Early Language Processing Efficiency
Early Language Processing Efficiency
Early Language Processing Efficiency
As rates of prematurity continue to rise, identifying which preterm children are at increased risk for
learning disabilities is a public health imperative. Identifying continuities between early and later skills
in this vulnerable population can also illuminate fundamental neuropsychological processes that support
learning in all children. At 18 months adjusted age, we used socioeconomic status (SES), medical
variables, parent-reported vocabulary, scores on the Bayley Scales of Infant and Toddler Development
(third edition) language composite, and children’s lexical processing speed in the looking-while-listening
(LWL) task as predictor variables in a sample of 30 preterm children. Receptive vocabulary as measured
by the Peabody Picture Vocabulary Test (fourth edition) at 36 months was the outcome. Receptive
vocabulary was correlated with SES, but uncorrelated with degree of prematurity or a composite of
medical risk. Importantly, lexical processing speed was the strongest predictor of receptive vocabulary
(r = −.81), accounting for 30% unique variance. Individual differences in lexical processing efficiency
may be able to serve as a marker for information processing skills that are critical for language learning.
Keywords: Language development; Children born preterm; Processing speed; Vocabulary outcomes;
Predictive validity.
Prematurity is a condition of utmost public health importance. While many children born
preterm develop typically, approximately 50% have mild to moderate disabilities that may
not be identified until the child enters school (Bhutta, Cleves, Casey, Cradock, & Anand,
2002; Marlow, 2004) and 10 to 25% of preterm children have a major disability such as
cerebral palsy or intellectual disability (Vohr et al., 2000). Given such variability in
developmental trajectories, the early identification of preterm children at risk for delays
is imperative in order to increase opportunities for early intervention. Moreover, identify-
ing continuities between early skills and later outcomes in this vulnerable population may
reveal fundamental neuropsychological processes that support learning in all children.
Thanks to the families who participated, Melanie Ashland, Mofeda Dabado, and the staff at the Language
Learning Lab.
No potential conflict of interest was reported by the authors.
This work was supported by the NIH (grant number R01 HD069150).
Address correspondence to Virginia A. Marchman, Department of Psychology, Stanford University, 450
Serra Mall, Stanford, CA 94305, USA. E-mail: [email protected]
and identify the referent are assessed with precision in relation to relevant points in the
speech signal, yielding reliable measures of reaction time (RT) and accuracy. Children
who were more efficient in spoken-word recognition at the age of 18 months, as measured
by faster RTs, showed more rapid vocabulary growth, higher IQ scores, and stronger
working memory skills at 8 years (Fernald et al., 2006; Marchman & Fernald, 2008).
Efficient language processing also predicted “catch up” in vocabulary in late-talking
infants at risk for language delays (Fernald & Marchman, 2012). Thus, how quickly
and reliably young children interpret familiar spoken words in real time is thought to
reflect efficiency in language processing that is critical for learning new vocabulary words
and that shapes longer-term trajectories of cognitive and language development.
Do individual differences in lexical processing efficiency by preterm infants at
18 months predict their later receptive vocabulary knowledge at 3 years? Finding such
relations would provide further evidence that early efficiency in language processing
supports later learning, and that these critical information processing skills are vulner-
able in populations with neurodevelopmental risk. In order to control for global imma-
turity associated with premature birth, we assessed the infants when they were
18 months but with age adjusted for degree of prematurity (AA). Outcomes were
assessed at or around the child’s third birthday (36 months). Two standard scores
were computed, one based on unadjusted (chronological) age, and the other based on
age after adjusting for degree of prematurity.
Evaluating our hypothesis required comparing the role of early language processing
efficiency to other factors that have been linked to adverse outcomes in preterm popula-
tions. Many studies have found that gestational age (GA) and birth weight (BW) are
negatively linked to neurodevelopmental functioning (Allen, 2008). For example, GA at
birth, particularly at the limits of viability, has been associated with mortality as well as
morbidity, including the probability of disability (Allen, 2008) and associated neural
markers of adverse outcomes, such as reduced gray matter volume and increased cere-
brospinal fluid volume (Inder, Warfield, Wang, Hüppi, & Volpe, 2005). However, poor
outcomes after preterm birth have been linked to a number of medical complications, such
as seizure disorder, respiratory distress syndrome and retinopathy that are generally also
associated with gestational age and birth weight (Stoll et al., 2010). Neurological condi-
tions, such as decreased brain volume, microstructure abnormalities, and alterations in
neural connectivity linked to shorter gestational periods tend to persist to school age and
are associated with learning challenges (Kidokoro et al., 2014). Importantly, recent studies
have shown that the presence of neural injuries, particularly those involving white matter,
assessed through magnetic resonance imaging (MRI), predict increased risk of poor
outcomes in older children and adolescents using tasks that specifically assess processing
speed (Feldman, Lee, Yeatman, & Yeom, 2012; Soria-Pastor et al., 2008). Of particular
interest in the current study is whether a global index of the degree of prematurity, such as
GA and BW, or cumulative risk of medical complications of preterm birth would be
associated with early measures of processing speed in preterm toddlers as well as later
language outcomes in preschool. To our knowledge, no standardized composite measure
is routinely used to assess the degree of medical risk. Therefore, we created a composite
score that catalogued the presence of 11 conditions, capturing medical risk that extends
beyond global health variables such as GA and BW.
Socioeconomic status (SES) has also been associated with children’s language
outcomes in preterm as well as full-term children (Johnson, Beitcheman, & Brownlie,
2010; Ramey & Ramey, 2004). The higher incidence of preterm birth in lower SES
652 V. A. MARCHMAN ET AL.
populations has remained stable over four decades (Glinianaia et al., 2013), with rates of
preterm birth in disadvantaged families nearly twice that of affluent families (Smith,
Draper, Manktelow, Dorling, & Field, 2007). Taken together, children born preterm
may experience cumulative risks for language delays due to a variety of medical and
social factors.
In this study, we evaluated several predictors of 3-year receptive vocabulary out-
comes in a sample of very preterm children. We first examined first-order correlations
between SES, global measures of prematurity (BW and GA), and a composite of medical
risk. We next explored the predictive validity of traditional estimates of early language
knowledge using parent report and standardized tests. We then explored the contributions
of an experimental measure of early language processing efficiency. We hypothesized that
processing efficiency, as indexed by accuracy and processing speed (RT) in the LWL task,
would predict receptive vocabulary outcomes in preterm children, based on associations
seen previously in children born full term (Fernald et al., 2006) and on findings of
compromised processing in older preterm children (Lee et al., 2011; Soria-Pastor et al.,
2008). We addressed the following questions:
● How much variance in 3-year receptive language was accounted for by family SES and
medical/birth variables?
● Did children’s vocabulary size and performance on a standardized developmental
assessment at 18 months AA predict later receptive language?
● Did experimental measures of real-time language processing in infancy account for
additional variance in later receptive vocabulary beyond SES and standardized
language assessments typically used to assess outcomes in children born preterm?
METHOD
Participants
Participants were 30 children (15 males, 15 females), all born preterm, with gesta-
tional age (GA) ≤32 weeks and birth weight (BW) <1800 grams. Most families were
recruited from the neonatal intensive care unit and the High-Risk Infant Follow-up Clinic
of the local children’s hospital. Other families were recruited through a research registry
and an online newsletter from a local mothers’ group. Exclusionary criteria were medical
conditions that would prevent participants from understanding and actively participating
in the study’s tasks, such as a history of meningitis, a seizure disorder requiring medica-
tion, a ventriculoperitoneal shunt, a genetic disorder, or visual or hearing impairments. All
children were primarily English speaking and heard less than 25% of another language at
home. The research protocol was approved by a university institutional review board
(IRB) and parents gave signed consent at each visit.
Children were tested at 18 months AA following the routine clinical practice of
adjusting the age of assessment for preterm children younger than 2 years of age. Date
of test was determined by adding number of days premature to each participant’s 18-
month birthday. In 80% of the cases, children were tested within 30 days of this target
date. Mean AA at test was 18.6 months (range = 18.0–20.3 months; chronological
age: M = 21.1, range = 20.3–22.8 months). Follow-up language measures were
administered when the children were 36 months unadjusted age (chronological age:
M = 36.5, range = 35.4–38.3 months).
PROCESSING SPEED PREDICTS OUTCOMES IN PRETERMS 653
Table 1 Descriptive Statistics for Sociodemographic, Birth History, and Medical Risk (n = 30).
M (SD) Range
Family Demographics
Maternal Education (years) 16.5 (1.7) 12–18
HI: SESa 57.3 (7.7) 35–66
Birth History
Birth Weight (g) 1231.5 (263.8) 710–1660
Gestational Age (weeks) 29.4 (1.8) 26.3–32.7
% (n)
Twin 53.3 (16)
Medical Risk Compositeb (11 max.) % (n)
Hearing Loss 0.0 (0)
Seizure Disorder 0.0 (0)
Small for gestational age: Birth weight < 10th percentile 6.7 (2)
Periventricular Leukomalacia: Injury to the white matter 6.7 (2)
around the ventricles in the brain
Necrotizing Enterocolitis: Injury to intestines 13.3 (4)
Intraventricular Hemorrhage 16.7 (5)
Grade I–II 80.0 (4)
Grade III–IV 20.0 (1)
Bronchopulmonary Dysplasia: Oxygen requirement at 36.7 (11)
28 days of life
Patent Ductus Arteriosis: Persistent abnormal circulation 36.7 (11)
through heart and major blood vessels
Retinopathy of Prematurity: Injury to the retina of the eye 43.3 (13)
Respiratory Distress Syndrome: Breathing problems due 76.7 (23)
to immature lung development
Hyperbilirubinemia: Jaundice: 86.7 (26)
M (SD) rangeb 3.2 (1.8) 0–7
Note. aHollingshead Four Factor Index of Socioeconomic Status (HI; Hollingshead, 1975; range = 8–66);
b
Composite score based on presence/absence of 11 conditions derived from inspection of discharge notes.
Table 1 shows the sociodemographic background of the sample. Most mothers had
completed a college degree and the mean years of maternal education was >16 years. SES
was also classified using the Hollingshead Four Factor Index (HI; Hollingshead, 1975), a
composite based on both parents’ education and occupation (possible range = 8–66). The
mean Hollingshead Index score of 57 reflects that participants were primarily from mid-
to high-SES backgrounds.
discharge summaries from the neonatal intensive care unit (NICU). Table 1 shows the
proportion of children with each condition. For each child, the presence or absence of
each condition was coded and summed to yield a medical risk score (max = 11). The
mean per child was just over 3 (range = 0–7). In this relatively healthy group of children
born preterm, none of the children had a hearing or visual impairment that precluded
participation in the study. Based on either an MRI or a head ultrasound conducted during
hospitalization, only two children were identified as having periventricular leukomalacia,
a white matter injury associated with preterm birth. Only five children were identified as
having an intraventricular hemorrhage (Volpe, 2001), most of which were mild grades.
Measures
Vocabulary Size. Expressive vocabulary was assessed at 18 months AA with the
MacArthur-Bates CDI: Words and Gestures (CDI: W&G), a reliable and valid parent
report instrument appropriate for children 8 to 18 months (Fenson et al., 2007). Parents
marked words that their child “understands and says” and a total vocabulary score was
derived (396 max). Percentile scores were computed based on adjusted age and gender.
Receptive and Expressive Language. When the children were 18 months AA,
a trained examiner administered the Bayley Scales of Infant and Toddler Development,
third edition (BSID-III; Bayley, 2005), a standardized developmental assessment for
children aged 16 days to 42 months. Testing typically occurred over the course of two
visits, conducted approximately one week apart. We chose to use the language composite
in our analyses because this measure is commonly applied in clinical protocols and it
offers a comprehensive evaluation of a broad range of children’s language skills. Children
point to a named picture, follow simple and complex commands, and produce words.
Scaled scores were summed and converted to a standard full-scale score, based on the
child’s adjusted age.
between 4 filler trials, yielding 64 experimental trials. Trials on which the parent reported
that the child did not understand the target word were excluded on a child-by-child basis.
All children were reported to know at least five of the target words, and about two thirds
were reported to know all eight words.
All LWL sessions were later prescreened and coded offline by trained research
assistants blind to target side. Trials where the participant was identified as inattentive or
where there was parental interference (M = 6.9 trials; range = 0–17) were not coded. On
codable trials, eye gaze was identified for each 33-ms interval as either fixed on one of the
images (left or right), or shifting between pictures or away (off). Trials were later
designated as target-initial (T-initial) or distracter-initial (D-initial) based on the child’s
fixation at target noun onset.
Two measures of language processing efficiency were derived. Accuracy was the
mean proportion looking to the target image divided by the total looking time to either
image from 300–1800 ms after noun onset on all trials (M = 39.7 trials; range = 17–58).
Reaction time (RT) was the mean latency (in ms) of shifts from the distracter to the target
image after target noun onset on D-initial trials (M = 15.2 trials; range = 2–29). Shifts
were excluded if they occurred before 300 ms or after 1800 ms from target noun onset.
25% of the sessions were randomly selected and recoded for reliability. Inter-coder
agreement was 96% for the proportion of frames within 300–1800 ms from noun onset
identified as target vs distracter. Proportion of trials on which first-shift latency agreed
within one 33-ms frame was 100%.
RESULTS
Overview
We first present descriptive statistics for all variables at 18 months AA and
36 months. To establish which predictors were relevant to our outcome measure, we
next document first-order correlations between all predictors and the outcome variable of
interest. A series of multiple regression models then examine the shared and unique
contributions of each significant predictor to receptive vocabulary outcomes on the
PPVT-4, focusing on the shared and unique contributions of language processing
efficiency.
Table 2 Descriptives (M, SD, range) for Measures at 18 months AA and 36 months (n = 30).
M (SD) range
18 months AA
CDI: Vocabulary Sizea
Raw 61.2 (71.9) 0–330
Percentile 28.6 (28.7) 0–97
LWL: Online Language Processing
Accuracyb 0.61 (0.12) 0.29–0.81
Reaction Time (ms)c 804.0 (170.4) 531–1159
BSID-III: Language Composited 98.6 (12.5) 68–127
36 months
PPVT-4: Receptive Vocabularye
Raw Score 49.6 (23.3) 5–111
Standard Score (Unadjusted) 104.9 (20.5) 61–154
Standard Score (Adjusted) 109.1 (20.7) 65–158
Note. aReported number of words produced on the MacArthur-Bates CDI: Words & Gestures at 18 months AA;
b
Mean proportion looking to target on all distracter- and target-initial trials, computed from 300 to 1800 ms
from target noun onset;
c
Mean latency to shift to target picture on distracter-initial trials, including only shifts that occurred between
300 and 1800 ms from target noun onset;
d
Standard scores on a composite of the Expressive and Receptive subscales of the BSID-III (Bayley, 2005);
e
Raw scores on the Peabody Picture Vocabulary Test (PPVT-4; Dunn & Dunn, 2007) at 36 months, and
standard scores both unadjusted and adjusted for degree of prematurity.
with about 60% accuracy, on average, reliably looking to the target picture at above-chance
levels, t(29) = 5.4, p < .001, [CI: .58–.66]. Children initiated a shift within about 800 ms of
target word onset, on average, yet some children had mean RTs less than 600 ms and others
were nearly twice as slow.
1. 2. 3. 4. 5. 6. 7. 8.
1. PPVT-4: Receptive –
Vocabularya
2. HI: SESb 0.52** –
3. Gestational Age (weeks) 0.25 0.36 –
4. Birth Weight (g) 0.08 0.22 0.85*** –
5. Medical Risk Compositec −0.28 −0.19 −0.63*** −0.50** –
6. CDI: Vocabulary Sized 0.44* 0.04 0.16 0.08 0.06 –
7. BSID-III: Language 0.67*** 0.56** 0.32 0.11 −0.22 0.57** –
Compositee
8. LWL: Accuracyf 0.57** 0.39* 0.17 0.08 −0.30 0.15 0.55** –
9. LWL: Reaction Time (RT)g −0.81*** −0.31 −0.08 0.01 0.31 −0.23 −0.46* −0.71***
to PPVT-4 scores (all rs < .28). In contrast, CDI vocabulary and BSID-III scores were
significantly correlated with PPVT-4, accounting for 20 to 44% of the variance. In
addition, both accuracy and RT in the LWL task were significantly correlated with
PPVT-4, with RT showing a particularly strong link. Speed of language processing
accounted for more than 65% of the variance (Figure 1).
We next explored patterns of relations among the predictors and PPVT-4 in a series
of multiple regression models, focusing on those correlates that had significant first-order
relations (SES, vocabulary size, BSID-III, accuracy, RT). In all models, SES was entered
first. In Table 4, the baseline model with only SES accounted for 27% of the variance in
PPVT-4. Model 1 indicated that vocabulary size adds an additional 17% after controlling
for SES, and together these factors accounted for nearly 45% of the variance. In Model 2,
in relation to the baseline model, the additional contribution of the BSID-III was slightly
greater than that of vocabulary, just over 20%, and SES was no longer significant. Model
3 examined the additional contribution of these two predictors taken together. Here, the
model accounted for more than 50% of the variance in later receptive language; however,
none of the predictors contributed significant unique variance.
The models in Table 5 assessed the predictive strength of the two processing
measures from the LWL task. Due to their high collinearity and similarity of the
constructs measured, accuracy and RT were evaluated as predictors in separate models.
Model 4 showed that accuracy accounted for nearly 16% additional variance beyond SES
and these factors together accounted for about 43% of the variance. Model 5 confirmed
that RT strongly correlated with later PPVT-4, contributing close to 50% additional
variance beyond SES. Together, RT and SES accounted for almost 75% of the variance
658 V. A. MARCHMAN ET AL.
Figure 1 Scatterplot of first-order correlation between speed of lexical processing (RT in ms) on the LWL task at
18 months AA and receptive language (PPVT-4 standard scores) at 36 months (n = 30).
Table 4 Multiple Regression Models Predicting Receptive Vocabulary (PPVT-4) at 36 months from 18 months
AA Demographic and Language Measures. Unstandardized B (standard error, SE).
HI: SESa 1.37 (0.42)** 1.32 (0.38)** 0.56 (0.44) 0.77 (0.48)
CDI: Vocabulary Sizeb – 0.12 (0.04)** – 0.06 (0.05)
BSID-III: Language Compositec – – 0.90 (0.27)** 0.64 (0.36)
r2 change 17.4%** 20.9%** 23.2%**
Total R2 27.0%** 44.4%*** 48.0%*** 50.3%***
Note. *p < .05, **p < .01, ***p < .001; r2 change for Models 1 to 3 in reference to Baseline model;
a
Hollingshead Four Factor Index of Socioeconomic Status (HI; Hollingshead, 1975);
b
Reported words produced on the MacArthur-Bates CDI: Words & Gestures at 18 months AA;
c
Standard scores on a composite of the Expressive and Receptive subscales of the BSID-III (Bayley, 2005).
in 3-year PPVT-4. Note that both SES, RT and accuracy each contributed significant
unique variance.
Finally, Models 6 and 7 explored the contributions of accuracy and RT over and
above vocabulary and BSID-III, controlling for SES. In Model 6, the four predictors
together accounted for nearly 60% of variance in PPVT-4 scores, although none
accounted for significant unique variance. In Model 7, mean RT, the measure of online
PROCESSING SPEED PREDICTS OUTCOMES IN PRETERMS 659
Table 5 Multiple Regression Models Predicting Receptive Vocabulary (PPVT-4) at 36 months from 18 months
AA Demographic, Language and Processing Measures. Unstandardized B (SE).
HI: SESa 0.92 (0.42)* 0.78 (0.27)** 0.75 (0.46) 0.66 (0.30)*
CDI: Vocabulary Sizeb – – 0.08 (0.05) 0.06 (0.03)
BSID-III: Language Compositec – – 0.29 (0.39) 0.19 (0.24)
LWL: Accuracyd 74.40 (27.13)* – 54.80 (27.73) –
LWL: RTe – −0.09 (0.01)*** – −0.08 (0.01)***
r2 change 15.9%* 46.7%*** 6.7% 31.1%***
Total R2 42.9%** 73.7%*** 57.0%*** 81.4%***
Notes. *p < .05, **p < .01, ***p < .001; r2 change for Models 4 and 5 are in reference to Baseline model
(Table 4); r2 change for Models 6 and 7 are in reference to Model 3 (Table 4).
a
Hollingshead Four Factor Index of Socioeconomic Status (HI; Hollingshead, 1975);
b
Reported number of words produced on the MacArthur-Bates CDI: Words & Gestures at 18 months AA.
c
Standard scores on a composite of the Expressive and Receptive subscales of the BSID-III (Bayley, 2005);
d
Mean proportion looking to target in the LWL task, computed from 300 to 1800 ms from target noun onset
and averaged across all distracter- and target-initial trials;
e
Mean latency to shift (RT) from distracter to target picture on the LWL task, including only shifts that
occurred between 300 and 1800 ms from target noun onset.
processing that reflects how efficiently children processed words in real time, accounted
for over 30% additional variance beyond SES, CDI and BSID-III. Including all four
factors accounted for more than 80% of the variance in 3-year receptive language skill,
with RT and SES remaining the only significant predictors.
Follow-up analyses indicated that including overall general health and degree of
prematurity as control variables in addition to SES did not change the general pattern of
results reported in Tables 4 and 5. Most notably, adding GA, BW and the medical risk
composite to Model 7 did not increase the overall model fit (Total R2 = 81.8%). In
addition, RT remained a significant predictor of PPVT-4, contributing about 26.6%
unique variance (p < .001). Interestingly, including these additional variables in the
model reduced the contribution of SES such that its unique contribution was no longer
statistically significant (p = .08). Thus, a task of real-time spoken language comprehen-
sion captured meaningful individual differences in toddlers’ processing speeds that were
uniquely related to later receptive vocabulary, above and beyond degree of prematurity
and medical risk, as well as SES, vocabulary size and overall language knowledge.
We can also note that our results were not influenced by the fact that our population
contained a large number of multiple births. Follow-up analyses were conducted on two
sub-samples of children, each including all of the singletons and one randomly-selected
twin from each pair. In both sub-samples, the results were strikingly similar to those from
the overall analyses, with RT making a substantial unique contribution to the prediction of
3-year PPVT-4: sub-sample 1: unique r2 = 19.4, B = −.07 (.01), p < .0001; sub-sample 2:
unique r2 = 26.3%, B = −.08 (.01), p < .001.
Finally, we can illustrate the strong relation between speed of real-time language
processing and receptive vocabulary outcomes in Figure 2. This graph plots the time
course of the mean proportion looking to the target picture on distracter-initial trials at
18 months AA for children sub-grouped by PPVT-4 scores (median = 107.5). Children
with higher PPVT-4 scores increased their mean proportion looking to the target picture
660 V. A. MARCHMAN ET AL.
Figure 2 Time course of spoken language comprehension at 18 months AA in preterm children with higher and
lower PPVT-4 (standard scores) scores at 36 months. Error bars represent 95% CI.
sooner in the sentence and reached a higher mean proportion looking than children with
lower PPVT-4 scores. This advantage is also reflected in mean RT. Children with higher
PPVT-4 scores were significantly faster to shift from the distracter to the target image on
average (M = 699 ms; SD = 99, [CI: 649–745]) than those with lower PPVT-4 scores
(M = 909; SD = 163 [CI: 828–988]), t(28) = 4.3, p < .001, ŋ2p = .39.
DISCUSSION
In this prospective longitudinal study, we examined the predictive validity of early
language processing efficiency in relation to preschool receptive vocabulary in preterm
infants. Language processing efficiency at 18 months AA not only strongly predicted
receptive vocabulary at 3 years when considered in isolation, but also captured variation
beyond that attributable to SES, vocabulary size, the BSID-III language scale, and medical
factors. Those infants who were faster and more accurate in real-time language processing
were those who also scored higher in receptive vocabulary at 3 years, with RT adding
more than 30% additional variance over knowledge-based assessments. When processing
speed, BSID-III, and vocabulary size were all considered together, in addition to SES, the
full model accounted for nearly 80% of the variance in later scores and processing speed
and SES remained the only significant predictor. This finding indicates that processing
speed not only accounts for, but also adds to, the continuities between traditional tests of
vocabulary knowledge and later outcomes in this population.
These results support our hypothesis that experimental measures of early neuropsy-
chological processes known to support learning in full-term children are also strongly
PROCESSING SPEED PREDICTS OUTCOMES IN PRETERMS 661
accounting altogether for over 50% of the variance in children’s later receptive language
abilities. But given that measures of processing efficiency at 18 months AA significantly
improved prediction over these traditional measures, it would be beneficial to explore
ways to incorporate processing-based tasks into early clinical protocols (Feldman et al.,
2012).
The study is not without limitations. First, the current study involved a relatively
small sample of children born preterm. While the results here were quite robust and
paralleled those reported in earlier studies with children born full term, ongoing data
collection will allow us to directly compare these relations in preterm children to those in
matched full-term controls. A second limitation is that this sample included a large
number of children who were twins; however, the analyses indicated that our results
were not different when only one of each twin pair was included. A final limitation is that
our outcome measure was limited to receptive vocabulary at 3 years. In our continuing
studies, we plan to explore longer-term links between early processing speed and later
outcomes utilizing comprehensive language and IQ-related assessments just prior to when
the children are entering school.
In conclusion, speed of early lexical processing at 18 months is highly predictive of
receptive vocabulary outcomes at age 3. Thus, differences among infants in their efficiency
in interpreting language in real time are linked to differences in skills critical for building
linguistic knowledge, in children born preterm as well as in full-term children. Future
research that explores the neuropsychological and environmental correlates of early language
processing deficits in typically developing and at-risk populations will deepen our under-
standing of neuropsychological mechanisms that support receptive language learning.
REFERENCES
Allen, M. C. (2008). Neurodevelopmental outcomes of preterm infants. Current Opinion in
Neurology, 21(2), 123–128. doi:10.1097/WCO.0b013e3282f88bb4
Back, S. A., & Miller, S. P. (2014). Brain injury in premature neonates: A primary cerebral
dysmaturation disorder? Annals of Neurology, 75, 469–486. doi:10.1002/ana.v75.4
Barre, N., Morgan, A., Doyle, L. W., & Anderson, P. J. (2011). Language abilities in children who
were very preterm and/or very low birth weight: A meta-analysis. The Journal of Pediatrics,
158, 766–774. doi:10.1016/j.jpeds.2010.10.032
Bayley, N. (2005). Bayley scales of infant and toddler development: Bayley-III. San Antonio, TX:
Harcourt Assessment, Psychological Corporation.
Bhutta, A. T., Cleves, M. A., Casey, P. H., Cradock, M. M., & Anand, K. J. S. (2002). Cognitive and
behavioural outcomes of school-age children who are born preterm. Journal of the American
Medical Association, 288, 728–737.
Caskey, M., Stephens, B., Tucker, R., & Vohr, B. (2011). Importance of parent talk on the
development of preterm infant vocalizations. Pediatrics, 128(5), 910–916. doi:10.1542/
peds.2011-0609
Dale, P. S., Price, T. W., Bishop, D. V. M., & Plomin, R. (2003). Outcomes of early language
DelayI. Predicting persistent and transient language difficulties at 3 and 4 Years. Journal of
Speech, Language, and Hearing Research, 46, 544–560.
PROCESSING SPEED PREDICTS OUTCOMES IN PRETERMS 663
Dollaghan, C. A., Campbell, T. F., Paradise, J. L., Feldman, H. M., Janosky, J. E., Pitcairn, D. N., &
Kurs-Lasky, M. (1999). Maternal education and measures of early speech and language.
Journal of Speech, Language, and Hearing Research, 42(6), 1432–1443.
Duff, F. J., Reen, G., Plunkett, K., & Nation, K. (2015). Do infant vocabulary skills predict school‐age
language and literacy outcomes? Journal of Child Psychology and Psychiatry. Advance online
publication. doi:10.1111/jcpp.12378
Dunn, L. M., & Dunn, D. M. (2007). The Peabody Picture Vocabulary Test (PPVT-4) (4th ed.).
Bloomington, MN: Pearson Assessments.
Feldman, H. M., Lee, E. S., Yeatman, J. D., & Yeom, K. W. (2012). Language and reading skills in
school-aged children and adolescents born preterm are associated with white matter properties
on diffusion tensor imaging. Neuropsychologia, 50(14), 3348–3362. doi:10.1016/j.
neuropsychologia.2012.10.014
Fenson, L., Bates, E., Dale, P. S., Marchman, V. A., Reznick, J. S., & Thal, D. J. (2007). MacArthur-
Bates communicative development inventories. Baltimore, MD: Paul H. Brookes Publishing.
Fernald, A., & Marchman, V. A. (2012). Individual differences in lexical processing at 18 months
predict vocabulary growth in typically developing and late talking toddlers. Child
Development, 83, 203–222. doi:10.1111/j.1467-8624.2011.01692.x
Fernald, A., Perfors, A., & Marchman, V. A. (2006). Picking up speed in understanding: Speech
processing efficiency and vocabulary growth across the 2nd year. Developmental Psychology,
42(1), 98–116. doi:10.1037/0012-1649.42.1.98
Fernald, A., Zangl, R., Portillo, A. L., & Marchman, V. A. (2008). Looking while listening: Using eye
movements to monitor spoken language. In I. A. Sekerina, E. M. Fernandez, & H. Clahsen (Eds.),
Developmental psycholinguistics: On-line methods in children’s language processing (pp. 97–135).
Philadelphia, PA: John Benjamins.
Glinianaia, S. V., Ghosh, R., Rankin, J., Pearce, M. S., Parker, L., & Pless-Mulloli, T. (2013). No
improvement in socioeconomic inequalities in birthweight and preterm birth over four decades:
A population-based cohort study. BMC Public Health, 13(1), 345–353.
Goldenberg, R. L., Culhane, J. F., Iams, J. D., & Romero, R. (2008). Epidemiology and causes of
preterm birth. The Lancet, 371, 75–84. doi:10.1016/S0140-6736(08)60074-4
Guarini, A., Sansavini, A., Fabbri, C., Alessandroni, R., Faldella, G., & Karmiloff-Smith, A. (2009).
Reconsidering the impact of preterm birth on language outcome. Early Human Development,
85(10), 639–645. doi:10.1016/j.earlhumdev.2009.08.061
Hollingshead, A. B. (1975). Four factor index of social status. New Haven, CT: Yale University
Press.
Inder, T. E., Warfield, S. K., Wang, H., Hüppi, P. S., & Volpe, J. J. (2005). Abnormal cerebral structure
is present at term in premature infants. Pediatrics, 115(2), 286–294. doi:10.1542/peds.2004-0326
Johnson, C. J., Beitchman, J. H., & Brownlie, E. B. (2010). Twenty-year follow-up of children with
and without speech-language impairments: Family, educational, occupational, and quality of
life outcomes. American Journal of Speech-Language Pathology, 19, 51–65.
Kidokoro, H., Anderson, P. J., Doyle, L. W., Woodward, L. J., Neil, J. J., & Inder, T. E. (2014).
Brain injury and altered brain growth in preterm infants: Predictors and prognosis. Pediatrics,
134(2), e444–e453. doi:10.1542/peds.2013-2336
Lee, E. S., Yeatman, J. D., Luna, B., & Feldman, H. M. (2011). Specific language and reading skills
in school-aged children and adolescents are associated with prematurity after controlling for IQ.
Neuropsychologia, 49(5), 906–913. doi:10.1016/j.neuropsychologia.2010.12.038
Månsson, J., & Stjernqvist, K. (2014). Children born extremely preterm show significant lower
cognitive, language and motor function levels compared with children born at term, as
measured by the Bayley-III at 2.5 years. Acta Pædiatrica, 103, 504–511. doi:10.1111/
apa.2014.103.issue-5
Marchman, V. A., & Fernald, A. (2008). Speed of word recognition and vocabulary knowledge in
infancy predict cognitive and language outcomes in later childhood. Developmental Science, 11
(3), F9–F16. doi:10.1111/j.1467-7687.2008.00671.x
664 V. A. MARCHMAN ET AL.
Marlow, N. (2004). Neurocognitive outcome after very preterm birth. Archives of Disease in
Childhood-Fetal and Neonatal Edition, 89(3), F224–F228. doi:10.1136/adc.2002.019752
Potharst, E. S., Van Wassenaer-Leemhuis, A. G., Houtzager, B. A., Livesey, D., Kok, J. H., Last, B.
F., & Oosterlaan, J. (2013). Perinatal risk factors for neurocognitive impairments in preschool
children born very preterm. Developmental Medicine & Child Neurology, 55, 178–184.
doi:10.1111/dmcn.2013.55.issue-2
Ramey, C. T., & Ramey, S. L. (2004). Early learning and school readiness: Can early intervention
make a difference? Merrill-Palmer Quarterly, 50(4), 471–491. doi:10.1353/mpq.2004.0034
Rescorla, L., & Dale, P. S. (2013). Late talkers: Language development, interventions, and out-
comes, Baltimore, MD: Brookes.
Rose, S. A., Feldman, J. F., & Jankowski, J. (2002). Processing speed in the 1st year of life: A
longitudinal study of preterm and full-term infants. Developmental Psychology, 38, 895–902.
doi:10.1037/0012-1649.38.6.895
Rose, S. A., Feldman, J. F., & Jankowski, J. J. (2009). A cognitive approach to the development of
early language. Child Development, 80(1), 134–150. doi:10.1111/cdev.2009.80.issue-1
Sansavini, A., Guarini, A., Alessandroni, R., Faldella, G., Giovanelli, G., & Salvioli, G. (2007). Are
early grammatical and phonological working memory abilities affected by preterm birth?
Journal of Communication Disorders, 40, 239–256. doi:10.1016/j.jcomdis.2006.06.009
Sansavini, A., Pentimonti, J., Justice, L., Guarini, A., Savini, S., Alessandroni, R., & Faldella, G.
(2014). Language, motor and cognitive development of extremely preterm children: Modeling
individual growth trajectories over the first three years of life. Journal of Communication
Disorders, 49, 55–68. doi:10.1016/j.jcomdis.2014.02.005
Shafto, C. L., Conway, C. M., Field, S. L., & Houston, D. M. (2012). Visual sequence learning in
infancy: Domain-general and domain-specific associations with language. Infancy, 17(3), 247–
271. doi:10.1111/infa.2012.17.issue-3
Smith, L. K., Draper, E. S., Manktelow, B. N., Dorling, J. S., & Field, D. J. (2007). Socioeconomic
inequalities in very preterm birth rates. Archives of Disease in Childhood. Fetal and Neonatal
Edition, 92(1), F11–F14. doi:10.1136/adc.2005.090308
Soria-Pastor, S., Gimenez, M., Narberhaus, A., Falcon, C., Botet, F., Bargallo, N., . . . Junque, C.
(2008). Patterns of cerebral white matter damage and cognitive impairment in adolescents born
very preterm. International Journal of Developmental Neuroscience, 26(7), 647–654.
doi:10.1016/j.ijdevneu.2008.08.001
Stoll, B. J., Hansen, N. I., Bell, E. F., Shankaran, S., Laptook, A. R., Walsh, M. C., . . . Higgins, R.
D. (2010). Neonatal outcomes of extremely preterm infants from the NICHD Neonatal
Research Network. Pediatrics, 126(3), 443–456. doi:10.1542/peds.2009-2959
Stolt, S., Haataja, L., Lapinleimu, H., & Lehtonen, L. (2009). The early lexical development and its
predictive value to language skills at 2 years in very-low-birth-weight children. Journal of
Communication Disorders, 42(2), 107–123. doi:10.1016/j.jcomdis.2008.10.002
Taylor, H. G., Klein, N., Minich, N. M., & Hack, M. (2000). Verbal memory deficits in children
with less than 750 g birth weight. Child Neuropsychology, 6(1), 49–63. doi:10.1076/0929-7049
(200003)6:1;1-B;FT049
US Preventive Services Task Force. (2014, November). Draft recommendation statement: Speech and
language delay and disorders in children age 5 and younger: Screening. Retrieved from http://
www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementDraft/
speech-and-language-delay-and-disorders-in-children-age-5-and-younger-screening
van Noort-van der Spek, I. L., Franken, M. C., & Weisglas-Kuperus, N. (2011). Language functions
in preterm children: A systematic review and meta-analysis. Pediatric Research, 70, 357–357.
doi:10.1038/pr.2011.582
Vohr, B. (2014). Speech and language outcomes of very preterm infants. Seminars in Fetal and
Neonatal Medicine, 19(2), 78–83. WB Saunders. doi:10.1016/j.siny.2013.10.007
Vohr, B. R., Wright, L. L., Dusick, A. M., Mele, L., Verter, J., Steichen, J. J., . . . Kaplan, M. D.
(2000). Neurodevelopmental and functional outcomes of extremely low birth weight infants in
PROCESSING SPEED PREDICTS OUTCOMES IN PRETERMS 665
the National Institute of Child Health and Human Development Neonatal Research Network,
1993–1994. Pediatrics, 105(6), 1216–1226. doi:10.1542/peds.105.6.1216
Volpe, J. J. (2001). Neurobiology of periventricular leukomalacia in the premature infant. Pediatric
Research, 50(5), 553–562. doi:10.1203/00006450-200111000-00003
Weisleder, A., & Fernald, A. (2013). Talking to children matters: Early language experience
strengthens processing and builds vocabulary. Psychological Science, 24(11), 2143–2152.
doi:10.1177/0956797613488145
Wolke, D., Samara, M., Bracewell, M., & Marlow, N. (2008). Specific language difficulties and
school achievement in children born at 25 weeks of gestation or less. The Journal of Pediatrics,
152, 256–262. doi:10.1016/j.jpeds.2007.06.043
Copyright of Child Neuropsychology is the property of Psychology Press (UK) and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.