This article was downloaded by: [University of Sussex Library] On: 28 October 2014, At: 10:54 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Child Neuropsychology: A Journal on Normal and Abnormal Development in Childhood and Adolescence Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ncny20

Children born with very low birth weight show difficulties with sustained attention but not response inhibition a

b

c

Katherine A. Johnson , Elaine Healy , Barbara Dooley , Simon P. d

Kelly & Fiona McNicholas

b

a

School of Psychological Sciences, University of Melbourne, Parkville, Australia b

Saint John of God Lucena Clinic Services, Dublin, Ireland

c

School of Psychology, University College Dublin, Dublin, Ireland

d

Department of Biomedical Engineering, City College of the City University of New York, New York, USA Published online: 24 Oct 2014.

To cite this article: Katherine A. Johnson, Elaine Healy, Barbara Dooley, Simon P. Kelly & Fiona McNicholas (2014): Children born with very low birth weight show difficulties with sustained attention but not response inhibition, Child Neuropsychology: A Journal on Normal and Abnormal Development in Childhood and Adolescence, DOI: 10.1080/09297049.2014.964193 To link to this article: http://dx.doi.org/10.1080/09297049.2014.964193

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions

Child Neuropsychology, 2014 http://dx.doi.org/10.1080/09297049.2014.964193

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Children born with very low birth weight show difficulties with sustained attention but not response inhibition Katherine A. Johnson1, Elaine Healy2, Barbara Dooley3, Simon P. Kelly4, and Fiona McNicholas2 1

School of Psychological Sciences, University of Melbourne, Parkville, Australia Saint John of God Lucena Clinic Services, Dublin, Ireland 3 School of Psychology, University College Dublin, Dublin, Ireland 4 Department of Biomedical Engineering, City College of the City University of New York, New York, USA 2

Children born with very low birth weight perform poorly on executive function and attention measures. Any difficulties with sustained attention may underpin impairments in performance on tasks measuring higher order cognitive control. Previous sustained attention research in very low birth weight cohorts has used tasks that involve arousing stimuli, potentially spoiling the measure of sustained attention. The aim of this study was to compare the performance of very low birth weight and normal birth weight children on a well-controlled task of sustained attention. The Fixed and Random versions of the Sustained Attention to Response Task were given to 17 very low birth weight and 18 normal birth weight children. The very low birth weight group performed the Fixed and Random Sustained Attention to Response Tasks in a similar manner as the normal birth weight group on all measures except for the omission error and Slow Frequency Area under the Spectra variables on the Fixed Sustained Attention to Response Task. These measures index lapses in sustained attention that may be underpinned by declining arousal. The very low birth weight group showed no response inhibition deficits. Omission errors and slow-timescale response-time variability on predictable tasks may thus present sensitive indices of difficulties with sustained attention and arousal associated with premature birth and low birth weight. Keywords: Very low birth weight; Response time; Gaussian; Arousal; Fourier analysis.

Survival rates of infants born with very low birth weight (VLBW), defined as a birth weight of less than 1500 g, have increased in accordance with improvements in obstetric and neonatal care (Allen, 2008). VLBW may be the result of premature birth or restricted fetal intrauterine growth (United Nations Children’s Fund and World Health Organization, 2004). These children We wish to thank the children and parents who participated in this study, Dr. Martin White and Dr. Margaret Sheridan for referring the children, Susie Coakley and Niamh O’Connor for data collection, and Associate Professor Peter Anderson for reading a draft of this work. St. John of God Hospitaller Services funded this study. There are no conflicts of interest. Address correspondence to Katherine A. Johnson, School of Psychological Sciences, University of Melbourne, Parkville, Australia. E-mail: [email protected]

© 2014 Taylor & Francis

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

2

K. A. JOHNSON ET AL.

are at increased risk of cerebral palsy and motor, cognitive, and sensory impairments (Allen, 2008). These impairments may be due to disturbances in brain function caused by interruption to normal in-utero brain development, the development of focal brain injuries, and/or ischemic or hemorrhagic lesions following very preterm birth (Bracewell & Marlow, 2002). A recent meta-analysis of studies of general cognitive ability in children born with VLBW found evidence of moderate-to-severe deficits in mathematics (weighted average effect size –0.60), reading (–0.48), spelling (–0.76), verbal fluency (–0.57), working memory (–0.36), and cognitive flexibility (–0.49), where an effect size of 1.0 represents one standard deviation from the average of a control group’s performance (Aarnoudse-Moens, Weisglas-Kuperus, van Goudoever, & Oosterlaan, 2009). Difficulties were also noted in parent (–0.59) and teacher (–0.43) reports of attention from the Child Behavior Checklist (Aarnoudse-Moens et al., 2009). Greater time spent focusing and sustaining attention to objects in preterm infancy is associated with greater cognitive abilities at 2 (Kopp & Vaughn, 1982) and 4 and 5 years (Lawson & Ruff, 2004) of age. Attention control difficulties may underpin deficits in higher order cognitive skills (Anderson & Doyle, 2004). A recent meta-analysis of studies investigating attention control in children born with VLBW found moderate impairment in sustained attention (weighted average effect size 0.45), response inhibition (0.25), and selective attention (0.38) (Mulder, Pitchford, Hagger, & Marlow, 2009). All of the nine sustained attention studies represented in this meta-analysis measured sustained attention through performance on classical Continuous Performance Tasks (CPT). In the standard CPT, the participant is asked to view a series of letters and to respond with a button press when a target letter appears. The Mulder et al. (2009) study used the number of omission errors made, a count of missed targets, as the measure of sustained attention. This study also used the number of commission errors, a count of responses made to nontargets, on the CPT as one of several measures of response inhibition. Alongside omission and commission errors, there are a number of other measures that can be taken from CPT-type tasks. The mean response time (RT) provides a measure of information-processing speed and variability in RT provides a measure of the integrity of the brain’s information-processing systems (MacDonald, Nyberg, & Bäckman, 2006). The balance of omission and commission errors can be used in signal detection theory-based measures of performance on the CPT. These include d-prime, a measure of perceptual sensitivity to the target, and criterion, a measure of a person’s response bias (Johnson, Kelly, et al., 2008). An updated summary of the findings from studies of sustained attention in premature and/or low birth weight participants is provided in Table 1 (Anderson et al., 2011; Bayless & Stevenson, 2007; Elgen, Lundervold, & Sommerfelt, 2004; Grunau, Whitfield, & Fay, 2004; Jakobson, Frisk, & Downie, 2006; Katz et al., 1996; Kulseng et al., 2006; Litt et al., 2012; Martel, Lucia, Nigg, & Breslau, 2007; Mulder, Pitchford, & Marlow, 2011; Pyhala et al., 2011; Taylor, Hack, & Klein, 1998; WilsonChing et al., 2013). Given the moderate effect size from the meta-analysis, a surprising conclusion drawn from a review of these 13 previous studies is the scarcity of significant differences in performance between the (V)LBW and control groups. Significant differences were noted in only two of five studies measuring omission errors (Elgen et al., 2004; Katz et al., 1996), and two of seven studies measuring commission errors (Katz et al., 1996; Taylor et al., 1998). One of four studies measuring RT noted a slowing in response time by the VLBW group (Kulseng et al., 2006). These inconsistent findings seem not to be due to insufficient statistical power, as the experimental group sizes were large. Rather, the inconsistencies may lay with the exogenously arousing nature of the tasks used. The standard CPT, which was designed to place demands on the endogenous, internally derived, maintenance of task-directed focus, nevertheless provides a degree of

5 to 9 years

•68 children born 61 children born at full term weighing < 750g (mean (mean birth weight = birth weight = 670 g; mean 3300 g) gestational age = 25.7 weeks) •65 children born weighing between 750 and 1499 g (mean birth weight = 1174 g, mean gestational age = 29.4 weeks)

Taylor et al. (1998)

Age at testing 6 to 8 years

Control group

64 children born prematurely 40 full-term children (26–34 weeks gestation) (mean birth weight = 1228 g, mean gestation = 29.2 weeks)

Study group

Katz et al. (1996)

Study

Microcomputer Test of Attention, a CPT-like task

CPT

Test

Result

(Continued )

Premature group made significantly more omission errors than the control group, particularly the younger children. Premature group made significantly more commission errors than the control group. Mean interstimulus interval VLBW group (ISI) (increases with performed with more incorrect responses significantly and decreases following greater ISI correct responses) compared with Mean number of LBW and control commission errors groups. VLBW group made significantly more commission errors than the LBW and control groups.

Measure Mean number of omission errors Mean number of commission errors

Table 1 A review of previous studies investigating sustained attention in premature and/or very low birth weight participants.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

SUSTAINED ATTENTION IN VLBW 3

Study group

Control group

6 years

14 years

43 children born prematurely 19 controls (birth weight (23 to 30 weeks) (mean range 2858–4366 g, birth weight = 883 g, mean gestation range 37–41 gestation = 26.6 weeks) weeks)

•54 adolescents born (≤ 1500 83 controls (mean birth g) (mean birth weight = weight = 3690 g, mean 1178 g, mean gestation = gestation = 39.6 weeks) 28.9 weeks) •60 adolescents born small for gestational age (mean birth weight = 2920 g, mean gestation = 39.5 weeks)

Kulseng et al. (2006)

17 years

Jakobson et al. (2006)

42 controls (mean birth weight = 3506 g, mean gestation = 40 weeks)

53 adolescents born ≤ 800 g birth weight (mean birth weight = 719 g, mean gestation = 25.8 weeks)

11 years

Age at testing

Grunau et al. (2004)

Elgen et al. (2004) 129 children born weighing < 129 children born 2000 g (mean birth weight weighing > 3000 g and = 1537 g, mean gestation = gestation > 37 weeks 32 weeks)

Study

Table 1 (Continued).

Mean number of omission errors Mean number of commission errors Mean RT Change in mean RT across blocks Mean number of commission errors Variability in RT Mean d-prime Mean criterion Total score (mean number of omission and commission errors)

Measure

Gordon Diagnostic System vigilance test (CPT-type task) CPT Mean number of omission errors Mean number of commission errors Mean RT Change in Mean RT across blocks Standard Error change across blocks

CPT

CPT

Test

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

(Continued )

Preterm group performed significantly more poorly than control group. No group difference No group difference VLBW group significantly slower than SGA and control groups. No group difference No group difference

LBW children made significantly more omission errors than control group. No group difference No group difference No group difference No group difference No group difference No group difference No group difference

Result

4 K. A. JOHNSON ET AL.

Pyhala et al., (2011)

Mulder et al. (2011)

Anderson et al. (2011)

Martel et al. (2007) Bayless and Stevenson (2007)

Study

105 controls (mean birth weight 3,609 g, mean gestation = 40.1 weeks)

173 controls (> 36 weeks gestation and birth weight > 2,499 g) (mean birth weight = 3507 g, mean gestation = 39.3 weeks) 22 controls

41 controls

40 premature (< 32 weeks gestation) (mean birth weight = 1201 g, mean gestation = 28.5 weeks) 189 EP/ELBW (gestational age = 22–27 weeks and/ or < 1,000 g birth weight) (mean birth weight = 833 g, mean gestation = 26.5 weeks) 56 very premature (< 31 weeks gestation) (mean gestation = 27.6 weeks gestation) 103 VLBW (< 1,500 g birth weight) (mean birth weight = 1,140 g, mean gestation = 29.3 weeks)

Control group 350 controls

Study group

473 ELBW (≤ 2,500g)

Table 1 (Continued).

25 years

9 to 10 years

8 years

6 to 12 years (mean 8 years)

6 years

Age at testing

CPT

Score! subtest from the TEACh

Score! subtest from the TEACh

Score! subtest from the TEACh

CPT

Test

Measure

Omission errors Mean number of Commission errors Perseveration Mean RT

Number of correct Score games

Number of correct Score games

Mean d-prime Mean beta (criterion) Number of correct Score games

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

(Continued )

No group difference except for an increased odds ratio of belonging to the group making 5 or more omission errors for the VLBW group. No group difference No group difference No group difference

No group difference

EP/VLBW group significantly less correct games than control group.

No group difference No group difference No group difference

Result

SUSTAINED ATTENTION IN VLBW 5

Measure

Rapid Visual Mean correct responses Information (Hits) Processing from the CANTAB (CPT-type task) TOVA Mean number of omission errors Mean number of commission errors Mean RT SDRT

Test

No No No No

group group group group

difference difference difference difference

ELBW group significantly less correct responses than control group.

Result

Notes: beta (criterion) = a measure of response bias based on signal detection theory; CANTAB = Cambridge Neuropsychological Test Automated Battery; CPT = Continuous Performance Task; d-prime = a measure of sensitivity to the presence of a target based on signal detection theory; EP = extremely premature; ELBW = extremely low birth weight; RT = response time; SDRT = standard deviation in response time; SGA = small for gestational age; TEA-Ch = Test of Everyday Attention for Children; TOVA = Test of Variables of Attention.

17 years

228 extremely premature and/ 166 controls or ELBW (≤ 1,000 g) (mean birth weight = 884 g, mean gestation = 26.6 weeks)

Age at testing

Wilson-Ching et al. (2013)

Control group

181 ELBW (< 1,000 g) (mean 115 controls (gestation > 36 14 years birth weight = 815 g, mean weeks) (mean birth weight gestation = 26.4 weeks) = 3260 g)

Study group

Litt et al. (2012)

Study

Table 1 (Continued).

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

6 K. A. JOHNSON ET AL.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

SUSTAINED ATTENTION IN VLBW

7

exogenous, external alerting support for the participant because the infrequent, salient targets capture attention, thus potentially hindering the measurement of sustained attention (Johnson, Kelly, et al., 2007). The Sustained Attention to Response Task (SART) is an alternative computerized neuropsychological Go/No-Go task used to measure sustained attention and response inhibition abilities. The SART involves the presentation of a series of single digits (Robertson, Manly, Andrade, Baddeley, & Yiend, 1997). The participant is asked to respond to every digit presented except to Digit 3, to which a response is required to be withheld. In the fixed version of the task, the Digits 1 to 9 are presented in a predictable, ascending, and repeating pattern. In the alternative random version, the digits are pseudo-randomly presented. Thus, the participant must withhold a response to an infrequent target and the “go” response is primed. The fixed SART engages the frontalparietal brain network involved in the voluntary maintenance of performance (Manly et al., 2003). The SART may provide greater insight into the neurocognitive outcomes associated with VLBW than the CPT. New analysis methods have been recently applied to the RT data gathered from neuropsychological tasks (Castellanos et al., 2005; Leth-Steensen, King Elbaz, & Douglas, 2000), which parse RT variability in insightful ways beyond the simple calculation of variance. The application of the Fast Fourier Transform (FFT) to RT time series breaks down overall RT variability into different timescales and has been used to dissociate fluctuations in performance on a moment-to-moment basis from slower changes in performance across the entire task (Johnson, Kelly, et al., 2007). Another way to analyze RT variability that has been increasingly adopted in clinical work is to model RT distributions as ex-Gaussian distributions (Geurts et al., 2008; Swick, Honzel, Larsen, & Ashley, 2013). This method allows for an assessment of extremely long RTs, which may indicate attention lapses (Leth-Steensen et al., 2000), whereas the traditional calculation of the mean RT condenses these exceptionally long RTs. Finally, time-on-task effects over the course of the task can be specifically tested for by simply halving the task and comparing the performances of the first and second halves. This split-half analysis has demonstrated significant time-on-task effects in children with attention deficit/hyperactivity disorder (ADHD) that might have been missed by examining performance on the whole task (Johnson, Kelly, et al., 2007). These new analysis techniques have proven useful in examining sustained attention performance in children with ADHD but have not yet been studied in the VLBW cohort. Such analysis may provide additional information in children born with VLBW. The aim of this study was to use a nonarousing sustained attention task with these new analysis techniques to deepen our understanding of sustained attention ability in a group of children born with VLBW and matched controls. Based on the meta-analysis results of Mulder et al. (2009) it was hypothesized that the VLBW group would have difficulties with sustained attention as measured by the omission error rate and the slowfrequency variability in RT measures, but would perform in a similar manner as the control group on the other SART measures.

METHOD Participants Twenty children in the very low birth weight (VLBW) group and 20 normal birth weight (NBW) children participated. Two children from the VLBW group were excluded from the analyses due to very low IQs (48, 41). One child from the VLBW and 2 children

8

K. A. JOHNSON ET AL.

Table 2 Demographics of the children in the very low and normal birth weight groups. Group Number of participants Age in years, mean (SD) Age in months, mean (SD) Gender, male/female IQ, mean (SD), range Birth weight in grams, mean (SD), range Gestational age in weeks, mean (SD), range

Very Low Birth Weight

Normal Birth Weight

17 12.3 (1.0) 152.3 (12.4) 8/9 91.1 (12.8), 72 to 117* 1177 (218), 710 to 1500** 29.6 (3.2), 21 to 34**

18 12.2 (0.8) 151.8 (10.4) 6/12 101.7 (11.0), 86–123 3528 (582), 2523–4508 40.2 (1.3), 36–42

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

*p < .05; **p < .001.

from the NBW group were excluded due to very high errors of omission on the experimental task, suggesting almost full disengagement from the task. The final sample consisted of 17 children in the VLBW and 18 children in the NBW groups (see Table 2). The VLBW children were recruited from a data set of children born with very low birth weight (VLBW; birth weight < 1500 g) between January 1, 1995 and December 31, 1997 in a major maternity hospital in Dublin, Ireland. The NBW children were recruited from a data set of children born next after a VLBW child from the same hospital. Exclusion criteria for participation in the current study included severe neurological, physical, or cognitive (< 70 on the Wechsler Intelligence Scale for Children [WISC; Wechsler, 2004]) dysfunction. In terms of the health of the mothers, one had a history of epilepsy and was on medication during pregnancy. No mother reported the use of illicit substances. Ten mothers of VLBW children and 4 mothers of the control group smoked during pregnancy, with rates ranging from occasional use to 30 cigarettes per day. Ten mothers of VLBW children and 14 mothers of control participants used alcohol during pregnancy and none were recorded as having an alcohol abuse problem. Of the VLBW group, 7 children were diagnosed with bronchopulmonary dysplasia, 9 received postnatal steroids, 6 were diagnosed with patent ductus arteriosus, and 4 had abnormal ultrasound scans suggesting intracranial bleeding. Five VLBW children were diagnosed with some visual disturbance but vision was corrected with the use of glasses. After complete description of the study to the participants, written, informed consent was obtained. The study was conducted under the approval of local Ethics Committees in accordance with the Declaration of Helsinki. There were no significant differences between the two groups in terms of age at assessment or gender. IQ was assessed using the full Wechsler Intelligence Scale for Children—IV (Wechsler, 2004), with the exception of one VLBW participant who completed a sample of subtests, for whom a full-scale IQ was derived using Sattler’s method (Sattler & Dumont, 2004). The full-scale IQ was significantly lower in the VLBW group compared with the NBW group, F(1, 33) = 6.965, p = .013, ηp2 = .17. An analysis of the effects of the difference in IQ was conducted by removing the three controls with the highest IQs, resulting in a nonsignificant difference in IQ between the groups. No significant changes in the analyses were detected with the slightly reduced control group (n = 15) except for one measure, the Slow Frequency Area Under the Spectra (SFAUS), in which the significant difference shifted to a trend significance. Due to definition, there was a significant difference in birth weight between the two groups, F(1, 33) = 244.906,

SUSTAINED ATTENTION IN VLBW

9

p < .001, ηp2 = .88. There was also a significant difference between the two groups in terms of gestational age, F(1, 33) = 162.579, p < .001, ηp2 = .83.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Apparatus and Procedure Participants performed both the fixed and random versions of the Sustained Attention to Response Task (SART), presented on a laptop computer (Robertson et al., 1997). In the fixed version, the Digits 1 to 9 were presented individually in fixed sequential order, and this sequence was repeated 25 times. In the random version, the same set of 225 digits was presented in a pseudorandom order. During the SARTs, a single digit appeared on the laptop screen for 313 ms, a mask was then presented for 125 ms, followed by a response cue (a bold cross) for 63 ms and a second mask for 375 ms, and then a fixation cross for 563 ms. The total interstimulus interval was 1439 ms (digit onset to digit onset). Participants were instructed to respond, using a mouse click, to every digit (Go trial) except “3” (No-Go trial). The response was to be made when the response cue appeared on the screen 125 ms after the digit was extinguished, or 438 ms from the start of the trial. This response cue was used to limit the impulsive response style observed in some children and to reduce any speed/accuracy trade-offs. Both the fixed and random SARTs consisted of 225 trials, lasting approximately 5.5 minutes. The order of presentation of the fixed and random SARTs was counterbalanced across participants. Data Analysis Commission and Omission Errors, d’, and Criterion. The method of Johnson, Kelly, et al. (2008) for analysis of the SART data was followed (Johnson, Kelly, et al., 2008). Errors of commission, where responses were made in response to the Digit 3 (No-Go), and errors of omission, a failure to respond to every other digit type (Go), were calculated. Sensitivity to the No-Go “3” target (d’) and response bias (criterion) measures were calculated using simple choice (yes/no) signal detection theory. The d’ represents a person’s sensitivity to stimulus identity: the higher the d’ value, the better the discrimination between a target and a nontarget. According the signal detection theory, the criterion represents the person’s bias in response: As the bias to respond “yes that was a target” to a stimulus increases, the criterion approaches 0 and an unbiased observer will have a criterion of 1. The criterion is independent of d’. Response Time Analyses. The mean and standard deviation of the response times (RTs) of the Go trials were calculated. Data Preparation for the FFTs and Ex-Gaussian Analyses. The sequence of 225 RTs for each SART version was analyzed using a Fast Fourier Transform (FFT) in Matlab (Johnson, Kelly, et al., 2008). A detailed description of data preparation and the procedures for derivation of the FFT spectra can be found in Johnson, Barry, et al. (2008), Johnson, Kelly, et al. (2007), Johnson, Kelly, et al. (2008), and Johnson, Robertson, et al. (2007). When commission errors were made, the RTs for Digit 3 and the next digit were both linearly interpolated from the immediately preceding and following RTs, as previous research has shown that the RT on the digit immediately following a commission error can be longer than normal (Molenberghs et al., 2009). In addition, RTs of less than 100 ms

10

K. A. JOHNSON ET AL.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

were linearly interpolated from the immediately preceding and following RTs. The individual RT time series were detrended for the fast frequency area under the spectra (FFAUS) measure to subtract out any linear components.

Derivation of the FFT Spectra. The RT data were analyzed according to Welch’s averaged, modified periodogram method, over the entire block (225 data points per individual). The time series was divided into seven segments of 75 data points with an overlap of 50. Each segment was Hamming-windowed and zero-padded to length 450. The FFT was then calculated for each segment, and each segment was averaged across the seven segments to provide a spectrum per individual. All RT data points were represented in this analysis. Any segments of 75 data points where there were over nine errors of omission (not necessarily consecutively) were excluded in the FFT; this occurred for 1 VLBW participant in the random SART. The FFT allows for a calculation of the area under the spectrum (AUS) over a broad band of interest and from this the power, or overall variance, in the RT signal may be measured. No information about the original RT series is lost following the FFT; indeed, if the power over the entire frequency range is integrated, this will equate to the overall variance in the data. The consistency of any pattern in the RT data is measured by the peak power at a particular point in the spectrum. The typical slowing in RT on Digit 1 relative to Digits 9 and 2 occurs during the fixed SART task, as the participant is preparing for the upcoming No-Go Digit 3. If the participant consistently produces this pattern of slowing, a peak in the spectra at 0.0772 Hz is found (reciprocal of 9 digits × 1.439 s, the interstimulus interval of the SART). This peak was used as a marker to divide the variability in responding into the Fast Frequency and Slow Frequency Area Under the Spectra (Johnson, Kelly, et al., 2007). The Fast Frequency Area Under the Spectra (FFAUS) encompasses all sources of variability faster than once per SART cycle (0.0772 Hz; area under curve to right of dotted line in Figures S1 and S2 in online supplementary material). Moment-to-moment variability in RT was captured in this measure. The Slow Frequency Area Under the Spectra (SFAUS) encompasses all sources of variability slower than once per SART cycle (area under the curve to the left of the dotted line in Figures S1 and S2). Variability in RT that occurred over any time point greater than one SART cycle was captured in this calculation. To ensure that all low frequencies were incorporated into the SFAUS, the time series was not divided into segments. Any RT time series where there were greater than nine errors of omission in a row were excluded in the FFT for the SFAUS measure; but this did not occur for any participants in this study. The data were not detrended in the calculation of the SFAUS measure to ensure that linear components of RT variation could contribute to this slow variability measure.

Data Analysis of the Ex-Gaussian Measures. The traditional RT measures of the central tendency (mean) and variability from the center (standard deviation of RT [SDRT]) are parameters of the Gaussian probability distribution (normal curve). RT data may alternatively be characterized by an exponential distribution and the parameter “Tau” is a measure of centrality of this form. An ex-Gaussian distribution characterizes a mix of the exponential and Gaussian distributions: The Mu is a measure of centrality, Sigma is a measure of the variability from the center, and Tau is a measure of the centrality of the exponential component. The parameters Mu, Sigma, and Tau were fitted to each individual SART RT data set using

SUSTAINED ATTENTION IN VLBW

11

an iterative-search-based maximum likelihood estimation procedure implemented in freely available MATLAB scripts provided by Lacouture and Cousineau (2008).

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Statistics All dependent variables were calculated per participant. The omission error and SFAUS measures were non-normal in distribution, as determined by the KolmogorovSmirnov test, and therefore nonparametric tests were used in their analyses. The MannWhitney test assessed for possible group differences on each condition. Friedman’s analysis of variation (ANOVA) assessed for differences in performance across conditions, for each group separately. The commission error, d’, criterion, mean RT, SDRT, FFAUS, Mu, Sigma, and Tau were normally distributed and were analyzed in a Group (VLBW, NBW) by SART Version (Fixed, Random) by Half (first half of the SART, second half of the SART) three-way mixed-factorial ANOVA design. The alpha level was set at .05. RESULTS A summary of the results in terms of averages and variability in data is portrayed in Table 3. Commission Errors There was no significant difference in the number of commission errors made by the VLBW and NBW groups, F(1, 33) = 0.001, p = .974, ηp2 = .001. The participants made significantly more commission errors during the random compared with the fixed SART, F(1, 33) = 44.730, p < .001, ηp2 = .575, and during the second compared with the first half of the SARTs, F(1, 33) = 12.956, p < .001, ηp2 = .282. There were no significant interactions. Omission Errors A Mann-Whitney test examined the possibility of a Group difference on the number of omission errors made on the fixed and random SART versions (see Figure 1). There was no significant difference in the number of omission errors made by the VLBW and NBW groups on the first, U = 129, ns, r = –.14, or second, U = 141, ns, r = –.07, halves of the fixed SART. Nor was there a significant difference between the two groups in the number of omission errors made in either the first, U = 135, ns, r = –.11, or second, U = 130, ns, r = –.13, halves of the random SART. Any difference in the number of omission errors made between the two halves of the fixed and random SARTs was analyzed using Friedman’s ANOVA for each group separately. There was a significant difference in the number of omission errors made between the two halves of the fixed and random SARTs for the VLBW group, Χ2(3) = 9.024, p = .026, but not for the NBW group, Χ2(3) = 7.252, p = .062. Wilcoxon’s Signed Rank Test with the Bonferroni adjustment for Type 1 error was used for post hoc comparisons, with a new alpha level of .025. For the fixed SART, the VLBW group made significantly more omission errors in the second compared with the first half, Z = –2.239, p = .014, r = –.54 (large effect size).

12

K. A. JOHNSON ET AL.

Table 3 The means and standard deviations of the very low birth weight (VLBW) and normal birth weight (NBW) groups for each parametric measure from the Sustained Attention to Response Task (SART) and the medians and interquartile ranges (IQR) of the nonparametric measures.

Measure Commission

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Mean RT

d prime

Criterion

SD RT

FFAUS

Mu

Sigma

Tau

SART

Half

VLBW Mean

VLBW SD

NBW Mean

NBW SD

Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random Fixed Fixed Random Random

First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second First Second

1.5 2.1 3.9 5.2 575 585 539 575 3.4 3.0 2.7 2.6 0.16 0.19 0.45 0.99 145 187 128 162 195 331 218 277 513 466 473 474 114 107 101 102 61 119 104 103

1.5 1.8 2.3 3.7 192 193 197 187 0.6 0.9 0.8 1.1 0.11 0.14 0.27 2.4 63 85 52 59 110 230 239 182 211 248 225 224 69 63 61 61 54 100 70 70

1.3 2.1 3.9 5.5 552 527 496 495 3.6 3.1 2.6 2.4 0.15 0.18 0.45 0.60 119 152 124 148 176 274 211 277 514 459 394 395 97 109 94 95 41 67 101 101

0.8 1.3 2.4 3.7 197 184 127 133 0.4 0.7 0.9 1.1 0.07 0.11 0.28 0.78 40 55 48 62 132 248 149 285 212 214 134 135 46 56 50 50 44 67 59 59

VLBW Median

VLBW IQR

NBW Median

NBW IQR

1.0 2.0 0.0 0.0 1297 714

2.5 4.5 1.5 3.0 1185 899

0.5 1.5 0.5 1.0 674 644

1.3 3.0 3.0 2.5 819 883

Measure

SART

Half

Omission

Fixed Fixed Random Random Fixed Random

First Second First Second

SFAUS

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

SUSTAINED ATTENTION IN VLBW

13

Figure 1 Count of omission errors made during the first and second halves of the fixed and random versions of the Sustained Attention to Response Task (SART) by the children born with very low birth weight (VLBW) and normal birth weight (NBW) control children.

For the random SART, the VLBW group made a similar number of omission errors across the two halves of the task, Z = –1.433, p = .086, r = –.35.

d’ There was no significant Group difference in sensitivity to the target “3” between the two groups, F(1, 33) = 0.004, p = .948, ηp2 = .001. Participants were significantly more sensitive to the target in the fixed SART compared with the random SART, F(1, 33) = 26.021, p < .001, ηp2 = .441. Participants were significantly more sensitive to the target in the first compared with the second half, F(1, 33) = 8.268, p = .007, ηp2 = .200. There were no significant interactions.

Criterion There was no significant Group difference in bias in response to the target “3” between the two groups, F(1, 33) = 0.363, p = .551, ηp2 = .011. Participants were significantly more likely to respond affirmatively to the target in the fixed SART compared with the random SART, F(1, 33) = 8.053, p < .008, ηp2 = .196. There was no significant effect of Half and there were no significant interactions.

Mean Response Time The VLBW and NBW groups performed the SART tasks with a similar mean response time, F(1, 33) = 0.865, p = .359, ηp2 = .026. There was no significant difference in mean RT between the fixed and random SARTs, F(1, 33) = 3.338, p = .077, ηp2 = .092, nor between the first and second half of the SARTs, F(1, 33) = 0.160, p = .692, ηp2 = .005. There were no significant interactions.

14

K. A. JOHNSON ET AL.

Standard Deviation of Response Time There was no significant difference in the SDRT of the VLBW and NBW groups, F(1, 33) = 1.757, p = .194, ηp2 = .051, nor between the fixed and random SARTs, F(1, 33) = 1.703, p = .201, ηp2 = .049. All participants performed the second half of the SARTs with significantly greater variability compared with the first half of the SARTs, F(1, 33) = 15.170, p < .001, ηp2 = .315. There were no significant interactions.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Fast Frequency Area Under the Spectra (FFAUS) Both the VLBW and NBW groups performed the SART tasks with a similar amount of moment-to-moment variability in RT (FFAUS), F(1, 33) = 0.136, p = .715, ηp2 = .004. There was no significant difference in FFAUS between the fixed and random SARTs, F(1, 33) = 0.011, p = .917, ηp2 = .001. All participants performed the second half of the SARTs with significantly greater variability compared with the first half of the SARTs, F(1, 33) = 12.738, p < .001, ηp2 = .279. There were no significant interactions. Slow Frequency Area Under the Spectra (SFAUS) A Mann-Whitney test examined if the VLBW and NBW groups differed significantly on the fixed and random SART SFAUS measures (see Figure 2). There was a significant difference between the two groups for the fixed SART SFAUS, U = 90, p = .019, r = –.35 (medium effect size), but there was no significant Group difference on the random SART SFAUS, U = 136, p = .292, r = –.09. For both groups, a Friedman’s ANOVA found no significant difference between the fixed and random SARTs. Mu There was no significant difference in Mu between the VLBW and NBW groups, F(1, 33) = 0.423, p = .520, ηp2 = .013. There was a significant SART by Half interaction,

Figure 2 Mean power of the Slow Frequency Area Under the Spectra (SFAUS), encompassing all sources of variability slower than once per Sustained Attention to Response Task (SART) cycle for the children born with very low birth weight (VLBW) and normal birth weight (NBW) control children, on the fixed and random versions of the SART.

SUSTAINED ATTENTION IN VLBW

15

F(1, 33) = 4.326, p = .045, ηp2 = .116. The Mu for the fixed SART in the first half was significantly slower than for the random SART, p = .011. There was no difference between the two SART types for Mu in the second half, p = .248. The Mu for the fixed SART was significantly greater in the first half compared with the second half, p = .045. There was no significant difference in Mu between the first and second halves of the random SART, p = .286.

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

Sigma There was no significant difference in Sigma between the VLBW and NBW groups, F(1, 33) = 0.183, p = .672, ηp2 = .006, nor between performance on the fixed and random SARTs, F(1, 33) = 0.901, p = .349, ηp2 = .027. The participants performed the first half of the SARTs with a similar amount of Sigma as the second half of the SARTs, F(1, 33) = 0.111, p = .741, ηp2 = .003. There were no significant interactions. Tau There was no significant difference in Tau between the VLBW and NBW groups, F(1, 33) = 1.184, p = .284, ηp2 = .035. There was a significant SART by Half interaction, F(1, 33) = 9.603, p = .004, ηp2 = .225. The Tau for the fixed SART in the first half was significantly smaller than for the random SART, p < .001. There was no difference between the two SART types for Tau in the second half, p = .485. The Tau for the fixed SART was significantly smaller in the first half compared with the second half, p = .004. There was no significant difference in Tau between the first and second halves of the random SART, p = .666. DISCUSSION The Very Low Birth Weight (VLBW) group performed the fixed and random SARTs in a similar manner as the Normal Birth Weight (NBW) group on all measures except for the omission error and Slow Frequency Area under the Spectra (SFAUS) variables. Whilst the two groups made a similar number of omission errors over the entire task, the VLBW group exhibited a significantly bigger increase in omission errors in the second half of the fixed SART compared with the first half. In contrast, the NBW group made a consistent number of omission errors across the task. For the random SART, both groups made a similar number of omission errors across the two halves. For the SFAUS measure, the VLBW group performed the fixed SART with a significantly greater amount of slow variability in response time (RT) compared with the NBW. The two groups did not differ significantly on the SFAUS of the random SART. The omission error and SFAUS measures index waning sustained attention (Johnson, Kelly, et al., 2008). Thus, through the use of very specific measures of sustained attention, a subtle deficit in sustained attention has been demonstrated in a group of normally intelligent VLBW children. Omission errors are made when a participant fails to respond to a “Go” stimulus and index a drifting of attention from the task (Johnson, Robertson, et al., 2007; O’Connell et al., 2009). This may be related to lowering arousal levels (Johnson, Kelly, et al., 2008) and may index a deficit in the alerting neural network (Petersen & Posner, 2012). There was no overall difference in the number of omission errors made by the two groups;

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

16

K. A. JOHNSON ET AL.

however, the half analysis indicated that the VLBW group made significantly more omission errors in the second half of the fixed SART compared with the first half. The VLBW group performed normally on all aspects of the random SART. The half analysis has not been used in previous analyses of sustained attention performance in VLBW cohorts (see Table 1), yet it highlights a deficit in attention over time in this group. The slow frequency area under the spectra (SFAUS) measure reflects slow changes in RT during the course of the 5-minute SART; it encapsulates all sources of variation in RT that occur over any time period greater than one SART cycle (the 1–9 stimulus cycle taking 12.9 seconds). Greater slow frequency variability may reflect deficits in endogenously produced arousal (Johnson, Barry, et al., 2008; Johnson, Kelly, et al., 2007), which may also index deficiencies in the alerting neural network (Petersen & Posner, 2012). The VLBW group showed greater variability in SFAUS during the fixed SART; during the random SART, the VLBW group performed normally. The random SART, with the unpredictable and exogenously alerting stimuli, is inherently more arousing than the entirely predictable presentation of the go and no-go stimuli of the fixed SART. The monotony of the fixed SART may have precipitated a decline in sustained attention to the task (O’Connell et al., 2009) in this VLBW group. This subtle difference, sustained attention during a dull or monotonous task, may be a specific weakness for VLBW cohorts. Educational tasks requiring sustained attention over long periods of time should be inherently arousing in order to maximize continuous attention in this group. In terms of the response time measures used in this study, the VLBW group showed a significant difference from the control group on only the SFAUS measure. The fast frequency area under the spectrum (FFAUS) measures the moment-to-moment variability in response time separately from the low-frequency components related to slowing in RT over the course of the task. The FFAUS, along with the number of commission errors, is argued to measure top-down cortical control of attention (Johnson, Kelly, et al., 2008). Like the commission error rate, the VLBW group performed both SARTs normally in terms of the FFAUS, indicating that the attentional control network (Petersen & Posner, 2012; Posner & Petersen, 1990) may be functioning normally in this group of children born with very low birth weight. The mean RT, standard deviation of RT, and the ex-Gaussian measures of RT were not sensitive to the subtle behavior exhibited by the VLBW group, as indexed by the SFAUS measure. The traditional RT measures and the probability density function calculations are not designed to scrutinize slow changes in RTs across the course of the task. An FFT or continuous wavelet transform is necessary to determine the power of periodic changes in RT at different temporal frequencies and this is the first application of this method to a VLBW cohort. Commission errors are a failure to withhold a response to a No-Go stimulus and index lapses in response inhibition (O’Connell et al., 2009), which may be driven by difficulties in top-down cortical control of attention (Johnson, Kelly, et al., 2008). The VLBW group performed the fixed and random SARTs normally in terms of commission errors, indicating that the prepotent response inhibition network (Wiecki & Frank, 2013) is functional in this VLBW group. This is a consistent finding, in that five of seven previous studies measuring commission errors also found no response inhibition deficits in premature and/or VLBW participants (see Table 1). The limitation of this study is the low number of participants, and this study may be considered as a pilot. Further research with a larger cohort of VLBW participants would further strengthen the claims within this research. Based on these research findings, it would be of great interest to directly measure levels of physiological arousal, such as

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

SUSTAINED ATTENTION IN VLBW

17

through galvanic skin response or pupil diameter, in children born with VLBW to identify if and when these levels drop in comparison with normal birth weight controls. Trying to identify and develop intrinsic systems that maximize physiological arousal, especially during tasks of boredom, may assist in the cognitive ability of these individuals. In conclusion, the omission error and SFAUS measures are sensitive measures of behavior associated with sustained attention and arousal during a deliberately boring task. When the task was monotonous and understimulating, the VLBW group produced lapses in attention and responded more slowly and variably to the task. When the task was stimulating, the VLBW group performed normally on the task. The educational implications of this research suggest that in order to help sustain attention to a task, material should be presented to children born with VLBW in a stimulating fashion, in order to help maintain their focus on the task.

Supplementary material Supplementary (Figures S1 and S2) is available via the “Supplementary” tab on the article’s online page (http://dx.doi.org/10.1080/09297049.2014.964193). Original manuscript received January 17, 2014 Revised manuscript accepted September 6, 2014 First published online October 10, 2014

REFERENCES Aarnoudse-Moens, C. S. H., Weisglas-Kuperus, N., van Goudoever, J. B., & Oosterlaan, J. (2009). Meta-analysis of neurobehavioral outcomes in very preterm and/or very low birth weight children. Pediatrics, 124(2), 717–728. doi:10.1542/peds.2008-2816 Allen, M. C. (2008). Neurodevelopmental outcomes of preterm infants. Current Opinion in Neurology, 21(2), 123–128. doi:10.1097/WCO.0b013e3282f88bb4 Anderson, P. J., De Luca, C. R., Hutchinson, E., Spencer-Smith, M. M., Roberts, G., & Doyle, L. W. (2011). Attention problems in a representative sample of extremely preterm/extremely low birth weight children. Developmental Neuropsychology, 36(1), 57–73. doi:10.1080/87565641.2011.540538 Anderson, P. J., & Doyle, L. W. (2004). Executive functioning in school-aged children who were born very preterm or with extremely low birth weight in the 1990s. Pediatrics, 114(1), 50–57. doi:10.1542/peds.114.1.50 Bayless, S., & Stevenson, J. (2007). Executive functions in school-age children born very prematurely. Early Human Development, 83(4), 247–254. doi:10.1016/j.earlhumdev.2006.05.021 Bracewell, M., & Marlow, N. (2002). Patterns of motor disability in very preterm children. Mental Retardation and Developmental Disabilities Research Reviews, 8(4), 241–248. doi:10.1002/ mrdd.10049 Castellanos, F. X., Sonuga-Barke, E. J. S., Scheres, A., Di Martino, A., Hyde, C., & Walters, J. R. (2005). Varieties of attention-deficit/hyperactivity disorder-related intra-individual variability. Biological Psychiatry, 57(11), 1416–1423. doi:10.1016/j.biopsych.2004.12.005 Elgen, I., Lundervold, A. J., & Sommerfelt, K. (2004). Aspects of inattention in low birth weight children. Pediatric Neurology, 30(2), 92–98. doi:10.1016/S0887-8994(03)00402-8 Geurts, H. M., Grasman, R. P. P. P., Verté, S., Oosterlaan, J., Roeyers, H., van Kammen, S. M., & Sergeant, J. (2008). Intra-individual variability in ADHD, autism spectrum disorders and Tourette’s syndrome. Neuropsychologia, 46(13), 3030–3041. doi:10.1016/j.neuropsychologia.2008.06.013

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

18

K. A. JOHNSON ET AL.

Grunau, R. E., Whitfield, M. F., & Fay, T. B. (2004). Psychosocial and academic characteristics of extremely low birth weight (≤800g) adolescents who are free of major impairment compared with term-born control subjects. Pediatrics, 114(6), e725–732. doi:10.1542/peds.2004-0932 Jakobson, L. S., Frisk, V., & Downie, A. L. S. (2006). Motion-defined form processing in extremely premature children. Neuropsychologia, 44(10), 1777–1786. doi:10.1016/j.neuropsychologia. 2006.03.011 Johnson, K. A., Barry, E., Bellgrove, M. A., Cox, M., Kelly, S. P., Dáibhis, A., … Gill, M. (2008). Dissociation in response to methylphenidate on response variability in a group of medication naïve children with ADHD. Neuropsychologia, 46(5), 1532–1541. doi:10.1016/j. neuropsychologia.2008.01.002 Johnson, K. A., Kelly, S. P., Bellgrove, M. A., Barry, E., Cox, M., Gill, M., & Robertson, I. H. (2007). Response variability in Attention Deficit Hyperactivity Disorder: Evidence for neuropsychological heterogeneity. Neuropsychologia, 45(4), 630–638. doi:10.1016/j.neuropsychologia.2006.03.034 Johnson, K. A., Kelly, S. P., Robertson, I. H., Barry, E., Mulligan, A., Daly, M., … Bellgrove, M. A. (2008). Absence of the 7-repeat variant of the DRD4 VNTR is associated with drifting sustained attention in children with ADHD but not in controls. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 147B, 927–937. doi:10.1002/ajmg.b.30718 Johnson, K. A., Robertson, I. H., Kelly, S. P., Silk, T. J., Barry, E., Dáibhis, A., … Bellgrove, M. A. (2007). Dissociation in performance of children with ADHD and high-functioning autism on a task of sustained attention. Neuropsychologia, 45(10), 2234–2245. doi:10.1016/j. neuropsychologia.2007.02.019 Katz, K. S., Dubowitz, L. M. S., Henderson, S., Jongmans, M., Kay, G. G., Nolte, C. A., & de Vries, L. (1996). Effect of cerebral lesions on continuous performance test responses of school age children born prematurely. Journal of Pediatric Psychology, 21(6), 841–855. doi:10.1093/ jpepsy/21.6.841 Kopp, C. B., & Vaughn, B. E. (1982). Sustained attention during exploratory manipulation as a predictor of cognitive competence in preterm infants. Child Development, 53(1), 174–182. doi:10.2307/1129650 Kulseng, S., Jennekens-Schinkel, A. A. G., Naess, P., Romundstad, P., Indredavik, M., Vik, T., & Brubakk, A.-M. (2006). Very-low-birthweight and term small-for-gestational-age adolescents: Attention revisited. Acta Paediatrica, 95(2), 224–230. doi:10.1111/j.1651-2227.2006.tb02211.x Lacouture, Y., & Cousineau, D. (2008). How to use MATLAB to fit the ex-Gaussian and other probability functions to a distribution of response times. Tutorials in Quantitative Methods for Psychology, 4(1), 35–45. Lawson, K. R., & Ruff, H. A. (2004). Early focused attention predicts outcome for children born prematurely. Journal of Developmental & Behavioral Pediatrics, 25(6), 399–406. doi:10.1097/ 00004703-200412000-00003 Leth-Steensen, C., King Elbaz, Z., & Douglas, V. I. (2000). Mean response times, variability, and skew in the responding of ADHD children: A response time distributional approach. Acta Psychologica, 104, 167–190. doi:10.1016/S0001-6918(00)00019-6 Litt, J. S., Taylor, H. G., Margevicius, S., Schluchter, M., Andreias, L., & Hack, M. (2012). Academic achievement of adolescents born with extremely low birth weight. Acta Paediatrica, 101(12), 1240–1245. doi:10.1111/j.1651-2227.2012.02790.x MacDonald, S. W. S., Nyberg, L., & Bäckman, L. (2006). Intra-individual variability in behavior: Links to brain structure, neurotransmission and neuronal activity. Trends in Neurosciences, 29(8), 474–480. doi:10.1016/j.tins.2006.06.011 Manly, T., Owen, A. M., McAvinue, L., Datta, A., Lewis, G. H., Scott, S. K., … Robertson, I. H. (2003). Enhancing the sensitivity of a sustained attention task to frontal damage: Convergent clinical and functional imaging evidence. Neurocase, 9(4), 340–349. doi:10.1076/neur.9.4.340.15553 Martel, M. M., Lucia, V. C., Nigg, J. T., & Breslau, N. (2007). Sex differences in the pathway from low birth weight to inattention/hyperactivity. Journal of Abnormal Child Psychology, 35(1), 87–96. doi:10.1007/s10802-006-9089-9

Downloaded by [University of Sussex Library] at 10:54 28 October 2014

SUSTAINED ATTENTION IN VLBW

19

Molenberghs, P., Gillebert, C. R., Schoofs, H., Dupont, P., Peeters, R., & Vandenberghe, R. (2009). Lesion neuroanatomy of the Sustained Attention to Response task. Neuropsychologia, 47(13), 2866–2875. doi:10.1016/j.neuropsychologia.2009.06.012 Mulder, H., Pitchford, N. J., Hagger, M. S., & Marlow, N. (2009). Development of executive function and attention in preterm children: A systematic review. Developmental Neuropsychology, 34(4), 393–421. doi:10.1080/87565640902964524 Mulder, H., Pitchford, N. J., & Marlow, N. (2011). Processing speed mediates executive function difficulties in very preterm children in middle childhood. Journal of the International Neuropsychological Society, 17(3), 445–454. doi:10.1017/S1355617711000373 O’Connell, R. G., Dockree, P. M., Bellgrove, M. A., Turin, A., Ward, S., Foxe, J. J., & Robertson, I. H. (2009). Two types of action error: Electrophysiological evidence for separable inhibitory and sustained attention neural mechanisms producing error on Go/No-go tasks. Journal of Cognitive Neuroscience, 21(1), 93–104. doi:10.1162/jocn.2009.21008 O’Connell, R. G., Dockree, P. M., Robertson, I. H., Bellgrove, M. A., Foxe, J. J., & Kelly, S. P. (2009). Uncovering the neural signature of lapsing attention: Electrophysiological signals predict errors up to 20s before they occur. Journal of Neuroscience, 29(26), 8604–8611. doi:10.1523/JNEUROSCI.5967-08.2009 Petersen, S. E., & Posner, M. I. (2012). The attention system of the human brain: 20 years after. Annual Review of Neuroscience, 35, 73–89. doi:10.1146/annurev-neuro-062111-150525 Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13, 25–42. doi:10.1146/annurev.ne.13.030190.000325 Pyhala, R., Lahti, J., Heinonen, K., Pesonen, A.K., Strang-Karlsson, S., Hovi, P., … Raikkonen, K. (2011). Neurocognitive abilities in young adults with very low birth weight. Neurology, 77(23), 2052–2060. doi:10.1212/WNL.0b013e31823b473e Robertson, I. H., Manly, T., Andrade, J., Baddeley, B. T., & Yiend, J. (1997). ‘Oops!’: Performance correlates of everyday attentional failures in traumatic brain injured and normal subjects. Neuropsychologia, 35(6), 747–758. doi:10.1016/S0028-3932(97)00015-8 Sattler, J. M., & Dumont, R. (2004). Assessment of children: WISC-IV and WPPSI-III supplement. San Diego, CA: Jerome M. Sattler Inc. Swick, D., Honzel, N., Larsen, J., & Ashley, V. (2013). Increased response variability as a marker of executive dysfunction in veterans with Post-Traumatic Stress Disorder. Neuropsychologia, 51(14), 3033–3040. doi:10.1016/j.neuropsychologia.2013.10.008 Taylor, G. H., Hack, M., & Klein, N. K. (1998). Attention deficits in children with