Altered functional connectivity of semantic processing in youths with autism

Written by  Monday, 23 July 2018 14:37

Abstract

     Individuals with autism spectrum disorders (ASD) have aberrant neural activity during semantic judgments. I aimed to examine age-dependent neural correlates of semantic processing in boys with ASD as compared to those in typically developing boys (TD). I used functional MRI to investigate 37 boys with ASD (mean age = 13.3 years, standard deviation = 2.4) and 35 age-, sex-, intelligence quotient (IQ)- and handedness-matched TD boys (mean age = 13.3 years, standard deviation = 2.7) from age 8 to 18 years. Participants had to indicate whether pairs of Chinese characters presented visually were related in meaning. Group (ASD, TD) x Age (Old, Young) ANOVA was used to examine the difference of age-related changes. Direct comparisons between the adolescent group and the child group were performed. Functional connectivity was also used to estimate the directional influence among brain regions for participants. The behavioral results showed that the ASD group had lower accuracy in the related condition relative to the TD group. The neuroimaging results showed greater activation in the cuneus and less activation in the left inferior frontal gyrus (IFG) in boys with ASD than TD boys. Children with ASD produced greater activation in the cuneus than TD children. Adolescents with ASD showed reduced left IFG activation as compared to TD adolescents. Moreover, stronger connectivity was found between the cuneus and middle temporal gyrus (MTG) for ASD children, while stronger connectivity was found between the IFG and MTG for TD adolescents. The findings suggest that TD boys may engage more in higher-level processing of retrieving or selecting semantic features while boys with ASD may rely more on lower-level visual processing during semantic judgments.

Keywords: Autism spectrum disorders (ASD), fMRI, semantic processing, age-dependent, connectivity
 

Introduction

     Autism spectrum disorder (ASD) is a pervasive, multi-factorial, highly heritable, clinically heterogeneous neuro-developmental disorder with prominent reciprocal social and communication impairment and restricted repetitive behavior or interest. Infantile autism and autistic psychopathology were first described by Kanner in 1943 and Asperger in 1944, respectively. The autistic symptoms often occur and are diagnosed in early childhood and continue throughout life [Chakrabarti & Fombonne, 2005]. In addition, children with ASD may have emotional or behavioral symptoms such as thought problems [Bolte et al., 1999], social problems [Constantino et al., 2003], attention/hyperactivity problems [Bolte et al., 1999], disruptive behaviors [Scattone et al., 2002], and maladaptive behaviors [Luteijn et al., 2000].

     “Dysconnectivity” can be an integrative psychological-neurobiological model to study ASD. Theoretically, several psychopathological models explain part of the whole behavioral and cognitive features of ASD. These models include deficit in theory of mind [Baron-Cohen et al., 1985], weak central coherence [Frith & Happe, 1994], executive dysfunction [Ozonoff et al., 1994], and hypo-empathizing with hyper-systemizing [Baron-Cohen, 2002]. However, none of the above theories accounts for the whole picture of ASD, leaving spaces for a more fundamental psychological and neurobiological explanation.

     A few studies have investigated the neuropsychological substrate of meaning processing through the use of neuroimaging techniques. In this way, the relationship between specific brain function and behaviors can be explored. Within this context, “dysconnectivity” within and among neural networks might be a basic feature in individuals with ASD [Geschwind & Levitt, 2007].

     Concerning brain functions related to language/communication in ASD, previous research provided empirical evidence of the dysfunctions of semantic processing in adults with ASD during performing sentence comprehension [Kana et al., 2006] and semantic matching tasks [Harris et al., 2006; Kana et al., 2006]. Recent studies on ASD have shown aberrant neural activity during semantic processing, for example, reduced activation in the left inferior frontal gyrus (IFG) in adults with ASD as compared to healthy controls [Harris et al., 2006; Just et al., 2004]. In contrast, increased activation in the left IFG was found for semantic functions in adolescents with ASD compared to typically developing (TD) adolescents [Knaus, 2008]. Moreover, individuals with ASD often have a perception-based preference with recruitment of primary sensory cortices [Kana et al., 2006; Pina-Camacho et al., 2012]. A meta-analysis study concluded that individuals with ASD are more likely to use the visual functions to support word processing [Samson, 2012]. Lai et al. (2014) summarized that in contrast to TD individuals who focused on top-down control processing with recruitment of the frontal cortices to deal with language/communication, especially semantic processing, individuals with autism usually engaged more in bottom-up process because they possessed enhanced sensory-perceptual processing for local features of the stimuli. At the neural level, individuals with autism showed enhanced activity of primary sensory cortices (i.e. primary visual cortex) and reduced activation in the frontal area which is involved in a top-down process [Lai et al., 2014]. Studies on semantic processing also demonstrated reduced activation in the frontal area and increased activation in the occipital region in the ASD group [Kana et al., 2006; Pina-Camacho et al., 2012]. Many studies of healthy adults have found brain activation in the ventral region of the left IFG that is positively related to the performance of selecting appropriate semantic features [Fletcher et al., 2000], or to the increased retrieval demands [Chou et al., 2006a; Chou et al., 2009; Wagner et al., 2001]. A study examined the effective connectivity between brain regions involved in Chinese semantic judgments also suggested that there is a modulatory effect of retrieving semantic representation from the left IFG to the left middle temporal gyrus [Fan, 2010]. These findings imply that unlike healthy adults, adults with ASD might recruit an immature, perception-based neural mechanism to support semantic processing instead of engaging in a frontal control mechanism [Kana et al., 2006; Pina-Camacho et al., 2012; Samson, 2012].

     For neural mechanisms of development, Brown et al. (2005) used cross-sectional and regression analyses to examine the development of brain function for word generation in TD children. They found a positive correlation between age and neural activity in the left frontal cortex but a negative correlation between age and brain activation in bilateral occipital cortices [Brown et al., 2005]. Moreover, the left IFG showed increased activation from TD children to TD adolescents during the semantic judgment task, indicating TD adolescents might be better able to retrieve/select semantic features than TD children [Lee et al., 2011; Lee et al., 2015]. Groen et al. (2010) examined auditory sentence comprehension of pragmatic information in adolescents with ASD. Compared to healthy controls, adolescents with ASD demonstrated reduced activation in the left frontal region for integrating social information with the speakers [Groen, 2010]. In addition, greater activation in the left extrastriate region was involved in auditory language comprehension in the ASD group [Groen, 2010]. To examine the developmental changes of neural activity during semantic categorization, Dunn and Bates (2005) used event-related potential to study children and adolescents with ASD [Dunn, 2005]. The ASD children showed a greater N4 effect on a wrong categorization as compared to TD children. In contrast, the ASD adolescents showed an attenuated N4 response as compared to TD adolescents. The authors suggested that overall changes in the amplitude of N4 may reflect changes in semantic expectancy. The attenuated N4 response in adolescents with ASD might be due to the fact that the adolescents have more experience with language use along with increasing age [Dunn, 2005]. Another study investigated the developmental difference in language processing by comparing adults with ASD and children with ASD [Williams et al., 2013]. During the processing of ironic texts, adults with ASD showed increasing activations in bilateral middle temporal gyri. However, children with ASD showed greater activations in triangularis of the left inferior gyrus and left medial frontal gyrus. These results suggested that there might be distinctive neural mechanisms accompanying developmental changes within the ASD group [Williams et al., 2013].

     In summary, ASD individuals tend to use visual areas to support semantic processing while TD participants usually recruit frontal areas. To explore such a visual versus frontal involvement in ASD, a semantic task is used in the present study. Previous studies demonstrated that there is a more direct mapping between orthography and semantics in Chinese using the same semantic task for healthy adults [Booth et al., 2006] and TD children [Lee et al., 2015]. In addition, this semantic task has been used to explore semantic development in terms of age-related selection or retrieval of semantic knowledge for TD children [Booth et al., 2006; Lee et al., 2011; Lee et al., 2015]. We therefore used the semantic task to examine the developmental difference between the ASD and the TD groups. For the age range, previous developmental studies on TD youths show that adolescents aged 13-17 might be better able to select and retrieve proper semantic representations when performing Chinese semantic judgment task than children aged 8-12 [Lee et al., 2011; Lee et al., 2015]. The evidence of age-dependent maturation in semantic processing in TD participants provides us the basis to further investigate the semantic processing in the ASD group and the developmental difference between the ASD and the TD groups.

     Given the lack of information about the developmental changes in neural activity in youths with ASD and comparisons between children with ASD and adolescents with ASD during semantic processing (as revealed by using functional magnetic resonance imaging), the primary goal of the current study is to fill this gap. Previous studies in healthy children showed developmental changes in maintaining multiple dimensions of concepts [Blakemore, 2006]. Adolescents aged 13-17 might be better able to perform semantic tasks than children aged 8-12 [Lee et al., 2011]. As language ability and deficits need to be assessed in the context of developmental stage [Williams et al., 2013], we aimed to explore the difference of age-dependent neural activity in related brain regions (i.e., IFG and occipital cortex) during semantic processing to understand the nature of impaired semantic processing of words in ASD in a developmental aspect. Moreover, our results in youths with ASD can also be compared to those reports based on adult ASD studies [Kana et al., 2006; Pina-Camacho et al., 2012].

     This study had several goals. First, I investigated the patterns of brain activation in the ASD and TD groups separately during semantic judgments. Second, I compared the neural activity of semantic processing between the ASD and TD groups. Third, I performed a Group (ASD, TD) x Age (Old, Young) ANOVA to examine whether there was a difference in age-related changes. Fourth, I investigated the location and extent of functional dysconnectivity and their changes in the ASD group as compared to the TD group during semantic processing. I further examined the developmental differences between the ASD and TD groups, further stratifying by the child and adolescent groups. My predictions are that compared to TD youths, youths with ASD may have an immature, perception-based neural mechanism during semantic processing instead of a frontal control mechanism. Moreover, the group x age interaction at the neural level may be revealed in both the extrastriate visual cortex and inferior frontal region.

 

Methods

Participants

     The sample consisted of 37 children and adolescents, who were clinically diagnosed with ASD according to the Diagnostic and Statistical Manual of Mental Disorder, 4th edition (DSM-IV) diagnostic criteria (mean age = 13.3 years, standard deviation, SD = 2.4) and 35 TD children and adolescents, who were recruited according to the distribution of age, sex, IQ and handedness of the ASD group (mean age = 13.3 years, SD = 2.7). Participants’ ages ranged from 8 to 18, all males (Table 1).

     Previous research showed that adolescents aged 13-17 might be better able to perform semantic tasks than children aged 8-12 [Lee et al., 2011]. Since developmental changes played a role in maintaining multiple dimensions of concepts [Blakemore, 2006], I further divided the participants into the adolescent (22 ASD and 21 TD, mean age = 15.07 years, age range = 13-18) and child groups (15 ASD and 14 TD, mean age = 10.67 years, age range = 8-12. All demographic variables were matched between the ASD and TD groups for both children and adolescents (Table 1). Participants and their parents were interviewed to ensure that the participants met the following inclusion criteria: (1) native Mandarin-Chinese speakers, (2) normal hearing and normal or corrected-to-normal vision, and (3) no clinical diagnosis of learning disability, attention-deficit/hyperactivity disorder, mood disorders, schizophrenia, schizoaffective disorder, or organic psychosis.

     The ASD group was recruited from the Department of Psychiatry, National Taiwan University Hospital and diagnosed by full-time board-certificated child psychiatrists of this hospital according to the DSM-IV and International statistical classification of disease and related health problems-10 (ICD) diagnostic criteria. The clinical diagnosis was further confirmed by interviewing the parents using the Chinese version of the Autism Diagnostic Interview-Revised (ADI-R) (mean score of verbal communication = 13.4, SD = 4.5) by Professor Gau [Gau et al., 2011; Rutter et al., 2003]. The TD group was recruited from similar school districts of the ASD group by teachers’ referral rather than advertisement. All the participants were clinically assessed and their parents were interviewed by using the Chinese Kiddie epidemiologic version of the Schedule for Affective Disorders and Schizophrenia (K-SADS-E) [Gau et al., 2005] by Professor Gau to ensure that they were free from any neuropsychiatric disorders, not taking medication affecting the central nervous system, no history of attention, reading, or verbal-language deficits, and no learning disability. Full-scale intelligence quotient (IQ) and sub-scales scores were assessed by using the Wechsler Intelligence Scale for Children, 3rd edition (WISC–III) [Wechsler, 1991] for participants aged 16 and younger, and the Wechsler Adult Intelligence Scale (WAIS-IV) [Wechsler, 2008] for participants older than 16. All participants were right handed according to Edinburgh Handedness Inventory [Oldfield, 1971]. This study was approved by the Ethics Research Committee at the National Taiwan University Hospital before implementation. All the participants and their parents provided written informed consent.

 

Functional Activation Tasks

     Two visual Chinese characters were presented sequentially on the screen, and the participant had to determine whether the character pair was related in meaning. Trials lasted 4500 milliseconds (ms) and consisted of a solid square (500ms), followed by the first character (800ms), a 200ms blank interval, then the second character (3000ms) (Figure 1). The participant was instructed to make a response quickly and accurately during the presentation of the second character. 48 character pairs were semantically related and 24 character pairs were semantically unrelated. The number of syllables, visual complexity, frequency, word class, and semantic relation were controlled across the related and unrelated conditions [Lo et al., 2013].

     There were 24 pairs of non-characters in the perceptual control condition created by replacing radicals of real characters with other radicals that did not form Chinese characters. For the perceptual controls, trials consisted of a solid square (500ms), followed by the first non-character (800ms), a 200ms blank interval, and the second non-character (3000ms). The participants determined whether the pair of stimuli was identical or not. There were also 24 baseline events in which the participant was instructed to press a button when a solid square (1300ms) at the center of the visual field turned into a hollow square (3000ms) after a blank interval (200ms). The related condition was used to investigate semantic processing [Chou et al., 2009] and the perceptual control condition was chosen as baseline to control for non-language components such as response demands [Chou et al., 2006a].

 

MRI Data Acquisition

     Participants lay in the scanner with their head position secured. An optical response box was placed in the participants’ right hands. Participants viewed visual stimuli via a mirror inside the scanner. This study adopted an event-related design. Each participant performed two functional runs, which took 4.7 minutes. All images were acquired using a 3-Tesla Siemens Trio scanner. Gradient-echo localizer images were acquired to determine the placement of the functional slices. Functional images were acquired with echo planar imaging method to detect the blood oxygenation level-dependent (BOLD) signal. The scanning parameters were the following: repetition time (TR) = 2000 ms; echo time (TE) = 24ms; flip angle = 90º; matrix size = 64 × 64; field of view = 25.6cm; slice thickness = 3mm; number of slices = 34. We acquired 272 images. A high-resolution, T1-weighted three dimensional image was also acquired (TR = 1560ms; TE = 3.68ms; flip angle = 15º; matrix size = 256 × 256; field of view = 25.6cm; slice thickness = 1mm). The task was administered in a pseudorandom order for all participants, in which the order of related, unrelated, perceptual, and baseline trials was optimized for event-related design [Burock et al., 1998; Lo et al., 2013].

 

Image Data Analysis

     Data analysis was performed using Statistical Parametric Mapping (SPM 8). The functional images were corrected for the differences in slice-acquisition time to the middle volume and were realigned to the first volume in the scanning session using affine transformations. No participant had more than 3mm of movement in any plane. Co-registered images were normalized to the Montreal Neurological Institute (MNI) average template. Statistical analyses were calculated on the smoothed data (10mm isotropic Gaussian kernel), with a high pass filter (128 seconds cutoff period) to remove low frequency artifacts.

     Data from each participant was entered into a general linear model using an event-related analysis procedure. Character pairs were treated as individual events for analysis and modeled using a canonical Hemodynamic Response Function (HRF). There were four event types: related, unrelated, perceptual, and baseline. Parameter estimates from contrasts of the canonical HRF in single subject models were entered into random-effects analysis using one-sample t-tests across all participants to determine whether activation during a contrast was significant (i.e., parameter estimates were reliably greater than 0). I used the contrast of the related condition versus perceptual control condition for the following analyses because this contrast allowed us to observe semantic and orthographic processes [Liu et al., 2010]. The incorrect trials were included, considering the statistical power should be equal between conditions with different accuracies for fMRI analyses [Bitan et al., 2007].

     Both within-group and between-group contrasts were based on a whole brain analysis. Reported areas of activation marked in bold indicate the significance using p < .05 familywise error (FWE) corrected. To examine whether there was a difference of age-related changes, we performed a Group (ASD, TD) x Age (Old, Young) ANOVA. Reported areas of activation marked in bold indicate the significance using p < .05 FWE corrected with the use of two anatomical regions of interest (ROIs, including the left cuneus and the left IFG) based on the results of the between-group analysis, using the WFU Pickatlas [Maldjian, 2003]. I further conducted bi-directional comparisons between the ASD and the TD groups for the adolescent group and for the child group in order to understand the differences in brain activation between children with ASD and TD children, as well as between adolescents with and without ASD.  Finally, several correlation analyses were performed. First, to explore the relation between brain activation and task performance, I extracted the signal intensity (beta-values) in the left cuneus for the ASD group and in the left IFG for the TD group, and then computed correlations with task performance (accuracy and response times) in each group. Second, I correlated the scores of Perceptual Organization Index of WISC-III [Wechsler, 1991] or WAIS-IV [Wechsler, 2008] associated with visual perception with left cuneus activation in the ASD group to investigate the role of cuneus in primary visual processing. Third, to examine the age-dependent ability to perform the task, we correlated age with behavioral performance for the ASD group, for the TD group, and for all the participants pooled together for the related condition.

 

Connectivity analysis

     Functional connectivity analysis was carried out using Psychophysiological Interaction (PPI) analyses as implemented in SPM 8. In general, PPI tests the condition-specific connectivity without assuming a direction of influence. In the present study, I selected the seed regions in the left cuneus and the left IFG based on the results of the results of the between-group analysis. The size of a seed region was defined as a sphere with a radius of 6 mm. Then, the PPI analysis estimates functional connectivity between the signal of the time series of the seed voxels and those of all other voxels in the whole brain. Specifically, PPI explains variation in brain activity through the interplay of two factors: (a) the input of a selected seed region (the physiological variable, Y regressor) and (b) a cognitive process as defined by a contrast (the psychological variable, P regressor). PPI analyses create an interaction factor consisting of the product of a vector representing the time series of brain activation in the region of interest (i.e. the physiological variable) and a vector representing the condition contrast (i.e. the psychological variable).

 

Results

Behavioral Results

     The accuracy and response times of the meaning judgment task for the ASD and TD groups are presented in Table 2. Independent two sample t-tests revealed that the ASD group had significantly lower accuracy than the TD group in the related (t (70) = -2.32, p < .05), unrelated (t (70) = -5.20, p < .05), perceptual (t (70) = -3.03, p < .05), and baseline event (t (70) = -2.30, p < .05) conditions. For response times, independent two sample t-tests showed that the ASD group had significantly longer response time than the TD group in the unrelated (t (70) = 2.21, p < .05) and perceptual (t (70) = 2.47, p < .05) conditions.

 

Imaging Results

     Table 3 summarizes the results of brain activation during semantic processing based on the within-group and between-groups analyses.

Within-group contrasts

     For the within-group analysis, I found commonly significant activation in the left IFG (BA 47) in both the ASD (Figure 2a) and TD groups (Figure 2b). Greater activation was found in the cuneus (BA 18) in the ASD group (Figure 2a). Moreover, there was no significant correlation between the 6 motion parameters and brain activation in the cuneus under the threshold of p < .001 uncorrected.

Between-group contrasts

     For the between-groups analysis, the ASD group showed greater activation in the cuneus (Figure 2c) and reduced activation in the left IFG as compared to the TD group (Figure 2d).

Group x Age ANOVA

     I performed a Group (ASD, TD) x Age (Old, Young) ANOVA analysis for brain imaging. According to the interaction between group and age, significant activations were shown in the left cuneus and left IFG (Table 4, Figure 3a). I also examined the bi-directional comparisons between the ASD and the TD groups for the adolescent group and for the child group, respectively. The results showed that children with ASD had greater activation in the cuneus than TD children while no difference was found in the IFG activation between the child groups (Table 4, Figure 3b). In addition, adolescents with ASD showed reduced left IFG activation as compared to TD adolescents. There was no difference in cuneus activation for the two adolescent groups (Table 4, Figure 3b).

 

Correlation Analyses

Brain activation in ASD/ TD and task performance

     I extracted the signal intensity (beta-values) in the left cuneus for the ASD group and in the left IFG for the TD group, and then computed correlations with task performance (accuracy and response times) in each group. The results showed that in the ASD group left cuneus activation was positively correlated with the response times of related pairs, r = .37, p < .05. The ASD participants showed higher left cuneus activation with slower response times.

Brain activation in ASD and perceptual Organization Index

     I correlated the scores and the sub-scores of Perceptual Organization Index of WISC-III or WAIS-IV which is associated with visual perception, with the left cuneus activation in the ASD group. The left cuneus activation was positively correlated with the score of block design, r = .40, p < .05.

Age of ASD/ TD and task performance

     We correlated age with behavioral performance in the ASD group, in the TD group, and in all participants pooled together for the related condition. In the ASD group, age was positively correlated with accuracy, r = .54, and negatively correlated with response times, r = -.35. In the TD group, age was positively correlated with accuracy, r = .42. When all participants were pooled together, age was positively correlated with accuracy, r = .44, and negatively correlated with response times, r = -.33. All p-values < .05.

 

Connectivity Results

     First, the functional connectivity between the cuneus and middle temporal gyrus (MTG, BA 21) was significant for ASD children, but was not significant for TD children. Second, the functional connectivity between the IFG and MTG was significant for TD adolescents, but was not significant for ASD adolescents (Figure 4).

 

Discussion

     Functional brain imaging methods have emerged as a means to investigate the neural correlates of language processing. However, complex cognitive functions are subserved by large-scale distributed networks whereby individual components may act as nodal points for integrating and distributing information among other regions in a network. It is crucial to use functional connectivity to compare the interactions between brain regions to understand their language abilities for ASD vs. TD youths.

     The present study showed different functional connectivity between the ASD and TD youths, and the differential connectivity was age-dependent. For ASD children, greater connectivity was found between the cuneus and MTG, suggesting that younger ASD participants may use lower-level visual information to have access to the storage of semantic representations during meaning judgments. In contrast, TD children did not show this tendency to fulfill task requirements. For TD adolescents, greater connectivity was found between the IFG and MTG, suggesting that older TD participants may engage more in higher-level processing of retrieving or selecting semantic representations. In contrast, ASD adolescents did not show this tendency during meaning judgments.

     The finding of greater activation in the cuneus in the ASD group than the TD group is in line with the evidence of atypical reliance on the visual cortex in individuals with ASD [Kana et al., 2006; Pina-Camacho et al., 2012], suggesting that boys with ASD tend to have the perception-based preference and to use the bottom-up cognitive resources during semantic processing. Previous studies indicated that individuals with ASD often showed a perception-based preference with recruitment of primary sensory cortices during language processing [Lai et al., 2014] and tended to recruit the visual cortex more than the inferior frontal cortex (BA 47) for word processing [Samson, 2012]. Adults with ASD might rely on recruitment of visual areas in the occipital lobe for comprehending language [Kana et al., 2006]. In addition, Kim et al. (2011) investigated the cognitive components of semantic networks in a semantic decision task in healthy adults, identifying bilateral cuneus activation related to visual perception and retrieving stored mental imagery of word stimuli [Kim et al., 2011]. In the ASD group, the left cuneus activation was positively correlated with the response time of the semantic task and with the score of block design. The increased cuneus activation might be related to processing local and detailed visual information in ASD [Frith, 2003] that enhanced the performance of the block design task. Specifically, the block design task is designed for investigating locally oriented visual processing [Caron et al., 2006]. However, the aberrant engagement of local visual processing might interfere with the mapping from orthography to semantics in Chinese that required global visual processing during semantic judgments [Li et al., 2015]. The functional orthographic unit of recognizing Chinese characters is rooted in integral stroke patterns [Chen, 1996]. Thus, the interference with global visual information may result in slower response time during the semantic task in the ASD group.

     Previous studies provided evidence that adults with ASD tend to use visuo–spatial regions to compensate for the dysfunction of higher-order cortical regions [Koshino et al., 2005], that an atypical involvement of visual functions might be related to insufficient functions in the prefrontal cortex during semantic processing [Samson, 2012], and of the involvement of the left extrastriate region in auditory language comprehension in adolescents with ASD [Groen, 2010]. However, this study is the first to demonstrate similar results in a larger sample size recruiting both children and adolescents with ASD. The results of ANOVA analysis, the bi-directional group comparison, and connectivity analysis suggest that children with ASD tend to include primary visual cortex for semantic processing and to rely more on visual processing to perform semantic judgments. However, for adolescents with ASD, this atypical dependency on primary visual cortex disappeared, implying that youths with ASD may have a delayed and immature function during semantic processing compared to TD youths. Combining the adult ASD data [Koshino et al., 2005] and our youth data, individuals with ASD are more likely to use the early visual pathway to support semantic processing. Despite the dependency on visual processing in children but not in adolescents with ASD as found in this study, adults with ASD still demonstrate a greater reliance on visual processing during lexical tasks as compared to healthy adults [Kana et al., 2006; Pina-Camacho et al., 2012]. However, such an assumption needs to be validated by including adults with ASD or a longitudinal study design from childhood to adulthood in our population to test whether a pattern of decreasing brain activation in the cuneus with age is noted from childhood to adulthood and whether adults with ASD still have greater cuneus activation than TD adults.

     The TD group showed greater activation in the left IFG during semantic processing as compared to the ASD group. The finding of connectivity analysis also demonstrated that TD adolescents but not children showed more activation in the left IFG for the related pairs as compared to the ASD group. Instead of the recruitment of primary sensory cortices, TD individuals showed top-down retrieval and selection during semantic processing with recruitment of frontal cortices [Lai et al., 2014]. The IFG activation is thought to be related to the control of retrieving long-term semantic knowledge and increased demands on retrieving and selecting semantic knowledge in healthy adults [Chou et al., 2006a; Fan, 2010; Wagner et al., 2001]. Along with development, increasing modulation from the prefrontal cortex to posterior brain regions was found, resulting in the enhanced processing of task-relevant information [Chou et al., 2006b]. Thus, the finding of significant left IFG activation in TD adolescents indicated that the older TD participants may engage in higher level retrieval/selection of semantic knowledge to achieve the task requirement along the developmental stage from childhood to late adolescence [Chou et al., 2006b; Chou et al., 2009].

     Another new contribution of the current study is that I found distinct patterns of age-dependent neural correlates for semantic processing between the two groups. The Group (ASD, TD) x Age (Old, Young) ANOVA analysis showed significant activation for the interaction between group and age in the left cuneus and left IFG, indicating significant group differences in functional organizations of the semantic system. For the bi-directional comparisons, I examined whether there was a difference of age-related changes between children with ASD and TD children, as well as between adolescents with and without ASD. The ASD children produced greater activation in the cuneus than TD children. Adolescents with ASD showed reduced left IFG activation as compared to TD adolescents. Brown et al. (2005) found increased activation in the frontal cortex with age in TD children [Brown et al., 2005]. Moreover, TD adolescents showed increased activation in the left IFG when compared to TD children during the semantic judgment task, indicating that TD adolescents might be better able to retrieve/select semantic features along with the development [Lee et al., 2011; Lee et al., 2015]. Thus, the TD group may engage more in the control processing of retrieval/selection in adolescence, while the ASD group may have more perceptual-based preference with younger age during the youth period. The results imply that youths with ASD may possess a different functional mechanism of semantic processing from that of TD youths. Children with ASD may rely on the occipital lobe for semantic processing. However, from childhood to adolescence, the lower-level visual dependency for semantic processing disappeared without developing into a control process of retrieving/selecting in the ASD group, implying that there is still a difference along with age between the ASD and the TD groups. In addition to the diagnostic group difference in neural mechanisms for children and adolescents, I found that there was a correlation between age and behavioral performance for the related condition in the ASD and TD groups. Age was positively correlated with accuracy in both groups yet was negatively correlated with response times only in the ASD group. Moreover, age was positively correlated with accuracy and negatively correlated with response times across groups. The findings suggest that despite differential underlying neural mechanisms, improving performance of semantic judgments with age was noted in both ASD and TD youths.

     In conclusion, the results showed different neural correlates of semantic processing and the distinct age-dependent functional neural mechanism during semantic processing for youths with ASD and TD youths. The TD group showed a controlled process of retrieving/selecting semantic knowledge as compared to the ASD group and the use of higher-level neural source in adolescence. However, youths with ASD showed atypical semantic processing as compared to TD youths by demonstrating a differential neural mechanism, i.e., relying more on lower-level visual processing during semantic processing at a younger age. Moreover, ASD youths have functional dysconnectivity in semantic processing and demonstrate differential development trajectory of connectivity from TD youths. The group differences in neural mechanism imply that ASD youths have different functional organizations of the semantic system than TD youths do.

 

Reference

American Psychiatric Association. (2013). DSM-5: Diagnostic and statistical manual of mental disorders (4 ed.). Washington, DC: American Psychiatric Association.

Baron-Cohen, S. (2002). The extreme male brain theory of autism. Trends in Cognitive Sciences, 6(6), 248-254.

Baron-Cohen, S., Leslie, A.M., & Frith, U. (1985). Does the autistic child have a "theory of mind"? Cognition, 21(1), 37-46.

Bitan, T., Burman, D.D., Chou, T.L., Lu, D., Cone, N.E., Cao, F., et al. (2007). The interaction between orthographic and phonological information in children: An fMRI study. Human brain mapping, 28, 880-891.

Blakemore, S.J., & Choudhury, S. (2006). Development of the adolescent brain: implications for executive function and social cognition. Journal of child psychology and psychiatry, 47, 296-312.

Bolte, S., Dickhut, H., & Poustka, F. (1999). Patterns of parent-reported problems indicative in autism. Psychopathology, 32(2), 93-97.

Booth, J.R., Dong, L., Burman, D.D., Chou, T.-L., Jin, Z., Pengd, D.-L., et al. (2006). Specialization of phonological and semantic processing in Chinese word reading. Brain Research, 1071, 197-207.

Brown, T.T., Lugar, H.M., Coalson, R.S., Miezin, F.M., Petersen, S.E., & Schlaggar, B.L. (2005). Developmental changes in human cerebral functional organization for word generation. Cereb Cortex, 15, 275-290.

Burock, M.A., Buckner, R.L., Woldorff, M.G., Rosen, B.R., & Dale, A.M. (1998). Randomized event-related experimental designs allow for extremely rapid presentation rates using functional MRI. NeuroReport, 9, 3735-3739.

Caron, M.-J., Mottron, L., Berthiaume, C., & Dawson, M. (2006). Cognitive mechanisms, specificity and neural underpinnings of visuospatial peaks in autism. Brain, 129, 1789-1802.

Chen, Y. (1996). What are the functional orthographic units in Chinese word recognition: The stroke or the stroke pattern? The Quarterly Journal of Experimental Psychology: Section A, 49, 1024-1043.

Chou, T.L., Booth, J.R., Bitan, T., Burman, D.D., Bigio, J.D., Cone, N.E., et al. (2006a). Developmental and skill effects on the neural correlates of semantic processing to visually presented words. Human brain mapping, 27, 915-924.

Chou, T.L., Booth, J.R., Burman, D.D., Bitan, T., Bigio, J.D., Lu, D., et al. (2006b). Developmental changes in the neural correlates of semantic processing. NeuroImage, 29, 1141-1149.

Chou, T.L., Chen, C.W., Fan, L.Y., Chen, S.Y., & Booth, J.R. (2009). Testing for a cultural influence on reading for meaning in the developing brain: The neural basis of semantic processing in Chinese children. Frontiers in Human Neuroscience, 3, 1-9.

Chakrabarti, S., & Fombonne, E. (2005). Pervasive developmental disorders in preschool children: confirmation of high prevalence. American Journal of Psychiatry, 162(6), 1133-1141.

Constantino, J.N., Hudziak, J.J., & Todd, R.D. (2003). Deficits in reciprocal social behavior in male twins: evidence for a genetically independent domain of psychopathology. Journal of the American Academy of Child & Adolescent Psychiatry, 42(4), 458-467.

Dunn, M.A., & Bates, J. C. (2005). Developmental Change in Neutral Processing of Words by Children with Autism. Journal of autism and developmental disorders, 35, 361-376.

Fan, L.Y., Lee, S. H., & Chou, T. L. (2010). Interaction between brain regions during semantic processing in Chinese adults. Lang. Linguist, 11, 159-182.

Fletcher, P.C., Shallice, T., & Dolan, R.J. (2000). Sculpting the response space - an account of left prefrontal activation at encoding. NeuroImage, 12, 404-417.

Frith, C. (2003). What do imaging studies tell us about the neural basis of autism. Autism: Neural basis and treatment possibilities, 149-176.

Frith, U., & Happé, F. (1994). Autism: beyond "theory of mind". Cognition, 50(1), 115-132.

Gau, S.S.-F., Lee, C.-M., Lai, M.-C., Chiu, Y.-N., Huang, Y.-F., Kao, J.-D., et al. (2011). Psychometric properties of the Chinese version of the Social Communication Questionnaire. Research in Autism Spectrum Disorders, 5, 809-818.

Gau, S.S., Chong, M.Y., Chen, T.H., & Cheng, A.T. (2005). A 3-year panel study of mental disorders among adolescents in Taiwan. Am J Psychiatry, 162, 1344-1350.

Geschwind, D.H., & Levitt, P. (2007). Autism spectrum disorders: developmental disconnection syndromes. Current Opinion Neurobiology, 17(1), 103-111.

Groen, W.B., Tesink, C. M. J. Y., Petersson, K. M., Van Berkum, J., Van der Gaag, R. J., Hagoort, P., & Buitelaar, J. K. (2010). Semantic, factual, and social language comprehension in adolescents with autism: an FMRI study. Cereb Cortex, 20, 1937-1945.

Harris, G.J., Chabris, C.F., Clark, J., Urban, T., Aharon, I., Steele, S., et al. (2006). Brain activation during semantic processing in autism spectrum disorders via functional magnetic resonance imaging. Brain and Cognition, 61, 54-68.

Just, M.A., Cherkassy, V.L., Keller, T.A., & Minshew, N.J. (2004). Cortical activation and synchronization during sentence comprehension in high-functioning autism: Evidence of underconnectivity. Brain, 127, 1811-1821.

Kana, R.K., Keller, T.A., Cherkassky, V.L., Minshew, N.J., & Just, M.A. (2006). Sentence comprehension in autism: thinking in pictures with decreased functional connectivity. Brain, 129, 2484-2493.

Kim, K.K., Karunanayaka, P., Privitera, M.D., Holland, S.K., & Szaflarski, J.P. (2011). Semantic association investigated with functional MRI and independent component analysis. Epilepsy & Behavior, 20, 613-622.

Knaus, T.A., Silver, A. M., Lindgren, K. A., Hadjikhani, N., & Tager-Flusberg, H. (2008). fMRI activation during a language task in adolescents with ASD. fMRI activation during a language task in adolescents with ASD, 14, 967-979.

Koshino, H., Carpenter, P.A., Minshew, N.J., Cherkassky, V.L., Keller, T.A., & Just, M.A. (2005). Functional connectivity in an fMRI working memory task in high-functioning autism. NeuroImage, 24, 810-821.

Lai, M.C., Lombardo, M.V., & Baron-Cohen, S. (2014). Autism. Lancet, 383, 896-910.

Lee, S.-H., Booth, J.R., Chen, S.-Y., & Chou, T.-L. (2011). Developmental changes in the inferior frontal cortex for selecting semantic representations. Developmental Cognitive Neuroscience, 1, 338-350.

Lee, S.-H., Booth, J.R., & Chou, T.-L. (2015). Developmental changes in the neural influence of sublexical information on semantic processing. Neuropsychologia, 73, 25-34.

Li, M.-F., Lin, W.-C., Chou, T.-L., Yang, F.-L., & Wu, J.-T. (2015). The role of orthographic neighborhood size effects in Chinese word recognition. Journal of psycholinguistic research, 44, 219-236.

Liu, L., Vira, A., Friedman, E., Minas, J., Bolger, D., Bitan, T., et al. (2010). Children with reading disability show brain differences in effective connectivity for visual, but not auditory word comprehension. PLoS One, 5, e13492.

Lo, Y.-C., Chou, T.-L., Fan, L.-Y., Gau, S.S.-F., Chiu, Y.-N., & Tseng, W.-Y.I. (2013). Altered Structure–Function Relations of Semantic Processing in Youths with High-Functioning Autism: A Combined Diffusion and Functional MRI Study. Autism Research, 6, 561-570.

Luteijn, E., Luteijn, F., Jackson, S., Volkmar, F., & Minderaa, R. (2000). The children's Social Behavior Questionnaire for milder variants of PDD problems: evaluation of the psychometric characteristics. Journal of Autism and Developmental Disorders, 30(4), 317-330.

Maldjian, J.A., Laurienti, P. J., Kraft, R. A., & Burdette, J. H. (2003). An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. NeuroImage, 19, 1233-1239.

Oldfield, R.C. (1971). The Assessment and Analysis of Handedness: The Edinburgh Inventory. Neuropsychologia, 9, 97-113.

Ozonoff, S., Strayer, D.L., McMahon, W.M., & Filloux, F. (1994). Executive function abilities in autism and Tourette syndrome: an information processing approach. Journal of Child Psychology and Psychiatry, 35(6), 1015-1032.

Pina-Camacho, L., Villero, S., Fraguas, D., Boada, L., Janssen, J., Navas-Sánchez, F.J., et al. (2012). Autism spectrum disorder: does neuroimaging support the DSM-5 proposal for a symptom dyad? A systematic review of functional magnetic resonance imaging and diffusion tensor imaging studies. Journal of autism and developmental disorders, 42, 1326-1341.

Rutter, M., Le Couteur, A., & Lord, C. (2003). Autism diagnostic interview-revised. Los Angeles, CA: Western Psychological Services.

Samson, F., Mottron, L., Soulieres, I., & Zeffiro, T. A. (2012). Enhanced visual functioning in autism: An ALE meta‐analysis. Human brain mapping, 33, 1553-1581.

Scattone, D., Wilczynski, S.M., Edwards, R.P., & Rabian, B. (2002). Decreasing disruptive behaviors of children with autism using social stories. Journal of Autism and Developmental Disorders, 32(6), 535-543.

Tager-Flusberg, H., Paul, R., & Lord, C. (2005). Language and communication in autism. Handbook of autism and pervasive developmental disorders, 1, 335-364.

Wagner, A.D., Pare-Blagoev, E.J., Clark, J., & Poldrack, R.A. (2001). Recovering meaning: left prefrontal cortex guides controlled semantic retrieval. Neuron, 31, 329-338.

Wechsler, D. (1991). WISC-III Wechsler Intelligence Scale for Children–third edition: Manual. San Antonio, TX: The Psychological Corporation.

Wechsler, D. (2008). Wechsler adult intelligence scale–Fourth Edition (WAIS–IV). San Antonio, TX: NCS Pearson.

Williams, D.L., Cherkassky, V.L., Mason, R.A., Keller, T.A., Minshew, N.J., & Just, M.A. (2013). Brain function differences in language processing in children and adults with autism. Autism Res, 6, 288-302.

 

Table 1 Demographic characteristics of youths with autism spectrum disorder (ASD) and typically developing youths (TD).

The ASD and TD groups were matched on demographic variables. Independent sample t-test was used for test statistics. Handedness was assessed by using Edinburgh Handedness Inventory. Verbal intelligence quotient (VIQ), performance intelligence quotient (PIQ) and full intelligence quotient (FIQ) are measured by using Wechsler Intelligence Scale for Children, 3rd version. ADI-R: Autism Diagnostic Interview-Revised.

 

Table 2 Accuracy and response time of the meaning judgment task for the ASD and TD groups.

ASD: autism spectrum disorders; TD: typically developing. Independent sample t-test was used for test statistics. *P value < .05.

 

Table 3 Areas of activation for the related versus perceptual conditions for the ASD and TD groups.

ASD: autism spectrum disorders; TD: typically developing; H: hemisphere; L: left; R: right; BA: Brodmann’s area. Voxels: number of voxels in cluster at p < .005 uncorrected, only clusters greater than or equal to 10 in a whole brain analysis are presented. Regions surviving p < .05 FWE (familywise error) corrected are indicated in bold. Coordinates of activation peak(s) within a region based on a z test are given in the MNI stereotactic space (x, y, z).

 

Table 4 Group (ASD, TD) x Age (Old, Young) ANOVA analysis and subsequent comparisons for the adolescent group and for the child group.

ASD: autism spectrum disorders; TD: typically developing; H: hemisphere; L: left; R: right; BA: Brodmann’s area. Voxels: number of voxels in cluster at p < .005 uncorrected, only clusters greater than or equal to 10 in a whole brain analysis are presented. Regions surviving p < .05 FWE (familywise error) corrected with the regions of interest (ROIs, including left cuneus and left inferior frontal gyrus) based on the between-group analysis, using the WFU Pickatlas [Maldjian, 2003] are indicated in bold. Coordinates of activation peak(s) within a region based on a z test are given in the MNI stereotactic space (x, y, z).

 

figure1_300dpi.tif

 

Figure 1 The procedure of the meaning judgment task.

 

figure2_300dpi.tif

Figure 2 Results of within-group and between-group analyses of ASD group and TD group. Both within-group and between-group contrasts were based on a whole brain analysis. The colors indicated the brain activation of within-group analysis, blue for the ASD group (a) and green for the TD group (b), respectively. For between-group analysis, blue indicated greater activation in the cuneus (c) and green indicated reduced activation in left inferior frontal gyrus (IFG, BA 45) (d) in the ASD group, as compared to the TD group. All reported areas of activation were significant using p < .005 uncorrected with clusters greater than or equal to 10 voxels.

 

figure3_300dpi.tif

Figure 3 Results of the Group (ASD, TD) x Age (Old, Young) ANOVA analysis.

(a) The activation of interaction effect. (b) Children with ASD had greater cuneus activation than TD children (blue) while adolescents with ASD showed reduced left IFG activation as compared to TD adolescents (green). All reported areas of activation were regions surviving p < .05 FWE (familywise error) corrected with clusters greater than or equal to 10 voxels with the use of two anatomical regions of interest (the left cuneus and the left IFG).

 

圖片1.jpg

Figure 4 Stronger connectivity between the cuneus (V) and middle temporal gyrus (MTG) for ASD children, while stronger connectivity between the inferior frontal gyrus (IFG) and MTG for TD adolescents.

 

 

Read 57 times