Reading Development in Typically Developing Children and Children With Prenatal or Perinatal Brain Lesions: Differential School Year and Summer Growth

Abstract

Summer slide, uneven growth of academic skills during the calendar year, captures the fact that the learning gains children make during the school year do not continue at the same pace over the summer, when children are typically not in school. We compared growth of reading skills during the school year and during the summer months in children with prenatal or perinatal brain lesions (PL) and typically developing (TD) children from varying socioeconomic-status (SES) backgrounds as a new way to probe the role of structured environmental support in functional plasticity for reading skills in children with PL. Results showed that children with PL performed worse than TD children on both reading decoding and reading comprehension. Group differences were primarily driven by children with larger lesions and children with right-hemisphere lesions (RH). For reading comprehension, children with RH showed greater growth during the school year but more slide during the summer months than both TD children and children with left-hemisphere lesions, implicating a particularly strong role of structured input in supporting reading comprehension in this group. TD children from lower-SES backgrounds fell behind their TD peers from higher-SES backgrounds on decoding and reading comprehension but did not show differential patterns of school year and summer growth. Overall, results highlight the importance of considering the role of a host of factors interacting at multiple levels of analyses, including biological and environmental, in influencing developmental trajectories of TD children and atypically developing children.

The uneven growth of academic skills during the calendar year, characterized by greater growth during the school year than during the summer, is a well-known phenomenon. This phenomenon is typically referred to as “summer slide,” which captures the fact that the learning gains children make during the school year do not continue during the summer, but instead learning tends to flatten or even fall off during this period when children are not in school. Summer slide has been interpreted to reflect the effects of the richer, structured instruction children receive during the academic year compared with during the summer months, when the academic instruction children receive is on average lower in quantity and quality and almost certainly more variable (Borman & Boulay, 2004; Cooper, Nye, Charlton, Lindsay, & Greathouse, 1996; Gershenson, 2013; J. Huttenlocher, Levine, & Vevea, 1998).

In the current study, we compared growth of reading skills during the school year and during the summer months as a new way to probe the role of structured environmental support in the functional plasticity exhibited by children with prenatal or perinatal focal brain injury (P. R. Huttenlocher, 2009; Levine, Raja Beharelle, Demir, & Small, 2015). The finding of greater “summer slide” in children with prenatal or perinatal brain lesions (PL) compared with typically developing (TD) children would lend support to the hypothesis that environmental input plays an even more important role in the development of academic skills in children with early brain injury than in TD children (Levine et al., 2015; Rowe, Levine, Fisher, & Goldin-Meadow, 2009; Stiles, Reilly, & Levine, 2012).

The existing literature supports the conclusion that children with early focal brain injury exhibit remarkable plasticity for language functions. In particular, unlike individuals with focal left-hemisphere injuries incurred in adulthood, children with prenatal or perinatal stroke show normal or near-normal language functions, even when lesions involve classic language networks. However, research on language plasticity has largely focused on early-developing language functions and has shown that after an initial delay in getting language off the ground, young children with PL on average show language-learning trajectories that are similar to those of TD children (Bates, 1999; Bates et al., 1997; Feldman, Holland, Kemp, & Janosky, 1992; Rowe et al., 2009). Much less is known about what happens later—whether this remarkable plasticity extends to later-emerging, more complex language skills and reading skills—and few studies have examined this question (Aram & Ekelman, 1988; Ballantyne, Spilkin, Hesselink, & Trauner, 2008). Recent studies have shown that plasticity for early language functions might not extend to later-developing complex language tasks, such as narrative production, at least in some children with PL (Demir, Fisher, Levine, & Goldin-Meadow, 2014; Demir, Levine, & Goldin-Meadow, 2010; Reilly, Bates, & Marchman, 1998; Reilly, Wasserman, & Appelbaum, 2013). Relatively little work has examined reading skills in children with PL, and those researchers who have done so report that the average performance of this group of children falls below the level of their TD peers (Aram & Ekelman, 1988; Ballantyne et al., 2008; Frith & Vargha-Khadem, 2001).

As is the case for TD children, there are many reasons why children with PL may vary in their levels of academic skill. However, one source of variability that is unique to children with PL is their lesions, which vary in terms of many characteristics including lesion size, location, and laterality. For example, previous studies of language development in this group have shown that lesion size predicts children’s language development trajectory (Brasky, Nikolas, Meanwell, Levine, & Goldin-Meadow, 2005; Levine, Brasky, & Nikolas, 2005; Rowe et al., 2009), whereas lesion laterality does not (Bates et al., 2001; Demir et al., 2010; Rowe et al., 2009). Studies focusing on the effects of lesion characteristics on reading development have mainly focused on lesion laterality and have reported inconsistent findings: Aram and Ekelman (1988) reported lower reading performance in children with right-hemisphere lesions, Frith and Vargha-Khadem (2001) reported lower performance in those with left-hemisphere lesions, and Ballantyne et al. (2008) reported no lesion laterality effects.

Multiple factors may account for these inconsistent findings, including heterogeneity among children included in these studies with respect to their age at the time of lesion, their age at the time of assessment, and the size of their lesions—all of which are typically not considered. Further, existing studies have used a variety of measures of reading skill and have not considered decoding and comprehension separately, and instead, they have tended to rely on composite measures. Longitudinal studies of reading development are completely lacking despite the fact that such studies would be informative in addressing whether children with prenatal or perinatal focal brain lesions have difficulty with reading at particular points in the learning trajectories.

In the current study, we aimed to address these issues and extend the existing literature in multiple ways: We examined multiple lesion characteristics and controlled for lesion size when considering laterality; we separately assessed reading decoding and reading comprehension; we focused only on children with PL, rather than on children who incurred their brain injury at different ages, often due to different etiologies; and finally, we assessed children’s reading development from kindergarten to second grade, while taking a longitudinal perspective. In addition, our study is the first to examine the role of environmental input on reading achievement in children with PL, an approach we have taken in previous studies examining the language development trajectories of children with PL (Rowe et al., 2009).

With regard to the role of input, studies examining the relation of parental language input and early language development in children with PL have shown that parental talk is even more related to certain aspects of language development in this group than in TD children. For example, Rowe et al. (2009) found that parental vocabulary input is a stronger predictor of preschoolers’ syntactic growth for children with PL than TD children, whereas input effects did not vary for the vocabulary growth of children in these groups. Similarly, Demir et al. (2015) found that parental talk about decontextualized topics (explanations, narratives, pretend play) when children were 30 months of age or older strongly predicted the kindergarten narrative skill, but not the vocabulary or syntax, of children with PL compared with TD children. Considered together, these studies highlight that parental language input may play a larger role in the language development of children with PL than in the language development of TD children and that the differential role of input might be specific to the aspects of language development that are more challenging at the age at which the skills are assessed.

The current study is the first to examine the role of input in the academic achievement of children with PL once these children reach school age. Based on our prior research showing larger parental input for children with PL than for TD children with respect to syntactic (but not vocabulary) development (Rowe et al., 2009), we hypothesized that the structured, formal instruction provided during the school year, compared with during the summer months when children are typically out of school, may be more important for the reading development of children with PL than that of TD children (Alexander, Entwisle, & Olson, 2001). Alternatively, children with PL might not be able to benefit from school input to the same degree as their TD peers, and input might play a weaker role for the reading development of children with PL than for TD children. As was the case for different aspects of early language development (i.e., vocabulary and syntax and narrative), relative growth during school versus during the summer period for TD children versus children with PL might also vary depending on the particular reading skill assessed (i.e., decoding and comprehension).

In the current study, we asked whether the growth patterns of reading achievement between the fall of kindergarten and the spring of second grade vary for children with prenatal or perinatal unilateral brain lesions compared with TD children and whether this variation was found for reading decoding and/or reading comprehension. We also asked whether this variation is associated with children having particular lesion characteristics. We addressed these questions by comparing the growth of reading decoding and reading comprehension during the school year and the summer for children with PL and TD children. If input is more important for children with PL than for TD children, we expected the children with PL to show a greater difference in the school year versus summer growth rates compared to the TD children. If true for more challenging aspects of reading, as we found for language development, we expected to find the summer slide or plateau to be greater for the children with PL than for TD children for reading comprehension skill, which continues to develop over time even in adulthood, but perhaps not for reading decoding skill, which most children acquire successfully in early elementary school (Paris, 2005).

Another goal of the current study was to examine developmental trajectories of reading skill in children with PL and TD children from varying socioeconomic-status (SES) backgrounds. Previous research has shown that TD children from lower-SES backgrounds perform worse than their peers from higher-SES backgrounds (Alexander et al., 2001; Brooks-Gunn & Duncan, 1997; National Center for Education Statistics [NCES], 2011). By comparing the reading developmental trajectories of children with biological challenges (children with PL) to those of children with environmental challenges (TD children from lower-SES backgrounds), we also aim to provide insight into how environmental and biological factors contribute to shaping reading achievement trajectories across these different populations.

METHOD

Participants

Typically Developing Children

The reading skills of TD children (n = 57; 32 girls) enrolled in a longitudinal language project in the Greater Chicago area were assessed starting from the beginning of kindergarten and continuing until the end of second grade. Only monolingual English-speaking families were recruited for the study. The children were 14 months of age when they were first enrolled in the study. The children and their families were recruited from the Chicago area via mailings to families and via an advertisement in a free parent magazine. The families were interviewed, and the sample was selected to represent the socioeconomic diversity of the Chicago area. Based on parental report, 35 children were Caucasian, 11 were African American, 5 were White Hispanic/Latino, and 6 were Mixed Race.

We measured SES based on the education level of the primary caregiver and the annual family income level. In both cases, the data were collected categorically from parents via a questionnaire. Parent education responses were transformed into a continuous scale by using the total number of years of schooling (e.g., “high school or GED” was scored as 12 years, “bachelor’s degree” as 16 years, etc.). On this scale, the average number of years of primary caregiver education was 16 years (SD = 2 years, range = 10–18 years). Parent income responses were transformed into a continuous scale by using the midpoint of each category (e.g., the category $15,000–$35,000 was scored as $25,000). On this scale, the average family income of the participating families was $69,152 (SD = $30,288, range = $7,500–$100,000).

Children With Prenatal or Perinatal Unilateral Brain Lesions

The reading skills of children with unilateral prenatal or perinatal lesions (n = 34, 19 girls) who were enrolled in the longitudinal language project were assessed starting from the beginning of kindergarten and continuing until the end of second grade. We recruited the children with PL by contacting physicians in the Greater Chicago area and by establishing relationships with parent support groups in the area (Childhood Stroke and Hemiplegia Connections of Illinois, Pediatric Stroke Network, and Children’s Hemiplegia and Stroke Association). We included every family that was interested as long as the child had experienced a unilateral prenatal or perinatal brain lesion and was monolingual English-speaking, regardless of SES. Thirty-one children with PL were reported by parents to be Caucasian, and 3 were reported to be Mixed Race. The average number of years of primary caregiver education was 16 years (SD = 2 years, range = 12–18). The average income was $82,121 (SD = 19,616, range = $42,500–$100,000).

Family Socioeconomic Status

In our sample of TD children and children with PL, income and education were significantly correlated (r = .40, p < .001). Thus, parent education and income were combined in a composite score of SES. The composite was generated using principal components analysis. The first principal component weighted education and income positively and equally. This component accounted for approximately 70% of the original variance. The mean score of the composite was 0 (SD = 1.0). Families of children with PL had significantly higher income than parents of TD children, t(90) = 2.21, p = .03. TD children and children with PL did not significantly differ from each other in terms of parental education, t(90) = 0.06, p = .95. Although there was a tendency for children with PL to have a higher composite SES score, the difference did not reach significance, t(90) = 1.33, p = .18.

Coding Characteristics of Brain Lesions

Lesion information was obtained from prior magnetic resonance imaging (MRI) scans (for 18 children), from MRI scans that we obtained for this study (for 12 children), or from detailed medical reports provided by families (for 4 children). All of the clinical and experimental scans were evaluated by two pediatric neurologists who coded the lesions according to lesion laterality (left, right), size (small, medium, large), and type (periventricular infarct [PV], cerebrovascular infarct [CI]).

Regarding lesion size, lesions were categorized as small, medium, or large based on the following criteria: Small lesions affected only one lobe or minimally affected subcortical regions. Medium lesions extended into more than one lobe or subcortical region. Large lesions affected three or four lobes and were all CIs; these lesions affected multiple cortical areas and involved the thalamus and subcortical regions. Regarding lesion type, CI lesions are primarily infarcts of the middle cerebral artery territory and tend to affect the inferior frontal, parietal, and/or superior temporal regions, with the lesion mainly impacting gray matter. PV lesions are primarily subcortical and involve white matter tracts, the thalamus, basal ganglia, and/or the medial temporal lobe. All of the children with PV lesions showed evidence of subcortical injury, enlarged ventricles, or reductions in white matter tracts (especially the internal capsule). Although periventricular leukomalacia in very low birth-weight prematurely born children (before 32 weeks) has been the focus of much previous literature, periventricular lesions also occur in full-term children (Krägeloh-Mann & Horber, 2007). All children in our sample were born at or near term according to parental report (gestational age at birth of at least 36 weeks). Thus, our sample of children with periventricular lesions differed from samples of very premature children with periventricular leukomalacia. Lesion characteristics for each participant are reported in Table 1.¹

Of the children with PL, 22 had left-hemisphere lesions (LH) and 12 had right-hemisphere lesions (RH). Prior research indicates that LH are more common in this group, and these proportions are consistent with the literature (Reilly et al., 1998). Nineteen of the children had lesions characterized as CI and 15 as PV. Twelve of the children had small lesions, 6 had medium lesions, and 16 had large lesions. There was a trend in the direction of lesion type and size being correlated: Ten of the 15 children with PV lesions had small or medium lesions, whereas 11 of the 19 children with CI lesions had large lesions, χ²(2, N = 34) = 3.83, p = .14. Neither lesion type and laterality, χ²(1, N = 33) = 1.52, p > .10, nor lesion size and laterality were related to each other, χ²(2, N = 33) = 4.44, p > .10. Because of the association between lesion laterality and type, we report analyses on lesion laterality. However, the pattern of results remains the same using lesion type instead of lesion size.

Measures

Reading Decoding

Reading decoding was assessed using the Word Attack and Letter-Word Identification subtests of the Woodcock-Johnson Tests of Achievement (WJ-III) administered twice a year (fall and spring) between kindergarten and second grade (Woodcock, McGrew, & Mather, 2001). The Letter-Word Identification subtest measures word-reading skills, and the Word Attack subtest measures skill in applying phonic and structural analysis skills to reading nonwords. Together, the two subtests constitute the Basic Reading Skills Cluster of the WJ-III, which served as our measure of decoding. W scores, which are standardized scores based on Rasch analyses, were used because they are most suitable for measuring progress over time.

Reading Comprehension

Reading comprehension was assessed using the Passage Comprehension subtest administered twice a year (fall, spring) between the beginning of first grade and the end of second grade (WJ-III; Woodcock et al., 2001). The Passage Comprehension subtest measures children’s ability to read and understand increasingly more complex, brief texts. As for reading decoding, W scores were assigned to children at each time point based on their test performance.

Procedure

The data included in this study were part of a larger longitudinal study examining the relation of parent language input to children’s language and literacy development. In this study, we focused on measures of the children’s reading decoding and reading comprehension skills administered during visits that occurred when the children were in kindergarten through second grade. The measures are described in the next section. Some of the tests, particularly reading comprehension at the first-grade fall visit, were not administered because of time limitations during the session. Missed visits due to difficulty in scheduling also led to variations in the number of children assessed at each session. Table 2 shows the number and average age of TD children and children with PL for each visit and task.

Table 1 Lesion characteristics for children with prenatal or perinatal lesions

Participant Number	Sex	Lesion Size	Lesion Laterality	Lesion Type	Affected Areas
1	Female	Large	Left	CI	F, T, P, O, subcortical
2	Female	Large	Right	PV	WM, insula, subcortical
3	Female	Medium	Left	PV	subcortical
4	Female	Large	Right	CI	F, T, P, subcortical
5	Male	Medium	Right	PV	subcortical
6	Female	Small	Right	PV	T, P, IC, subcortical
7	Male	Small	Left	CI	F, P, subcortical
8	Male	Large	Left	CI	T, P, O, IC, subcortical
9	Female	Large	Right	CI	F, T, P, subcortical
10	Male	Large	Right	PV	T, subcortical
11	Male	Medium	Left	PV	subcortical
12	Female	Small	Left	PV	subcortical
13	Female	Large	Left	PV	O, IC, subcortical
14	Female	Small	Right	CI	F, P
15	Female	Medium	Left	CI	F, T, P, subcortical
16	Female	Small	Right	CI	F, T, P
17	Female	Large	Left	CI	F, T, P, IC, subcortical
18	Male	Large	Left	CI	F, T, P, subcortical
19	Male	Small	Left	PV	F, T, subcortical
20	Female	Large	Left	CI	F, T, P, O, subcortical
21	Female	Medium	Left	CI	F, T, P
22	Female	Small	Left	PV	WM, subcortical
23	Female	Large	Left	PV	F, T, P, O, subcortical
24	Female	Large	Right	CI	F, T, P, subcortical
25	Female	Small	Left	PV	no MRI
26	Male	Medium	Left	CI	F, P, subcortical
27	Male	Small	Right	PV	T, subcortical
28	Male	Small	Right	PV	WM
29	Male	Large	Right	PV	F, T, P, subcortical
30	Female	Large	Left	CI	F, T, P, subcortical
31	Male	Medium	Left	CI	T, P, subcortical
32	Male	Large	Left	CI	F, T, P
33	Male	Large	Left	CI	F, T, P, O, IC, subcortical
34	Male	Small	Left	CI	–

Note. F = frontal; P = parietal; O = occipital; T = temporal; subcortical = involved subcortical areas; PV = periventricular infarct; MRI = magnetic resonance imaging.

Table 2 Average reading decoding and comprehension score (W score) in kindergarten (KG), first grade (1G), and second grade (2G) in typically developing (TD) children and children with prenatal or perinatal brain lesion (PL)

	TD			PL			TD–PL differences
	Mean	SD	n	Mean	SD	n	F
Decoding
KG Fall	403.3	29.3	56	400.9	34.4	27	0.52
KG Spring	435.2	32.2	57	437.0	30.0	26	0.18
1G Fall	450.5	28.4	55	440.8	33.5	26	3.98*
1G Spring	476.8	24.5	57	469.1	29.3	24	3.26
2G Fall	481.7	22.2	59	472.2	31.4	26	4.39*
2G Spring	494.2	18.0	58	485.2	27.3	25	5.36*
Reading comprehension
1G Fall	455.5	19.5	33	449.4	22.0	19	2.48
1G Spring	474.1	16.3	57	465.0	26.6	24	5.71*
2G Fall	479.4	17.1	58	467.8	28.5	24	9.35*
2G Spring	486.1	12.9	58	477.2	27.3	25	6.80*

Methodological Approach

To assess the growth of children’s reading skills over time, we built two-level statistical models for child-specific growth using hierarchical linear modeling (HLM): one for decoding and another for reading comprehension (Raudenbush & Bryk, 2002). An advantage of HLM over the ordinary least squares approach is its flexibility in dealing with missing data by incorporating all participants who have been observed at least once (Raudenbush & Bryk, 2002). Group (TD, PL) and SES were included as predictors of growth estimates. Lesion characteristics (laterality and size) were included as predictors of growth estimates in follow-up models. The Level 1 model accounted for the variation in repeated measures of reading decoding and reading comprehension within each child. The Level 2 model accounted for the variation between children as a function of group (TD, PL) and parental SES (composite measure).

Reading Decoding

At Level 1 (within children), we represented children’s reading decoding trajectory using a piecewise two-level HLM of child reading growth. We modeled summer growth separately from academic year growth based on findings showing steeper growth in reading during the academic year than over the summer (McCoach, O’Connell, Reis, & Levitt, 2006). We modeled three separate academic year growth slopes: The first was from the fall of kindergarten to the spring of kindergarten (kindergarten growth), the second was from the fall of first grade to the spring of first grade (first-grade growth), and the third was from the fall of second grade to the spring of second grade (second-grade growth). The fourth slope modeled summer growth—a combined slope based on summer growth between the spring of kindergarten and the fall of first grade and between the spring of first grade and the fall of second grade. The results did not change using separate summer slopes.

To estimate the growth of reading decoding during the four periods, we created four time-varying Level 1 variables: kindergarten exposure, first-grade exposure, second-grade exposure, and summer exposure. Because children were administered the reading decoding assessments at different times during the fall and spring periods, exposure was assessed by the number of months a given child had already spent during a given period (e.g., at school during the school year or out of school during summer vacation) on the day of the assessment. All Level 1 variables were grand-centered, so our intercept is the average reading decoding skill in the midpoint of the fall of kindergarten and the spring of second grade. Thus, we had for each child i at time t:

where π_0i is child i’s average reading decoding status in the midpoint of kindergarten (KG) fall and second grade (2G) spring, π_1i is child’s slope or the monthly learning rate during kindergarten, π_2i is the child’s slope or the monthly learning rate during first grade, π_3i is the child’s slope or the monthly learning rate during second grade, and π_4i is the child’s slope or the monthly learning rate during the two summers combined. The residual e_ti represents the portion of children’s reading decoding score that is not predicted by their academic year or summer exposure.At Level 2, we tested whether an individual child’s status and slope over the four time periods described previously are predicted by group (TD, PL) and/or parental SES. For the intercept and each of the reading decoding slopes, we had a separate Level 2 equation for each Level 1 coefficient, π_pi, where p = 0, 1, 2, 3, 4:

where π_pi is the pth growth parameter from the Level 1 model, β_p0, β_p1, β_p2 are linear regression coefficients, and r_pi is a random effect.

Reading Comprehension

We built parallel models to represent children’s reading comprehension trajectory between the fall of first grade and the spring of second grade. At Level 1 (within children), we represented children’s reading comprehension trajectory using a piecewise two-level HLM of child reading growth. We modeled three separate growth slopes. For modeling growth during the academic year, we used two slopes: The first was from the fall of first grade to the spring of first grade (first-grade growth), and the second was from the fall of second grade to the spring of second grade (second-grade growth). The third slope modeled summer growth during the summer between the spring of first grade and the fall of second grade. Thus, we had three time-varying Level 1 variables for reading comprehension: first-grade exposure, second-grade exposure, and summer exposure. As for reading decoding, because children’s reading comprehension was assessed at different times during the fall and spring periods, exposure was assessed by the number of months a given child had already spent in given period (e.g., at school during the school year or out of school during summer vacation) on the day of the assessment. Level 1 variables were grand-centered. Thus, our intercept was reading comprehension performance at the midpoint of the fall of first grade and the spring of second grade.

For each child i at time t:

where π_0i is child i’s average reading comprehension status at the midpoint of the fall of first grade (1G) and the spring of second grade (2G), π_1i is child’s slope or the monthly learning rate during first grade, π_2i is the child’s slope or the monthly learning rate during second grade, and π_3i is the child’s slope or the monthly learning rate during the summer between those two grades. The residual e_ti represents the portion of children’s reading comprehension score that is not predicted by their academic year or summer exposure. At Level 2, we tested if an individual child’s status and slope over the two time periods described earlier are predicted by group status (TD, PL) and/or parental SES. For the intercept and academic and summer slopes, we had a separate Level 2 equation for each Level 1 coefficient, π_pi, where p = 0, 1, 2, 3:

where π_pi is the pth growth parameter from the Level 1 model, β_p0, β_p1, β_p2 are linear regression coefficients, and r_pi is a random effect. To examine the effect of lesion characteristics, we included lesion size and lesion laterality as Level 2 predictors.

RESULTS

Descriptive Statistics and Univariate Analyses

Table 2 shows the average reading decoding and reading comprehension scores for the TD children and the children with PL during the different time periods in each analysis. Prior to carrying out HLM analyses, we carried out an analysis of covariance (ANCOVA) with group (TD, PL) and parental SES as independent variables and reading decoding and reading comprehension at each time point as the dependent variable to examine group differences at different time points. These analyses showed that parental SES was a significant predictor of children’s reading decoding skill and their reading comprehension skill at each assessment time point between the fall of kindergarten and spring of second grade (all ps <.05). Controlling for SES, the children with PL performed significantly lower on the reading decoding compared with the TD children during the fall of first grade and the fall and the spring of second grade, but not during kindergarten. Similarly, for reading comprehension, controlling for SES, the children with PL significantly lagged behind their TD peers in the spring of first grade as well as in the fall and the spring of second grade but not in the fall of first grade. However, these results include different subgroups at each data point, whereas HLM incorporates all of the participants who have been observed at least once in the same model. In addition, as shown in the HLM analyses detailed in the next section, these differences vary depending on lesion characteristics.

HLM Analyses for Reading Decoding

Unconditional Models

We first developed a Level 1 model that best describes the growth rates of the individual children. Table 3 presents the unconditional growth models for reading decoding for all children in the study, TD children and children with PL combined. The fixed effects for intercept, academic year slopes, and summer slope are all statistically significant. These effects tell us that the intercept is significantly different from 0 and that children’s reading decoding significantly grows over time. The estimated intercept was 454, which is the average reading W-score at the midpoint of the fall of kindergarten and spring of second grade, which is approximately the middle of first grade (SD = 24.4). The estimated intercept of our sample for reading decoding was similar to statistics of the normative sample of 7-year-olds on the WJ-III Tests of Achievement (M = 460.6, SD = 31.9; McGrew & Woodcock, 2001). The children’s linear monthly rate of growth was 4.1 points per month during kindergarten, 3.4 points per month during first grade, 1.5 during second grade, and 1.7 during the summers. Average rate of growth during the academic year was significantly higher than growth during the summer (p = .02). The significant Level 2 random effects for intercept and kindergarten growth reveal sufficient variation in the intercept and kindergarten growth rates explained by Level 2 predictor variables (Table 3).

Table 3 Estimates of fixed and random effects from individual growth models to predict average reading decoding and comprehension score (intercept) and kindergarten (KG), first-grade (1G), second-grade (2G), and summer rate of change (simultaneously) in child reading decoding and comprehension in the combined sample of typically developing TD children and children with prenatal or perinatal brain lesions (PL)

	Parameter Estimate (standard error)
	Reading Decoding	Reading Comprehension
Fixed Effects
Intercept	453.86***	462.35***
	2.59	2.26
KG	4.1***	—
	0.29	—
1G	3.41***	2.79***
	0.23	0.33
2G	1.52***	0.89***
	0.19	0.2
Summer	1.65**	1.13
	0.51	0.72
Variance Components
Level 2: Intercept	594.44***	388.78***
	24.38	19.72
Level 2: Slope (KG)	3.85***	—
	1.96	—
Level 2: Slope (1G)	0.94	6.46***
	0.97	2.54
Level 2: Slope (2G)	0.37	3.19***
	0.61	1.78
Level 2: Slope (Summer)	0.86	40.22***
	0.92	6.34
Goodness of fit
−2LL	4,147.37	2,418.87

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

Lesion Group, SES, and Reading Decoding Growth

To determine whether group (TD, PL) or parental SES predict reading decoding growth, we included both as Level 2 predictors in our growth models (Table 4). We first examined whether the intercept for reading decoding score varied as a function of group (TD, PL), while controlling for SES. The models predicting reading decoding showed an effect of group on the intercept, controlling for parental SES (p = .03). On average, the children with PL performed 11.8 points lower than their TD peers at the midpoint time between the fall of kindergarten and the spring of second grade (approximately the middle of first grade). This finding corresponds to a 0.5 standard deviation difference in the reading decoding scores of TD children and the children with PL at this time point. According to the WJ-III manual (Schrank & Woodcock, 2002), W score differences from the normative sample of 0 to 6 correspond to age-appropriate performance within normal limits, differences of 7 to 13 points correspond to a mild delay that is within normal limits, differences of 14 to 30 points correspond to a mild delay outside of normal limits, differences of 31 to 50 points correspond to a moderate delay outside of normal limits, and differences of 50 or more points indicate a severe delay. Using these descriptions, the children with PL were mildly delayed within normal limits in reading decoding compared with their TD peers, while controlling for SES.

Table 4 Estimates of fixed and random effects from individual growth models in which socioeconomic status (SES; parent education), group (children with prenatal or perinatal brain lesion [PL] versus typically developing [TD] children), predict average reading decoding and comprehension (intercept) and kindergarten (KG), first grade (1G), second grade (2G), and summer rate of change (simultaneously) in child reading decoding and comprehension production in the combined sample of TD children and children with PL

		Parameter Estimate (standard error)
		Reading Decoding	Reading Comprehension
Fixed Effects	Intercept	457.45***	465.39***
		3.01	2.61
	KG	4.3***	—
		0.34	—
	1G	3.40***	2.84***
		0.28	0.41
	2G	1.38***	0.80***
		0.22	0.24
	Summer	1.93**	1.33
		0.60	0.87
SES
	Intercept	8.25**	6.30**
		2.46	2.14
	KG	0.08	—
		0.28	—
	1G	0.15	−0.22
		0.24	0.34
	2G	−0.01	−0.01
		0.19	0.21
	Summer	−0.94	0.59
		0.51	0.74
Brain Lesion
	Intercept	−11.81*	−11.49*
		5.21	4.64
	KG	−0.78	—
		0.64	—
	1G	0.05	−0.12
		0.53	0.74
	2G	0.40	0.38
		0.41	0.45
	Summer	−0.88	−0.86
		1.13	1.62
Goodness of fit
	−2LL	4,068.45	2,359.87

*p < .05; **p < .01; ***p < .001.

We also examined the relations of group (TD, PL) to academic year and summer growth for reading decoding. Neither academic year nor summer growth in reading decoding significantly differed as a function of group. All groups showed greater rates of growth during the academic year than during the summer, but this finding did not vary as a function of group (all ps >.05; Table 4, Figure 1a).

Figure 1 Predicted growth curves for (a) reading decoding in children with prenatal or perinatal lesion (PL), low socioeconomic-status (SES) typically developing (TD–LOW SES) children and high-SES TD (TD–HIGH SES) children, and (b) reading comprehension in children with PL, low-SES TD children, and high-SES TD children. Note. KG = kindergarten; 1G = first grade; 2G = second grade.

Lesion Laterality and Reading Decoding Growth

Because the children with PL do not form a homogeneous group and differ widely in their lesion characteristics, we next examined the relation of laterality group, while controlling for SES, in predicting children’s reading decoding growth. We included RH and LH presence as dummy variables, as well as parental SES, at Level 2, comparing each of these groups to TD children.

We first examined if the intercept varied as a function of lesion laterality group. Results showed that the children with LH on average performed 7.8 points lower than their TD peers, but this difference was not statistically significant (p = .20), controlling for SES. The children with RH on average performed 19.6 points lower than their TD peers (p = .01), controlling for SES. This finding corresponded to a 0.8 standard deviation difference in reading decoding scores between TD children and the children with RH. Using W difference interpretations provided in the WJ-III manual, the children with RH were mildly delayed outside normal limits in their reading decoding as compared with their TD peers. However, even though the children with RH, but not LH, differed from the TD children, the children with LH and RH did not differ from each other in their intercepts (p = .17). We next examined if academic year, summer growth, or the difference in growth during the academic year versus summer varied as a function of lesion laterality group, and we found no differences (all ps >.05; see Figure 2a).

Figure 2 (a) Empirical (left) and predicted (right) growth curves for reading decoding for typically developing children (TD), children with left-hemisphere lesions (LH), and children with right-hemisphere lesions (RH). (b) Empirical (left) and predicted (right) growth curves for reading comprehension for TD children, children with LH, and children with RH. Note. KG = kindergarten; 1G = first grade; 2G = second grade.

Lesion Size and Reading Decoding Growth

We next included lesion size, as well as parental SES, at Level 2. We first examined whether the intercept for reading decoding varied as a function of lesion size. Results showed that lesion size was a significant predictor of the intercept of reading decoding (p = .02); a one-category increase in lesion size was associated with a 5.4-point decrease in reading decoding scores. In other words, the children with small lesions were estimated to score 5.4 points lower than the TD children, the children with medium lesions were estimated to score 10.8 points lower, and the children with large lesions were estimated to score 16.2 points lower than their TD peers. This finding corresponded to a 0.7 standard deviation difference in reading decoding score between the TD children and the children with large lesions. Using W difference interpretations provided by the WJ-III manual, the children with small lesions were within the normal range, the children with medium lesions were mildly delayed within the normal limits, and the children with large lesions were mildly delayed outside the normal range as compared with TD peers in their reading decoding. Again, there were no group differences in academic year growth, summer growth, or the difference between academic year and summer growth as a function of lesion size (all ps >.05).

Socioeconomic Status

Because our sample of children with PL, on average, came from higher-SES backgrounds than our sample of TD children, we controlled for parental SES differences in all models presented. However, because the children with PL had less variability in their SES characteristics and did not show SES effects in previous studies or in our preliminary analyses, we did not examine effects of SES on reading growth in this group. Examining the effects of SES for the TD children where we did have wide variability, we found that SES was significantly related to the intercept of reading decoding (p < .01). On average, a 1 standard deviation increase in the composite SES score was associated with an 8.3-point increase in children’s reading decoding score, which corresponds to a 0.3 standard deviation advantage in children’s reading decoding score at the midpoint between the fall of kindergarten and the spring of second grade (approximately the middle of first grade). A 1 standard deviation difference in the composite SES score corresponds roughly to a $27,000 difference in income and a 2-year difference in parental education. Using the W score difference interpretations provided in the WJ-III manual, TD children whose families were 1 standard deviation below the mean SES of our TD sample were mildly delayed within normal limits in their reading decoding compared with their higher-SES peers.

We also examined the relations of parental SES to academic year, summer growth, and difference in growth rate during the academic year versus summer in reading decoding and did not find a pattern that differed by SES (all ps >.05; Table 4, Figure 1a).

Reading Decoding in Children With PL Versus TD Children From Low-SES Backgrounds

Finally, we compared reading decoding development in the two groups of children who experience challenges, either biological (children with PL) or environmental (children from lower-SES backgrounds). We defined the TD children from low-SES backgrounds as those from families who were at least 1 standard deviation below the average SES of families in the study. Using this definition, the reading decoding intercept of the children with PL was significantly lower than that of lower-SES children from the TD group (1 standard deviation below the mean SES; p < .01), but there were no differences in growth rates during the academic years or summer.

Interim Summary

Results on our reading decoding measure revealed that children with PL performed worse than their TD peers (i.e., they had significantly lower intercepts). The worse performance of the children with PL was primarily driven by the children with RH and children with larger lesions. However, although the RH group differed from the TD group in terms of their intercept for reading decoding and the LH group did not, the two laterality groups did not significantly differ from each other. In addition, TD children from lower-SES backgrounds performed worse on the intercept measure than their higher-SES TD peers, but growth in reading decoding was not associated with group (TD, PL) or parental SES. Finally, children with PL showed a lower reading decoding intercept than TD children who were at least 1 standard deviation below the SES mean of our TD sample.

HLM Analysis for Reading Comprehension

Unconditional Models

Table 3 presents the unconditional growth models for reading comprehension for all children in the sample (TD children and children with PL combined). The fixed effects for intercept and academic year slopes were statistically significant. The effects tell us that the intercept was significantly different from 0 and that children’s growth in reading comprehension changed significantly during the academic year. The fixed effect for summer slope was not significant, but the significant Level 2 random effects for summer slope revealed significant variation to be explained by Level 2 variables. The estimated intercept was 462, the average reading comprehension W score at the midpoint of the fall of first grade and spring of second grade, which corresponds approximately to the reading comprehension scores at the end of first grade and beginning of second grade (SD = 19.72). The estimated intercept of our sample for reading comprehension was similar to statistics of the normative sample of WJ-III Tests of Achievement (7-year-olds, M = 474.03, SD = 20.7; McGrew & Woodcock, 2001). On average, children gained 2.8 points per month during first grade, 0.9 point per month during second grade, and 1.1 points per month during summer. The average rate of growth during the first grade was marginally significantly higher than growth during the summer (estimate = 1.6, p = .07), while the growth rate during second grade and growth rate during the summer did not differ from each other (estimate = 0.1, p > .05). The significant Level 2 random effects for intercept, first-grade growth, and second-grade growth reveal that there was sufficient variation in these parameters by Level 2 predictor variables.

Lesion Group and SES in Relation to Reading Comprehension Growth

To determine whether group (TD, PL) and parental SES predict reading decoding growth, we included group and parental SES as Level 2 predictors in our growth models (Table 4). We first examined if the intercept varied as a function of group, while controlling for parental SES. The models showed a significant effect of group (p = .02), such that children with PL scored 11 points per month lower than their TD peers. This finding corresponded to a 0.6 standard deviation difference in reading comprehension score between the TD children and the children with PL. Using the W score difference descriptions in the WJ-III manual, a difference of 11 points corresponds to a mild delay within normal limits.

We also examined the relation of brain lesion group to academic year growth, summer growth, and the difference between the two for reading comprehension and found no differences as a function of group (Table 4, Figure 1b).

Lesion Laterality and Reading Comprehension Growth

Next we examined the role of lesion laterality by including RH and LH as dummy variables, while controlling for parental SES, at Level 2. We first examined if the intercept varied as a function of lesion laterality. Results showed that, when controlling for SES, children with RH performed significantly lower than their TD peers and scored 18.3 points below them (p = .01), whereas children with LH did not differ from TD children (p = .14). This finding corresponded to a 0.9 standard deviation difference in the reading comprehension intercept between TD children and children with RH (Figure 2b). Using the W score interpretations provided in the WJ-III manual, a score difference of 18 corresponds to a mild reading comprehension delay in children with RH compared with their TD peers. The intercept did not significantly differ between children with RH and children with LH (p = .18)

We then examined if academic year or summer growth varied as a function of lesion laterality. In terms of growth during the academic year and summer periods, children with RH lesions showed greater growth during second grade than their TD peers and greater loss during the summers, whereas the growth of children with LH did not differ from that of TD children in terms of growth during any period. Children with RH gained 1.7 points per month more than their TD peers in reading comprehension scores during second grade (p < .01), but they lost 7 points per month more than their TD peers over the summers (p = .02). Children with RH also exhibited greater growth during second grade (p < .01) and greater loss during the summer periods (p < .01) compared with children with LH. Children with RH grew significantly more in reading comprehension during second grade than during the summer periods (p = .03). The differences reported remained the same when controlling for lesion size in the same model. No other relations to lesion laterality reached significance (Figure 2b).

Lesion Size and SES in Relation to Reading Comprehension Growth

Finally, we examined lesion size and parental SES as Level 2 predictors. We first examined whether reading comprehension intercepts varied as a function of lesion size, while controlling for parental SES. The results showed that lesion size was associated with the intercept of the reading comprehension score (p < .01). On average, a one-category increase in lesion size (e.g., small to medium lesions, medium to large lesions) was associated with 5.6 points fewer in children’s reading comprehension scores. In other words, children with small lesions were estimated to score 5.6 points lower than TD children, children with medium lesions were estimated to score 11.2 points lower, and children with large lesions were estimated to score 16.8 points lower than their TD peers. This finding corresponds to a 0.8 standard deviation difference in reading decoding scores between TD children and children with large lesions. Using W score difference interpretations provided in the WJ-III manual, children with small lesions were within the normal range, children with medium lesions were mildly delayed within normal limits, and children with large lesions were mildly delayed outside of normal limits in their reading comprehension. Neither academic year nor summer growth varied as a function of lesion size (all ps > .05).

Parental SES

Similar to reading decoding, parental SES effects were examined in TD children only. The reading comprehension intercept varied as a function of parental SES (p < .01). A 1 standard deviation increase in the composite SES score was associated with a 6-point increase in the reading comprehension intercept in TD children, which corresponds to a 0.3 standard deviation increase in the reading comprehension score. Using W difference interpretations provided in the WJ-III manual, children from families that were 1 standard deviation below the mean in their parental SES were within the age-appropriate normal limits. Growth during the school year and summer periods and the difference in growth between these periods did not vary as a function of SES background (all ps > .05; Table 4, Figure 1b).

Reading Comprehension in Children with PL Compared to TD Children From Low-SES Backgrounds

Finally, we compared reading comprehension development in the two groups of children who experienced the greatest challenges, either biological (children with PL) or environmental (TD children from lower-SES backgrounds). We defined TD children from low-SES backgrounds as those from families who were at least 1 standard deviation below the average SES of families in the study. Using this definition, the reading comprehension intercept of the children with PL was significantly lower than that of TD children from lower-SES backgrounds (1 standard deviation below the mean SES; p < .01), but there were no differences in growth rates (p > .05).

Interim Summary

In summary, results on reading comprehension revealed that children with early prenatal or perinatal lesions performed worse in reading comprehension than their TD peers (i.e. they had significantly lower intercepts). As for reading decoding, the worse performance of children with PL was mainly driven by children with RH and children with larger lesions. In addition, among TD children, children from lower-SES backgrounds scored lower than their higher-SES TD peers. Different than our results for reading decoding, growth in reading comprehension during the academic year and during the summer periods was differentially associated with lesion laterality, such that children with RH showed significantly greater growth during the academic year and significantly greater summer slide compared with both children with LH and TD children.

DISCUSSION

We compared the reading decoding and reading comprehension growth of children with PL and TD children by examining their growth during the academic year and summer months. We used differential growth during the academic year versus summer months to examine whether the structured instruction provided during periods of schooling is particularly important for children who have biological risks for lower academic achievement. We found that children with PL as a group performed worse than their TD peers on both reading decoding and reading comprehension. The performance of the children with PL was delayed compared with our TD sample. Nevertheless, performance of the children with PL was within the normal range according to the norms of the WJ-III. Thus, the remarkable plasticity children with PL exhibit for oral language functions extends to reading decoding and comprehension, even when lesions involve classic language networks. This pattern of plasticity contrasts with that seen when brain injury is incurred in adulthood as these later lesions are associated with substantial difficulties in reading decoding and comprehension (Price et al., 2003).

Our findings also showed that a subset of the children with PL experience greater difficulties in reading in early elementary school compared with TD children and other children with PL—in particular, those who had larger lesions and those with RH. For reading decoding, these two subgroups had lower intercepts than the TD group but did not show differential growth during the academic year or the summer compared with other groups. For reading comprehension, these two subgroups also had lower intercepts than the TD group. Moreover, children with RH had significantly greater growth during second grade and greater summer slide than the TD children and the children with LH. Their more jagged growth profile suggests that the structured learning environment of school may play a more crucial role in the development of reading comprehension skills for children with RH than for either children with LH or TD children.

Earlier studies have shown that children with PL initially lag behind their TD peers on various language measures such as first words, vocabulary size, word combinations, syntactic complexity, and narrative development, but they catch up and eventually perform, on average, within the low normal to normal range on each of these aspects of language development (Bates, 1999; Bates et al., 1997; Rowe et al., 2009). Our findings show that children with PL with certain lesion characteristics lag behind their peers in reading decoding and comprehension during the early years of elementary school but do not provide information as to whether they eventually catch up to TD children, a pattern that would mirror prior findings on language development.

In terms of lesion characteristics, consistent with previous research, we found that children with larger lesions have more difficulty with reading. These findings suggest that the functional plasticity children with PL display for reading is limited by lesion size, perhaps because larger lesions compromise the wide neural networks that reading decoding and comprehension typically engage (Dehaene et al., 2010; Levine et al., 2005; Shaywitz et al., 1998). What remains to be determined is whether these findings reflect a difficulty in getting reading off the ground with subsequent catch-up or a more long-lasting difficulty.

Different from adults with brain injury, children with PL generally do not show strong effects of lesion laterality in terms of language functioning (Bates et al., 1997; Levine et al., 2015). For reading, there are inconsistent findings in the small number of existing studies with some reporting RH are more detrimental (Aram & Ekelman, 1988; Woods & Carey, 1979), others reporting LH are more detrimental (Frith & Vargha-Khadem, 2001), and still others reporting no laterality effects (Ballantyne et al., 2008). In contrast to other studies, we examined multiple aspects of reading, examined multiple lesion characteristics, and controlled for differences in socioeconomic characteristics of the children’s families. We found that children with RH showed lower intercepts in reading decoding and reading comprehension than TD children but did not significantly differ from children with LH. Further, for reading comprehension, the RH group showed a more jagged, time-varying profile of growth characterized by greater growth during the school year and more slide during the out-of-school summer months than both TD children and children with LH.

The patterns of difficulty we observed in reading decoding and comprehension for the children with RH might be related to laterality differences in domain-general executive function skills, such as focused attention or inhibition, which could negatively impact reading skills (Blair & Razza, 2007). Both adults and children with RH have been reported to have executive function difficulties compared with individuals with no injury (Champagne-Lavau & Joanette, 2009; Garavan, Ross, & Stein, 1999). A second possibility is that the reading comprehension difficulties of the children with RH are related to laterality differences in discourse processing. Right-hemisphere networks are activated in the processing of oral discourse in both children and adults, and right-hemisphere injury has been associated with discourse-processing difficulties in adults (Chapman, Max, Gamino, McGlothlin, & Cliff, 2003; Eviatar & Just, 2006; Kaplan, Brownell, Jacobs, & Gardner, 1990). Children with RH, like adults with RH, might experience difficulties in processing oral discourse. Such difficulties might explain our finding of greater reading comprehension difficulties among children with prenatal or perinatal RH, given that oral discourse comprehension forms the basis for reading comprehension (Storch & Whitehurst, 2002). These possibilities are not mutually exclusive, and both of these factors—executive functioning and discourse processing—might contribute to the reading difficulties of the children with RH. A note of caution is that our sample of children with RH was small. Future behavioral and neuroimaging studies with larger sample sizes should replicate our findings and probe these various explanations for the reading difficulties we observed.

A major question we asked in the current study was whether environmental input plays a greater role in supporting the development of reading skills, particularly the more challenging aspects of reading in children with PL than in TD children. Previous findings have supported the more important role of input for language development in this group (Demir et al., 2015; Rowe et al., 2009). In the current study, we found support for this hypothesis only among children with RH and only for reading comprehension, as the difference between reading comprehension growth during the school year versus the summer was greater for this group than for children with LH or TD children. These findings suggest that the richer input provided in school may be particularly important to the development of the reading comprehension skills of children with RH. A similar pattern has been shown for challenging aspects of language development for children with PL regardless of lesion laterality. In particular, parental language input is a stronger predictor of preschoolers’ syntactic growth for preschool children with PL than for TD children (Rowe et al., 2009). Similarly, parental talk about decontextualized topics played a larger role in predicting narrative skills of children with PL than TD children (Demir et al., 2015). Our results extend these findings to the relation of formal schooling versus summer periods to the growth of reading comprehension but not to decoding skills for children with RH, implicating a particularly strong role of structured input in supporting reading comprehension, which is the most challenging aspect of reading and one that develops over an extended period of time (Paris, 2005).

We also compared developmental trajectories of reading skill in children with PL to those of TD children from varying parental SES backgrounds. TD children from lower-SES backgrounds exhibited achievement gaps in reading from kindergarten to second grade in both reading decoding and reading comprehension compared with their higher-SES TD peers, as indicated by their lower intercepts for these two aspects of reading (as shown by both ANCOVAs and HLM analyses). The lower reading development profile of children from lower-SES backgrounds is consistent with many studies documenting SES-related academic achievement gaps (Bradley & Corwyn, 2002; Brooks-Gunn & Duncan, 1997; NCES, 2011). For example, lower-SES children’s reading performance has been shown to be three grade levels below that of their higher-SES peers by the end of fifth grade (Cooper, Borman, & Fairchild, 2010). Previous studies also have shown greater summer loss in children from disadvantaged socioeconomic backgrounds (Alexander et al., 2001). In the current study, we did not find that summer slide was significantly related to the SES background of TD children. Although our TD sample was demographically diverse, it did not include children from the very low end of the SES continuum, a group that was included in prior studies showing greater summer slope for lower-SES children. Thus, lower-SES children included in our study might have received sufficient amounts of environmental support so that they did not experience differential summer slide. In addition, summer slide differences related to SES are more prominent in the later years of school as compared with the early years, which were the focus of the current study (Kieffer, 2010).

The reading difficulties observed in children with RH and in TD children from lower-SES backgrounds emerged at different time points. For TD children from lower-SES backgrounds, differences were present at school entry (as shown by the ANCOVA analyses) but there were no differences in growth trajectories as a function of SES (as shown by the HLM analyses), implicating the contribution of experiences prior to school entry in these differences. In contrast, for children with early brain injury, particularly those with RH, there were no differences in reading at school entry (as shown by the ANCOVA), but the growth of reading during school versus summer months (as shown by the HLM analyses) contributed to group differences, thereby implicating the impact of school versus home input on learning.

Our findings extend prior discussions in the literature focusing on the dynamic, ongoing interplay between the child and the environment in developmental psychopathology and physical health to academic development (Lewis & Mayes, 2012; Sameroff & Chandler, 1975). In particular, in both TD children and children with PL, they highlight the important interactions between children’s biological characteristics and the environmental factors they experience (Sameroff, 2010). Comparing children’s reading development over multiple time periods that offer varying supports, rather than comparing groups at a single time point, provides insight into the role of environmental and biological factors in shaping the reading trajectories of TD children and children with PL. Specifically, examining children’s performance both at school entry and during periods of schooling and summer breaks provides a more complete picture of the similarities and differences between the challenges faced by different groups of children and the factors that influence their developmental trajectories.

The current study has several limitations, which impacted our ability to address certain questions but also provide opportunities for future research. Notably, our study was limited by our sample size of children with PL, particularly with respect to our subgroup analyses. For example, fewer children in our sample had RH than LH, reflecting the prevalence of these lesions in children with PL (Reilly et al., 1998). Although children with RH differed from those with LH in terms of academic and summer growth in reading comprehension, the two groups overall had quite similar reading trajectories—specifically in terms of their intercepts for both reading decoding and comprehension. Thus, our lesion laterality results need to be replicated with a larger sample. Additionally, we had little SES variation among the families of children with PL, which precluded examining the joint effects of different kinds of challenges. Another limitation is that we only focused on reading development. Thus, we do not know whether the patterns we found are specific to reading or whether they extend to other academic areas, such as mathematics. Finally, we only considered early reading development—from the start of kindergarten through the spring of second grade. By extending beyond second grade, future studies could determine whether children with early brain lesion catch up with their peers in later years or whether they experience more long-lasting difficulties.

In sum, the aim of the current special issue is to foster a dialogue between cognitive developmental scientists who work on typical development and those who work on atypical development. Studies of atypically developing children extend the variability in children’s performance and allow us to ask how biological variations that are not possible to manipulate in typical children impact development and how they interact with experiential variations. The current findings suggest that the phenotype observed for reading decoding and comprehension, in both children with PL and TD children, represents the combined effect of factors interacting at multiple levels of analyses, including whether the brain is neurotypical or not at birth, other individual differences associated with the child’s biological characteristics, the particular aspect of reading under consideration, and environmental factors, which are often ignored in considering developmental trajectories in atypical groups (Demir et al., 2015; Karmiloff-Smith, 1998; Levine et al., 2015; Rowe et al., 2009). Although the outcomes of children who are subject to different types of challenges (internal, external) might seem similar, the factors influencing growth as well as the specifics of their growth patterns, including start levels, patterns of growth over time, and the factors that influence growth, may differ. As shown in our study, such differences can be observed by examining children’s growth trajectories during periods that vary in environmental support. Studying how learning trajectories differ for TD children and children with early brain injury who have differing challenges holds promise for informing our understanding of development as well as for guiding intervention efforts aimed at supporting children’s development.

Acknowledgments

We thank participating families and children; Chicago Language Developmental Project research assistants for help in collecting and transcribing the data; and Kristi Schonwald, Jodi Khan, and Jason Voigt for administrative and technical assistance.

Funding

This research was supported by P01HD40605 from the National Institute of Child Health and Human Development to S. C. Levine.

Additional information

Funding

This research was supported by P01HD40605 from the National Institute of Child Health and Human Development to S. C. Levine.

Notes

¹ Detailed lesion location information was not available for one participant whose lesion information was obtained from medical reports provided by the family.