Longitudinal Language Changes Associated with MRI Anatomy in Children with Autism Spectrum Disorder

Keywords

Autism; Language; MRI; Anatomy; DTI; Longitudinal; Planum Temporale, Broca’s area

Abstract

Background: Language ability is one of the strongest predictors of prognosis and developmental course in Autism Spectrum Disorder (ASD). A range of language abilities occur in ASD and although many have delays in language it remains unclear why some children’s language continues to lag, while others do not. Abnormal anatomy and function of language-related regions has been found in ASD, however, how these differences relate to language development over time is undetermined.

Methods: This study examined longitudinal changes in language functions in children with ASD and investigated whether cortical language region anatomy was related to these changes in language. Eighteen boys with ASD, 2-8 years old were evaluated (Time 1) and re-examined about 3.5 years later (Time 2) at ages 7-10. MRIs were collected at Time 2 to evaluate gray matter volume of anterior (Pars Triangularis, PTR; pars opercularis, POP) and posterior (Planum Temporale, PT; Posterior Superior Temporal Gyrus, pSTG) language regions and the microstructure of the arcuate fasciculus.

Results: Eleven boys had relative decline in language functions (decline group) and 7 boys had no relative change in language (no change group). The no change group had larger PT and right PTR volume relative to the decline group. In addition, the right PTR was correlated with the language change score, with larger right PTR associated with less language decline. There was a trend for non-right-handers to have more language decline than right-handers.

Conclusions: Results suggest differences in cortical language anatomy may play a role in language development, with further studies warranted.

Introduction

Autism Spectrum Disorder (ASD) is characterized by deficits in social interaction and communication, as well as the presence of restricted interests or repetitive behaviors [1]. ASD is often diagnosed in young children, usually between the ages of 2 and 5 years and language delay is one of the most common signs that a young child may have ASD. Language impairments can range from a complete lack of functional language to above average ability on standardized tests [2]. The developmental course of the disorder is highly variable, with likely many contributing factors; however, language ability is one of the strongest predictors of prognosis and developmental course in ASD [3]. A number of studies have examined various behavioral predictors of language development in ASD, with many different factors being identified, such as early motor imitation skills; pretend play skills, and commenting abilities [4-7]. Although many individuals with ASD have significant delays in language it remains unclear why some children show improvements in language over time, while others do not.

There is speculation that language networks are atypically organized in children with ASD. Several studies have found atypical frontal and/or temporal activation during language processing with a reduction or reversal of laterality being the most consistent finding, particularly in frontal regions [8-13]. A recent study of 2-3 year olds with ASD found a positive correlation between right fronto-temporal activation and receptive language skill, and that autism severity negatively correlated with left fronto-temporal activity [14]. These results suggest that the right hemisphere may be recruited to a greater extent in children with ASD with early right hemisphere recruitment associated with better language skill.

Studies examining the anatomy of cortical language regions have shown differences in children with ASD in both temporal and frontal language regions [15,16]. Only a few studies have investigated associations between anatomy and function or behavior. One study showed that groups with language impairment (ASD and specific language impairment) had rightward asymmetry of frontal language regions, while those without language impairment had leftward asymmetry [17]. There was also a positive correlation between verbal IQ and degree of leftward asymmetry of the inferior frontal gyrus in the ASD group. In another study, we found increased frontal language volume in ASD relative to controls [15]. Similar to de Fosse et al. [17], in the younger ASD group (7-11 years), the right pars triangularis volume was negatively correlated with language scores, suggesting better language was associated with a smaller right hemisphere volume. In addition, the left pars triangularis volume was positively correlated with autism severity. In contrast, a study of young ASD children (4-7 years), found higher language ability associated with increased rightward pars opercularis asymmetry [16]. In typically developing children (6-11 years), however, a leftward asymmetry of the planum temporale was associated with more advanced language ability. In a study of adolescent boys with and without ASD, we found that the group with leftward (typical) language laterality had smaller frontal gray matter volumes and higher Fractional Anisotropy (FA) of the arcuate fasciculus compared to the atypical language laterality group [18]. Laterality groups, however, did not differ on language ability and language scores for most ASD individuals were in the average range and did not differ from controls.

In summary, although both fMRI and structural MRI studies have found associations between right brain measures and language ability in children and adolescents with ASD, the results from the anatomical studies have been less clear. Anatomical differences, however, have been identified in the anatomy of language areas in individuals with ASD compared to controls, and one earlier study did examine agerelated changes in language regions using a cross-sectional design [18]. Although we were not able to compare MRI changes overtime within-subjects in the current study, we were able to compare language ability over-time. Given the important role of language abilities in predicting prognosis and developmental course [3], the current study was designed to examine the relationship between the size of cortical language regions and language development in boys with ASD. There were two main goals of this study. First, was to determine if children with ASD change in their individual language abilities over time. We predicted that we would find two subgroups within our ASD sample, one with improved language over time and one without language improvement. The second major goal of this study was to determine whether gray matter volume and/ or white matter integrity predicted changes in language abilities in boys with ASD who were examined longitudinally. A group of boys who were initially evaluated for ASD between 2 and 8 years of age returned for re-evaluation about 3½ years later to participate in an anatomical MRI study that included a structural MRI sequence and Diffusion Tensor Imaging (DTI). We hypothesized that there would be differences in gray matter volume in those individuals who showed improvement in language abilities compared to those who continued to show a significant delay in language development. Given the left hemisphere’s predominant role in language, we hypothesized that larger left gray matter volume of language regions, particularly in frontal language zones, would be associated with improved language over time. In addition, we predicted improved language ability would be associated with increased integrity of the arcuate fasciculus, particularly in the left hemisphere.

Materials and Methods

Subjects

Subjects included 18 boys with ASD, 7-10 years old, (mean = 8.80, SD = 1.00) with a previous evaluation 1-6 years prior (mean = 3.54, SD = 1.64) at 2-8 years of age (mean = 5.26, SD = 1.37). Information was collected from the parent regarding services that the child had been receiving from the time of the previous evaluation until participation in the study. The majority of subjects were receiving speech therapy through the school system (30 minutes once or twice per week), as well as occupational therapy, with some also receiving private speech and occupational therapy. Only 2 participants received ABA therapy and several had other therapies including counseling, social skills group, and adapted physical education or physical therapy. There were 2 children who received no speech therapy. Individuals with frank neurological damage, with a known genetic disorder, who were born prematurely (less than 36 weeks), who had had seizures within the last 3 years, or who were on anti-seizure medication were excluded from the study. All subjects had English as their first language.

Parents and participants were informed of the procedures and parents gave written consent prior to the child’s participation in the study. Children who were able, also provided written assent prior to participation. All data in this manuscript were collected in compliance with the Louisiana State University Health Sciences Center and Children’s Hospital Institutional Review Boards.

Standardized Tests

Time 1: An ASD diagnosis was given by a licensed clinical psychologist, based on DSM-IV criteria [19] using the Autism Diagnostic Interview-Revised (ADI-R) [20] and the Autism Diagnostic Observation Schedule (ADOS) [21]. Most children were administered the Mullen Scales of Early Learning (Mullen) [22] to assess developmental abilities, however, some children who were over the Mullen upper age limit, 5 years and over (n = 5) received the Wechsler Preschool and Primary Scale of Intelligence (WPPSI) [23] or the Wechsler Intelligence Scale for Children (WISC) [24]. These tests all include verbal and nonverbal subtests.

Time 2: Current ASD diagnosis was confirmed with the ADOS. The Leiter International Performance Scale-Revised (Leiter) [25] was administered to assess nonverbal IQ. Since children were older at Time 2, the Mullen could not be repeated, therefore, all subjects received the Oral and Written Language Scales (OWLS) [26] to assess receptive and expressive language. The Almli Handedness Assessment [27], which has individuals perform 10 unimanual tasks, was also administered to all subjects to assess hand preference.

Language Change: Both the Mullen and OWLS have a receptive language scale which assesses the understanding of spoken language and an expressive language scale which measures the use of spoken language. Receptive and expressive age equivalent scores were averaged to create a total language measure. Age equivalent scores, rather than standard scores, were used due to floor effects on the Mullen. For the 5 subjects who did not receive the Mullen at Time 1, age equivalents for the verbal subtests of the Wechsler scale were used. For both Times 1 and 2, each subject’s age equivalent score was subtracted from his chronological age. In order to determine withinsubject change in language function, the difference in these scores from Time 1 to Time 2 (Time 1 minus Time 2) was calculated. Individuals who had a decrease of more than 12 months were considered to have a relative decline in language function (decline group), while those who had a decrease or increase of 12 months or less were considered to have no change in language function (no change group). There were no individuals who had an increase of greater than 12 months.

MRI Acquisition

MRI data was collected at Time 2 and all participants were trained in a mock scanner prior to the actual MR scanning. Volumetric MR images were acquired on a Siemens 3 T Verio scanner. T1-weighted images were obtained as a series of 160, 1 mm gapless sagittal images. MPRage was used, with technical factors of: TR = 1900 ms, TE = 2.48, 256×256 pixel matrix, 250 mm field of view, and 9° flip angle. Data sets were rotated into alignment in the sagittal, axial, and coronal planes in order to eliminate any head rotation and MRI scans were maintained in real space. Each MRI scan series was assigned a blind number to assure subject confidentiality and to ensure that all measurements were performed blind to subject. Axial diffusion-weighted images, aligned parallel to the intercommissural plane, were acquired using echo planar imaging, as a series of 36, 3 mm contiguous images. The following parameters were used: b-value = 1000 s/mm2, 30 gradient directions plus 1 reference image (b = 0), pixel matrix = 128x128, FOV = 230 mm.

Volume Measurements

Four Regions of Interest (ROIs) were measured by a single investigator, including: two posterior language zones (planum temporale and posterior superior temporal gyrus) and two frontal language zones (pars triangularis and pars opercularis). 3D Slicer [28] was used to manually trace the gray matter volume of each ROI in each hemisphere (Figure 1). Boundaries were the same as those used in our prior studies and inter-rater reliability for each of these measures has previously been established, with Intra-Class Correlations (ICCs) of 0.83 or higher for each ROI [29, 30]..

Figure 1: ROI measurements in a single subject. Top row: Planum Temporale (yellow) and posterior Superior Temporal Gyrus (blue), middle row: Pars Triangularis, bottom row: Pars Opercularis.

Planum Temporale (PT): The anterior boundary was defined as Heschl’s sulcus, and when present the second Heschl’s gyrus was included in the planum measure. In the coronal plane, this image was where Heschl’s was fully visible and there was a small amount of white matter lateral to it. The posterior boundary was defined in the sagittal plane, as the point where the horizontal ramus of the Sylvian fissure turns upward into the Posterior Ascending Ramus (PAR), so neither the PAR nor posterior descending ramus was included in the planum measurement. In cases where the Sylvian fissure gently sloped upward, the knifecut method [31] was utilized. If there was no PAR, the end of the horizontal portion of the Sylvian fissure was used as the posterior boundary. This boundary was defined in the coronal plane as the most posterior slice where the Sylvian fissure was clearly visible, before it became intermixed with white matter. However, if the Sylvian fissure extended into parietal regions, the posterior boundary was defined as the image just anterior to the one where the intraparietal sulcus appeared to encircle the Sylvian fissure. The PT was measured in the coronal plane, with Heschl’s sulcus, when present, as the medial boundary. Laterally, the boundary for the planum was the edge of the Sylvian fissure, not including the lateral wall of the posterior superior temporal gyrus.

Posterior Superior Temporal Gyrus (pSTG): The anterior boundary was the same as for the PT. The posterior boundary was the most posterior point of the Sylvian fissure, as defined for the PT. The pSTG was bounded superiorly by the horizontal ramus of the Sylvian fissure and inferiorly by the superior temporal sulcus, including the superior bank of the superior temporal sulcus. If a PAR was present, the superior boundary was defined in the sagittal plane as the point at which the Sylvian fissure turned upward into the PAR, so as not to include parietal regions in the pSTG measure. When the pSTG was discontinuous and there were overlapping areas, only the superior bank of the most superior sulcus was included in the measure.

Pars Triangularis (PTR): The sagittal plane was predominantly used to measure this region because it provides the clearest view of the PTR. The anterior boundary was the anterior horizontal ramus and the posterior boundary was the anterior ascending ramus. Thus, the posterior/superior bank of the anterior horizontal ramus and the anterior/superior bank of the anterior ascending ramus were included in the PTR measurement. When extra sulci, internal notches, occurred between these two rami, the banks of these sulci were included in the measurement. The lateral boundary was defined as the most lateral sagittal image, prior to the rami being cut off, so that surface gray matter was not included. At times, only one of the sulci, either the anterior horizontal ramus or the anterior ascending ramus, was cut off, while the full extent of the other sulcus remained, in which case the bank of the remaining sulcus was still measured on images as far lateral as possible, until the sulcus was cut off. The medial boundary was defined as the most medial sagittal image in which the insula was clearly defined, prior to the white matter intruding. At this boundary, the insula appeared with clearly defined strips of white matter. The superior boundary was the inferior frontal sulcus.

Pars Opercularis (POP): The sagittal plane provides the best view of the POP, so this plane of section was primarily used to measure this region. In the sagittal plane, the anterior boundary was defined as the anterior ascending ramus and the posterior boundary was defined as the precentral sulcus. Thus, the posterior/inferior bank of the anterior ascending ramus and the anterior bank of the precentral sulcus, up to the inferior frontal sulcus, were included in the measurement. The inferior frontal sulcus, also defined in the sagittal plane, was used as the superior boundary. While on some images the precentral sulcus connects with the Sylvian fissure, it was not continuous with the Sylvian fissure throughout the full extent of the POP. In addition, in some brains, the precentral sulcus never fully connected with the Sylvian fissure. When the precentral sulcus did not connect with the Sylvian fissure, the POP was measured to the point where the Sylvian fissure would have connected with the precentral sulcus and the anterior bank of the precentral sulcus was measured from the inferior frontal sulcus, down until it disappeared. The lateral boundary was defined on sagittal images as the most lateral image prior to the rami being cut off, so that surface gray matter was not included. At times, only one of the sulci, the anterior ascending ramus or the precentral sulcus was cut off, while the other one remained, in which case the bank of the remaining sulcus was still measured on images as far lateral as possible, until the sulcus was cut off. The medial boundary was defined as one sagittal image lateral to the medial boundary of the PTR. The diagonal sulcus was measured separately and added into the POP measure. Other extra sulci that appeared between the anterior ascending ramus and the precentral sulcus, were also included.

Diffusion Tensor Imaging (DTI) Measures

The volumetric measurements described above, which included the PTR and POP and pSTG and PT in both hemispheres, were used for probabilistic tractography. Each region for each subject was edited to include a small amount of white matter, 1-2 voxels on each side [32]. Tractography was performed separately in the left and right hemispheres with the temporal language areas (pSTG + PT) used as the seeding mask and the frontal language areas (PTR + POP) as the termination region in order to examine the arcuate fasciculus.

Analyses

Behavioral: Group differences in age, ADOS scores, language, and IQ were examined. To examine age, ANOVAs were performed with group (decline, no change) as the independent variable and Time 2 age, Time 1 age, or time between visits as the dependent variable. In order to investigate ADOS scores, a MANOVA was done with group as the independent variable and ADOS social and ADOS communication scores as the dependent variables. For language, ANOVAs were performed. Group was the independent variable and Time 1 language score or Time 2 language score was the dependent variable. The language scores were based on age equivalent scores and the difference between that and chronological age. To assess IQ, an ANOVA was done with group as the dependent variable and nonverbal IQ as the independent variable. Pearson correlations between change in language scores and change in ADOS social, communication, and total scores were also calculated.

Anatomical Volume Measurements: To examine group differences in ROI volumes, MANOVA was used with hemisphere as the withinsubjects independent variable, group (decline, no change) as the between-subjects independent variable, and PT, pSTG, PTR, and POP volume as the dependent variables. Pearson correlations were performed between change in language scores and ROI volumes (left and right hemisphere).

Probabilistic Tractography: FMRIB’s Diffusion Toolbox (FDT) which is part of FSL [33] was used for all analyses; detailed methods have been described previously [34]. Briefly, the diffusion and T1 data were first skull-stripped using the BET tool [35]. Diffusion data were then transformed, using affine registration, to a reference volume (the first volume) to correct for eddy currents and head motion. The DTI and T1 data were aligned using affine registration. Bayesian techniques were used to create a probability distribution of fiber direction for each voxel. Probability connectivity distributions between seed and termination points were created by repeatedly sampling from the distributions on voxel-wise diffusion directions. This resulted in each voxel having a value representing the probability of connection to the masks. This connection probability image was then binarized and thresholded by 10. See Figure 2 for tractography results from a single subject. To calculate FA, this mask was multiplied by individual FA maps and the mean FA of the tract was calculated. Similarly, for MD, the mask was multiplied by In order to examine differences in white matter microstructure, MANOVA was used with hemisphere as the within-subjects independent variable, group (decline, no change) as the betweensubjects independent variable, and arcuate fasciculus FA and MD as the dependent variables. Pearson correlations between language change scores and right and left FA and MD were calculated.

Figure 2: The results of probabilistic tractography, showing the arcuate fasciculus in the left hemisphere of a single subject.

Handedness: Based on the Almli handedness score, subjects were categorized as right-handers or non-right-handers, which included individuals who were left or mixed-handed. To examine handedness effects, we performed an ANOVA with handedness group (right, non-right) as the independent variable and language change score as the dependent variable. To examine differences in the number of right- and non-right-handers in each group (decline, no change), a chi-square was computed. Pearson correlations between handedness score and language change score was also calculated.

Results

In our sample, there were 11 (61.1%) children who showed a relative decline in language functions, with language change scores ranging from -13 to -53 months, and 7 (38.9%) with no change in language function, with language change scores ranging from -11 to +7 months (Table 1). There were no significant group differences in current age, previous age, or time between visits (p’s > .05). There were also no group differences in current or prior ADOS social or communication scores (p’s >.05). For language abilities, there were no group differences in Time 1 language measures (p>.05), however, for current language, there was a significant group effect (F1,16 = 18.11, p = .001), with the no change group having significantly higher scores than the decline group. Most subjects, however, had very impaired current language with a mean composite language score of 62.78 (SD = 15.34). There were no group differences in non-verbal IQ (p>.05). See Table 2 for mean behavioral measures by group. We found a significant positive correlation between change in language and change in total ADOS score (r = .572, p = .021). When we looked at the services subjects were receiving between Time 1 and Time 2, this did not seem to be driving the change in language over time. Individuals who received the most amount of speech therapy, with a combination of private and through the school system, all showed relative decline in language. Of the 2 individuals who did not receive any speech therapy, 1 was in the no change group and had the highest language change score of +7 months, while the other was in the decline group.

Table 1: Change in language scores from Time 1 to Time 2 in months for each subject by group

Table 2: Means and standard deviations of demographic and behavioral measures for Time 1 and Time 2 for each group. Language scores are the difference between age equivalent score and chronological age, with higher scores, therefore, indicating more impairment.

Anatomical Volume Measures

Refer to Table 3 for mean volume by group. The MANOVA examining group differences in ROI volumes revealed a significant group effect (F4,13 = 3.86, p = .028), hemisphere effect (F4,13 = 5.18, p = .010), and hemisphere by group interaction (F4,13 = 3.50, p = .038). At the univariate level, the group effect was significant for the PT (F1,16 = 11.35, p = .004), indicating significantly larger PT volume in the no change group than in the decline group. The hemisphere effect was also significant for the PT (F1,16 = 9.38, p = .007), indicating overall larger left than right volume for the PT. The interaction was significant for the PTR (F1,16 = 13.62, p = .002), indicating no group differences for left PTR, but larger right PTR for the no change group compared to the decline group. Stated another way, in the decline group, the left PTR volume was larger than the right, while in the no change group the right volume was greater than the left. The only significant positive correlation was between change in language function and right PTR volume (r = .575, p = .013) (Figure 3).

Table 3: Means and standard deviations of ROI gray matter volume in each hemisphere for each group.

Figure 3: Correlation between right pars triangularis (PTR) gray matter volume and change in language score.

Probabilistic Tractography

See Table 4 for mean FA and MD for each group. At the multivariate level, there was a significant hemisphere effect (F2,15 = 6.53, p = .009). This was significant for both FA (F1,16 = 13.26, p = .002) and MD (F1,16 = 7.74, p = .013) at the univariate level, indicating higher FA and lower MD in the right relative to the left hemisphere. Correlations between language change scores and FA and MD were not significant.

Table 4: Means and standard deviations of Arcuate Fasciculus (AF) Fractional Anisotropy (FA) and Mean Diffusivity (MD) in each hemisphere for each group.

Handedness

Our sample included 10 right-handers and 8 non-right-handers, including 4 left-handers and 4 with mixed handedness. Although there were no significant differences based on handedness, we did find some interesting trends which warrant comment. There was a trend for the non-right-handers to have more language decline than the right-handers (F1,16 = 4.24, p = .056). In the decline group, 54.5% were non-right-handers, whereas in the no change group, only 28.6% were non-right-handers. The correlation between handedness and language change was not significant.

Discussion

There were three main findings in this study. First, two groups emerged, with approximately 2/3s of the boys with ASD showing a relative decline in language function over time compared to a second group that showed no major change in language function over time. Second, there were significant differences in the anatomy of language cortex in these two language groups. The group with no change in language had a significantly larger overall PT gray matter volume and larger right PTR gray matter volume compared to the group that showed a decline in language function. Right PTR volume was significantly correlated with change in language scores. Third, there was a trend for non-right-handers to be over-represented in the group that showed a decline in language function. Each of these findings will be discussed below.

As predicted, two language-change subgroups emerged within our ASD sample. We had expected to find one group with improved language skill and another group with no language improvement. Instead, we found that some of the boys with ASD (n=11) had a relative decline in language function over time, while others (n=7) had no relative change in language abilities during this same timeframe. There are two important factors to consider. First, many of our study participants had very impaired language scores at baseline; only one subject had an overall language score in the average range when examined at Time 1. To our knowledge, there are few studies investigating language anatomy in individuals with ASD that have included children evaluated at such a young age and most have not included individuals with such impaired language. Secondly, there were differences in how language changes were measured in this study. The majority of prior studies have not formally examined changes in language ability within-subjects, using standardized.

tests. In the current study we used standardized language tests and examined changes in the difference between test age equivalent and chronological age. This approach gives a more accurate indication of whether the child’s language is becoming closer to typically developing peers or whether he is falling further behind. Empiric studies have shown that early language ability is an important predictor of the developmental course of language [36], so a group with improved language may have emerged, if we had included more individuals with higher initial language abilities. Results from the current study suggest that young ASD children with a significant language delay are unlikely to show significant language improvements over time. We found a subgroup whose language delay remained relatively stable over time, and a group whose language fell even further behind their peers.

Results, however, may be different in a group who received earlier diagnoses and earlier and more interventions. For most subjects, visit 1 was their initial ASD diagnosis and the average age of visit 1 was 5.26 years. It is important to note, however, that in our sample, language changes did not seem to be accounted for by differences in services received or behavioral measures, such as age, time between visits, autism severity, prior language ability, or current nonverbal IQ. This is consistent with a prior study showing that the amount of intervention (number of hours) was not related to outcome, as measured by autism symptom severity, speech level, and adaptive behavior [37]. These results suggest that other factors, such as genetic/environment and/ or brain anatomy and organization, may be important in determining developmental course. The significant correlation between language changes and change in ADOS scores, with a greater decline in language function associated with an increase in autism severity over time supports the notion that language ability plays an important role in the developmental course of ASD. Longitudinal studies examining language trajectory in ASD are warranted to examine these observed effects in larger samples with a range of language functions.

specific regions implicated, however, differed from our prediction. The PT, across both hemispheres, was larger in the no change relative to the decline group. The PT consists of auditory association cortex and is involved in higher order auditory speech and language processing. It is, therefore, important in receptive language abilities; our measure of changes in language included a receptive component. Some prior studies have found differences in PT anatomy in ASD with most showing decreased leftward PT asymmetry [15]. Our current results, however, indicated that smaller overall PT volume was associated with more decline in language abilities. The PT volume of the no change group in the current study was very similar to the volume we found in our prior study [15] in a group of typically developing children (7-11 years) using similar methods (mean left PT = 2.21, mean right PT = 1.99), suggesting that the group differences may be due to a reduced PT volume in the language decline group, rather than enlarged volumes in the no change group. This finding may relate to Leonard et al.’s [38] work examining an anatomical risk index in children with developmental dyslexia and specific language impairment, in which they found that individuals with smaller, more symmetrical brain structures had impairments in language functions, including comprehension, while children with larger, more asymmetric structures had fewer deficits, with relatively spared receptive language.

The right PTR volume was larger in the no change group relative to the group with language decline over time. In addition, right PTR volume was correlated with the change in language score, with a larger right PTR associated with reduced decline in language function. The right PTR volume of the decline group was similar to right PTR volume of typically developing children (7-11 years) from our previous study (mean = .86) [15]. This finding suggests that group differences may be due to enlarged right PTR size in the no change group, as opposed to reduced volume in the decline group. It is interesting that we found this association with the right structure, as we had predicted that those with more improvements in language would have larger left language volume, given the left hemisphere’s predominant role in language and prior anatomical studies of associations with language ability. A larger structure in the right hemisphere, however, may relate to the ability of some individuals to compensate and may be related to previous work demonstrating reduced left lateralization of language in ASD [18]. In a prior study, we found more bilateral activation during a semantic task, specifically in frontal language regions in ASD, suggesting recruitment of right hemisphere structures to compensate and accurately complete the task [13]. Our current finding, implicating a larger right structure with less decline in language skills, is also consistent with several prior findings showing associations between right language areas and language abilities. Joseph et al. [16] found increased rightward POP asymmetry and bilaterally increased FA and decreased Regional Diffusivity (RD) of the arcuate fasciculus were associated with increased language level in young children with ASD. This finding suggests that how the right hemisphere develops may influence how language develops in children with ASD. An fMRI study examined 2-3 year olds with ASD during a speech perception task [14]. Laterality analysis showed greater right hemispheric recruitment in the ASD group compared to age-matched controls, who had greater left hemisphere activation. In ASD, there was a positive correlation between right fronto-temporal activation and receptive language skill and autism severity negatively correlated with left fronto-temporal activation. These results indicate that atypical rightward asymmetry of frontal language regions may be beneficial to language development in some individuals with ASD by allowing for compensation. This suggests that right frontal language zones may be a good target to improve language functions in ASD, however, further studies are needed.

Increased rates of left- or mixed-handedness have been reported in ASD [39-43] and consistent with this finding we found 44% of the boys in our sample were classified as non-right-handers. We also found a trend for non-right-handers to be over-represented in the group that showed a decline in language function over time compared to the group that showed no change. Although speculative, these results suggest that in ASD, a poorer language prognosis may be associated with left- or mixed-hand preference. Handedness and language are strongly associated, and our findings are consistent with a study in typically developing children which showed that righthanders performed significantly better than left-handers on language tests [44]. It is also in agreement with ASD studies. One study showed increased mixed-handedness in a group of ASD children with early disordered language compared to typically developing controls and to children with ASD without early disordered language [45]. In another study involving a large sample of ASD children, we found that, within the ASD group, receptive and expressive language scores were higher in right-handers compared to non-right-handers [43]. Further study of the associations of hand preference, praxis representations, and motor control in relation to language development would help to further our understanding of these functional networks.

There were several limitations to this study. First, although we had longitudinal behavioral data, we were not able to collect longitudinal neuroimaging data. It would be interesting to examine brain anatomy at a much younger age, and determine if changes in anatomy with age are predictive of language development, hand preference, and autism severity. It would also be interesting to study the anatomy of other language-related regions, such as the caudate, to determine if they are related to language development. Another limitation is the relatively small sample size and the limitation of the sample to boys. Finally, most of the subjects examined in this study had very impaired language function, both at Time 1 and at Time 2. Future studies should investigate these variables in larger samples including individuals with a greater range of language ability.

Conclusions

Two groups emerged based on changes in language ability over time; a group with relative decline in language and a group with no changes over time. The group with no language change had larger PT volume and right PTR volume compared to the decline group. In addition, larger right PTR volume was associated with less decline in language skills. The two language groups did not differ on current or previous age, time between visits, current or prior ADOS social or communication scores, or Time 1 language measures. There was a trend for non-right-handers to be over-represented in the group with language decline. Results suggest differences in language cortex anatomy may play an important role in language development in ASD, with further studies needed. In addition, results provide information on regions which may be important for future studies to target, such as the right PTR, in interventions to improve language abilities in individuals with ASD.

References

1. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association Press Washington, DC. 2013.

2. Tager-Flusberg H, Paul R, Lord CE, Volkmar F, Klin A. Language and communication in autism. Handbook of Autism and Pervasive Developmental Disorder, vol 3rd Wiley New York, pp. 2005; 335-364.

3. Venter A, Lord C, Schopler E. A follow-up study of high-functioning autistic children. J of Child Psychol Psychiatry. 1992; 33: 489-507.

4. Stone WL, Yoder PJ. Predicting spoken language level in children with autism spectrum disorders. Autism. 2001; 5: 341-361.

5 . McDuffie A, Yoder P, Stone W. Prelinguistic predictors of vocabulary in young children with autism spectrum disorders. J Speech, Lang Hear Research. 2005; 48: 1080-1097.

6. Smith V, Mirenda P, Zaidman-Zait A. Predictors of expressive vocabulary growth in children with autism. J Speech, Lang Hear Res. 2007; 50: 149-160.

7. Sigman M, Spence SJ, Wang AT. Autism from developmental and neuropsychological perspectives. Annu Rev Clin Psychol. 2006; 2: 327-355.

8. Boddaert N, Belin P, Chabane N, Poline JB, Barthélémy C, Mouren-simeoni MC, et al. Perception of complex sounds: Abnormal pattern of cortical activation in autism. Am J Psychiatry. 2003; 160: 2057-2060.

9. Boddaert N, Chabane N, Belin P, Bourgeois M, Royer V, Barthelemy C. et al. Perception of complex sounds in autism: Abnormal auditory cortical processing in children. Am J Psychiatry. 2004; 161: 2117-2120.

10. Gervais H, Belin P, Boddaert N, Leboyer M, Coez A, Marion et al. Abnormal cortical voice processing in autism. Nature Neuroscience. 2004; 7: 801-802.

11. Müller RA, Chugani DC, Behen ME, Rothermel RD, Chakraborty PK, Muzik O, et al. Impairment of dentato-thalamo-cortical pathway in autistic men: Language activation data from positron emission tomography. Neurosci Lett. 1998; 245: 1-4.

12. Müller RA, Behen ME, Rothermel RD, Chugani DC, Muzik O, Mangner TJ, et al. Brain mapping of language and auditory perception in high-functioning autistic adults: A PET study. J Autism Dev Disord. 1999; 29: 19-31.

13. Knaus TA, Silver AM, Lindgren KA, Hadjikhani N, Tager-Flusberg H. fMRI activation during a language task in adolescents with ASD. J Int Neuropsychol Soc. 2008; 14: 967-979.

14. Redcay E, Courchesne E. Deviant FMRI patterns of brain activity to speech in 2-3-year-old children with autism spectrum disorder. Biol Psychiatry. 2008; 64: 589-598.

15. Knaus TA, Silver AM, Dominick KC, Schuring MD, Shaffer N, Lindgren KA, et al. Age-related changes in the anatomy of language regions in autism spectrum disorder. Brain Imaging Beha. 2009; 3: 51-63.

16. Joseph RM, Fricker Z, Fenoglio A, Lindgren KA, Knaus TA, Tager-Flusberg H, et al. Structural asymmetries of language-related gray and white matter and their relationship to language function in young children with ASD. Brain Imaging Behav. 2014; 8: 60-72.

17.De Fossé L, Hodge SM, Makris N, Kennedy DN, Caviness VS, Mc Grath L, et al. Language-association cortex asymmetry in autism and specific language impairment. Ann Neurol. 2004; 56 : 757-766.

18. Knaus TA, Silver AM, Kennedy M, Lindgren KA, Dominick KC, Sieqel, et al. Language laterality in autism spectrum disorder and typical controls: A functional, volumetric, and diffusion tensor MRI study. BrainLang . 2010; 112: 113-120.

19. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association Press Washington, 1994.

20. Rutter M, Le Couteur A, Lord C, Autism Diagnostic Interview - Revised. Western Psychological Services Los Angeles, CA, 2003.

21.Lord C, Rutter M, DiLavore PC, Risi S. Autism Diagnostic Observation Schedule. Western Psychological Services Los Angeles, CA, 1999.

22. Mullen EM. Mullen Scales of Early Learning. AGS Publishing Bloomington, MN. 1995.

23. Wechsler D. Wechsler Preschool and Primary Scale of Intelligence. The Psychological Corporation New York. 2002.

24. Wechsler D. Wechsler Intelligence Scale for Children. The Psychological Corporation Toronto, ON. 2003.

25. Roid GH, Miller LJ. Leiter International Performance Scale-Revised. Western Psychological Services Torrance, CA. 1997.

26. Carrow-Woolfolk E. Oral and Written Language Scales( OWLS). American Guidance Services, Inc. Circle Pines, MN. 1995.

27. Almli CR. Handedness Test (Preschool Version). 2006.

28. Fedorov A, Beichel R, Kalpathy-Cramer J, Finet J, Fillion-Robin JC, Pujol S, et al. 3D Slicer as an image computing platform for the Quantitative Imaging Network. Magn Reson Imaging. 2012; 30: 1323-1341.

29. Knaus TA, Bollich AM, Corey DM, Lemen LC, Foundas AL. Sex-linked differences in the anatomy of perisylvian language cortex: A volumetric MRI study of gray matter volumes. Neuropsychology. 2004; 18: 738-747.

30. Knaus TA, Bollich AM, Corey DM, Lemen LC, Foundas AL. Variability in perisylvian brain anatomy in healthy adults. Brain Lang. 2006; 97: 219-232.

31. Witelson SF, Kigar DL. Sylvian fissure morphology and asymmetry in men and women: Bilateral differences in relation to handedness in men. J Comp Neurol. 1992; 323 : 326-340.

32. Parker GJ, Luzzi S, Alexander DC, Wheeler-Kingshott CA, Ciccarelli O, Lambon Ralph MA. et al. Lateralization of ventral and dorsal auditory- language pathways in the human brain. NeuroImage. 2005; 24: 656-666.

33. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, et al. Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage. 2004; 23: 208-219.

34. Behrens TE, Woolrich MW, Jenkinson M, Johansen-Berg H, Nunes RG, Clare S. et al. Characterization and propagation of uncertainty in diffusion- weighted MR imaging. Magn Reson Med. 2003; 50: 1077-1088.

35. Smith SM. Fast robust automated brain extraction. Human Brain Mapp. 2002; 17: 143-155.

36. Tek S, Mesite L, Fein D, Naigles L. Longitudinal analyses of expressive language development reveal two distinct language profiles among young children with autism spectrum disorders. J Autism Dev Disord. 2014; 44: 75- 89.

37. Darrou C, Pry R, Pernon E, Michelon C, Aussilloux C, Baghdadli A. et al. Outcome of young children with autism: Does the amount of intervention influence developmental trajectories? Autism. 2010; 14: 663-677.

38. Leonard C, Eckert M, Given B, Virginia B, Eden G. Individual differences in anatomy predict reading and oral language impairments in children. Brain.2006; 129: 3329-3342.

39. Cornish KM, McManus IC. Hand preference and hand skill in children with autism. J Autism Dev Disord. 1996; 26: 597-609.

40. Dane S, Balci N. Handedness, eyedness and nasal cycle in children with autism. Int J Dev Neuroscie. 2007; 25: 223-226.

41. Hauck JA, Dewey D. Hand preference and motor functioning in children with autism. J Autism Dev Disord. 2001; 31: 265-277.

42. Soper HV, Satz P, Orsini DL, Henry RR, Zvi JC, Schulman M. et al. Handedness patterns in autism suggest subtypes. J Autism Dev Disord. 1986; 16: 155-167.

43. Knaus TA, Kamps J, Foundas AL. Handedness in Children with Autism Spectrum Disorder. Percept Mot Skills. 2016; 122: 542-559.

44. Natsopoulos D, Kiosseoglou G, Xeromeritou A, Alevriadou A. Do the hands talk on mind’s behalf? Differences in language ability between left- and right- handed children. Brain Lang. 1998; 64: 182-214.

45. Escalante-Mead PR, Minshew NJ, Sweeney JA. Abnormal brain lateralization in high-functioning autism. J Autism Dev Disord. 2003; 33: 539-543.

Citation

Knaus TA, Kamps J and Foundas AL. Longitudinal Language Changes Associated with MRI Anatomy in Children with Autism Spectrum Disorder. SM J Neurol Neurosci. 2016; 2(1): 1004.