosophy, Psychology and Language Sciences, Psychology Department, Edinburgh, UK Center for Interdisciplinary Research (ZiF), Bielefeld University, Germany c University of Ulm, Department of Psychology and Pedagogy, Ulm, Germany b
article info
abstract
Article history:
When we view the visual world, our eyes move from one location to another about three
Received 1 August 2013
times each second. When looking at pictures of natural scenes, neurologically intact in-
Reviewed 18 September 2013
dividuals show a leftward bias in the direction of their first eye movement. The present
Revised 2 November 2013
study investigates the time course of this pseudoneglect and how it depends on task-
Accepted 18 November 2013
related control. Eye movements were recorded from 72 participants, each viewing 135
Action editor Paolo Bartolomeo
scenes under three different viewing instructions (memorization, esthetic preference
Published online 26 November 2013
judgment, object-in-scene search). In the memorization and preference tasks, pseudoneglect had a maximum extent of about 1 and lasted for about 1500 msec, or 5 fixations. The
Keywords:
effect was somewhat reduced in the preference task, which gave subjects free reign to
Scene perception
fixate anywhere they wanted to. During scene search, a task that is guided primarily by
Eye movements
top-down control, observers also showed a distinct pseudoneglect. Strikingly, a leftward
Attention
bias was present even when the search object was located in the right hemispace. Search
Pseudoneglect
performance was not affected by the observed spatial asymmetries. The effects likely arise
Hemispheric asymmetry
from a right-hemisphere dominance for visuo-spatial attention. ª 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
Humans do not attend equally to the left and the right sides of their visual world. Various systematic spatial asymmetries have been observed in healthy populations and neurological patients. Most prominently, a group of neurological patients shows a strong preference for objects located in the right side of space, known as visuo-spatial neglect (Bartolomeo & Chokron, 2002; Heilman, Watson, & Valenstein, 2003; Vallar, 1998). For example, in simple visual search tasks these patients neglect to fixate items on the left side of the display
(Husain et al., 2001; Lanyon & Denham, 2010). However, within the right hemispace they make leftward saccades as frequently as rightward ones (Niemeier & Karnath, 2000). According to the biased competition model of attention (Desimone & Duncan, 1995), the rightward spatial bias in neglect patients reflects lower “attentional weighting” for information presented in the left hemispace, which leads to a processing advantage for stimuli presented in the overattended (right) hemispace. Neurologically typical individuals also show a spatial processing bias, which is typically referred to as “pseudoneglect” (Bowers & Heilman, 1980). Notably, pseudoneglect is smaller
* Corresponding author. University of Edinburgh, Psychology Department, 7 George Square, Edinburgh EH8 9JZ, UK. E-mail address: [email protected] (A. Nuthmann). 1 Both authors contributed equally to this article. 0010-9452/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cortex.2013.11.007
114
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
in magnitude and is in the opposite direction: In contrast to neglect patients who show a pronounced rightward spatial bias, healthy individuals tend to exhibit a slight leftward spatial bias. A common tool for assessing asymmetries in visuo-spatial attention is the line bisection task, where participants are asked to select the middle point of a horizontal line (Jewell & McCourt, 2000; McCourt & Jewell, 1999; McCourt & Olafson, 1997). In this type of task, neurologically unimpaired individuals typically show a small but reliable error to the left (effect size between 0.37 and 0.44, Jewell & McCourt, 2000). A leftward bias has also been observed for visuo-spatial representations held in short-term memory (Della Sala, Darling, & Logie, 2010) and long-term memory (McGeorge, Beschin, Colnaghi, Rusconi, & Della Sala, 2007) such that participants reported more items from the left-hand side of remembered space. A few studies have reported a leftward bias in initial visual exploration through eye movements. In such experiments, observers initially fixate at the center of the screen, from which they could, in theory, move their eyes anywhere. However, participants undertaking a simple search task often start with a saccade to the left side of the display (Zelinsky, 1996). Similarly, when viewing pictures of real-world scenes, the very first saccade more often goes to the left than to the right side of the image (Dickinson & Intraub, 2009; Foulsham, Gray, Nasiopoulos, & Kingstone, 2013). Acknowledging that scene exploration is an active and dynamic process extending beyond the first saccade, the present study investigated the time course of pseudoneglect in scene perception and search: Is it a short-lived transient phenomenon or a longer lasting persistent one? In addition, we investigated how pseudoneglect depends on cognitive topdown influences, operationalized as probing effects of three different viewing tasks (e.g., Mills, Hollingworth, Van der Stigchel, Hoffman, & Dodd, 2011; Yarbus, 1967). First, a scene memorization task was included. Participants had to encode the scene to prepare for an old/new recognition test administered at the end of the experiment. In the memory test, half of the old and new scenes were mirrored horizontally. This served as a control condition ensuring that observed biases could not be attributed to left-right differences in the composition of the scenes (cf., Dickinson & Intraub, 2009). Second, an esthetic preference judgment was included in which participants rated how much they liked the scene. Third, a visual search task was employed in which participants had to look for a pre-specified object in the scene. Search objects were located on either the left or right side of the image. This allowed testing whether the bias toward the left side of space is also present when the search object is located in the right hemispace. In sum, the search task was experimenter-directed such that the target object was specified in the instructions. In contrast, the memorization and preference tasks were participant-directed as neither task specified a particular target object to which gaze should be directed. The preference task gave subjects free reign to fixate anywhere they choose to. The memorization task was less constrained than the search task but more constrained than the preference task, as it required observers to encode sufficient information in preparation for the memory test.
2.
Methods
2.1.
Participants, apparatus, and materials
Analyses were based on a large corpus of eye movements during scene viewing (Nuthmann & Henderson, 2010; Pajak & Nuthmann, 2013). Seventy-two participants (mean age ¼ 22.6 years, 34 males) each viewed 135 unique full-color photographs of real-world scenes from a variety of categories (indoor and outdoor). The 92 indoor scenes came from different sub-categories, ranging from common rooms in one’s house (e.g., living room, kitchen) to images from shops, garages etc. Scenes subtended 25.78 horizontally 19.34 vertically at a viewing distance of 90 cm. A chin rest with a head support was used to minimize head movement. The chin rest was placed such that the participant’s midpoint between the centers of the eyes was close to the vertical midline of the screen. In addition, table height was adjusted such that participants’ straight-ahead view was at the midpoint of the screen. Eye movements were recorded using an SR Research EyeLink 1000/2K system. Data from the right eye were analyzed.
2.2.
Design and procedure
Participants viewed each of the 135 scenes once, 45 scenes in each of the three viewing tasks (memorization, preference judgment, search). The order of tasks and the scenes used in each task were rotated across 9 subject groups (with 8 subjects each) using a dual Latin-square design. This ensured that every order of task and combination of scene with task was represented at least once across the 9 subject groups. All scenes were presented for 8 sec. Responses were collected using a controller with two side triggers and four front buttons. In the search task, participants indicated that they found the target by pressing either the left or right trigger button of the centrally located controller. In the preference task, participants responded by using the four buttons on the controller’s face (AeD for 1 dislike to 4 like). At the beginning of each trial a fixation point was presented at the center of the screen and acted as a fixation check. In the search task, prior to the fixation check a text label describing the target (e.g., basket) was presented for 800 msec. For details on the selection of search targets and their properties see Nuthmann and Henderson (2010). To keep viewing time constant across tasks, the scene remained on the screen until the 8 sec were over. However, analyses of spatial biases only considered fixations made until the button press terminating the search. After participants completed the three viewing tasks, the memory test was administered. It included 44 scenes from the memorization task, but half of them were mirrored horizontally. A further 22 new scenes were presented while half of these had also been flipped horizontally. Each scene was presented for 3 sec. Participants’ task was to identify whether a scene was (1) old, (2) altered (old scenes which have been mirrored horizontally), or (3) new. Participants responded by using three front buttons of the controller.
115
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
2.3.
Data analysis
Gaze raw data were converted into a fixation sequence matrix using SR Research Data Viewer. Data were further processed and analyzed using the R system for statistical computing (version 2.15.2; R Development Core Team, 2012) under the GNU General Public License (Version 2, June 1991). For analyses of the search data, scenes with search objects that covered the vertical midline of the image were excluded (n ¼ 6). Two of the three sets of scenes had more search objects on the left side of the image than on the right side (set 2: 24 vs 20, set 3: 25 vs 18). To ensure that the target appeared with equal probability in either the left or right visual field, randomly drawn scenes with left-side targets (set 2: 4, set 3: 7) were excluded from analyses.
3.
Results and discussion
Analogous to the line bisection task, during which the bisection error is measured as deviation from the middle of the line, we quantified pseudoneglect by calculating horizontal fixation deviation (FixDevx, in degrees of visual angle). To this end, horizontal fixation locations were analyzed as deviations from the vertical midline of the image. Fixations on the left/right side of the scene therefore have negative/positive horizontal deviations. If, at a given point in time, the deployment of attention and fixation was symmetric, we should observe a mean deviation of zero. Negative deviations are indicative of a
a
leftward bias in the distribution of attention. For time-course analysis, each fixation was further characterized by its position in the fixation sequence (ordinal number of fixation) as well as by the time passed between scene onset and the end of the current fixation (viewing time). In scene viewing, there is considerable variation in fixation durations (Nuthmann, Smith, Engbert, & Henderson, 2010), with individual differences explaining some of this variability. Observers who make comparatively long fixations make fewer fixations within a given trial. Other participants make many short fixations, and only these participants contribute data points to the higher fixation numbers. Therefore, the main analyses considered viewing time as a dependent variable. A generalized additive model (GAM) was fitted to each set of FixDevx raw data, using a penalized cubic regression spline. The implementation of the gam function from the mgcv package (Wood, 2011) supplied in R was used, called through the smoothing parameter of the ggplot function from the ggplot2 package (Wickham, 2009). The first analysis considered the data from the memorization and preference tasks, where observers made on average 27 (memorization) and 30 (preference) fixations during the 8sec viewing period. The GAM fit to the data describes the change of horizontal fixation deviation over the course of viewing (Fig. 1). For both tasks, there was an initial leftward bias, which can be characterized by its amplitude (maximal extent) and duration. The duration of the leftward bias is given by the point in time (Fig. 1a) or number of fixation (Fig. 1b) at which the regression line intersects with the point of zero deviation. First fixation denotes eye position following the first
b 8000
Task Memorization Preference
7000
Ordinal Number of Fixation
Viewing Time (ms)
6000
5000
4000
3000
2000
1000
0 −1.5
−1
−.5
0
.5
1
1.5
Horizontal Fixation Deviation (°)
19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 −1.5
−1
−.5
0
.5
1
1.5
Horizontal Fixation Deviation (°)
Fig. 1 e Pseudoneglect in the memorization and preference tasks. Spatial asymmetries are described as a change of horizontal fixation deviation ( ) over time. Time was measured in terms of viewing time (panel a), or ordinal number of fixation (panel b). Presented is the GAM fit to the data from the memorization (blue solid line) and preference (red broken line) tasks, with the shaded area depicting the 95% confidence interval. Negative deviations are indicative of a leftward bias in the distribution of attention and fixation. Similarly, positive deviations denote a rightward bias. See text for details.
116
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
a Mirrored during
b Not Mirrored during
Memory Test
Memory Test
3000
Viewing Time (ms)
2500
2000 Task Memory Encoding
1500
Memory Test 1000
500
0 −2
−1
0
1
2 −2
−1
0
1
2
Horizontal Fixation Deviation (°) Fig. 2 e Pseudoneglect in the memory-encoding and memory-test tasks. During the memory test, the scenes from the memory-encoding (memorization) block were repeated, with half of them mirror reversed. Panel a depicts how horizontal fixation deviation ( ) changes over time for the mirror-reversed scenes during the memory test (red broken line) and their original counterparts during memory encoding (blue solid line). For comparison, panel b shows the corresponding data for the scenes that were not mirrored during the memory test. During the memory test, scenes were presented for 3 sec. For correspondence, analysis of the memory-encoding data considered the initial 3 sec of the 8 sec viewing period.
saccade. The data displayed in Fig. 1b show that pseudoneglect was not restricted to the first eye movement. It lasted for approximately 1500 msec (Fig. 1a) or 5 fixations (Fig. 1b). The amplitude of the effect was around 1 . The leftward bias was larger and longer lasting in the memorization than the preference task and was followed by a rightward bias. Could the observed bias be attributed to lefteright differences in the composition of the scenes? Various image statistics were derived for our scenes, including contrast and edge density. The composed scenes tended to show the common bias toward having more visual features and objects2 in their center (e.g., Tatler, 2007), with no apparent lefteright asymmetries. The data from the memory test allowed us to assess more directly whether effects were due to idiosyncrasies of the scenes. During the memory test, the scenes from the memory-encoding (memorization) block were repeated, with half of them mirror reversed. For the old (not mirrored) and altered (mirrored) scenes, mean accuracies were 84.3% and 67%, respectively. The control analysis, including both correct and incorrect trials, compared the eye-movement data for the mirror-reversed scenes in the memory test with the eye-movement data for their original counterparts from memory encoding. If pseudoneglect was due to lefteright differences in the composition of the scenes, it should 2
This refers to all objects in the scenes.
disappear for the mirror-reversed scenes. This was not the case, as shown in Fig. 2a. There was, however, an effect of task such that the amplitude and duration of pseudoneglect were smaller during the memory test than during memory encoding. This effect was also present for scenes that were not mirrored during the memory test (Fig. 2b). The final set of analyses considered the data from the experimenter-directed scene search task. The average likelihood of correctly locating the target was 88.7%; the average search time for correct trials was 2856 msec. Analysis of spatial biases included only correct trials. In trials during which search time exceeded 3 sec, only fixations made during the initial 3 sec were analyzed. Pooling data from all scenes revealed a distinct pseudoneglect that lasted for about 1750 msec (Fig. 3a). The shortest distance between object center and the vertical midline was, on average, the same for left-side and right-side targets, t(116) ¼ 0.88, p ¼ .38. Towards the end of the search process, the mean deviation was therefore zero (Fig. 3a). A further analysis contrasted scenes in which the target was located on the left versus right side of the image (Fig. 3b). Not surprisingly, fixations on scenes with left-side targets showed a sustained bias toward the left side of space (solid line). Strikingly, this initial leftward bias was also present when the search object was located in the right hemispace (broken line). The effect was attenuated, as is evident from its reduced amplitude
117
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
b 3000
3000
2500
2500
2000
2000
Viewing Time (ms)
Viewing Time (ms)
a
1500
1000
500
1500
1000
Target location
500
left right 0
0 −2
−1
0
1
2
Horizontal Fixation Deviation (°)
−2
−1
0
1
2
Horizontal Fixation Deviation (°)
Fig. 3 e Pseudoneglect in the scene search task. Panel a depicts how horizontal fixation deviation ( ) changes over time for data from 118 search scenes in which the target appeared with equal probability on either side of the scene. In (panel b) the data are depicted separately for scenes in which the target was located on the left (blue solid line) or right (red broken line) side of the image.
(0.5 ). Notably, it still lasted for about 1 s. The search data allowed testing whether the observed spatial asymmetries affected performance. Does the initial leftward bias lead to longer search times for targets on the right side of the scene? Results from a linear mixed model specifying target location (left vs right) as fixed effect and subjects and scenes as (varying intercept) random effects suggested that this was not the case (fixed effect target location: t ¼ 0.36, p ¼ .72); the analysis was conducted using the lme4 package (Bates, Maechler, & Bolker, 2012) in R.
4.
General discussion
Previous research has reported a bias to start visual exploration of a natural scene with a saccade to the left side of the image (Dickinson & Intraub, 2009; Foulsham et al., 2013). The present study extends this research in two important ways. First, we provide a detailed time-course investigation of pseudoneglect during scene perception. Second, we show how pseudoneglect depends on cognitive top-down influences, operationalized by different viewing tasks. Fitting GAMs to the fixation raw data, a novel approach, allowed quantifying the amplitude and duration of pseudoneglect for a given task. In all tasks (memorization, preference judgment, scene search), participants’ gaze data were indicative of a pseudoneglect, which extended beyond the first eye movement. For
example, in the memorization task pseudoneglect had a maximum extent of about 1 and lasted for about 1500 msec, or 5 fixations. In the memorization and preference tasks, the initial leftward bias was followed by a rightward bias, possibly representing a compensatory mechanism to further visual exploration of previously unattended regions. During later scene viewing, the deployment of attention and fixation was symmetric, sometimes showing slight oscillations. The most striking aspect of the search data was the presence of a leftward bias even for right-sided targets. Notably, the amplitude and duration of pseudoneglect differed across tasks. In terms of viewing instructions, the preference task was the least constrained, followed by the memorization task. The search task was experimenterdirected and guided primarily by top-down control. Pseudoneglect was larger and longer lasting in the memorization than the preference task, and there was a distinct pseudoneglect in the search task. Collectively, the present data tentatively suggest that pseudoneglect may be more pronounced for tasks that evoke more top-down control. Pseudoneglect was smaller during the memory test than during memory encoding, which could be due to participants’ familiarity with the queried scenes or the shorter presentation time. What are the mechanisms that give rise to pseudoneglect in scene viewing? We argue that pseudoneglect is an attentional phenomenon rather than an oculomotor bias. We will then discuss the involvement of lateralized attention networks in the brain.
118
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
A number of studies have investigated the relative direction of saccades (relative to the previous fixation) in scene viewing and visual search. Averaged across the whole trial, rightward saccades are made as frequently as leftward ones (cf., Foulsham et al., 2013; Niemeier & Karnath, 2000),3 suggesting that pseudoneglect in scene viewing and search is not an oculomotor bias favoring the programming of left-directed saccades. Eye movements are an overt behavioral manifestation of the allocation of attention. A large body of literature has established that attentional orienting either precedes or coincides with saccade programming and that there is a mandatory link between the two (e.g., Deubel & Schneider, 1996). Interestingly, the general pattern of results reported heredthat is, an initial leftward bias followed by a rightward biasdcoincides well with results from a paradigm requiring covert attention shifts (Matthias et al., 2010) rather than overt attention shifts through eye movements. Most accounts of pseudoneglect focus on the advantage of the right hemisphere in directing spatial attention (Corbetta & Shulman, 2011; Foxe, McCourt, & Javitt, 2003; Mesulam, 1999; Mort et al., 2003). Considering the general principle of contralaterality in the brain, a right hemispheric specialization causes a superiority of stimuli presented in the left visual hemifield (Siman-Tov et al., 2007). Thus, pseudoneglect in tasks involving covert and/or overt attention shifts is likely to reflect cortical asymmetries with respect to the distribution of spatial attention. Attentional selection can occur as the result of top-down (cognitive) or bottom-up (stimulus-driven) processes. Accordingly, the model of selective attention by Corbetta and Shulman (2002) postulates two attentional networks: The bilateral IPs-FEF network is involved in top-down control, whereas the right-lateralized TPJ-VFC network is involved in bottom-up control.4 Crucially, the two networks interact such that the IPs provides the TPJ with task relevant information, either directly, or indirectly through top-down modulation of the visual cortex. Applied to the present scenario, viewing tasks that provoke more top-down control should lead to a stronger engagement of the IPs-FEF network and a stronger functional interaction between the IPs and TPJ. This may lead to enhanced activation of the right-lateralized bottom-up TPJVFC network, which in turn leads to an even stronger “overt orienting” of visuo-spatial attention towards the left hemispace. Investigation of this account within functional imaging studies will be instructive.
Acknowledgments We thank John M. Henderson for making the eye-tracking corpus data available and Tim J. Smith for his help with data pre-processing. The present article reports partial analyses of 3
This analysis ignores whether the fixation following the saccade is placed on the left or right half of the stimulus. The exception is the very first saccade, which is launched from screen center. 4 IPs ¼ intraparietal sulcus; FEF ¼ frontal eye field; TPJ ¼ temporoparietal junction, VFC ¼ ventral frontal cortex.
data collected under a grant from the Economic and Social Research Council (ESRC) of the UK to JMH (RES-062-23-1092).
references
Bartolomeo, P., & Chokron, S. (2002). Orienting of attention in left unilateral neglect. Neuroscience and Biobehavioral Reviews, 26(2), 217e234. http://dx.doi.org/10.1016/s0149-7634(01)00065-3. Bates, D. M., Maechler, M., & Bolker, B. (2012). lme4: Linear mixedeffects models using S4 classes (R package version 0.999999-0). http://CRAN.R-project.org/package¼lme4. Bowers, D., & Heilman, K. M. (1980). Pseudoneglect: effects of hemispace on a tactile line bisection task. Neuropsychologia, 18(4e5), 491e498. http://dx.doi.org/10.1016/0028-3932(80) 90151-7. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3(3), 201e215. http://dx.doi.org/10.1038/nrn755. Corbetta, M., & Shulman, G. L. (2011). Spatial neglect and attention networks. Annual Review of Neuroscience, 34, 569e599. http://dx.doi.org/10.1146/annurev-neuro-061010-113731. Della Sala, S., Darling, S., & Logie, R. H. (2010). Items on the left are better remembered. Quarterly Journal of Experimental Psychology, 63(5), 848e855. http://dx.doi.org/10.1080/17470211003690672. Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193e222. http://dx.doi.org/10.1146/annurev.neuro.18.1.193. Deubel, H., & Schneider, W. X. (1996). Saccade target selection and object recognition: evidence for a common attentional mechanism. Vision Research, 36(12), 1827e1837. http:// dx.doi.org/10.1016/0042-6989(95)00294-4. Dickinson, C. A., & Intraub, H. (2009). Spatial asymmetries in viewing and remembering scenes: consequences of an attentional bias? Attention Perception & Psychophysics, 71(6), 1251e1262. http://dx.doi.org/10.3758/app.71.6.1251. Foulsham, T., Gray, A., Nasiopoulos, E., & Kingstone, A. (2013). Leftward biases in picture scanning and line bisection: a gazecontingent window study. Vision Research, 78, 14e25. http:// dx.doi.org/10.1016/j.visres.2012.12.001. Foxe, J. J., McCourt, M. E., & Javitt, D. C. (2003). Right hemisphere control of visuospatial attention: line-bisection judgments evaluated with high-density electrical mapping and source analysis. NeuroImage, 19(3), 710e726. http://dx.doi.org/10.1016/ s1053-8119(03)00057-0. Heilman, K. M., Watson, R. T., & Valenstein, E. (2003). Neglect and related disorders. In K. M. Heilman, & E. Valenstein (Eds.), Clinical neuropsychology (pp. 296e346). Oxford: Oxford University Press. Husain, M., Mannan, S., Hodgson, T., Wojciulik, E., Driver, J., & Kennard, C. (2001). Impaired spatial working memory across saccades contributes to abnormal search in parietal neglect. Brain, 124(5), 941e952. http://dx.doi.org/10.1093/brain/ 124.5.941. Jewell, G., & McCourt, M. E. (2000). Pseudoneglect: a review and meta-analysis of performance factors in line bisection tasks. Neuropsychologia, 38(1), 93e110. http://dx.doi.org/10.1016/ s0028-3932(99)00045-7. Lanyon, L. J., & Denham, S. L. (2010). Modelling visual neglect: computational insights into conscious perception. PLoS ONE, 5(6), e11128. http://dx.doi.org/10.1371/journal.pone.0011128. Matthias, E., Bublak, P., Mu¨ller, H. J., Schneider, W. X., Krummenacher, J., & Finke, K. (2010). The influence of alertness on spatial and nonspatial components of visual attention. Journal of Experimental Psychology: Human Perception and Performance, 36(1), 38e56. http://dx.doi.org/10.1037/a0017602.
c o r t e x 5 2 ( 2 0 1 4 ) 1 1 3 e1 1 9
McCourt, M. E., & Jewell, G. (1999). Visuospatial attention in line bisection: stimulus modulation of pseudoneglect. Neuropsychologia, 37(7), 843e855. http://dx.doi.org/10.1016/ s0028-3932(98)00140-7. McCourt, M. E., & Olafson, C. (1997). Cognitive and perceptual influences on visual line bisection: psychophysical and chronometric analyses of pseudoneglect. Neuropsychologia, 35(3), 369e380. http://dx.doi.org/10.1016/s0028-3932(96)00143-1. McGeorge, P., Beschin, N., Colnaghi, A., Rusconi, M. L., & Della Sala, S. (2007). A lateralized bias in mental imagery: evidence for representational pseudoneglect. Neuroscience Letters, 421(3), 259e263. http://dx.doi.org/10.1016/j.neulet.2007.05.050. Mesulam, M. M. (1999). Spatial attention and neglect: parietal, frontal and cingulate contributions to the mental representation and attentional targeting of salient extrapersonal events. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 354(1387), 1325e1346. http://dx.doi.org/10.1098/rstb.1999.0482. Mills, M., Hollingworth, A., Van der Stigchel, S., Hoffman, L., & Dodd, M. D. (2011). Examining the influence of task set on eye movements and fixations. Journal of Vision, 11(8), 17, 1e15. http://dx.doi.org/10.1167/11.8.17. Mort, D. J., Malhotra, P., Mannan, S. K., Rorden, C., Pambakian, A., Kennard, C., et al. (2003). The anatomy of visual neglect. Brain, 126(9), 1986e1997. http://dx.doi.org/10.1093/brain/awg200. Niemeier, M., & Karnath, H. O. (2000). Exploratory saccades show no direction-specific deficit in neglect. Neurology, 54(2), 515e518. Nuthmann, A., & Henderson, J. M. (2010). Object-based attentional selection in scene viewing. Journal of Vision, 10(8), 20, 1e19. http://dx.doi.org/10.1167/10.8.20.
119
Nuthmann, A., Smith, T. J., Engbert, R., & Henderson, J. M. (2010). CRISP: a computational model of fixation durations in scene viewing. Psychological Review, 117(2), 382e405. http:// dx.doi.org/10.1037/a0018924. Pajak, M., & Nuthmann, A. (2013). Object-based saccadic selection during scene perception: evidence from viewing position effects. Journal of Vision, 13(5), 2, 1e21. http://dx.doi.org/10. 1167/13.5.2. Siman-Tov, T., Mendelsohn, A., Schonberg, T., Avidan, G., Podlipsky, I., Pessoa, L., et al. (2007). Bihemispheric leftward bias in a visuospatial attention-related network. Journal of Neuroscience, 27(42), 11271e11278. http://dx.doi.org/10.1523/ jneurosci.0599-07.2007. Tatler, B. W. (2007). The central fixation bias in scene viewing: selecting an optimal viewing position independently of motor biases and image feature distributions. Journal of Vision, 7(14), 4, 1e17. http://dx.doi.org/10.1167/7.14.4. Vallar, G. (1998). Spatial hemineglect in humans. Trends in Cognitive Sciences, 2(3), 87e97. http://dx.doi.org/10.1016/s13646613(98)01145-0. Wickham, H. (2009). ggplot2: Elegant graphics for data analysis. New York: Springer. Wood, S. N. (2011). Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society Series BStatistical Methodology, 73(1), 3e36. http://dx.doi.org/10.1111/ j.1467-9868.2010.00749.x. Yarbus, A. L. (1967). Eye movements and vision. New York: Plenum. Zelinsky, G. J. (1996). Using eye saccades to assess the selectivity of search movements. Vision Research, 36(14), 2177e2187. http://dx.doi.org/10.1016/0042-6989(95)00300-2.
