Infant Behavior & Development 37 (2014) 225–234
Contents lists available at ScienceDirect
Infant Behavior and Development
Eighteen-month-olds’ ability to make gaze predictions following distraction or a long delay Linda Forssman ∗ , Gunilla Bohlin, Claes von Hofsten Department of Psychology, Uppsala University, Sweden
a r t i c l e
i n f o
Article history: Received 3 September 2013 Received in revised form 16 December 2013 Accepted 24 January 2014 Available online 15 March 2014
Keywords: Attentional control Anticipatory gaze Children Eye-tracking
a b s t r a c t The abilities to flexibly allocate attention, select between conflicting stimuli, and make anticipatory gaze movements are important for young children’s exploration and learning about their environment. These abilities constitute voluntary control of attention and show marked improvements in the second year of a child’s life. Here we investigate the effects of visual distraction and delay on 18-month-olds’ ability to predict the location of an occluded target in an experiment that requires switching of attention, and compare their performance to that of adults. Our results demonstrate that by 18 months of age children can readily overcome a previously learned response, even under a condition that involves visual distraction, but have difficulties with correctly updating their prediction when presented with a longer time delay. Further, the experiment shows that, overall, the 18-month-olds’ allocation of visual attention is similar to that of adults, the primary difference being that adults demonstrate a superior ability to maintain attention on task and update their predictions over a longer time period. © 2014 Elsevier Inc. All rights reserved.
1. Introduction In the first years of life, the development of attentional control provides an important mechanism for children’s exploration and learning about their environment. Attentional control involves the ability to flexibly allocate attention and suppress conflicting stimuli that interfere with the task at hand. The early development of this ability has been suggested to underpin the development of more complex skills, such as emotional and social regulation (Posner, Rothbart, Sheese, & Voelker, 2012) executive functions (e.g., working memory and inhibition; see Garon, Bryson, & Smith, 2008 for a review), and language development (Salley, Panneton, & Colombo, 2012). Further, deficits in the early development of attentional control may be related to increased risk for developmental disorders (e.g., attention deficit hyperactivity disorder and autism spectrum disorder) (Johnson, 2012). It is generally believed that brain maturation and increased functional connectivity are accompanying developmental improvements in behavioral control of attention. Resting-state fMRI studies indicate that the neural networks, supporting basic forms of attention control, become functional in infancy. They show a strong increase in connectivity over the first 2 years of life, and follow different developmental trajectories (Gao et al., 2013, 2009; Uddin, Supekar, & Menon, 2010). Functional connectivity between brain areas that support resolution of attentional conflict, such as the suppression of interfering stimuli (e.g., connectivity between the frontal cortex and anterior cingulate gyrus), emerges after 6 months of age and has a protracted development that lasts throughout childhood (Petersen & Posner, 2012; Posner & Fan, 2008).
∗ Corresponding author. Current address: Tampere Center for Child Health Research, School of Medicine, University of Tampere, FIN-33014 Tampere, Finland. Tel.: +358 40 190 1358; fax: +358 3 213 4473. E-mail address: Linda.Forssman@uta.fi (L. Forssman). 0163-6383/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.infbeh.2014.01.007
226
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
One method to study attentional control in infancy and toddlerhood is the assessment of anticipatory gaze, that is, gaze shifts occurring prior to the presentation of an event or stimulus. Studies measuring eye movements have shown that infants can make anticipatory gaze shifts around 4–9 months of age (Johnson, Posner, & Rothbart, 1991; Nelson, 1971; Sheese, Rothbart, Posner, White, & Fraundorf, 2008), depending on the difficulty of the task. In contrast to reactive gaze shifts, anticipatory eye movements are generally thought to rely on voluntary attentional processes and to be an early marker of attentional control. By the end of the first year infants demonstrate improvements in their ability to resolve visual attentional conflict as seen in their enhanced performance on tasks that involve switching gaze response, such as in looking versions of the A-not-B task (e.g., Cuevas & Bell, 2010; Watanabe, Forssman, Green, Bohlin, & von Hofsten, 2012). The A-not-B task requires the ability to cope with conflicting mental representations (i.e., previous hiding location vs. current hiding location) and the suppression of a previously learned response in favor of an updated prediction. Thus, successful performance includes flexibly shifting direction of attention when the target’s hiding location is switched from location A to location B, which in turn has been associated with increased activation in the frontal and parietal areas in infancy (Baird et al., 2002; Cuevas & Bell, 2011). Studies using manual search versions of the A-not-B task have shown that by the end of the first year and throughout toddlerhood, children are increasingly able to deal with longer delays between hiding and searching (Marcovitch & Zelazo, 1999), or increasing conflict (Schutte, Spencer, & Schöner, 2003), but they can still be found to make search errors on this task at pre-school age (Espy, Kaufmann, McDiarmid, & Glisky, 1999). In a recent study (Watanabe et al., 2012), we used a looking version of the A-not-B task to assess 10- and 12-monthold infants’ ability to correctly anticipate the reappearance of a hidden target during both pre- (A) and post-switch (B) trials. By using eye-tracking we were able to measure the infants’ visual attention to both the correct and incorrect location quantitatively throughout the task. The study showed that an age-related improvement in attentional control takes place between 10 and 12 months of age. This age-related improvement was particularly reflected in less perseverative anticipatory looking (i.e., less anticipatory looking at the incorrect hiding location on the B trials) in the older age group. In one condition the infants were presented with a visual distractor that preceded the reappearance of the target on the B trials and thereby increased the level of attentional conflict of the task. The infants in the visual distractor condition showed more perseverative anticipatory looking compared to the infants in the control condition where no visual distractor was presented. The result also indicated that the 12-month-olds were better than the 10-month-olds at handling the distractor, but also that older infants’ ability to overcome distraction was not yet sufficiently developed. This finding suggests that attentional control and the ability to overcome attentional conflict is still under development by the end of the first year. Thus, to further our understanding of this development it would be fruitful to further examine the developmental course of this ability. Current research suggests that marked improvements in the ability to control attention take place between 18 and 24 months of age (Clohessy, Posner, & Rothbart, 2001; Garon et al., 2008; Posner et al., 2012). In line with Colombo’s (2001) view, that research on cognitive development gains from focusing on age periods where rapid improvements take place, in the present study we focus on 18-month-old children’s ability to control attention and we contrast our results to that from adults and from our previous study on infants. For the current study we adapted our previously used looking version of the A-not-B task (Watanabe et al., 2012), which included a control and visual distractor condition. Eighteen-month-old children’s and adult’s anticipatory looking, on four A (pre-switch) and two B (post-switch) trials, was examined following a 3.5 s hiding delay. The A trials were identical, whereas a visually distracting stimulus was added in the visual distractor condition on the B trials. In addition, we included a long delay condition where the duration of the target’s hiding delay was extended to 10 s on the B trials. The purpose of adding a longer delay was to assess how the increased time needed to keep information in mind (i.e., the target’s correct hiding location) would affect the participants’ allocation of anticipatory looking. We presumed that this condition would be more challenging for the participants in terms of working memory demand compared to the control condition. Working memory is an ability that is closely linked to attentional control and these two cognitive processes rely on overlapping brain regions (e.g., parietal and prefrontal cortex) (Corbetta & Shulman, 2002; McNab & Klingberg, 2008). Whereas attentional control is important for selecting between and suppressing conflicting information, working memory is necessary for actively maintaining and retaining information (Kastner et al., 2007). To our knowledge, no previous study has examined how both distraction and an increase in time delay affect 18-month-olds’ ability to make correct anticipatory gaze predictions on a task that requires the ability to control visual attention. Considering our previous findings from 10- to 12-month-olds (Watanabe et al., 2012), we expected the 18-montholds to correctly anticipate the target’s location on the A trials and also on the B trials in the control condition (i.e., a 3.5 s empty hiding interval). In the same study Watanabe et al. (2012) found that the introduction of a distractor deteriorated anticipatory looking performance in 10–12-month-olds. We used the same distractor in the present study to investigate whether the ability to suppress this conflict has matured by 18 months of age. In the long delay condition, we predicted the 18-month-olds to display less accurate anticipatory looking on B trials compared to the children in the control condition. This could be revealed in both more perseverative anticipatory looking and/or less correct anticipatory looking. Finally, given the central role of the maturation of the attentional control system for the current task we expected that the adults would outperform the 18-month-olds across trials (A and B) and conditions. By comparing how the allocation of attention is managed for adults and for children at the age when they presumably are starting to master the task we hope to improve our understanding of the development of attentional control.
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
227
Fig. 1. (left) Screen images of the presented movies and an illustration of the areas of interest; (right) schematic diagram of A and B trials in the different conditions. Please note that the A trials where the same across participants and that a visual distractor was presented for 2 s (.5 s after the target’s disappearance) in the Distractor Condition on the B trials.
2. Methods 2.1. Participants Sixty-one typically developing 18-month-olds (M = 548.75 days, SD = 8.52 days, 38 girls and 30 boys) and thirty-six adults (M = 25.10 years; SD = 5.38 years; 20 women; 16 men; 81% undergraduate students) participated in the experiment. All 18month-olds were born full-term and lived in the area of a university town in central Sweden. An additional 18 (control condition n = 6; distractor condition n = 7; long delay condition n = 5) 18-month-olds were tested, but excluded from the analysis because of fussiness, insufficient data (i.e. less than 50% recorded gaze data of the experimental session) or technical difficulties. In the adult group, an additional three adults were tested but excluded due to technical difficulties. 2.2. Apparatus and stimuli Movie clips were presented on a 17-inch monitor that was part of a cornea reflection eye-tracking system (Tobii T120, Tobii Technology, Stockholm, Sweden). The eye-tracking system records the reflection of near infrared light from the pupil and cornea of both eyes at 60 Hz. The participants were seated at a distance of approximately 60 cm from the monitor, and at this distance the display subtended an angle of 30 × 24 visual degrees. The participants were presented with six short movie clips, interspersed with brief attention-grabbing animations. The presentation of the six movie clips was based on the A-not-B paradigm and consisted of four pre-switch (A) trials followed by two post-switch (B) trials. Screen images of the movies and schematic illustrations of trials and conditions are presented in Fig. 1. The A trials were the same across all participants to clarify that the participants understood the task (i.e., anticipated the target’s reappearance) and to establish that the equivalence of performance between the three groups (i.e., assigned conditions for the B trials: Control Condition, Visual Distractor Condition, and Long Delay Condition). At the beginning of each trial, an interesting target (a Mickey Mouse figure, subtending a 3.3 × 5.2 visual degree) was first positioned at the center of the display and then moved to disappear behind one of two occluders (A or B, subtending 8.4 × 7.2 visual degrees),
228
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
accompanied by a child-friendly instrumental melody. In the delay period, when the target was occluded, no sounds were presented. After the delay period, a sound cue (a chime) was presented and thereafter the target reappeared accompanied by the same melody as previously presented. The A and B occluders were located to the left and right of the center, and the distance between the occluders subtended a horizontal angle of 8.9 visual degrees. The locations (left or right) of the A and B occluders were counterbalanced across participants. 2.2.1. A and B trials During the first four movie clips, the target disappeared completely behind occluder A after 5.5 s. When 3.5 s had passed a sound cue was presented and 2 s later the target reappeared from behind occluder A and moved to the center of the display. The four A trials were followed by two B trials. On these trials, the target disappeared behind the opposite occluder (B) after 5.5 s. In the first B trial, the time period between presentation of the sound cue and the target’s reappearance was extended to 9 s (referred to as extended time period) in all three conditions. The extended time period was included on the first B trial to allow for measurement of anticipatory gaze following the presentation of the sound cue over a longer time interval (i.e., 9 s in contrast to 2 s) without the influence of learning. In the second B trial, the target reappeared 2 s after presentation of the sound cue as in the A trials (all conditions). 2.2.2. Conditions For the B trials the participants had been assigned to three different conditions: A control condition (n = 21 children, n = 12 adults), a visual distractor condition (n = 21 children, n = 12 adults), or a long delay condition (n = 19 children, n = 12 adults). In the control condition, no distractor was presented (i.e., an empty time interval), and we used a short time delay (3.5 s) between the target’s disappearance and presentation of the sound cue, as in the A trials. In the visual distractor condition, a visual distractor (a bouncing ball) was presented during the delay in the center of the screen for 2 s, 0.5 s after the target’s disappearance during the two B trials. The time delay between the target’s disappearance and the presentation of the sound cue was the same as in the control condition (3.5 s). In the long delay condition, the time delay between the target’s disappearance and the presentation of the sound cue was extended to 10 s (no visual distractor was presented in this condition). 2.3. Definition of areas of interest Identical to our previous study (Watanabe et al., 2012) we analyzed looking time in two areas of interest consisting of a left AOI (11.8 × 12.8 visual degrees) and right AOI (11.8 × 12.8 visual degrees). The left and right AOIs covered the left and right occluders, but were somewhat larger than the actual size of each occluder, as the participants’ anticipatory gaze could be at the border of each occluder. 2.4. Procedure Families with children of the appropriate age were contacted in a letter with a description of the study and an invitation to participate. Parents who decided to participate were contacted by telephone and an appointment was made. Once in the lab, the parent(s) received a verbal description of the study’s procedure and purpose and signed a consent form. The adult participants were recruited on the campus area of a university. They were given a verbal description of the procedure (i.e., “Following an eye-tracking calibration procedure you will be presented with short movie clips and we ask you to attend to these”) and signed a consent form before the study started. They were informed about the purpose of the study only afterwards. The study was approved by the Swedish ethics committee and was conducted in accordance with ethical standards specified in the 1964 Declaration of Helsinki. The experiment took place in a dimly lit room at the lab, and all participants were seated approximately 60 cm from the monitor. The 18-month-olds sat in a car seat on the parent’s lap or directly on the parent’s lap. Before the experiment began, a calibration procedure was performed. During the calibration, five short expanding–contracting checkerboard patterned spheres with accompanying sounds appeared, one at a time, in each corner and in the center of the screen. Any unsuccessfully calibrated point was recalibrated. After a successful calibration procedure, the experimental session started and the participant was presented with the six video clips, which all together were about 2.5 min in duration. Each participant spent approximately 10–15 min in the lab and received a token worth $13 as compensation. 3. Results Following the approach of our previous study (Watanabe et al., 2012) we analyzed absolute looking times to the correct and incorrect occluders during the time windows of interest. Descriptive data of mean anticipatory looking time at the correct and incorrect occluder during 2 s (i.e., a time window following the sound cue and before the target’s reappearance) on the A and B trials and during 9 s (the extended time period) on the first B trial are presented in Table 1 for each condition and age group. For descriptive purposes mean looking times to the central area (i.e., an aligned area between the two occluders) and looking times elsewhere (i.e., looking times elsewhere on the display and/or outside the display) are also presented in Table 1. In the analyses that follow we examine 18-month-olds’ anticipatory looking before and after the switch of hiding location in the three conditions and compare these effects to those observed in adults.
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
229
Table 1 Mean anticipatory looking time (s) at the correct and incorrect occluder, c-area (central area located between the correct and incorrect occluder), and elsewhere (on the display or outside the display) during 2 s (following the sound cue) on the A and B trials and during the extended time period (9 s divided into 3 time intervals: 0-3 s, 3-6 s, and 6-9 s, following the sound cue) on the first B trial in each condition and age group. 18-month-olds Control condition M
Adults Distractor condition
SD
M
Long delay condition SD
Distractor condition
Long delay condition
SD
M
SD
M
SD
M
SD
.82 .22 .15 .81
.35 .22 .13 .40
1.34 .20 .13 .32
.55 .25 .19 .47
1.64 .14 .04 .17
.32 .22 .05 .22
1.32 .14 .19 .35
.55 .20 .43 .37
.49 .43 .40 .59
.36 .27 .12 1.26
.44 .37 .22 .53
1.51 .11 .25 .13
.62 .12 .53 .28
1.14 .09 .56 .21
.50 .21 .46 .21
1.51 .03 .15 .31
.58 .08 .37 .34
1.05 .51 .43 1.01
.93 .79 .67 .99
.77 .66 .16 1.46
.82 .90 .28 .99
2.27 .35 .21 .17
.83 .55 .53 .62
1.72 .15 .73 .40
.86 .31 .86 .48
2.36 .07 .28 .30
.72 .16 .58 .41
.91 .58 .25 1.00
.92 .29 .30 1.51
.82 .57 .39 .99
.42 .47 .26 1.85
.57 .48 .40 .87
2.49 .27 .09 .15
.47 .35 .17 .30
1.90 .21 .57 .32
.94 .28 .86 .46
2.00 .20 .26 .54
.93 .39 .62 .68
.66 .55 .17 .94
.48 .17 .22 2.14
.73 .28 .44 .87
.53 .41 .17 1.89
.54 .41 .32 .93
2.32 .40 .13 .16
.52 .50 .17 .33
2.03 .32 .32 .33
.84 .45 .62 .42
1.97 .24 .33 .46
.90 .58 .71 .55
A trials Correct Incorrect C-area Elsewhere
.90 .27 .13 .68
.37 .27 .13 .50
.96 .22 .14 .63
.58 .20 .20 .47
B trials Correct Incorrect C-area Elsewhere
.72 .24 .08 .94
.54 .22 .11 .52
.66 .31 .39 .64
1st B trial (0-3 s) Correct Incorrect C-area Elsewhere
1.21 .49 .14 1.10
.96 .52 .24 .91
1st B trial (3-6 s) Correct Incorrect C-area Elsewhere
.92 .43 .13 1.46
1st B trial (6-9 s) Correct Incorrect C-area Elsewhere
.59 .38 .10 1.91
M
Control condition
3.1. Anticipatory looking on A trials For illustrative purposes, Fig. 2 shows the 18-month-olds’ and adults’ mean looking time at the correct and incorrect occluders continuously during the A trials (averaged across the four trials). An inspection of the figure suggests that the adults tend to keep their gaze at the correct occluder (A), whereas the children show a more distinct decrease in looking at occluder A following the target’s disappearance. A general decrease in looking at the AOIs (occluder A and B) means that the participant looked somewhere else on the screen or outside the screen. Relevant findings related to anticipatory looking time at the two occluders and age effects are presented in the analyses below. Mean looking time during the 2 s interval after the sound cue and before the reappearance of the target, averaged over the four A trials, was analyzed in a 3 × 2 × 2 mixed repeated measures analysis of variance (ANOVA) with condition (between factors; control vs. visual distractor vs. long delay; the assigned
Fig. 2. An illustration of mean looking in seconds averaged across the four A trials for the 18-month-olds (left) and the adults (right) throughout the whole trials. The solid and dotted lines represent looking time at the correct (A) and incorrect (B) occluders, respectively. Error bars represent ±standard error.
230
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Adults
18-month-olds 2,00
2,00
Occluder
Occluder
Correct Incorrect
Correct Incorrect 1,50
Mean Looking Time (s)
Mean Looking Time (s)
1,50
1,00
1,00
0,50
0,50
0,00
0,00
Control
Distractor
Long Delay
Condition
Control
Distractor
Long Delay
Condition
Fig. 3. 18-month-olds’ (left) and adults’ (right) mean anticipatory looking time (s) during 2 s after the sound cue in the B trials. The solid and dotted lines represent looking time at the correct (B) and incorrect (A) occluders, respectively. Error bars represent ±standard error.
conditions during the B trials that followed), area (within factors; A vs. B), and age group (between factors; 18-month-olds vs. adults) as independent variables. The ANOVA showed that the main effects of area, F(1, 91) = 246.85, p = .000, partial Á2 = .73, and age group, F(1, 91) = 25.96, p = .000, partial Á2 = .22, were significant. The participants looked more at the correct occluder (M = 1.10 s) than the incorrect occluder (M = .21 s) and the adults (M = 1.60 s) had higher total looking time than the children (M = 1.14 s). The ANOVA also revealed a significant interaction between area and age group, F(1, 91) = 25.12, p = .000, partial Á2 = .22, all other Fs < 2.14 and ps > .123. Planned follow-up analyses demonstrated that the adults had more anticipatory looking at the correct occluder than the children, t(95) = 5.53, p = .000, Madults = 1.44 s, M18-month-olds = .90 s, but there was no significant difference in looking at the incorrect occluder, t(95) = 1.65, p = .102, Madults = .16 s, M18-month-olds = .24 s. These results show that the participants in both age groups correctly anticipated the target’s reappearance and that the adults’ higher total looking time was reflected in more looking at the correct occluder. 3.2. Anticipatory looking on B trials We next conducted the same analysis on the B trials to clarify whether the children would correctly anticipate the target’s reappearance following the switch of hiding location and also to examine possible effects of conditions and age. As no significant interaction was found between trial (First B trial vs. Second B trial) and condition, or trial and area, in either age group (ps > .20), we examined mean anticipatory looking time during a 2 s interval immediately after the sound cue, averaged over the two B trials. The results of the ANOVA revealed significant main effects of area, F(1, 91) = 130.41, p = .000, partial Á2 = .59, and age group, F(1, 91) = 31.92, p = .000, partial Á2 = .26. The participants had more anticipatory looking at the correct occluder (B) compared to the incorrect occluder (A) and the adults displayed higher total looking time compared to the children, see Fig. 3. These main effects were modified by significant interactions between area and age group, F(1, 91) = 49.71, p = .000, partial Á2 = .35 and a borderline significant interaction between condition and age group, F(2, 91) = 3.06, p = .052, partial Á2 = .06, all other Fs < 2.05 and ps > .134. Planned follow-up analysis of the effect of age group on anticipatory looking, using one-way ANOVAs, demonstrated that the adults displayed more correct anticipatory looking, F(95) = 50.00, p = .000, and less incorrect anticipatory looking, F(95) = 10.33, p = .002, than the 18-month-olds. To further understand the interaction effects and to test our a priori hypotheses of effects of condition on the 18-month-olds’ anticipatory looking, we conducted planned comparisons of looking time at the incorrect and correct occluder separately for the age groups. One-way ANOVAs revealed no significant effects of condition on anticipatory looking in the adult age group, whereas a borderline significant effect of condition on anticipatory looking at the correct occluder was found for the 18-month-old age group, F(2, 58) = 3.08, p = .054, all other Fs < 1.68, ps > .20. As indicated by Fig. 3 the children displayed less correct looking in the long delay condition compared with the other two conditions. Pairwise comparisons revealed a significant difference for the comparison of the long delay with the control condition (p = .023) and borderline significant difference for the comparison of the long delay with the distractor condition (p = .058), all other ps > .516. These results show that the adults’ anticipatory looking was similar irrespective of condition, whereas the 18-month-olds showed less anticipatory looking at the correct occluder in the long delay condition compared to the control condition. As supplemental information, to clarify age-related improvements in 18-month-old children’s ability to suppress conflicting distraction, we analyzed data on 10-, 12- (from our previous study; Watanabe et al., 2012), and 18-month-old’s anticipatory looking to the correct and incorrect location in the distractor condition on B trials. These results suggest that age-related improvements in the ability to control attention take place between 10 and 18-months of age (see supplemental information and SI Fig. 1). 3.3. Anticipatory looking during the extended time period in the first B trial To allow for an examination of how anticipatory looking is maintained over time, the time period between the presentation of the sound cue and the target’s reappearance was extended to 9 s in the first B trial. We examined looking time to the correct and incorrect area and also possible effects of conditions, age group, and time interval. The ANOVA revealed significant main effects for area, F(1, 91) = 38.16, p = .000, partial Á2 = .30, and age group, F(1, 91) = 69.14, p = .000, partial Á2 = .43, and a borderline significant main effect for condition, F(2, 91) = 3.09, p = .050, partial Á2 = .06. The significant main effects indicated higher total looking time at the correct occluder compared to the incorrect occluder and also that the adults had overall higher total looking time compared to the 18-month-olds across the 9 s extended period (see Fig. 4). The borderline significant main effect of condition reflected a tendency to overall more looking time in the control condition compared to the other two conditions (Mcontrol = 5.50 s; Mdistractor = 4.48 s; Mlong delay = 4.61 s).
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
231
Fig. 4. 18-month-olds’ (left) and adults’ (right) mean anticipatory looking time (s) at the correct (solid line) and incorrect (dotted line) occluder in the three conditions during the extended time period (9 s, divided into three time intervals: 0–3 s, 3–6 s, and 6–9 s) in the first B trial following the presentation of the sound cue. Error bars represent ±standard error. The ANOVA also showed significant interactions between area and age group, F(1, 91) = 74.05, p = .000, partial Á2 = .45, area and time interval, F(2, 91) = 135.78, p = .000, partial Á2 = .60, and area, time interval and age group, F(1, 91) = 53.82, p = .000, partial Á2 = .37, all other Fs > 2.15 and ps < .076. To clarify these interactions and to further understand the ability of the 18-month-olds’ ability to maintain anticipatory looking during the extended time interval pairwise comparisons of looking time to the correct and incorrect occluder in the three conditions and the three time intervals were performed. The 18-month-olds in the control condition looked more to the correct occluder during the first two time intervals (0–3 s: p = .017; 3–6 s: p = .061; 6–9 s: p = .247); in the distractor condition they looked more to the correct occluder in the second time interval (0–3 s: p = .110; 3–6 s: p = .021; 6–9 s: p = .091); and in the long delay condition they looked equally to the correct and incorrect occluder in all three time intervals (0–3 s: p = .839; 3–6 s: p = .825; 6–9 s: p = .264). These findings are illustrated in Fig. 5, where it can first be seen that in all three conditions the 18-month-olds’ looking time to the correct occluder decreases after the target’s disappearance. In the control and distractor condition the children display a relative increase in looking time at the correct occluder following the presentation of the sound cue. This increase in looking time at the correct occluder is most pronounced (but slightly delayed) in the distractor condition and can be explained by the fact that the children’s gaze is drawn to the center of the screen during the presentation of the visual distractor, whereas the children in the control condition to a greater extent maintain their gaze at the correct occluder. The children in the long delay condition, on the other hand, show a more distinct increase in looking time at the incorrect occluder compared to the correct occluder following the sound cue, although the amount of looking time at the two occluders is relatively equal. Finally, we conducted follow-up analyses to clarify interactions with age-group. One-way ANOVAs on looking time at the correct and incorrect occluder in each time interval, with age group as a between-subjects factor, revealed that the 18-month-olds looked more at the incorrect occluder in the first time interval (0–3 s), and less at the correct occluder in all time intervals (0–3 s, 3–6 s, 6–9 s) compared to the adults, Fs > 7.23 and ps < .009. No significant difference was found between the age groups in looking time at the incorrect occluder in the last two time intervals (3–6 s, 6–9 s), Fs < 2.79, ps > .098 (see Fig. 4).
4. Discussion We investigated the effects of different conditions on the allocation of visual attention in a task that assesses the ability to make anticipatory gaze predictions and voluntary control of visual attention. Whereas both the 18-month-olds and adult participants made correct anticipatory gaze predictions in the pre-switch phase (A trials) of the experiment, we observed important behavioral differences in the 18-month-old group on the post-switch (B) trials. The children in the visual distractor and control conditions readily suppressed the increased attentional conflict caused by the previously learned response and looked more to the correct location. The children in the long delay condition, on the other hand, showed difficulties in correctly anticipating the target’s reappearance on the B trials. As the children’s performance in the three groups was comparable on the A trials, this result indicates that the longer time delay challenged their ability to actively keep the target’s
232
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Fig. 5. Illustrates 18-month-olds’ mean looking time in the first B trial across the first 22 s in the control (left) and distractor (middle) conditions, and across 28 s in the long delay (right) condition. The solid and dotted lines represent looking time at the correct occluder (B) and incorrect occluder (A), respectively. In the control and visual distractor conditions, the sound cue is presented 3.5 s after the target’s disappearance, whereas in the long delay condition the presentation of the sound cue take place 10 s after the target’s disappearance. In all there conditions the target reappears 9 s after the presentation of the cue signal (extended time period). Note that in the visual distractor condition a central distractor (between occluder A and B) is presented (for 2 s) before the cue signal. Error bars represent ±standard error.
new location in mind in order to correctly predict the target’s reappearance. Our data (e.g., see Figs. 2 and 5) suggest that the children’s attentional focus on the target’s location fades following the occlusion and continue to do so over time as seen during the extended hiding period in the first B trial. Whereas the children in the control and visual distractor conditions appeared to recover their attentional focus on the correct occluder rapidly at the presentation of the sound cue, this recovery was not clearly displayed by the children in the long delay condition. Notably, the data from the first B trial indicate that even without the long delay, that is, in the control and distractor conditions the children had difficulty maintaining attention beyond the first 6 s after the sound cue. These data could be interpreted to mean that the attention span of 18-month-olds in a rather eventless situation is about 6 s. The finding that the children in the visual distractor condition were able to overcome the increased attentional conflict caused by the distractor and performed equally to the children in the control condition is in contrast to our results from 10- and 12-month-olds’ performance on the same task (Watanabe et al., 2012). In our previous study, the infants in the visual distractor condition, particularly the 10-month-olds, made more incorrect (perseverative) anticipatory gaze predictions. Taken together, these two studies indicate that children acquire better control over their visual attention between 12 and 18 months of age. This conclusion is further supported by our supplementary analyses of looking at the correct and incorrect occluder within the distractor condition, which indicates a development from 10 months of age, toward a decrease in perseveration, but little improvement in correct anticipatory gaze predictions at 12 months of age, to correct anticipatory looking at 18 months of age. This development is interpreted to involve improvements in an inhibitory process of the attention system, which enables suppressing a gazing response toward the incorrect hiding location on the B trials (c.f., Watanabe et al., 2012), as well as improvements in the ability to update a hidden target’s location in order to correctly anticipate the target’s reappearance following a brief time delay. Importantly, the continuous measure used in this and our previous study (Watanabe et al., 2012), yields data for anticipatory gaze indicating that developmental improvements in voluntary control of visual attention is better described as gradual than as dichotomous. One implication of this suggestion, from a methodological perspective, is that research on young children’s development of attentional control (or cognitive development in general) can gain new insights from using more sensitive measures (e.g., continuous measures) of their performance, rather than using outcome measures that define behavioral improvements as something children either can do or cannot do. The behavioral improvement in attentional control is presumably closely tied to the neurological maturation of the attention control system, such as the increase in connectivity between the frontal cortex and anterior cingulate gyrus during the second year of life (e.g. Gao et al., 2013, 2009; Posner & Fan, 2008), and children’s growing explorations of their environment through independent locomotion by the end of the first year and forward. This view is clearly in line with our demonstration of a gradual improvement in infants’ attentional control. The developmental literature (e.g., Rothbart & Posner, 2001) suggests that important improvements in the attention systems are in place around 18 to 24 months of age, particularly in terms of suppressing competing information that interferes with the task at hand. The results of the present study would date these developmental changes to the period just before 18 months, but longitudinal data are necessary if we are to improve our understanding of this development. Our data also shed light on the 18-month-olds’ developmental level of attentional control through our comparison with adults’ performance on the same task. With the exception of the performance of the children in the long delay condition on the B trials, the 18-month-old’s allocation of visual attention was strikingly similar to that of the adults. However, a difference between the 18-month-olds’ and the adult participants’ performance was seen in the adults’ longer looking times
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
233
throughout the experiment, mainly reflected in more correct anticipatory looking on both A and B trials. As to incorrect looking and the tendency to perseverate under conditions of conflict, it should perhaps be noted that whereas there was no difference between the adults’ and the children’s looking at the incorrect occluder on the A trials, the 18-month-olds looked more than the adults at the incorrect occluder on B trials. This may be interpreted to mean that the 18-month-olds encountered more problems in switching than did the adults even though they did not perseverate in the general sense of showing more incorrect than correct looking. This view is further supported by looking at the results from the perspective of proportion of “on-task-behavior”, which suggests that while the higher amount of looking at the correct occluder for the adults to a large extent could be explained by the larger amount of on-task-behavior, their looking at the incorrect occluder would be even lower (e.g., see Figs. 2 and 3). The ability to maintain attention on the task is clearly not fully matured by 18 months of age and this was particularly evident during the time periods of the experiment when the target was occluded and no sound or moving stimuli were presented on the screen, such as during the long delay between the target’s disappearance and the presentation of the sound cue signaling the target’s reappearance. The abilities to control and flexibly allocate visual attention, tested in the current study, are believed to lay the foundation for more complex cognitive skills and are essential in order to predict future events, act in a goal-directed manner, and in interactions with the external world. Because a compromised development of attentional control has been suggested to be an important risk factor for the development of attentional disorders (e.g., Johnson, 2012), it is of great importance that research continues to clarify the developmental process of attentional control. If we can improve our insights into the early development of attention and how it can go astray, it will also improve our possibilities of providing the right measures for treating children with attention deficits. To summarize, our study has demonstrated that by 18 months of age, children can efficiently suppress distracting information and make correct anticipatory predictions, even following a switch in the requirements of the task. This finding, together with data from our previous study on 10- and 12-month-olds (Watanabe et al., 2012), demonstrate that important age-related improvements in attentional control are taking place in a gradual fashion from the end of the first year to 18 months of age. Further, by 18 months of age children’s performance show an adult-like pattern, the main difference being that adults demonstrate a better ability to maintain attention on the task and to update their predictions following a longer time period. Acknowledgements We gratefully acknowledge the efforts of the children, their parents, and the adult participants who took part in the studies. We thank Stefan Boström for assistance in data collection. This research was financially supported by the Swedish Research Council. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.infbeh. 2014.01.007. References Baird, A. A., Kagan, J., Gaudette, T., Walz, K. A., Hershlag, N., & Boas, D. A. (2002). Frontal lobe activation during object permanence: Data from near-infrared spectroscopy. NeuroImage: 16., 120–126. Clohessy, A. B., Posner, M. I., & Rothbart, M. K. (2001). Development of the functional visual field. Acta Psychologica: 106., 51–68. Colombo, J. (2001). The development of visual attention in infancy. Annual Review of Psychology: 52., 337–367. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience: 3., 201–215. Cuevas, K., & Bell, M. A. (2010). Developmental progression of looking and reaching performance on the A-Not-B task. Developmental Psychology: 46., 1363–1371. Cuevas, K., & Bell, M. A. (2011). EEG and ECG from 5 to 10 months of age: Developmental changes in baseline activation and cognitive processing during a working memory task. International Journal of Psychophysiology: 80., 119–128. Espy, K. A., Kaufmann, P. M., McDiarmid, M. D., & Glisky, M. L. (1999). Executive functioning in preschool children: Performance on A-not-B and other delayed response format tasks. Brain and Cognition: 41., 178–199. Gao, W., Gilmore, J. H., Dinggang, S., Smith, J. K., Zhu, H., & Lin, W. (2013). The synchronization within and interaction between the default and dorsal attention networks in early infancy. Cerebral Cortex: 23., 594–603. Gao, W., Zhu, H., Giovanello, K. S., Smith, J. K., Shen, D., Gilmore, J. H., et al. (2009). Evidence of the emergence of the brain’s default network from 2-week-old to 2-year-old healthy pediatric subjects. Proceedings of the National Academy of Sciences of the United States of America: 106., 6790–6795. Garon, N., Bryson, S. E., & Smith, I. M. (2008). Executive function in preschoolers: A review using an integrative framework. Psychological Bulletin: 134., 31–60. Johnson, M. H. (2012). Executive function and developmental disorders: The flip side of the coin. Trends in Cognitive Science: 16., 454–457. Johnson, M. H., Posner, M. I., & Rothbart, M. K. (1991). Components of visual orienting in early infancy: Contingency learning, anticipatory looking and disengaging. Journal of Cognitive Neuroscience: 3., 335–344. Kastner, S., DeSimone, K., Konen, C. S., Szczepanski, S. M., Weiner, K. S., & Schneider, K. A. (2007). Topographic maps in human frontal cortex revealed in memory-guided saccade and spatial working-memory tasks. Journal of Neurophysiology: 97., 3494–3507. Marcovitch, S., & Zelazo, P. D. (1999). The A-not-B error: Results from a logistic meta-analysis. Child Development: 70., 1297–1313. McNab, F., & Klingberg, T. (2008). Prefrontal cortex and basal ganglia control access to working memory. Nature Neuroscience: 11., 103–107. Nelson, K. E. (1971). Accommodation of visual tracking patterns in human infants to object movements patterns. Journal of Experimental Child Psychology: 12., 182–196. Petersen, S. E., & Posner, M. I. (2012). The attention system of the human brain: 20 years after. Annual Review of Neuroscience: 35., 73–89.
234
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Posner, M. I., & Fan, J. (2008). Attention as an organ system. In J. R. Pomerantz (Ed.), Topics in integrative neuroscience (pp. 31–61). New York; NY: Cambridge University Press. Posner, M. I., Rothbart, M., Sheese, B. E., & Voelker, P. (2012). Control networks and neuromodulators of early development. Developmental Psychology: 48., 827–835. Rothbart, M., & Posner, M. (2001). Mechanism and variation in the development of attentional networks. In C. Nelson, & Luciana (Eds.), Handbook of developmental cognitive neuroscience (pp. 353–363). Cambridge, MA: MIT Press. Salley, B., Panneton, R. K., & Colombo, J. (2012). Separable attentional predictors of language outcome. Infancy, http://dx.doi.org/10.1111/ j.1532-7078.2012.00138.x Schutte, A. R., Spencer, J. P., & Schöner, G. (2003). Testing the dynamic field theory: Working memory for locations becomes more spatially precise over development. Child Development: 74., 1393–1417. Sheese, B. E., Rothbart, M. K., Posner, M. I., White, L. K., & Fraundorf, S. H. (2008). Executive attention and self-regulation in infancy. Infant Behavior and Development: 31., 501–510. Uddin, L., Supekar, K., & Menon, V. (2010). Typical and atypical development of functional human brain networks: Insights from resting-state fMRI. Frontiers in System Neuroscience: 4., http://dx.doi.org/10.3389/fnsys.2010.00021 Watanabe, H., Forssman, L., Green, D., Bohlin, G., & von Hofsten, C. (2012). Attention demands influence 10- and 12-month-old infants’ perseverative behavior. Developmental Psychology: 48., 44–56.
Contents lists available at ScienceDirect
Infant Behavior and Development
Eighteen-month-olds’ ability to make gaze predictions following distraction or a long delay Linda Forssman ∗ , Gunilla Bohlin, Claes von Hofsten Department of Psychology, Uppsala University, Sweden
a r t i c l e
i n f o
Article history: Received 3 September 2013 Received in revised form 16 December 2013 Accepted 24 January 2014 Available online 15 March 2014
Keywords: Attentional control Anticipatory gaze Children Eye-tracking
a b s t r a c t The abilities to flexibly allocate attention, select between conflicting stimuli, and make anticipatory gaze movements are important for young children’s exploration and learning about their environment. These abilities constitute voluntary control of attention and show marked improvements in the second year of a child’s life. Here we investigate the effects of visual distraction and delay on 18-month-olds’ ability to predict the location of an occluded target in an experiment that requires switching of attention, and compare their performance to that of adults. Our results demonstrate that by 18 months of age children can readily overcome a previously learned response, even under a condition that involves visual distraction, but have difficulties with correctly updating their prediction when presented with a longer time delay. Further, the experiment shows that, overall, the 18-month-olds’ allocation of visual attention is similar to that of adults, the primary difference being that adults demonstrate a superior ability to maintain attention on task and update their predictions over a longer time period. © 2014 Elsevier Inc. All rights reserved.
1. Introduction In the first years of life, the development of attentional control provides an important mechanism for children’s exploration and learning about their environment. Attentional control involves the ability to flexibly allocate attention and suppress conflicting stimuli that interfere with the task at hand. The early development of this ability has been suggested to underpin the development of more complex skills, such as emotional and social regulation (Posner, Rothbart, Sheese, & Voelker, 2012) executive functions (e.g., working memory and inhibition; see Garon, Bryson, & Smith, 2008 for a review), and language development (Salley, Panneton, & Colombo, 2012). Further, deficits in the early development of attentional control may be related to increased risk for developmental disorders (e.g., attention deficit hyperactivity disorder and autism spectrum disorder) (Johnson, 2012). It is generally believed that brain maturation and increased functional connectivity are accompanying developmental improvements in behavioral control of attention. Resting-state fMRI studies indicate that the neural networks, supporting basic forms of attention control, become functional in infancy. They show a strong increase in connectivity over the first 2 years of life, and follow different developmental trajectories (Gao et al., 2013, 2009; Uddin, Supekar, & Menon, 2010). Functional connectivity between brain areas that support resolution of attentional conflict, such as the suppression of interfering stimuli (e.g., connectivity between the frontal cortex and anterior cingulate gyrus), emerges after 6 months of age and has a protracted development that lasts throughout childhood (Petersen & Posner, 2012; Posner & Fan, 2008).
∗ Corresponding author. Current address: Tampere Center for Child Health Research, School of Medicine, University of Tampere, FIN-33014 Tampere, Finland. Tel.: +358 40 190 1358; fax: +358 3 213 4473. E-mail address: Linda.Forssman@uta.fi (L. Forssman). 0163-6383/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.infbeh.2014.01.007
226
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
One method to study attentional control in infancy and toddlerhood is the assessment of anticipatory gaze, that is, gaze shifts occurring prior to the presentation of an event or stimulus. Studies measuring eye movements have shown that infants can make anticipatory gaze shifts around 4–9 months of age (Johnson, Posner, & Rothbart, 1991; Nelson, 1971; Sheese, Rothbart, Posner, White, & Fraundorf, 2008), depending on the difficulty of the task. In contrast to reactive gaze shifts, anticipatory eye movements are generally thought to rely on voluntary attentional processes and to be an early marker of attentional control. By the end of the first year infants demonstrate improvements in their ability to resolve visual attentional conflict as seen in their enhanced performance on tasks that involve switching gaze response, such as in looking versions of the A-not-B task (e.g., Cuevas & Bell, 2010; Watanabe, Forssman, Green, Bohlin, & von Hofsten, 2012). The A-not-B task requires the ability to cope with conflicting mental representations (i.e., previous hiding location vs. current hiding location) and the suppression of a previously learned response in favor of an updated prediction. Thus, successful performance includes flexibly shifting direction of attention when the target’s hiding location is switched from location A to location B, which in turn has been associated with increased activation in the frontal and parietal areas in infancy (Baird et al., 2002; Cuevas & Bell, 2011). Studies using manual search versions of the A-not-B task have shown that by the end of the first year and throughout toddlerhood, children are increasingly able to deal with longer delays between hiding and searching (Marcovitch & Zelazo, 1999), or increasing conflict (Schutte, Spencer, & Schöner, 2003), but they can still be found to make search errors on this task at pre-school age (Espy, Kaufmann, McDiarmid, & Glisky, 1999). In a recent study (Watanabe et al., 2012), we used a looking version of the A-not-B task to assess 10- and 12-monthold infants’ ability to correctly anticipate the reappearance of a hidden target during both pre- (A) and post-switch (B) trials. By using eye-tracking we were able to measure the infants’ visual attention to both the correct and incorrect location quantitatively throughout the task. The study showed that an age-related improvement in attentional control takes place between 10 and 12 months of age. This age-related improvement was particularly reflected in less perseverative anticipatory looking (i.e., less anticipatory looking at the incorrect hiding location on the B trials) in the older age group. In one condition the infants were presented with a visual distractor that preceded the reappearance of the target on the B trials and thereby increased the level of attentional conflict of the task. The infants in the visual distractor condition showed more perseverative anticipatory looking compared to the infants in the control condition where no visual distractor was presented. The result also indicated that the 12-month-olds were better than the 10-month-olds at handling the distractor, but also that older infants’ ability to overcome distraction was not yet sufficiently developed. This finding suggests that attentional control and the ability to overcome attentional conflict is still under development by the end of the first year. Thus, to further our understanding of this development it would be fruitful to further examine the developmental course of this ability. Current research suggests that marked improvements in the ability to control attention take place between 18 and 24 months of age (Clohessy, Posner, & Rothbart, 2001; Garon et al., 2008; Posner et al., 2012). In line with Colombo’s (2001) view, that research on cognitive development gains from focusing on age periods where rapid improvements take place, in the present study we focus on 18-month-old children’s ability to control attention and we contrast our results to that from adults and from our previous study on infants. For the current study we adapted our previously used looking version of the A-not-B task (Watanabe et al., 2012), which included a control and visual distractor condition. Eighteen-month-old children’s and adult’s anticipatory looking, on four A (pre-switch) and two B (post-switch) trials, was examined following a 3.5 s hiding delay. The A trials were identical, whereas a visually distracting stimulus was added in the visual distractor condition on the B trials. In addition, we included a long delay condition where the duration of the target’s hiding delay was extended to 10 s on the B trials. The purpose of adding a longer delay was to assess how the increased time needed to keep information in mind (i.e., the target’s correct hiding location) would affect the participants’ allocation of anticipatory looking. We presumed that this condition would be more challenging for the participants in terms of working memory demand compared to the control condition. Working memory is an ability that is closely linked to attentional control and these two cognitive processes rely on overlapping brain regions (e.g., parietal and prefrontal cortex) (Corbetta & Shulman, 2002; McNab & Klingberg, 2008). Whereas attentional control is important for selecting between and suppressing conflicting information, working memory is necessary for actively maintaining and retaining information (Kastner et al., 2007). To our knowledge, no previous study has examined how both distraction and an increase in time delay affect 18-month-olds’ ability to make correct anticipatory gaze predictions on a task that requires the ability to control visual attention. Considering our previous findings from 10- to 12-month-olds (Watanabe et al., 2012), we expected the 18-montholds to correctly anticipate the target’s location on the A trials and also on the B trials in the control condition (i.e., a 3.5 s empty hiding interval). In the same study Watanabe et al. (2012) found that the introduction of a distractor deteriorated anticipatory looking performance in 10–12-month-olds. We used the same distractor in the present study to investigate whether the ability to suppress this conflict has matured by 18 months of age. In the long delay condition, we predicted the 18-month-olds to display less accurate anticipatory looking on B trials compared to the children in the control condition. This could be revealed in both more perseverative anticipatory looking and/or less correct anticipatory looking. Finally, given the central role of the maturation of the attentional control system for the current task we expected that the adults would outperform the 18-month-olds across trials (A and B) and conditions. By comparing how the allocation of attention is managed for adults and for children at the age when they presumably are starting to master the task we hope to improve our understanding of the development of attentional control.
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
227
Fig. 1. (left) Screen images of the presented movies and an illustration of the areas of interest; (right) schematic diagram of A and B trials in the different conditions. Please note that the A trials where the same across participants and that a visual distractor was presented for 2 s (.5 s after the target’s disappearance) in the Distractor Condition on the B trials.
2. Methods 2.1. Participants Sixty-one typically developing 18-month-olds (M = 548.75 days, SD = 8.52 days, 38 girls and 30 boys) and thirty-six adults (M = 25.10 years; SD = 5.38 years; 20 women; 16 men; 81% undergraduate students) participated in the experiment. All 18month-olds were born full-term and lived in the area of a university town in central Sweden. An additional 18 (control condition n = 6; distractor condition n = 7; long delay condition n = 5) 18-month-olds were tested, but excluded from the analysis because of fussiness, insufficient data (i.e. less than 50% recorded gaze data of the experimental session) or technical difficulties. In the adult group, an additional three adults were tested but excluded due to technical difficulties. 2.2. Apparatus and stimuli Movie clips were presented on a 17-inch monitor that was part of a cornea reflection eye-tracking system (Tobii T120, Tobii Technology, Stockholm, Sweden). The eye-tracking system records the reflection of near infrared light from the pupil and cornea of both eyes at 60 Hz. The participants were seated at a distance of approximately 60 cm from the monitor, and at this distance the display subtended an angle of 30 × 24 visual degrees. The participants were presented with six short movie clips, interspersed with brief attention-grabbing animations. The presentation of the six movie clips was based on the A-not-B paradigm and consisted of four pre-switch (A) trials followed by two post-switch (B) trials. Screen images of the movies and schematic illustrations of trials and conditions are presented in Fig. 1. The A trials were the same across all participants to clarify that the participants understood the task (i.e., anticipated the target’s reappearance) and to establish that the equivalence of performance between the three groups (i.e., assigned conditions for the B trials: Control Condition, Visual Distractor Condition, and Long Delay Condition). At the beginning of each trial, an interesting target (a Mickey Mouse figure, subtending a 3.3 × 5.2 visual degree) was first positioned at the center of the display and then moved to disappear behind one of two occluders (A or B, subtending 8.4 × 7.2 visual degrees),
228
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
accompanied by a child-friendly instrumental melody. In the delay period, when the target was occluded, no sounds were presented. After the delay period, a sound cue (a chime) was presented and thereafter the target reappeared accompanied by the same melody as previously presented. The A and B occluders were located to the left and right of the center, and the distance between the occluders subtended a horizontal angle of 8.9 visual degrees. The locations (left or right) of the A and B occluders were counterbalanced across participants. 2.2.1. A and B trials During the first four movie clips, the target disappeared completely behind occluder A after 5.5 s. When 3.5 s had passed a sound cue was presented and 2 s later the target reappeared from behind occluder A and moved to the center of the display. The four A trials were followed by two B trials. On these trials, the target disappeared behind the opposite occluder (B) after 5.5 s. In the first B trial, the time period between presentation of the sound cue and the target’s reappearance was extended to 9 s (referred to as extended time period) in all three conditions. The extended time period was included on the first B trial to allow for measurement of anticipatory gaze following the presentation of the sound cue over a longer time interval (i.e., 9 s in contrast to 2 s) without the influence of learning. In the second B trial, the target reappeared 2 s after presentation of the sound cue as in the A trials (all conditions). 2.2.2. Conditions For the B trials the participants had been assigned to three different conditions: A control condition (n = 21 children, n = 12 adults), a visual distractor condition (n = 21 children, n = 12 adults), or a long delay condition (n = 19 children, n = 12 adults). In the control condition, no distractor was presented (i.e., an empty time interval), and we used a short time delay (3.5 s) between the target’s disappearance and presentation of the sound cue, as in the A trials. In the visual distractor condition, a visual distractor (a bouncing ball) was presented during the delay in the center of the screen for 2 s, 0.5 s after the target’s disappearance during the two B trials. The time delay between the target’s disappearance and the presentation of the sound cue was the same as in the control condition (3.5 s). In the long delay condition, the time delay between the target’s disappearance and the presentation of the sound cue was extended to 10 s (no visual distractor was presented in this condition). 2.3. Definition of areas of interest Identical to our previous study (Watanabe et al., 2012) we analyzed looking time in two areas of interest consisting of a left AOI (11.8 × 12.8 visual degrees) and right AOI (11.8 × 12.8 visual degrees). The left and right AOIs covered the left and right occluders, but were somewhat larger than the actual size of each occluder, as the participants’ anticipatory gaze could be at the border of each occluder. 2.4. Procedure Families with children of the appropriate age were contacted in a letter with a description of the study and an invitation to participate. Parents who decided to participate were contacted by telephone and an appointment was made. Once in the lab, the parent(s) received a verbal description of the study’s procedure and purpose and signed a consent form. The adult participants were recruited on the campus area of a university. They were given a verbal description of the procedure (i.e., “Following an eye-tracking calibration procedure you will be presented with short movie clips and we ask you to attend to these”) and signed a consent form before the study started. They were informed about the purpose of the study only afterwards. The study was approved by the Swedish ethics committee and was conducted in accordance with ethical standards specified in the 1964 Declaration of Helsinki. The experiment took place in a dimly lit room at the lab, and all participants were seated approximately 60 cm from the monitor. The 18-month-olds sat in a car seat on the parent’s lap or directly on the parent’s lap. Before the experiment began, a calibration procedure was performed. During the calibration, five short expanding–contracting checkerboard patterned spheres with accompanying sounds appeared, one at a time, in each corner and in the center of the screen. Any unsuccessfully calibrated point was recalibrated. After a successful calibration procedure, the experimental session started and the participant was presented with the six video clips, which all together were about 2.5 min in duration. Each participant spent approximately 10–15 min in the lab and received a token worth $13 as compensation. 3. Results Following the approach of our previous study (Watanabe et al., 2012) we analyzed absolute looking times to the correct and incorrect occluders during the time windows of interest. Descriptive data of mean anticipatory looking time at the correct and incorrect occluder during 2 s (i.e., a time window following the sound cue and before the target’s reappearance) on the A and B trials and during 9 s (the extended time period) on the first B trial are presented in Table 1 for each condition and age group. For descriptive purposes mean looking times to the central area (i.e., an aligned area between the two occluders) and looking times elsewhere (i.e., looking times elsewhere on the display and/or outside the display) are also presented in Table 1. In the analyses that follow we examine 18-month-olds’ anticipatory looking before and after the switch of hiding location in the three conditions and compare these effects to those observed in adults.
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
229
Table 1 Mean anticipatory looking time (s) at the correct and incorrect occluder, c-area (central area located between the correct and incorrect occluder), and elsewhere (on the display or outside the display) during 2 s (following the sound cue) on the A and B trials and during the extended time period (9 s divided into 3 time intervals: 0-3 s, 3-6 s, and 6-9 s, following the sound cue) on the first B trial in each condition and age group. 18-month-olds Control condition M
Adults Distractor condition
SD
M
Long delay condition SD
Distractor condition
Long delay condition
SD
M
SD
M
SD
M
SD
.82 .22 .15 .81
.35 .22 .13 .40
1.34 .20 .13 .32
.55 .25 .19 .47
1.64 .14 .04 .17
.32 .22 .05 .22
1.32 .14 .19 .35
.55 .20 .43 .37
.49 .43 .40 .59
.36 .27 .12 1.26
.44 .37 .22 .53
1.51 .11 .25 .13
.62 .12 .53 .28
1.14 .09 .56 .21
.50 .21 .46 .21
1.51 .03 .15 .31
.58 .08 .37 .34
1.05 .51 .43 1.01
.93 .79 .67 .99
.77 .66 .16 1.46
.82 .90 .28 .99
2.27 .35 .21 .17
.83 .55 .53 .62
1.72 .15 .73 .40
.86 .31 .86 .48
2.36 .07 .28 .30
.72 .16 .58 .41
.91 .58 .25 1.00
.92 .29 .30 1.51
.82 .57 .39 .99
.42 .47 .26 1.85
.57 .48 .40 .87
2.49 .27 .09 .15
.47 .35 .17 .30
1.90 .21 .57 .32
.94 .28 .86 .46
2.00 .20 .26 .54
.93 .39 .62 .68
.66 .55 .17 .94
.48 .17 .22 2.14
.73 .28 .44 .87
.53 .41 .17 1.89
.54 .41 .32 .93
2.32 .40 .13 .16
.52 .50 .17 .33
2.03 .32 .32 .33
.84 .45 .62 .42
1.97 .24 .33 .46
.90 .58 .71 .55
A trials Correct Incorrect C-area Elsewhere
.90 .27 .13 .68
.37 .27 .13 .50
.96 .22 .14 .63
.58 .20 .20 .47
B trials Correct Incorrect C-area Elsewhere
.72 .24 .08 .94
.54 .22 .11 .52
.66 .31 .39 .64
1st B trial (0-3 s) Correct Incorrect C-area Elsewhere
1.21 .49 .14 1.10
.96 .52 .24 .91
1st B trial (3-6 s) Correct Incorrect C-area Elsewhere
.92 .43 .13 1.46
1st B trial (6-9 s) Correct Incorrect C-area Elsewhere
.59 .38 .10 1.91
M
Control condition
3.1. Anticipatory looking on A trials For illustrative purposes, Fig. 2 shows the 18-month-olds’ and adults’ mean looking time at the correct and incorrect occluders continuously during the A trials (averaged across the four trials). An inspection of the figure suggests that the adults tend to keep their gaze at the correct occluder (A), whereas the children show a more distinct decrease in looking at occluder A following the target’s disappearance. A general decrease in looking at the AOIs (occluder A and B) means that the participant looked somewhere else on the screen or outside the screen. Relevant findings related to anticipatory looking time at the two occluders and age effects are presented in the analyses below. Mean looking time during the 2 s interval after the sound cue and before the reappearance of the target, averaged over the four A trials, was analyzed in a 3 × 2 × 2 mixed repeated measures analysis of variance (ANOVA) with condition (between factors; control vs. visual distractor vs. long delay; the assigned
Fig. 2. An illustration of mean looking in seconds averaged across the four A trials for the 18-month-olds (left) and the adults (right) throughout the whole trials. The solid and dotted lines represent looking time at the correct (A) and incorrect (B) occluders, respectively. Error bars represent ±standard error.
230
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Adults
18-month-olds 2,00
2,00
Occluder
Occluder
Correct Incorrect
Correct Incorrect 1,50
Mean Looking Time (s)
Mean Looking Time (s)
1,50
1,00
1,00
0,50
0,50
0,00
0,00
Control
Distractor
Long Delay
Condition
Control
Distractor
Long Delay
Condition
Fig. 3. 18-month-olds’ (left) and adults’ (right) mean anticipatory looking time (s) during 2 s after the sound cue in the B trials. The solid and dotted lines represent looking time at the correct (B) and incorrect (A) occluders, respectively. Error bars represent ±standard error.
conditions during the B trials that followed), area (within factors; A vs. B), and age group (between factors; 18-month-olds vs. adults) as independent variables. The ANOVA showed that the main effects of area, F(1, 91) = 246.85, p = .000, partial Á2 = .73, and age group, F(1, 91) = 25.96, p = .000, partial Á2 = .22, were significant. The participants looked more at the correct occluder (M = 1.10 s) than the incorrect occluder (M = .21 s) and the adults (M = 1.60 s) had higher total looking time than the children (M = 1.14 s). The ANOVA also revealed a significant interaction between area and age group, F(1, 91) = 25.12, p = .000, partial Á2 = .22, all other Fs < 2.14 and ps > .123. Planned follow-up analyses demonstrated that the adults had more anticipatory looking at the correct occluder than the children, t(95) = 5.53, p = .000, Madults = 1.44 s, M18-month-olds = .90 s, but there was no significant difference in looking at the incorrect occluder, t(95) = 1.65, p = .102, Madults = .16 s, M18-month-olds = .24 s. These results show that the participants in both age groups correctly anticipated the target’s reappearance and that the adults’ higher total looking time was reflected in more looking at the correct occluder. 3.2. Anticipatory looking on B trials We next conducted the same analysis on the B trials to clarify whether the children would correctly anticipate the target’s reappearance following the switch of hiding location and also to examine possible effects of conditions and age. As no significant interaction was found between trial (First B trial vs. Second B trial) and condition, or trial and area, in either age group (ps > .20), we examined mean anticipatory looking time during a 2 s interval immediately after the sound cue, averaged over the two B trials. The results of the ANOVA revealed significant main effects of area, F(1, 91) = 130.41, p = .000, partial Á2 = .59, and age group, F(1, 91) = 31.92, p = .000, partial Á2 = .26. The participants had more anticipatory looking at the correct occluder (B) compared to the incorrect occluder (A) and the adults displayed higher total looking time compared to the children, see Fig. 3. These main effects were modified by significant interactions between area and age group, F(1, 91) = 49.71, p = .000, partial Á2 = .35 and a borderline significant interaction between condition and age group, F(2, 91) = 3.06, p = .052, partial Á2 = .06, all other Fs < 2.05 and ps > .134. Planned follow-up analysis of the effect of age group on anticipatory looking, using one-way ANOVAs, demonstrated that the adults displayed more correct anticipatory looking, F(95) = 50.00, p = .000, and less incorrect anticipatory looking, F(95) = 10.33, p = .002, than the 18-month-olds. To further understand the interaction effects and to test our a priori hypotheses of effects of condition on the 18-month-olds’ anticipatory looking, we conducted planned comparisons of looking time at the incorrect and correct occluder separately for the age groups. One-way ANOVAs revealed no significant effects of condition on anticipatory looking in the adult age group, whereas a borderline significant effect of condition on anticipatory looking at the correct occluder was found for the 18-month-old age group, F(2, 58) = 3.08, p = .054, all other Fs < 1.68, ps > .20. As indicated by Fig. 3 the children displayed less correct looking in the long delay condition compared with the other two conditions. Pairwise comparisons revealed a significant difference for the comparison of the long delay with the control condition (p = .023) and borderline significant difference for the comparison of the long delay with the distractor condition (p = .058), all other ps > .516. These results show that the adults’ anticipatory looking was similar irrespective of condition, whereas the 18-month-olds showed less anticipatory looking at the correct occluder in the long delay condition compared to the control condition. As supplemental information, to clarify age-related improvements in 18-month-old children’s ability to suppress conflicting distraction, we analyzed data on 10-, 12- (from our previous study; Watanabe et al., 2012), and 18-month-old’s anticipatory looking to the correct and incorrect location in the distractor condition on B trials. These results suggest that age-related improvements in the ability to control attention take place between 10 and 18-months of age (see supplemental information and SI Fig. 1). 3.3. Anticipatory looking during the extended time period in the first B trial To allow for an examination of how anticipatory looking is maintained over time, the time period between the presentation of the sound cue and the target’s reappearance was extended to 9 s in the first B trial. We examined looking time to the correct and incorrect area and also possible effects of conditions, age group, and time interval. The ANOVA revealed significant main effects for area, F(1, 91) = 38.16, p = .000, partial Á2 = .30, and age group, F(1, 91) = 69.14, p = .000, partial Á2 = .43, and a borderline significant main effect for condition, F(2, 91) = 3.09, p = .050, partial Á2 = .06. The significant main effects indicated higher total looking time at the correct occluder compared to the incorrect occluder and also that the adults had overall higher total looking time compared to the 18-month-olds across the 9 s extended period (see Fig. 4). The borderline significant main effect of condition reflected a tendency to overall more looking time in the control condition compared to the other two conditions (Mcontrol = 5.50 s; Mdistractor = 4.48 s; Mlong delay = 4.61 s).
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
231
Fig. 4. 18-month-olds’ (left) and adults’ (right) mean anticipatory looking time (s) at the correct (solid line) and incorrect (dotted line) occluder in the three conditions during the extended time period (9 s, divided into three time intervals: 0–3 s, 3–6 s, and 6–9 s) in the first B trial following the presentation of the sound cue. Error bars represent ±standard error. The ANOVA also showed significant interactions between area and age group, F(1, 91) = 74.05, p = .000, partial Á2 = .45, area and time interval, F(2, 91) = 135.78, p = .000, partial Á2 = .60, and area, time interval and age group, F(1, 91) = 53.82, p = .000, partial Á2 = .37, all other Fs > 2.15 and ps < .076. To clarify these interactions and to further understand the ability of the 18-month-olds’ ability to maintain anticipatory looking during the extended time interval pairwise comparisons of looking time to the correct and incorrect occluder in the three conditions and the three time intervals were performed. The 18-month-olds in the control condition looked more to the correct occluder during the first two time intervals (0–3 s: p = .017; 3–6 s: p = .061; 6–9 s: p = .247); in the distractor condition they looked more to the correct occluder in the second time interval (0–3 s: p = .110; 3–6 s: p = .021; 6–9 s: p = .091); and in the long delay condition they looked equally to the correct and incorrect occluder in all three time intervals (0–3 s: p = .839; 3–6 s: p = .825; 6–9 s: p = .264). These findings are illustrated in Fig. 5, where it can first be seen that in all three conditions the 18-month-olds’ looking time to the correct occluder decreases after the target’s disappearance. In the control and distractor condition the children display a relative increase in looking time at the correct occluder following the presentation of the sound cue. This increase in looking time at the correct occluder is most pronounced (but slightly delayed) in the distractor condition and can be explained by the fact that the children’s gaze is drawn to the center of the screen during the presentation of the visual distractor, whereas the children in the control condition to a greater extent maintain their gaze at the correct occluder. The children in the long delay condition, on the other hand, show a more distinct increase in looking time at the incorrect occluder compared to the correct occluder following the sound cue, although the amount of looking time at the two occluders is relatively equal. Finally, we conducted follow-up analyses to clarify interactions with age-group. One-way ANOVAs on looking time at the correct and incorrect occluder in each time interval, with age group as a between-subjects factor, revealed that the 18-month-olds looked more at the incorrect occluder in the first time interval (0–3 s), and less at the correct occluder in all time intervals (0–3 s, 3–6 s, 6–9 s) compared to the adults, Fs > 7.23 and ps < .009. No significant difference was found between the age groups in looking time at the incorrect occluder in the last two time intervals (3–6 s, 6–9 s), Fs < 2.79, ps > .098 (see Fig. 4).
4. Discussion We investigated the effects of different conditions on the allocation of visual attention in a task that assesses the ability to make anticipatory gaze predictions and voluntary control of visual attention. Whereas both the 18-month-olds and adult participants made correct anticipatory gaze predictions in the pre-switch phase (A trials) of the experiment, we observed important behavioral differences in the 18-month-old group on the post-switch (B) trials. The children in the visual distractor and control conditions readily suppressed the increased attentional conflict caused by the previously learned response and looked more to the correct location. The children in the long delay condition, on the other hand, showed difficulties in correctly anticipating the target’s reappearance on the B trials. As the children’s performance in the three groups was comparable on the A trials, this result indicates that the longer time delay challenged their ability to actively keep the target’s
232
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Fig. 5. Illustrates 18-month-olds’ mean looking time in the first B trial across the first 22 s in the control (left) and distractor (middle) conditions, and across 28 s in the long delay (right) condition. The solid and dotted lines represent looking time at the correct occluder (B) and incorrect occluder (A), respectively. In the control and visual distractor conditions, the sound cue is presented 3.5 s after the target’s disappearance, whereas in the long delay condition the presentation of the sound cue take place 10 s after the target’s disappearance. In all there conditions the target reappears 9 s after the presentation of the cue signal (extended time period). Note that in the visual distractor condition a central distractor (between occluder A and B) is presented (for 2 s) before the cue signal. Error bars represent ±standard error.
new location in mind in order to correctly predict the target’s reappearance. Our data (e.g., see Figs. 2 and 5) suggest that the children’s attentional focus on the target’s location fades following the occlusion and continue to do so over time as seen during the extended hiding period in the first B trial. Whereas the children in the control and visual distractor conditions appeared to recover their attentional focus on the correct occluder rapidly at the presentation of the sound cue, this recovery was not clearly displayed by the children in the long delay condition. Notably, the data from the first B trial indicate that even without the long delay, that is, in the control and distractor conditions the children had difficulty maintaining attention beyond the first 6 s after the sound cue. These data could be interpreted to mean that the attention span of 18-month-olds in a rather eventless situation is about 6 s. The finding that the children in the visual distractor condition were able to overcome the increased attentional conflict caused by the distractor and performed equally to the children in the control condition is in contrast to our results from 10- and 12-month-olds’ performance on the same task (Watanabe et al., 2012). In our previous study, the infants in the visual distractor condition, particularly the 10-month-olds, made more incorrect (perseverative) anticipatory gaze predictions. Taken together, these two studies indicate that children acquire better control over their visual attention between 12 and 18 months of age. This conclusion is further supported by our supplementary analyses of looking at the correct and incorrect occluder within the distractor condition, which indicates a development from 10 months of age, toward a decrease in perseveration, but little improvement in correct anticipatory gaze predictions at 12 months of age, to correct anticipatory looking at 18 months of age. This development is interpreted to involve improvements in an inhibitory process of the attention system, which enables suppressing a gazing response toward the incorrect hiding location on the B trials (c.f., Watanabe et al., 2012), as well as improvements in the ability to update a hidden target’s location in order to correctly anticipate the target’s reappearance following a brief time delay. Importantly, the continuous measure used in this and our previous study (Watanabe et al., 2012), yields data for anticipatory gaze indicating that developmental improvements in voluntary control of visual attention is better described as gradual than as dichotomous. One implication of this suggestion, from a methodological perspective, is that research on young children’s development of attentional control (or cognitive development in general) can gain new insights from using more sensitive measures (e.g., continuous measures) of their performance, rather than using outcome measures that define behavioral improvements as something children either can do or cannot do. The behavioral improvement in attentional control is presumably closely tied to the neurological maturation of the attention control system, such as the increase in connectivity between the frontal cortex and anterior cingulate gyrus during the second year of life (e.g. Gao et al., 2013, 2009; Posner & Fan, 2008), and children’s growing explorations of their environment through independent locomotion by the end of the first year and forward. This view is clearly in line with our demonstration of a gradual improvement in infants’ attentional control. The developmental literature (e.g., Rothbart & Posner, 2001) suggests that important improvements in the attention systems are in place around 18 to 24 months of age, particularly in terms of suppressing competing information that interferes with the task at hand. The results of the present study would date these developmental changes to the period just before 18 months, but longitudinal data are necessary if we are to improve our understanding of this development. Our data also shed light on the 18-month-olds’ developmental level of attentional control through our comparison with adults’ performance on the same task. With the exception of the performance of the children in the long delay condition on the B trials, the 18-month-old’s allocation of visual attention was strikingly similar to that of the adults. However, a difference between the 18-month-olds’ and the adult participants’ performance was seen in the adults’ longer looking times
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
233
throughout the experiment, mainly reflected in more correct anticipatory looking on both A and B trials. As to incorrect looking and the tendency to perseverate under conditions of conflict, it should perhaps be noted that whereas there was no difference between the adults’ and the children’s looking at the incorrect occluder on the A trials, the 18-month-olds looked more than the adults at the incorrect occluder on B trials. This may be interpreted to mean that the 18-month-olds encountered more problems in switching than did the adults even though they did not perseverate in the general sense of showing more incorrect than correct looking. This view is further supported by looking at the results from the perspective of proportion of “on-task-behavior”, which suggests that while the higher amount of looking at the correct occluder for the adults to a large extent could be explained by the larger amount of on-task-behavior, their looking at the incorrect occluder would be even lower (e.g., see Figs. 2 and 3). The ability to maintain attention on the task is clearly not fully matured by 18 months of age and this was particularly evident during the time periods of the experiment when the target was occluded and no sound or moving stimuli were presented on the screen, such as during the long delay between the target’s disappearance and the presentation of the sound cue signaling the target’s reappearance. The abilities to control and flexibly allocate visual attention, tested in the current study, are believed to lay the foundation for more complex cognitive skills and are essential in order to predict future events, act in a goal-directed manner, and in interactions with the external world. Because a compromised development of attentional control has been suggested to be an important risk factor for the development of attentional disorders (e.g., Johnson, 2012), it is of great importance that research continues to clarify the developmental process of attentional control. If we can improve our insights into the early development of attention and how it can go astray, it will also improve our possibilities of providing the right measures for treating children with attention deficits. To summarize, our study has demonstrated that by 18 months of age, children can efficiently suppress distracting information and make correct anticipatory predictions, even following a switch in the requirements of the task. This finding, together with data from our previous study on 10- and 12-month-olds (Watanabe et al., 2012), demonstrate that important age-related improvements in attentional control are taking place in a gradual fashion from the end of the first year to 18 months of age. Further, by 18 months of age children’s performance show an adult-like pattern, the main difference being that adults demonstrate a better ability to maintain attention on the task and to update their predictions following a longer time period. Acknowledgements We gratefully acknowledge the efforts of the children, their parents, and the adult participants who took part in the studies. We thank Stefan Boström for assistance in data collection. This research was financially supported by the Swedish Research Council. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.infbeh. 2014.01.007. References Baird, A. A., Kagan, J., Gaudette, T., Walz, K. A., Hershlag, N., & Boas, D. A. (2002). Frontal lobe activation during object permanence: Data from near-infrared spectroscopy. NeuroImage: 16., 120–126. Clohessy, A. B., Posner, M. I., & Rothbart, M. K. (2001). Development of the functional visual field. Acta Psychologica: 106., 51–68. Colombo, J. (2001). The development of visual attention in infancy. Annual Review of Psychology: 52., 337–367. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience: 3., 201–215. Cuevas, K., & Bell, M. A. (2010). Developmental progression of looking and reaching performance on the A-Not-B task. Developmental Psychology: 46., 1363–1371. Cuevas, K., & Bell, M. A. (2011). EEG and ECG from 5 to 10 months of age: Developmental changes in baseline activation and cognitive processing during a working memory task. International Journal of Psychophysiology: 80., 119–128. Espy, K. A., Kaufmann, P. M., McDiarmid, M. D., & Glisky, M. L. (1999). Executive functioning in preschool children: Performance on A-not-B and other delayed response format tasks. Brain and Cognition: 41., 178–199. Gao, W., Gilmore, J. H., Dinggang, S., Smith, J. K., Zhu, H., & Lin, W. (2013). The synchronization within and interaction between the default and dorsal attention networks in early infancy. Cerebral Cortex: 23., 594–603. Gao, W., Zhu, H., Giovanello, K. S., Smith, J. K., Shen, D., Gilmore, J. H., et al. (2009). Evidence of the emergence of the brain’s default network from 2-week-old to 2-year-old healthy pediatric subjects. Proceedings of the National Academy of Sciences of the United States of America: 106., 6790–6795. Garon, N., Bryson, S. E., & Smith, I. M. (2008). Executive function in preschoolers: A review using an integrative framework. Psychological Bulletin: 134., 31–60. Johnson, M. H. (2012). Executive function and developmental disorders: The flip side of the coin. Trends in Cognitive Science: 16., 454–457. Johnson, M. H., Posner, M. I., & Rothbart, M. K. (1991). Components of visual orienting in early infancy: Contingency learning, anticipatory looking and disengaging. Journal of Cognitive Neuroscience: 3., 335–344. Kastner, S., DeSimone, K., Konen, C. S., Szczepanski, S. M., Weiner, K. S., & Schneider, K. A. (2007). Topographic maps in human frontal cortex revealed in memory-guided saccade and spatial working-memory tasks. Journal of Neurophysiology: 97., 3494–3507. Marcovitch, S., & Zelazo, P. D. (1999). The A-not-B error: Results from a logistic meta-analysis. Child Development: 70., 1297–1313. McNab, F., & Klingberg, T. (2008). Prefrontal cortex and basal ganglia control access to working memory. Nature Neuroscience: 11., 103–107. Nelson, K. E. (1971). Accommodation of visual tracking patterns in human infants to object movements patterns. Journal of Experimental Child Psychology: 12., 182–196. Petersen, S. E., & Posner, M. I. (2012). The attention system of the human brain: 20 years after. Annual Review of Neuroscience: 35., 73–89.
234
L. Forssman et al. / Infant Behavior & Development 37 (2014) 225–234
Posner, M. I., & Fan, J. (2008). Attention as an organ system. In J. R. Pomerantz (Ed.), Topics in integrative neuroscience (pp. 31–61). New York; NY: Cambridge University Press. Posner, M. I., Rothbart, M., Sheese, B. E., & Voelker, P. (2012). Control networks and neuromodulators of early development. Developmental Psychology: 48., 827–835. Rothbart, M., & Posner, M. (2001). Mechanism and variation in the development of attentional networks. In C. Nelson, & Luciana (Eds.), Handbook of developmental cognitive neuroscience (pp. 353–363). Cambridge, MA: MIT Press. Salley, B., Panneton, R. K., & Colombo, J. (2012). Separable attentional predictors of language outcome. Infancy, http://dx.doi.org/10.1111/ j.1532-7078.2012.00138.x Schutte, A. R., Spencer, J. P., & Schöner, G. (2003). Testing the dynamic field theory: Working memory for locations becomes more spatially precise over development. Child Development: 74., 1393–1417. Sheese, B. E., Rothbart, M. K., Posner, M. I., White, L. K., & Fraundorf, S. H. (2008). Executive attention and self-regulation in infancy. Infant Behavior and Development: 31., 501–510. Uddin, L., Supekar, K., & Menon, V. (2010). Typical and atypical development of functional human brain networks: Insights from resting-state fMRI. Frontiers in System Neuroscience: 4., http://dx.doi.org/10.3389/fnsys.2010.00021 Watanabe, H., Forssman, L., Green, D., Bohlin, G., & von Hofsten, C. (2012). Attention demands influence 10- and 12-month-old infants’ perseverative behavior. Developmental Psychology: 48., 44–56.
