A longitudinal study of the development of shifts of gaze to a peripheral stimulus in preterm infants with transient periventricular echogenicity

A longitudinal study of the developmentof shifts of gaze to a peripheral stimulus in

preterm infants withtransient periventricular echogenicity q

Phillipa R. Butcher, a,* Alex F. Kalverboer, a

Reint H. Geuze, a and Elizabeth F. Stremmelaar b

a Department of Developmental and Experimental Clinical Psychology,

University of Groningen, 9712 TS Groningen, The Netherlandsb Beatrix Children’s Clinic, University Hospital, Groningen, The Netherlands

Received 15 June 2000; revised 22 October 2001

Abstract

Shifts of gaze to peripheral targets were studied longitudinally, between 6 and 26

weeks corrected age, in full-term and very preterm infants with transient periventric-

ular echogenicity (PVE). Before 10 weeks, simple shifts of gaze were faster and more

frequent in preterms with PVE < 14 days than in full-terms, suggesting these pre-

terms profited from additional early visual experience. After 16 weeks, there were

subtle differences between full- and preterm infants in the development of shifts of

gaze requiring disengagement. The differences suggest that, after disengagement

had become established, its fine-tuning occurred more slowly in the preterms. Slower

fine-tuning of disengagement was not associated with duration of PVE, since it was

more marked in preterms with PVE < 14 days than in preterms with PVEP 14 days.

The differences in performance between full- and preterm infants were small. How-

ever, even small differences may affect the efficiency of visually guided behav-

iors. � 2002 Elsevier Science (USA). All rights reserved.

J. Experimental Child Psychology 82 (2002) 116–140

www.academicpress.com

qWe thank Paul van Geert for his constructive comments on the original manuscript, Tom

Snijders for his assistance with the statistical analysis, and the parents and infants for

participating in the study.*Corresponding author. Fax: +31-50-363-6304.

E-mail address: [email protected] (P.R. Butcher).

0022-0965/02/$ - see front matter � 2002 Elsevier Science (USA). All rights reserved.

PII: S0022 -0965 (02)00006 -1

Keywords: Infants; Preterm; Disengagement; Shifts of gaze

In the past 15 years, advances in cranial ultrasound technology havegreatly improved the detection of neonatal brain injury. It has become clearthat very preterm infants (<32 weeks gestation) frequently suffer from is-chemic brain injury. This is detectable as an increase in the density of theimage on neonatal ultrasound scans of the brain, referred to as echogenicity.Periventricular echogenicity (PVE) refers to an increase in echodensitywhich is most marked in the periventricular region, but frequently extendsinto occipital, parietal, and frontal regions (Fawer & Calame, 1991; Fazziet al., 1994). PVE is estimated to occur in 32–44% of very low birth weight(VLBW) infants (Ichord, 1993). In a minority of infants, PVE is severe andlong lasting, leading to the formation of cysts, or periventricular leukomal-acia, a condition closely associated with serious motor and visual impair-ment. In the majority of infants, PVE is transient and does not lead todetectable, lasting tissue damage. However, even transient PVE can affectthe quality of development. For example, PVE lasting 14 days or longerhas been associated with poorer outcomes on measures of cognition, atten-tion, and motor skills in later childhood (Appleton, Lee, & Hey, 1990; Fa-wer & Calame, 1991; Fazzi et al., 1994; Levene, Dowling, Graham, Galton,& Phillips, 1992).

Since PVE frequently extends into occipital, parietal, and frontal regions,processes mediated by these regions may be disturbed in infants with tran-sient PVE. The processes involved in disengaging attention and gaze fromtheir current focus in order to shift them to a new location are mediatedby areas in the parietal and frontal cortex. The development of disengage-ment may therefore be delayed or disturbed in infants with transient PVE.To determine if this is the case, we carried out an intensive, longitudinal in-vestigation of shifts of gaze in very preterm infants with transient PVE andfull-term infants with no history of pre- or perinatal medical complications.

The development of gaze and attention shifting

The efficient deployment of visual attention requires the ability to strike abalance between engaging attention on specific aspects of the visual environ-ment and shifting it in order to regulate the flow of incoming information orto monitor the environment (Krakow & Kopp, 1982; Tellinghuisen & Oa-kes, 1997). A number of recent studies have shown that infants become moreefficient in achieving a balance between engaging and shifting attention inthe first 6 months of life. In these studies, gaze is used as an index of visualattention because the focus of gaze generally corresponds with the focus of

P.R. Butcher et al. / Journal of Experimental Child Psychology 82 (2002) 116–140 117

visual attention. Visual attention is considered to be engaged at the locationon which gaze is focused, and shifts of gaze are thought to involve disengag-ing both attention and gaze from their current focus and then shifting themto a new focus, where they can become engaged, permitting detailed process-ing of new visual information (Kingstone & Klein, 1993; Posner, Inhoff,Friedrich, & Cohen, 1987). These studies have shown that, at 1–2 monthsof age, infants often have difficulty disengaging gaze and may stare at a par-ticular object for long periods, a phenomenon known as obligatory attention(Stechler & Latz, 1966) or sticky fixation (Hood et al., 1996). However, be-tween 2 and 4 months of age, disengaging gaze becomes increasingly easy.By 4 months infants are rarely captured by stimuli they are fixating and shiftgaze rapidly from one stimulus to another (Atkinson, Hood, Wattam-Bell,& Braddick, 1992; Butcher, Kalverboer, & Geuze, 2000; Johnson, Posner, &Rothbart, 1991).

Rapid, reliable disengagement of gaze indicates that infants are able toinhibit attention for one stimulus in order to direct it at another and is thusan important step toward more efficient deployment of visual attention. Theincrease in the efficiency of disengagement between 2 and 4 months has beenshown to be associated with better performance on tasks calling for flexibleshifts of visual attention. Fixations on habituation tasks, for example, arelongest at 2 months, when infants’ disengagement difficulties are greatest(Hood et al., 1996). At 3 and 4 months, latency to disengage is positively as-sociated with length of fixation (Frick, Colombo, & Saxon, 1999). Duringsocial interaction, infants younger than 3 months spend long periods simplystaring, while infants older than 3 months shift gaze easily between theirpartner’s face and the surroundings (Kaye & Fogel, 1980; van WulfftenPalthe, 1986). The increase in the efficiency with which gaze is disengagedand shifted during the first 6 months of life is thus important not only initself but also because it is crucial for the normal development of cognitionand communication (Johnson, 1998).

Shifts of gaze from an attended stimulus (disengagement) in preterm infants

with transient PVE

Studies of adult subjects have localized the operations involved in shiftingattention and gaze in particular areas or networks in the brain. Disengagingattention and gaze from their current focus is thought to be mediated by theposterior parietal and/or frontal cortex, shifting them to a new location isthought to be mediated by eye movement centers in the midbrain, while re-engaging them at that new location is thought to be mediated by the thala-mus (Atkinson et al., 1992; Posner et al., 1987). In adult subjects, damage toone of the areas of the brain involved in a shift of attention or gaze has beenshown selectively to disrupt the functioning of the operation which it

118 P.R. Butcher et al. / Journal of Experimental Child Psychology 82 (2002) 116–140

mediates (Posner et al., 1987). Transient PVE extending into the parietal andfrontal areas may therefore disrupt the development of disengagement invery preterm infants.

We know of no studies of disengagement in very preterm infants withPVE which could provide evidence on this issue. The evidence on disengage-ment in heterogeneous groups of preterms is inconsistent. Using a standarddisengagement task, Atkinson (2000) found shorter latencies to disengage in4- to 6-week old VLBW infants with no evidence of structural brain damage.However, on the basis of their performance on habituation and preferentiallooking tasks, other researchers have suggested that preterm infants may beless efficient in disengaging gaze from visual stimuli (Rose, Feldman,McCarton, & Wolfson, 1988; Sigman, Cohen, Beckwith, & Parmalee, 1986).

Shifts of gaze that do not require disengagement in preterm infants with

transient PVE

PVE is associated with injury to parietal and frontal cortex and maytherefore threaten the development of disengagement. It has not been asso-ciated with injury to the subcortical eye movement centers, which mediatethe shifting of attention and gaze, and should therefore not affect the effi-ciency of shifts of gaze that do not require disengagement. Subcortical eyemovement centers which mediate simple shifts of gaze mature early, andeye movements have been observed well before the term date in preterm in-fants. Additional early visual experience may then be associated with anearly increase in their efficiency in healthy preterms (Fielder, Foreman,Moseley, & Robinson, 1994). Alternatively, eye movement centers are farfrom fully mature at the time of a very preterm birth, and early exposureto the visual environment may damage still maturing visual structures(Als, 1995; Friedman, Jacobs, & Werthmann, 1981). Different outcomesare therefore possible. If additional early visual experience has a positive im-pact on development, simple shifts of gaze should be more frequent and fas-ter in preterms than in full-terms shortly after the expected date of birth,when the advantage of additional experience should be clearest. If, however,a very preterm birth in itself has a negative impact on visual development,then preterm infants should perform more poorly than full-term infants atall ages. Further, since medical complications may also depress performancein very preterms, performance within a group of very preterm infants mayvary with the seriousness of the medical complications.

We were also unable to find studies of simple shifts of gaze and attentionto peripheral stimuli in very preterm infants with PVE. However, a numberof studies have investigated simple shifts of gaze in diverse groups ofpreterm infants. Four studies have investigated attention-shifting—also re-ferred to as attention-getting (Landry, Leslie, Fletcher, & Francis, 1985)


or recruitment of attention (Rose et al., 1988)—in preterm infants in the firstfew weeks after the term date. Friedman et al. (1981) found that, at the termdate, low-risk preterms were slower to look to peripheral stimuli than werefull-term controls. Masi and Scott (1983) found that, at both 1 and 6 weekscorrected age, preterms were slower to look to a suddenly appearing facethan were 2 week older, full-term controls. However, two more recent stud-ies (Atkinson, 2000; Foreman, Fielder, Price, & Bowler, 1991) found that upto 4–6 weeks corrected age, preterms with no evidence of structural braindamage and preterms with diverse medical histories, respectively, lookedto peripheral stimuli more quickly than full-term controls.

Studies of the development of gaze shifting over the following severalmonths have also produced diverging results. Foreman et al. (1991) foundthat, from 1 to 10 months corrected age, response latencies of pre- andfull-terms were approximately equal. Landry et al. (1985) found that, at 7months corrected age, preterms with respiratory distress syndrome (RDS)but no intraventricular hemorrhage (IVH) looked to suddenly appearing pe-ripheral targets as quickly as full-term controls, while preterms with bothRDS and IVH (Grades I–IV) responded more slowly than full-term con-trols. Vervloed (1995) found that, at 9 months corrected age, relativelyhealthy preterms also responded more slowly than full-term controls.

To summarize, at the moment information on both disengaging andshifting attention and gaze in very preterm infants with transient PVE islacking. The information that is available on disengaging and shifting atten-tion and gaze in diverse groups of preterms is inconsistent.

Goals and expectations

The goal of this study was to compare the development of disengagingand shifting attention and gaze in preterm infants with transient PVE andfull-term infants. Both simple shifts of gaze and shifts of gaze requiring dis-engagement were measured. To determine if developmental differences wereassociated with the severity of PVE, the group of very preterm infants wassubdivided into a high- and a low-risk group on the basis of the duration ofthe PVE. Infants in whom PVE persisted beyond the 14th day, a commonlyused cutoff point for identifying infants at increased risk of subsequent de-velopmental abnormalities (Damman & Leviton, 1997; Fazzi et al., 1994;Levene et al., 1992), formed the high-risk group.

The most recent, and therefore the most comparable, studies of shifts ofgaze which do not require disengagement (Atkinson, 2000; Foreman et al.,1991) showed an early advantage for preterm infants with few medical com-plications. We therefore expected that the preterms with PVE < 14 dayswould shift gaze more frequently and faster than the full-terms in the earlysessions, but that their advantage would disappear gradually as the relative


difference in experience decreased. Since the very preterms in the groupPVEP 14 days had less favorable medical histories than the preterms withPVE < 14 days (see Method), we expected that any early advantage wouldbe smaller in the former group.

The parietal and frontal areas, which mediate disengagement of gaze andattention, develop during the second and fourth months after a full-termbirth (Hood & Atkinson, 1993; Johnson et al., 1991), so that additionalearly visual experience should not affect the development of disengagement.We therefore expected no early advantage for the preterms. However, giventhe association between transient PVE and injury to areas of the brain whichmediate disengagement, we expected that its development would be delayedor attenuated in the preterm groups or would differ in quality, for example,in the range of responses observed or the variability of response latencies.Further, since the duration of PVE may affect the degree to which atten-tional functioning is impaired, we expected any differences to be moremarked in preterms in whom PVE lasted 14 days or longer.

Method

Subjects

Sixteen full-term and 21 preterm infants took part in the study. The full-term infants were recruited through childbirth education classes. All wereborn after gestation periods between 38 and 42 weeks, weighed more than2500 g at birth, and had uneventful pre- and perinatal histories. Mean ges-tational ages and birthweights are presented in Table 1. The preterm infantswere born in, or admitted within 24 h of birth to, the neonatal intensive careunit at the University Hospital in Groningen between May 1995 and Janu-ary 1999. All were born before 32 weeks of gestation. All but one weighedless than 1500 g (range 660–1540 g). Two of the infants were small for ges-tational age. None suffered neonatal medical complications serious enoughto lead to a Nursery Neurobiologic Risk Score (NBRS) above the cutoffpoint indicating heightened risk of later developmental problems (Brazy,Eckerman, Oehler, Goldstein, & O’Rand, 1991). Routine screening showedthat none of the infants suffered from retinopathy of prematurity. Infantswith PVE frequently also suffer from germinal matrix or intraventricularhemorrhage (GM/IVH). To test the association between disengagementand PVE in isolation, infants with GM/IVH higher than grade 1, that isbleeding extending beyond the germinal matrix into the lateral ventricles(Papile, Burstein, Burstein, & Koffler, 1978), were excluded.

The presence of PVE was established from ultrasound scans made at ap-proximately weekly intervals beginning within a few days of birth. The scanswere interpreted by two radiologists experienced in neonatal radiology.


Scans in the first week of life showed PVE in all of the infants. The durationof the PVE ranged from less than 1 week to more than 12 weeks. Twogroups were formed, using duration of PVE beyond the 14th day as the cut-off point. There were 11 low-risk (PVE < 14 days) infants and 10 high-riskðPVEP 14 daysÞ infants. The low- and high-risk preterms did not differwith respect to birthweight, presence of grade 1 GM/IVH, or durationof assisted ventilation. However, infants with PVEP 14 days had shortergestations ðtð19Þ ¼ �3:18; p < :01Þ and higher total NBRS scoresðtð19Þ ¼ �2:72; p < :01Þ. The total NBRS scores included duration ofPVE. When this item was excluded, the two groups no longer differedðtð19Þ ¼ �:80; p > :1Þ. An overview of the medical characteristics of thetwo preterm groups is presented in Table 1.

The families of all full-term infants and all but one of the preterm infantswere intact. The parents of all full-term infants and all but one of the pre-term infants were Caucasian. Mothers of preterms tended to be less likelyto have followed a university education ðv2ð2Þ ¼ 4:74; p ¼ :09Þ.

Design

For 11 full-terms, measurements started at 6 or 7 weeks of age and con-tinued at 2-week intervals until approximately 25 weeks of age. For the re-maining six infants, measurements started at 8 or 9 weeks and continued toapproximately 26 weeks of age. Ages in both groups were calculated fromthe due date.

The session was broken off seven times because of fussing or sleepiness.In each case the infant was able to make the session up within 2 or 3 days.

Table 1

Birth history and medical variables for the full-term infants, preterm infants with PVE < 14

days and preterm infants with PVEP 14 days

Full-term PVE < 14 days PVEP 14 days

Gestational age (M (SD)) 40 weeks (�1) 30.3 weeks (�1) 28.7 weeks (�1)Birthweight (M (SD)) 3552 g (�404) 1234 g (�207) 1126 g (�259)Boys:girls 8:8 3:8 4:6

Assisted ventilation

None 36% 44%

<7 days 55% 33%

8–28 days 9% 22%

GM/IVHa 18% 60%

NBRSb (M (SD)) 2.64 (�1) 3.90 (�1)aThe percentage of preterms in each group with grade 1 germinal matrix or intraventricular

hemorrhage, that is bleeding that does not extend beyond the germinal matrix into the lateral

ventricles.bNBRS, Nursery neurobiologic risk score. Scores > 7 indicate heightened risk for develop-

mental disability as a result of neonatal medical complications such as infection, respiratory

problems, and hypoglycemia.


Thirteen full-terms completed all ten measurement sessions. Two completednine sessions. One completed eight sessions. The missing sessions were at 22weeks (two sessions) and 24 weeks (two sessions).

For 20 preterms, measurements started at 6 weeks and continued at 4-week intervals until 26 weeks of age. Illness delayed the first session forthe remaining infant until 8 weeks. Measurements were thus made in bothgroups at 6, 10, 14, 18, 22, and 26 weeks. The session was broken off ninetimes because of fussing or sleepiness. In three cases the session was madeup within 2 weeks. Sixteen preterms completed all six measurement sessions.Four completed five sessions. One completed four sessions. The missing ses-sions were at 6 weeks (five sessions), 10 weeks (one session), and 18 weeks(one session). The longer interval between measurements in the pretermgroups was chosen for practical reasons, namely to minimize the strain onthe infants and their parents. An unavoidable consequence was that anylearning effects may have been more pronounced in the full-term group thanin the preterm group.

Procedure

Infants were brought to the laboratory at a time when they were expectedto remain alert for 20–30min. Measurements were made with the infants instate 3 or 4 of Prechtl’s scale of alertness: awake, eyes open, some spontane-ous movements (Prechtl & Beintema, 1964). They carried out the disengage-ment task in an infant-seat in front of an arc of three computer monitors, ata distance of 45 cm from each monitor. Only the screens of the monitorswere visible. The monitors themselves, the equipment necessary to run thetasks and record the infant’s eye movements, and the experimenter wereconcealed by a gray screen which filled approximately 180� of the infant’svisual field. The infant’s eye movements and the display on the monitorswere recorded on an SVHS video monitor, allowing the experimenter torun through the task on the basis of the infant’s behavior. The video record-ing was scored off-line.

The disengagement task

StimuliTwo types of fixation stimuli were used. The first was a phase-reversing

schematic face similar to that used by Hood and Atkinson (1993).The sche-matic face consisted of four stripes, one for each eye, the nose, and themouth. The orientation of the stripes changed from horizontal to verticalsix times per second. Their color changed from trial to trial. The secondstimulus consisted of either six 2.5-cm diameter sets of red, yellow, and blueconcentric circles or six 2.5-cm diameter sets of green, purple, and white con-centric squares. The colors shifted six times per second. From a distance of


45 cm, fixation stimuli subtended a visual angle of 9� ðheightÞ � 6� (width).Half of the fixation stimuli in each session were schematic faces. The remain-ing half were concentric circles in the odd-numbered sessions and concentricsquares in the even-numbered sessions. The two stimulus types were distrib-uted equally across left and right peripheral targets and between the first andsecond half of the task.

The peripheral targets consisted of a rectangular grating above an oval.Peripheral targets flashed on and off six times per second and changed colorfrom trial to trial. They subtended an angle of 7� ðheightÞ � 4� (width), andtheir inner edge was 30� from the midline of the fixation stimulus. Most 4-week-olds look to targets of this size and eccentricity at above chance levelswhen the fixation stimulus disappears (Maurer & Lewis, 1998). However,under binocular viewing conditions, 6- to 8-week-old pre- and full-term in-fants look to peripheral targets at this eccentricity at lower than chance lev-els when the fixation stimulus persists (Mohn & van Hof-van Duin, 1986).Both the probability and the speed of correct refixations of peripheral tar-gets at this eccentricity are thus reduced by the infants’ difficulty in disengag-ing. Locating peripheral targets at this eccentricity therefore provided twoassociated, but separate, indices of the development of disengagement: a fre-quency index and a latency index.

Stimulus presentationThe steps in the trials of the disengagement task are shown in Fig. 1.

Competition and noncompetition trials were included. All trials began withthe appearance of the fixation stimulus on the center monitor, accompaniedby a brief melody. On competition trials, a peripheral target was added to

Fig. 1. The sequence of stimuli in the disengagement task. Each row of three panels depicts the

trio of computer monitors. The top row shows the first step in each task. The bottom row shows

the second step. The centering stimulus was a brightly colored pattern. The targets were a solid

oval above a rectangular grating. All colors changed six times per second.


the display when the infant had been fixating the stimulus for 1–2 s. After5 s, the peripheral target and the fixation stimulus disappeared simulta-neously. Following a 1.25-s inter-trial-interval with blank screens, the fixa-tion stimulus reappeared and the following trial began. Competition trialsthus required disengagement of gaze and attention from the fixation stimu-lus before an eye movement could be made to the peripheral target. Thesetrials measured the attentional processing required for disengagement in ad-dition to the sensory-motor processing involved in detecting and moving theeyes to the peripheral target. The frequency and latency of correct refix-ations on competition trials in each session provided an index of the effi-ciency with which infants disengaged.

On noncompetition trials, the fixation stimulus disappeared when theperipheral target appeared. Noncompetition trials thus required onlythe generation and execution of an eye movement. They measured thesensory-motor processing involved in detecting, and preparing and execut-ing an eye movement to a peripheral target. Changes in the frequency andlatency of correct refixations on noncompetition trials provided both an in-dex of the efficiency of simple shifts of gaze and a control for age-associatedchanges in their efficiency.

There were 20 competition trials and eight noncompetition trials. Thefirst two trials were noncompetition trials with left and right peripheral tar-gets, respectively. The remaining noncompetition trials were distributedpseudorandomly among the competition trials. Half of the peripheral tar-gets on each type of trial were presented to the left of the central stimulusand half to the right in pseudorandom order.

Scoring

Video recordings of each session were played back half-frame by half-frame (20-ms intervals), and the onset and direction of the first eye movementafter the peripheral target appeared were scored by psychology studentstrained by the first author. Looks away from the location of fixation stimulusending in the quadrant centered around the peripheral target were scored ascorrect refixations. Performance on noncompetition trials showed thatinfants of all ages were able correctly to refixate peripheral targets with thisdegree of accuracy. Looks away from the fixation stimulus that did not pro-duce correct refixations of the peripheral target were scored as errors. Thedirection of each error was categorized as vertical, opposite from the targetlocation, or the same as the target location. One session in five was alsoscored by the first author. The interscorer reliability for the onset of eyemovements was 94%. Cohen’s j for the category of eye movements averaged.81 (range .56–1.0). For the direction of errors it was .79 (range .7 to .85).

In total, 4303 trials were recorded in full-term infants, and 3506 in pre-term infants. Trials on which the infant’s gaze was not centered on the


fixation stimulus when the target appeared (6.8% full-terms, 8.5% preterms)were dropped from the analysis. Reaction times that were shorter than200ms (2.5% full-terms, 2.4% preterms) were considered anticipatory (Ha-ith, Hazan, & Goodman, 1988). The analysis was carried out on the remain-ing trials (3897 (90.6%) full-terms, 3228 (92%) preterms). Each full-terminfant contributed on average 244 trials (range 198–277). Each preterm in-fant contributed on average 153 trials (range 96–169).

Statistical analysis

The frequency and the median latency of correct refixations of targets oncompetition and noncompetition trials were calculated for each infant foreach session, as were the frequency and direction of the errors. Multilevelanalysis was used to determine if the frequency and the latency of correctresponses and the frequency and direction of errors changed significantlywith age and to determine if the developmental trajectories of the threegroups of infants differed over the age range studied.

Multilevel analysis is a regression procedure which is particularly suitedto the analysis of longitudinal data because it takes into account the factsthat measurement sessions are nested within individual infants and thatthe association between session scores and explanatory variables may differsystematically from infant to infant (Snijders & Bosker, 1999). As with stan-dard regression procedures, estimates of regression coefficients for the con-stant term or intercept and for the slope are provided. These are called fixedeffects. However, multilevel analysis also provides an estimate of the extentto which intercept and slope coefficients differ across infants. These arecalled random effects. A further advantage is that all available data sets, re-gardless of the number of measurement points, can be included in the mul-tilevel analysis.

Earlier research had shown that disengagement develops in a nonlinearfashion. Its frequency is consistently low before 2 months of age, increasesrapidly during the third and fourth months, then stabilizes at a high level inthe fifth month (see, for example, Butcher et al., 2000; Hicks & Richards,1998; Johnson et al., 1991). For the analysis of the frequency data, the totalage range was therefore split into three intervals, one corresponding to eachof the phases in development: 6–9 weeks, 9–16, and 16–26 weeks, and a sep-arate age function for each interval was included in the model. Such piece-wise age functions are both flexible and relatively insensitive to fluctuationsoutside their own age range (Snijders & Bosker, 1999). They are thus idealfor modeling nonlinear developmental trajectories where the tempo ofchange within one phase is not necessarily related to the tempo of changein other phases. The low frequency of accurate shifts of gaze at 6 weeksmeant that few full-term infants provided reliable latency data at that age.A model consisting of two piecewise linear age functions, 8–16 and 16–26


weeks, was therefore fit to the latency data. The data were centered around 8weeks, the youngest age at which data were available for all infants. Allmodels had three levels, infant, session, and type of fixation stimulus.

Paired t-tests were carried out to determine if differences between groupsin the mean frequency and latency of correct refixations were significant atthree points: the beginning (6 weeks), middle (14 weeks), and end (25 weeks)of the measurement period. Bonferroni corrections were used to keep alphaat .05 despite multiple testing (Stevens, 1992). Unless otherwise indicated(uncorrected p-values), the p-values were multiplied by the number of testscarried out. v2 were used to determine if the frequencies of the types of errorwithin the groups changed over time or differed across groups at a singlepoint in time.

Results

The results for simple shifts of gaze (noncompetition trials) will be pre-sented first, followed by the results for disengagement (competition trials).

Noncompetition trials

FrequencyThe mean frequencies of correct refixations on noncompetition trials for

each group in each session are presented in Fig. 2. In the 6-week session,both preterm groups correctly refixated the peripheral target on 85%ðSD ¼ 17Þ of the trials. Their frequency of correct refixations did not changebetween 6 and 26 weeks. In the 6-week session, full-terms made correct re-fixations on 63% ðSD ¼ 36Þ of the trials, significantly fewer than bothgroups of preterms ðtð31Þ ¼ �2:31; p < :05; tð32Þ ¼ �2:49; p < :05Þ. How-ever, their correct refixations increased rapidly, reaching 87% ðSD ¼ 18Þ inthe 10-week session and remaining close to 90% in all following sessions. Be-tween 6 and 9 weeks, the frequency of correct refixations on noncompetitiontrials increased more rapidly in the full-term group than in both pretermgroups, but only the contrast with preterms with PVE < 14 days reachedsignificance ðtð32Þ ¼ �2:0; p < :05Þ. No significant intergroup differenceswere found at 14 or 26 weeks.

LatencyThe means of the median latencies of correct refixations on noncompeti-

tion trials for each group in each session are presented in Fig. 3. Both pre-term groups responded faster than full-terms in the 6-week session(PVE < 14 days: M ¼ :85 s, SD ¼ :25; PVEP 14 days: M ¼ 1:00 s,SD ¼ :18; full-term: M ¼ 1:06 s, SD ¼ :32). Only the difference betweenfull-terms and preterms with PVE < 14 days reached significance


0

0.5

1

1.5

6 10 14 18 22 26

Corrected age in weeks

Sec

onds

full-termpreterm: pve<14 dayspreterm: pve > 13 days

Fig. 3. The mean and standard error of the latencies of correct refixations on noncompetition

trials between 6 and 26 weeks for full-term infants, preterm infants with PVE < 14 days, and

preterm infants with PVEP 14 days.

0

25

50

75

100

6 10 14 18 22 26


Per

cen

tage

full-termpreterm pve < 14dayspreterm pve > 13 days

Fig. 2. The mean and standard error of the frequency of correct refixations on noncompetition

trials between 6 and 26 weeks for full-term infants, preterm infants with PVE < 14 days, and

preterm infants with PVEP 14 days.


ðtð34Þ ¼ 2:32; p < :05Þ. Latency to respond dropped significantly in allthree groups between the 6- and 14-week sessions ðtð34Þ ¼�14:5; p < :001Þ and then leveled off between 18 and 26 weeks. Therewas no effect of group in the multilevel analysis, indicating that the develop-mental trajectories were similar in all groups. No significant group differ-ences were found in the 14- or 26-week sessions.

Competition trials

FrequencyThe mean group frequencies of correct refixations in the competition con-

dition in each session are shown in Fig. 4. The developmental trajectories forthe three groups were highly similar. In the 6-week session, the frequency ofcorrect refixations was approximately equal in all groups. It increased rap-idly in all groups between the 6- and the 14-week sessionsðtð32Þ ¼ 8:05; p < :001Þ, then more gradually between the 14- and the 26-week sessions ðtð32Þ ¼ 2:65; p < :01Þ. The multilevel analysis showed no ef-fect of group. No significant intergroup differences were found at 14 or 26weeks.

However, while the preterm groups did not differ from the full-terms infrequency of correct refixations on competition trials, responses on the trialsthat did not produce correct refixations suggested that disengagement

0

25

50

75

100

6 10 14 18 22 26


Per

cen

tag

e

full-termpreterm pve < 14dayspreterm pve > 13 days

Fig. 4. The mean and standard error of the frequency of correct refixations on competition tri-

als between 6 and 26 weeks for full-term infants, preterm infants with PVE < 14 days, and pre-

term infants with PVEP 14 days.


emerged more slowly in preterm than in full-term infants. On these trials,infants either failed to look away from the fixation stimulus or looked some-where other than the peripheral target (an error). Errors could be looksaway from both fixation and peripheral target stimuli or looks to the sideof the peripheral target that did not end within the quadrant centeredaround the peripheral target. (It is important to remember here that the per-formance of even the youngest infants on noncompetition trials showed thatthey could localize peripheral targets rather accurately.) The frequencies ofthe types of response on the competition trials that did not produce correctrefixations changed significantly between the 6- and 26-week sessions (seeButcher et al., 2000). At 6 and 10 weeks, failure to look away from the fix-ation stimulus or a look to the target side that ended outside the area des-ignated a correct refixation were significantly more frequent than looksaway from both fixation stimulus and peripheral target. The strongest influ-ence on looking behavior at these ages was apparently the task stimuli.When infants shifted gaze from the central stimulus it was to look, success-fully or unsuccessfully, to the peripheral target. From 18 weeks onward,looks somewhere other than the peripheral target had become significantlymore frequent ðv2ð5Þ ¼ 424:24, p < :001Þ. On at least some of the trials, in-fants were now able to resist the pull of the highly salient task stimuli in or-der to look to another, apparently more attractive location. Failure to lookaway from the fixation stimulus and looks to the target side that ended out-side the area designated a correct refixation were therefore considered im-mature responses, while looks away from both fixation stimulus andperipheral target were considered mature responses.

Preterms and full-terms differed in both frequency of errors and type oferror made. Both preterm groups made fewer errors than the full-term in-fants did throughout the measurement period (PVE < 14 days:tð31Þ ¼ �3:53, p < :001; PVEP 14 days: tð31Þ ¼ �2:26, p < :025). The dif-ferences were greatest after the 14-week session, when errors were matureresponses, that is, clear looks away from both fixation and target stimuli(PVE < 14 days: tð97Þ ¼ 2:80, p < :01; PVEP 14 days: tð96Þ1:957 ¼p < :05).

The frequency of immature and mature responses on the competition tri-als that did not produce correct refixations in each session are presented foreach group in Fig. 5a and b. The frequency of immature responses decreasedmore slowly in the preterm groups. In the 14-, 18-, and 22-week sessions,they were significantly more frequent in both preterm groups than in thefull-term group (PVE < 14 days: v2ð1Þ ¼ 25:01; p < :001; PVEP 14 days:v2ð1Þ ¼ 7:41; p < :006).

LatencyThe means of the median latencies of correct refixations on competition

trials in each group in each session are presented in Fig. 6. In the 10-week


a

b

Fig. 5. The relative frequencies of types of response on competition trials that did not lead to

correct refixations of the peripheral target. (a) Immature responses (failure to look away from

the fixation stimulus and inaccurate looks to the side of the target). (b) Mature responses (looks

away from both fixation stimulus and peripheral target).


session, the earliest session in which a meaningful intergroup comparisoncould be made, preterms with PVE < 14 days responded significantly morequickly than full-terms ðtð40Þ ¼ 2:30; p < :04Þ. Latency to respond droppedsignificantly in all groups between the 10- and the 14-week sessionsðtð30Þ ¼ �5:4; p < :001Þ and again between the 18- and the 26-week ses-sions ðtð30Þ ¼ �4:11; p < :001Þ. A significant age by group interaction,however, showed that latency did not decline as rapidly during this periodin the preterms with PVE < 14 days ðtð31Þ ¼ 2:6; p < :01Þ, and, while theseinfants responded as quickly as the full-terms at 14 weeks, they had becomesignificantly slower by 26 weeks ðtð19Þ ¼ �2:52; p < :05Þ.

The picture for variability of the latencies was similar to that for the la-tencies themselves. Significant decreases occurred in all groups between the8- and the 14-week sessions ðtð32Þ ¼ �5:45; p < :001Þ and again betweenthe 18- and the 26-week sessions ðtð32Þ ¼ �2:56; p < :01Þ in the full-termsand preterms with PVEP 14 days. However, a significant age by group in-teraction showed that the variance in latency to respond declined less rap-idly during this period in the preterms with PVE < 14 daysðtð32Þ ¼ 2:24; p < :025Þ. At 26 weeks, variance in this group was signifi-cantly higher than in the full-term group ðtð43Þ ¼ �2:56; p < :025Þ.

As can be seen in Fig. 3, latency to respond on noncompetition trials alsodecreased between 6 and 26 weeks. The multilevel analysis showed that thisdecrease contributed significantly to the decrease in latency on competitiontrials ðtð12Þ ¼ 2:84; p < :01Þ. To isolate the decrease in the time required by

Fig. 6. The mean and standard error of the latencies of correct refixations on competition trials

between 6 and 26 weeks for full-term infants, preterm infants with PVE < 14 days, and preterm

infants with PVEP 14 days.


processes specific to disengagement, the difference between the median laten-cies on competition and noncompetition trials was calculated for each infantfor each session. We called this disengagement cost. The multilevel analysiswas repeated with disengagement cost as the dependent variable. Like la-tency to respond, disengagement cost dropped significantly in all groups be-tween the 8- and the 14-week sessions (tð30Þ ¼ �3:69; p < :001; see Fig. 7).It dropped again between the 18- and the 26-week sessionsðtð30Þ ¼ �3:59; p < :001Þ in the full-terms and preterms with PVEP 14days. However, a significant age by group interaction showed that disen-gagement cost did not decline during this period in the preterms withPVE < 14 days ðtð30Þ ¼ 2:3; p < :025Þ. Intergroup comparisons showedno significant differences at 10 or 14 weeks. However, at 26 weeks, pretermswith PVE < 14 days tended to have higher disengagement costs than full-terms ðtð40Þ ¼ �2:21; p < :06Þ.

Discussion

Shifts of gaze to peripheral targets which did not require disengagement

In the 6- and the 10-week sessions, both groups of preterms correctly re-fixated peripheral targets more frequently and faster than full-terms, but the

Fig. 7. The mean and standard error of the disengagement cost per session between 6 and 26

weeks for full-term infants, preterm infants with PVE < 14 days, and preterm infants with

PVEP 14 days.


differences were significant only for the preterms with PVE < 14 days. From14 weeks on, there were no intergroup differences in latency or frequency ofshifts of gaze on noncompetition trials. Thus, although the preterms withPVE < 14 days did not maintain their early superiority, they were perform-ing as efficiently as the full-terms at the end of the measurement period.These results suggest that, for the first several weeks after the term date,the sensory-motor processing involved in the detection of, and preparationand execution of an eye movement to, a peripheral target was more efficientin the preterms with PVE < 14 days.

These results are similar to those of Atkinson (2000) and Foreman et al.(1991) whose cross-sectional studies found that low-risk preterms respondedfaster to peripheral targets than full-terms of the same corrected age until sev-eral weeks after the term date. After 4 weeks, Foreman et al. (1991) found nodifferences in response latencies. Landry et al. (1985) also found that low-riskpreterms responded as fast as full-terms at 7 months. However, they sug-gested that their failure to find a difference reflected lack of power. Sincewe also failed to find a difference with a larger sample of low-risk preterms,the evidence currently available is consistent in indicating that, between 3 and7 months of age, low-risk preterms perform comparably to full-terms on sim-ple shifts of gaze. Other investigations of the development of a range of visualfunctions have also failed to find differences between low-risk preterms andfull-term age mates (see, for example, van Hof-van Duin, Evenhuis-Leunen,Mohn, Baerts, & Fetter, 1989; van Hof-van Duin & Mohn, 1985).

The visual system is highly sensitive to environmental input during devel-opment, and it has been suggested that the visual systems of preterm infantsmay suffer from being exposed prematurely to the extrauterine environment(Als, 1995; Duffy, Als, & McAnulty, 1990; Friedman et al., 1981). Fielderet al. (1994), in contrast, suggested that, in preterms with few medical com-plications, looking behaviors that emerge early in development might benefitfrom the extra early visual experience. Our results are consistent with the lat-ter position. In this group of very preterm infants with few neonatal compli-cations, the processes involved in simple shifts of gaze did not suffer, butrather benefited temporarily from additional early visual experience. Pre-terms with slightly more serious complications also showed an early advan-tage. However, this was smaller and did not reach significance. Apparently,their more serious perinatal complications reduced their ability to take ad-vantage of their visual experience in the weeks following birth.

Shifts of gaze which required disengagement

Both indices of disengagement followed the same broad developmentalcourse in the three groups of infants. All groups showed a period of rapidincrease in the frequency and decrease in the latency of correct refixationsbetween the 6- and the 14-week session, followed by smaller increases in


frequency and decreases in latency between the 18- and the 26-week session.However, from the 14-week session onward, there were subtle differenceswith regard to the developmental changes in both the frequency and thelatency of responses in the competition condition between the full- and pre-term groups. The results for the low-risk preterms will be discussed first.

While the frequency of correct refixations in low-risk preterms and full-terms did not differ at any age, responses on the three to four trials in eachsession which did not lead to correct refixations of the peripheral targetdid. Staring at the fixation stimulus or inaccurate looks to the target, re-sponses typical of 6- and 10-week-old full-term infants, were more frequentin the preterms in the 14-, 18-, and 22-week sessions, while clear looks awayfrom both fixation stimulus and peripheral target emerged more slowly. Fur-ther, the low-risk preterms made fewer errors in all but the 26-week session. Alow error rate may seem to indicate superior performance. However, from 14weeks on, most of the errors made by the full-term infants were not inaccu-rate attempts to look to the target, but rather looks away from both fixationstimulus and target. In an earlier analysis of the full-term data (Butcher et al.,2000) we argued that this type of error is not a sign of incapacity, but ratheran indication that the development of disengagement has progressed furtherand infants are now able not only to inhibit attention to the fixation stimulus,but also to inhibit a response to a highly salient peripheral target in order tolook to another, apparently more attractive location. Together, the persis-tence of immature responses and the lower error frequency in the pretermssuggest that, while disengagement emerged as early and as rapidly in these in-fants, its fine-tuning, once it was established, took longer.

Latency to disengage developed quite differently in full-terms and low-riskpreterms. At 10 weeks, these infants disengaged more rapidly than the full-terms. At 26 weeks, they disengaged more slowly. Atkinson (2000) also foundshorter latencies in VLBW preterms with no neurological damage at 4–6weeks (corrected age). However, it is important to note that, in both studies,shifts of gaze in the noncompetition condition were also faster at the same agein these infants. In this study, the preterms’ disengagement costs were not sig-nificantly lower than those of the full-terms, showing that the processes spe-cific to disengagement were not faster and that the shorter response latenciesin the competition condition, as in the noncompetition condition, wereprimarily the result of the more efficient visual sensory-motor processing.

From the 14-week session onward, latency to disengage decreased moreslowly in low-risk preterms than in full-terms. At this stage, disengagementwas well established, with infants in all groups disengaging on more than80% of trials. The gradual decrease in latency in the full-term group seemedto reflect an increase in the efficiency of a looking behavior that had becomewell established (see Butcher et al., 2000). Its absence in the low-risk pre-terms then suggests the efficiency of disengagement in this group was not in-creasing. In the last two sessions, these infants disengaged more slowly than


the full-terms. The variance in latency to disengage in this group also de-creased more slowly between the 18- and the 26-week sessions and was sig-nificantly higher at 26 weeks. Further, latencies of correct refixations oncompetition trials were not becoming stable after 18 weeks in this group,while they were in the full-term group. These results are thus consistent withthe frequency results in indicating that the fine-tuning that followed the ra-pid development of disengagement occurred more slowly in low-risk pre-terms than in the full-terms. Such subtle differences in performance wereconsistent with our expectations for the group of preterms with onlyshort-lasting injury to the areas of the brain that mediate disengagement.

The results for the high-risk preterms, in contrast, were less consistentwith expectations. These infants occupied a position between the low-riskpreterms and the full-terms. While immature responses on trials which didnot lead to correct refixations of the peripheral target persisted longer,and the frequency of errors was lower in this group than in the full-terms,the differences in developmental changes in latency to disengage betweenthe 18- and the 26-week sessions did not reach significance.

It is difficult to explain why the group of preterms considered to be athigh-risk did not differ from full-terms on the latency measures. One possi-ble explanation is that the impact of PVE on the development of disengage-ment depends not only on the duration, but also on the location of the braininjury. The sample in this study, as in other studies of preterms with PVE,was homogeneous with regard to the duration but not with regard to the siteof the PVE. While all 10 high-risk infants showed persisting echodensities inthe occipital region, only two showed persisting echodensities in the parietalregion and six in the frontal region. If the impact on the development of dis-engagement of PVE lasting longer than 14 days is site specific, then the het-erogeneity of this sample would make it difficult to detect any associationbetween PVE and development.

We suggested above that the nature and extent of the differences betweenthe low-risk preterms and the full-terms was consistent with the effect ofshort-lasting injury to areas of the brain which mediate disengagement.The finding that preterms who were considered, on the basis of duration ofPVE, to be at high-risk differed less from full-terms than did preterms whowere considered to be at low-risk makes an association between durationof PVE and performance in low-risk preterms questionable. Another expla-nation is that poorer performance was associated with birth status. Even themost optimal very preterm births tend to be accompanied by a range of mi-nor complications. The preterm infants in this study, for example, sufferedfrom different combinations of respiratory problems, infections, and hyper-glycaemia, conditions which have been associated with brain injury in thevery preterm infant (Brazy et al., 1991) and which were not associated withduration of PVE. It has recently been proposed that the dorsal visualpathways, which mediate viewer-centered visuo-motor functions such as


disengagement, may be particularly vulnerable to disruption by, amongother factors, the cumulative impact of such minor complications (Atkinson,2000; see also Foreman, Fielder, Minshell, Hurrion, & Sergienko, 1997). It isnot understood why the dorsal visual pathways should be especially vulner-able to insult. However, developmental disorders in the visuo-motor func-tions mediated by these pathways are much more frequent than disordersin the perceptual functions mediated by the ventral visual pathways (Atkin-son, 2000). Further, recent studies have shown that, in very preterm infantswith optimal medical histories, visuo-motor functions mediated by the dorsalpathway are less efficient both in infancy (Atkinson, 2000) and in later child-hood (Foreman et al., 1997) than they are in full-term controls.

If a delay in the fine-tuning of disengagement following its rapid develop-ment is associated with the cumulative impact of minor complications ac-companying very preterm birth, then such a delay should have beenfound for the group of preterms as a whole. To test this, we reanalyzedthe data, treating the preterms as a single group. The reanalysis showed sig-nificant differences between the pre- and full-term infants on the frequencymeasures which had differentiated both high- and low-risk preterms fromfull-terms. The differences in the developmental trajectory and final levelof latency to disengage, though smaller than the differences between thelow-risk preterms and the full-terms, nevertheless reached significance.The results of this analysis then suggest that the differences between pre-terms and full-terms in the fine-tuning of disengagement between the 18-and the 26-week sessions that were observed in this study were more closelyassociated with the cumulative effect of complications attendant on very pre-term birth than with the duration of PVE.

The differences between pre- and full-terms on frequency measures haddisappeared by 26 weeks. While the differences in latency and variance per-sisted, they were small. At 26 weeks for example, the mean response latencyin the preterm group was approximately 10% longer than in the full-termgroup. However, shifts of gaze play an integral role in extracting informa-tion from the visual environment and guiding action upon the environment.As Rose (1983) pointed out, the impact on development of even small dif-ferences in the efficiency of such fundamental behaviors may not be negligi-ble. Further research is needed to disentangle the effects of the complicationswhich accompany very preterm birth and duration and location of PVE onthe development of disengagement.

References

Als, H. (1995). The preterm infant: a model for the study of fetal brain expectation. In J.-P.

Lecanuet, W. P. Fifer, N. A. Krasnegor, & W. P. Smotherman (Eds.), Fetal development: A

psychobiological perspective (pp. 439–469). Hillsdale, NJ: Erlbaum.


Appleton, R. E., Lee, R. E. J., & Hey, E. N. (1990). Neurodevelopmental outcome of transient

neonatal intracerebral echodensities. Archives of Disease in Childhood, 65, 27–29.

Atkinson, J. (2000). The developing visual brain. Oxford: Oxford University Press.

Atkinson, J., Hood, B., Wattam-Bell, J., & Braddick, O. J. (1992). Changes in infants’ ability to

switch visual attention in the first three months of life. Perception, 21, 643–653.

Brazy, J. E., Eckerman, C. O., Oehler, J. M., Goldstein, R. F., & O’Rand, A. M. (1991).

Nursery neurobiologic risk score: important factors in predicting outcome in very low birth

weight infants. Journal of Pediatrics, 118, 783–792.

Butcher, P. R., Kalverboer, A. F., & Geuze, R. H. (2000). Infants’ shifts of gaze from a central

to a peripheral stimulus: a longitudinal study of development between 6 and 26 weeks. Infant

Behavior and Development, 23, 3–21.

Damman, O., & Leviton, A. (1997). Duration of hyperechoic images of white matter in very-

low-birthweight infants: a proposed classification. Developmental Medicine and Child

Neurology, 39, 2–5.

Duffy, F., Als, H., & McAnulty, G. B. (1990). Behavioral and electrophysiological evidence for

gestational age effects in healthy preterm and full-term infants studied two weeks after

expected due date. Child Development, 61, 1271–1286.

Fawer, C. L., & Calame, A. (1991). Significance of ultrasound appearances in the neurological

development and cognitive abilities of preterm infants at 5 years. European Journal of

Pediatrics, 150, 515–520.

Fazzi, E., Orcesi, S., Caffi, L., Ometto, A., Rondini, G., Telesca, C., & Lanzi, G. (1994).

Neurodevelopmental outcome at 5–7 years in preterm infants with periventricular

leukomalacia. Neuropediatrics, 25, 134–139.

Fielder, A. R., Foreman, N., Moseley, M., & Robinson, J. (1994). Prematurity and visual

development. In K. Simon (Ed.), Early visual development, normal and abnormal (pp. 485–

504). New York: Oxford University Press.

Foreman, N., Fielder, A., Minshell, C., Hurrion, E., & Sergienko, E. (1997). Visual search.

Perception and visual-motor skill in ‘‘healthy ‘‘ children born at 27–32 weeks’ gestation.

Journal of Experimental Child Psychology, 64, 27–41.

Foreman, N., Fielder, A., Price, D., & Bowler, V. (1991). Tonic and phasic orientation in full-

term and preterm infants. Journal of Experimental Child psychology, 51, 407–422.

Frick, J. E., Colombo, J., & Saxon, T. F. (1999). Individual and developmental differences in

disengagement of fixation in early infancy. Child Development, 70, 537–548.

Friedman, S. L., Jacobs, B. S., & Werthmann, M. W. (1981). Sensory processing in pre- and

full-term infants in the neonatal period. In S. L. Friedman, & M. Sigman (Eds.), Preterm

birth and psychological development (pp. 159–178). New York: Academic Press.

Haith, M. M., Hazan, C., & Goodman, G. S. (1988). Expectation and anticipation of dynamic

visual events by 3.5-month-old babies. Child Development, 59, 467–479.

Hicks, J. M., & Richards, J. E. (1998). The effects of stimulus movement and attention on

peripheral stimulus localization by 8- to 26-week-old infants. Infant Behavior and

Development, 21, 571–589.

Hood, B. M., & Atkinson, J. (1993). Disengaging visual attention in the infant and adult. Infant

Behavior and Development, 16, 405–422.

Hood, B. M., Murray, L., King, F., Hooper, R., Atkinson, J., & Braddick, O. (1996).

Habituation changes in early infancy: longitudinal measures from birth to6months. Journal

of Reproductive and Infant Psychology, 14, 177–185.

Ichord, R. N. (1993). Neurologic complications. In F. R. Witter, & L. G. Keith (Eds.),

Textbook of prematurity: Antecedents, treatment and outcome (pp. 305–320). Boston: Little,

Brown and Company.

Johnson, M. H. (1998). The neural basis of cognitive development. In D. Kuhn, & R. Siegler

(Eds.), Handbook of child psychology. Vol. 2. Cognition perception and language (pp. 1–49).

New York: Wiley.


Johnson, M. H., Posner, M. I., & Rothbart, M. K. (1991). The development of visual attention

in infancy: contingency learning, anticipation and disengagement. Journal of Cognitive

Neuroscience, 3, 335–344.

Kaye, K., & Fogel, A. (1980). The temporal structure of face-to-face communication between

mothers and infants. Developmental Psychology, 16, 454–464.

Kingstone, A., & Klein, R. M. (1993). Visual offsets facilitate saccadic latency: Does

predisengagement of visuospatial attention facilitate this gap effect? Journal of Experimental

Psychology: Human Perception and Performance, 19, 1251–1265.

Krakow, J. B., & Kopp, C. B. (1982). Sustained attention in young Down’s syndrome children.

Topics in Early Childhood Special Education, 2, 32–42.

Landry, S. H., Leslie, N. A., Fletcher, J. M., & Francis, D. J. (1985). Visual attention skills of

premature infants with and without intraventricular hemorrhage. Infant Behavior and

Development, 8, 309–321.

Levene, M., Dowling, S., Graham, M., Galton, M., & Phillips, M. (1992). Impaired motor

function (clumsiness) in 5 year old children: correlation with neonatal ultrasound scans.

Archives of Disease in Childhood, 67, 687–690.

Masi, W. S., & Scott, K. G. (1983). Preterm and full-term infants visual responses to mothers’

and strangers’ faces. In T. M. Field, & A. Sostek (Eds.), Infants born at risk: physiological,

perceptual and cognitive processes (pp. 173–180). New York: Grune and Stratton.

Maurer, D., & Lewis, T. (1998). Overt orienting to peripheral stimuli: normal development and

underlying mechanisms. In J. E. Richards (Ed.), Cognitive Neuroscience of Attention (pp.

53–102). Mahwah, NJ: Erlbaum.

Mohn, G., & van Hof-van Duin, J. (1986). Development of the binocular and monocular visual

fields of human infants during the first year of life. Clinical Vision Science, 1, 51–64.

Papile, L., Burstein, J., Burstein, R., & Koffler, H. (1978). Incidence and evolution of

subependymal and intraventricular hemorrhage: a study of infants with birth weights less

than 1500 gm. Journal of Pediatrics, 92, 529–534.

Posner, M. I., Inhoff, A. W., Friedrich, F. J., & Cohen, A. (1987). Isolating attention systems: a

cognitive-anatomical analysis. Psychobiology, 15, 107–121.

Prechtl, H. F. R., & Beintema, D. (1964). The neurological examination of the full-term newborn

infant (Little Club Clinics in Developmental Medicine no. 12). London: Heineman.

Rose, S. A. (1983). Differential rates of information processing in full-term and preterm infants.

Child Development, 54, 1189–1198.

Rose, S. A., Feldman, J. F., McCarton, C. M., & Wolfson, J. (1988). Information processing in

seven-month-old infants as a function of risk status. Child Development, 59, 589–603.

Sigman, M., Cohen, S. E., Beckwith, L., & Parmalee, A. H. (1986). Infant attention in relation

to intellectual abilities in childhood. Developmental Psychology, 22, 788–792.

Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis: an introduction to basic and

advanced multilevel modelling. London: Sage.

Stechler, G., & Latz, E. (1966). Some observations on attention and arousal in the human

infant. Journal of the American Academy of Child Psychiatry, 5, 517–525.

Stevens, J. (1992). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Erlbaum.

Tellinghuisen, D. J., & Oakes, L. M. (1997). Distractibility in infancy: the effects of distractor

characteristics and type of attention. Journal of Experimental Child Psychology, 64, 232–

254.

van Hof-van Duin, J., Evenhuis-Leunen, A., Mohn, G., Baerts, W., & Fetter, W. P. F. (1989).

Effects of very low birth weight (VLBW) on visual development during the first year after

term. Early Human Development, 20, 255–266.

van Hof-van Duin, J., & Mohn, G. (1985). The development of visual function in preterm

infants. Ergebnisse der Experimentellen Medizin, 46, 350–361.

van Wulfften Pathe, T., 1986. Neural maturation and early social behaviour: a longitudinal study

of mother–infant interaction. Doctoral dissertation. Groningen: van Denderen.


Vervloed, M. P. J., 1995. Learning in preterm infants: habituation, operant conditioning

and their associations with motor development. Unpublished doctoral dissertation,

Rijksuniversiteit Groningen.


A longitudinal study of the development of shifts of gaze to a peripheral stimulus in preterm infants with transient periventricular echogenicity

Documents