Open Access
Add to BibSonomyAbstract
Autism spectrum disorder (ASD) is a neuro-developmental disorder, which has been associated with atypical neural synchronization. In this study, functional near infrared spectroscopy (fNIRS) was used to study the differences in functional connectivity in bilateral inferior frontal cortices (IFC) and bilateral temporal cortices (TC) between ASD and typically developing (TD) children between 8 and 11 years of age. As the first report of fNIRS study on the resting state functional connectivity (RSFC) in children with ASD, ten children with ASD and ten TD children were recruited in this study for 8 minute resting state measurement. Compared to TD children, children with ASD showed reduced interhemispheric connectivity in TC. Children with ASD also showed significantly lower local connectivity in bilateral temporal cortices. In contrast to TD children, children with ASD did not show typical patterns of symmetry in functional connectivity in temporal cortex. These results support the feasibility of using the fNIRS method to assess atypical functional connectivity of cortical responses of ASD and its potential application in diagnosis.
© 2014 Optical Society of America
1. Introduction
Optical methods have been demonstrated as a non-invasive way to probe human brain activities [1,2]. Functional near infrared spectroscopy (fNIRS), which measures the absorption of near infrared light (with a typical wavelength between 650nm and 950nm) through the scalp and skull [3–5], provide the information for concentration changes of oxy hemoglobin (HbO), deoxy hemoglobin (Hb) and total hemoglobin (HbT) in the superficial cortical regions of a human brain. fNIRS has high temporal resolution and reasonable spatial resolution. It also provides non-invasive and portable imaging environment which could be easily applied to clinical populations, children and neonates [6–8], especially on children with neurological disorder. So far, fNIRS has been applied to children with attention deficit hyperactivity disorder (ADHD), epilepsy, speech and language delay, preterm birth/early brain injury, ASD and infant at risk of ASD (for a review, see [9]), showing that fNIRS can be an effective tool to reveal the biological deficits underlying cognitive dysfunction of atypically developed brain and provide new perception for early diagnosis, assessment and treatment for the symptoms. fNIRS could also cover the whole brain to measure two or more regions of cerebral cortex simultaneously and thus provides an effective way to measure the functional connectivity of different brain cortices. Resting state functional connectivity (RSFC), presented as slow spontaneous oscillations (<0.1Hz, also known as low frequency fluctuation, LFF) during the resting, was found by functional magnetic resonance imaging (fMRI) in typical brains [10,11]. Atypical patterns of brain network and neuronal synchronization were shown to be pathological manifestations of many brain disorders, such as schizophrenia, ASD, Alzheimer's disease, Parkinson's and epilepsy (for a review, see [12]).
ASD is characterized by impaired social interactions, communication deficits and restricted, repetitive pattern of interests and behaviors [13]. Dysfunction and pathophysiology of ASD are linked to atypical brain development. There is growing evidence to support the important role of frontal lobe and temporal lobe in brain research of ASD. Anatomical studies have shown increased frontal cortex lobe volume [14,15] and atypical development in temporal gray matter volume of brain in children with ASD [16,17]. Furthermore, atypical functional brain activation and connectivity were found in specific frontal and temporal in brain of participants with ASD [18–27].
Inferior frontal cortex (IFC) is typically involved in inhibition behavior (right inferior frontal gyrus, RIFG) and language processing and comprehension (left inferior frontal gyrus, LIFG, commonly known as “Broca’s Area”). When performing inhibition task, adults with high-functioning autism showed an atypically inhibition circuitry, manifesting as reduced connectivity between inhibition network and right middle and inferior frontal regions and right inferior parietal regions [18]. Pars opercularis IFG was also shown to be one of the mirror neuron systems (MNS; including IFC, inferior parietal lobule and superior temporal sulcus) which were related to imitation, theory of mind and social communication [28]. It has been found that children with high-functioning autism showed no activity in IFG even though they performed equally well when compared to a typical developing (TD) group in tasks involving imitation and observation of emotional expressions [19].
Temporal cortex (TC) is typically involved in auditory processing, language processing and social cognition. ASD has been linked with atypical lateralization in TC, both anatomically [16,29] and functionally [20,21]. TC was also shown to be one of the most important regions of mentalizing network, including medial prefrontal cortex, superior temporal sulcus (STS) at the temporal-parietal junction and temporal poles. This network forms the neural basis for the ability to understand other people’s thoughts, desires and emotions [22]. When trying to describe the mental state of animated shapes’ movements, adults with ASD showed less activation than typical group in these regions as well as reduced functional connectivity between extrastriate cortex and STS [23].
IFC combined with TC often plays roles in both social and language processing. According to Brothers’ theory, the “social brain” network was comprised of the orbito-frontal cortex (OFC), superior temporal gyrus (STG) and amygdale [24]. With BOLD (blood oxygen level depended)-fMRI, Baron-Cohen et al. tested how this “social brain” network functioned in high-functioning autism and Asperger Syndrome. When adults with ASD performed social intelligence tasks, they showed less extensive activation in left amygdale, right insula, left IFG and a greater response in bilateral STG [25]. When performing the n-back tasks with faces, adults with ASD showed reduced activation than TD group in the right temporal area (including STG and middle temporal gyrus (MTG)) and left frontal area (inferior and middle frontal gyrus), suggesting a deficit in theory of mind and less verbal coding or verbally mediated processing of faces in ASD [26]. The inferior and middle frontal gyrus of ASD also showed lower connectivity with left and right fusiform gyrus (face area of the brain) [26]. This study suggested atypical activation and functional connectivity pattern for face processing of brains with ASD. During sentence comprehension, ASD showed less activation in left IFG and more activation in left STG than TDs [27].
A few studies have revealed that ASD is linked to an atypical pattern of cortical resting-state functional connectivity. With Electroencephalographic (EEG) measurement, adults with ASD showed increased local coherence in frontal and temporal regions in the theta (3-6Hz) frequency range and reduced coherence within FC and between frontal and all other regions [30]. By using fMRI, studies on adults and adolescents with ASD found altered functional connectivity in default model network [31–34], also known as task negative network (including medial prefrontal cortex, posterior cingulate/precuneus and left angular gyrus). Further, an ASD brain showed reduced interhemispheric correlation in sensorimotor cortex, anterior insula, fusiform gyrus, STG and superior parietal lobule [35]. Moreover, disrupted neural synchronization in STG and IFG were shown to emerge very early (12-24 months) in brain development, which could be an important early diagnostic indication for ASD [36]. However, most of these pieces of evidence were provided from studies with adults or adolescents, mostly because fMRI has limitation for child and infant studies, especially for awoken children with ASD.
fNIRS have been previously used for studying ASD patients under certain cognitive tasks [37–39]. So far, there is no report of fNIRS study on the RSFC in children with ASD. In this paper, we propose to use fNIRS as a cheaper and easy-to-operate neuro-imaging technique to find some characteristic features of RSFC neural activity of children with ASD. As IFC and TC have been shown by many previous studies to be two most relevant cortices of the brain of an individual with ASD, we suppose our study would provide more specific hemodynamic response (HbO and Hb) in these regions. As IFC and TC are very close to each other anatomically, inter-region, local (between IFC and TC in the same hemisphere) and long-distance (interhemisphere) brain connectivity could be studied in ASD and typically developed (TD) children.
2. Method
2.1 Participants and protocol
In order to compare the patterns of RSFC between ASD and typical children, we recruited 10 children (mean age = 9.0 ± 1.3) who were diagnosed as ASD and 10 TD children (mean age = 8.9 ± 1.4) for our study. All of the participants were boys and right-handed. During the experiment, children sat in a silent room with dim light. They were asked to close eyes and sit still for 8 minutes. After a participant kept stable for 2 minutes, we began to record the signals. The experiment protocol was approved by the Institutional Review Board of South China Normal University.
It might be difficult for children, especially children with ASD to keep perfectly still for 8 minutes. We believe NIRS measurement is less sensitive to head movement than MRI. Unlike MRI, optical sources and detectors are fixed on the scalp by helmet. Once the helmet is fixed, the relative measurement position is hardly changed by regular head movement. However, we still introduced several additional steps to remove the motion effect from our study and made sure the data were qualified for analysis and interpretation. First, in order to let children with ASD reach the standard of resting-state measurement, we had a training program before the experiment. The goal of this training was that the frequency of interference (such as any kind of small head movement and speaking) was less than once in a minute. Noticeable head movement and touching the optics fibers or helmet were not allowed. Although there was no training for TD children, we found they could easily reach the standard after clarify the instructions. We made sure all the children participating in this study reached the standard. In the process of experiment, we provided each child a soft pillow so that all children can bend forward and lean on the pillows to support their heads (thus they would not feel tired with neck and back). We were always watching the real-time recording of the signals as well as children. If any disturbance occurred, we would make a mark on the data immediately which enable us to identify the interference of motion. Secondly, in data preprocessing, we cut and removed fNIRS data with artifacts.
2.2 Experiment setup
44 channels of a fNIRS system (FOIRE-3000, Shimadzu Corporation, Kyoto, Japan) were used to assess the LFFs of neural activity of the bilateral IFC and TC. The absorptions of three wavelengths (780nm, 805nm, and 830nm) of near infrared light were measured with a sampling rate of 14.286 Hz (time resolution = 70ms) and then transformed into concentration changes of HbO and Hb by the modified Beer-Lambert law. The international 10-10 system [40] was adopted to locate the IFC and TC. We located Channel 4 in F7, Channel 9 in FT7 and Channel 14 in T7 in the left hemisphere. In the right hemisphere, we located Channel 26 in F8, Channel 31in FT8 and Channel 36 in T8. According to previous research on anatomical location of 10-10 system, F7 and F8 were most likely in IFG; FT7 and FT8 were mostly likely in STG; T7 and T8 were mostly likely located in MTG [40]. As shown in Fig. 1, there were 22 channels in each hemisphere covering IFC and TC. The source-detector distance was fixed at 3 cm.
Fig. 1 The position of the optical probes (7 sources: red circles, 8 detectors: blue circles) in the left hemisphere. A black line connecting a source and a detector presents a data channel, which has a number alongside. 22 channels were used to cover inferior frontal cortex (IFC) and temporal cortex (TC). There were 7 channels (left hemisphere: from Channel 1 to Channel 7, right hemisphere: from Channel 23 to Channel 29) covering IFC and 15 channels (left hemisphere: from Channel 8 to Channel 22, right hemisphere: from Channel 30 to Channel 44) covering TC. IFC was boxed with the blue line and TC with the red line in Figs. 1(a)–1(c). Channels 4, 9, 14 (26, 31, 36 in symmetry in the right hemisphere) were located in the F7, FT7 and T7 in the international 10-10 system, respectively. The settings of the optical probes in the right hemisphere were identical to those of the probes in the left hemisphere through the anatomical symmetry. Channels’ numbers were marked on the left hemisphere (Fig. 1(b)) and right hemisphere (Fig. 1(c)) of the brain. We used an image of standard brain to visualize where the channels were mostly likely located in the cortex.
2.3 Data analysis
First of all, only data of HbO and Hb, but not HbT, were included in analysis. Before calculating the time course correlation values and generating correlation maps, data of each channel of each individual were preprocessed in three steps. First, we removed artifacts from fNIRS signals. Since some motional events were marked in advance and an interference noise would appear like a “sudden” change (especially in Hb), artifacts due to some motional events could be easily recognized from the slow fluctuation of resting-state signals. We cut and removed the motion segment from the data. Secondly, in order to remove possible physiological noise and signals beyond our study, a band pass filtered between 0.009 and 0.08 Hz [6–9] was used. Thirdly, the global signal, which was estimated by averaging the time series over all channels, was removed by linear regression [41]. The residual generated from regression was used for later correlation analysis and mapping. In fMRI RSFC studies, removal of global signal from resting state correlation maps could enhance in detection of system-specific correlations and improve the correspondence between resting-state correlations and anatomy [42]. In fNIRS RSFC studies, globally systemic fluctuation would cause overestimation of the correlation value and broaden the regions with connectivity [8]. It has been shown that removal of global signal after band pass filter did not significantly change the power spectra of NIRS signal and thus gave the same pattern of correlation maps [8].
We used interhemispheric correlation to reveal the left-right connectivity of IFC and TC in ASD and TD children. Then, we analyzed the inter-region correlation with all regions of interest (ROIs) to investigate both the local connectivity within each ROI and interaction of different ROIs. At last, we generated correlation maps to visualize the pattern of connectivity of the ASD and TD children. For the interhemispheric correlation analysis, we calculated the Pearson correlation coefficient r between the time course of each channel and the corresponding symmetrical channel in the other hemisphere from the chosen ROI (overall, IFC and TC in this study) and then averaged all possible seeds in each ROI.
For inter-region correlations and correlation maps, we chose one or two channels from a ROI as seeds and calculated r between the time course of the seed and the time course of all other channels in all of the ROIs. The position of the seed was defined as the locator channel of each hemisphere (Channel 4, Channel 9, Channel 14 in the left hemisphere and Channel 26, Channel 31, Channel 36 in the right hemisphere). Each seed represented a possible position in the 10-10 international system (see Fig. 1). For example, Channel 4 was F7; Channel 9 was FT7; Channel 14 was T7; Channel 26 was F8; Channel 31 was FT8; Channel 36 was T8. As mentioned before, F7 and F8 were most likely in IFG; FT7 and FT8 were mostly likely in STG; T7 and T8 were mostly likely in MTG [40]. In inter-region correlation analysis, we calculated the maximum correlation value between a seed channel in one of the ROIs and all the other channels inside each ROI. For example, when the seed was Channel 4, we calculated the maximum correlation value of HbO/Hb between Channel 4 and all other channels in the left IFC (except Channel 4 itself), left TC, right IFC, and right TC. To generate correlation maps, the correlation values between the seed channel and other channels were mapped onto the channel geometry. Then, we used interpolation to fill the blank area between the adjacent channels.
The average correlation values of the ASD/TD group were estimated by converting r values to z values with Fisher’s r - z transformation for each participant, and then converting back to the averaged z values to obtain the averaged r values. In our statistic analysis, we used Bonferroni correction to counteract Type Ierrors caused by multiple comparisons. The significance level is α = 0.05/n, where n is the number of tests for each data.
3. Results
3.1 Interhemispheric correlation
Results of the correlation analysis are shown in Fig. 2. We used the Student’s t-Test to study the statistical significance (p value) of the difference between the average correlation values for ASD and TD children. If the p value obtained from the t-test is lower than the significance level α, we can infer that the difference is significant (smaller p value corresponds to a more significant result). Our results of the t-test (with α = 0.017, considering there were three independent tests for each observed data of each participant) indicated that children with ASD exhibited nearly significantly smaller overall left-right correlation values of HbO than the TDs (rASD = 0.144 ± 0.085, rTD = 0.318 ± 0.194, p = 0.018 ≈0.017). Children with ASD showed weaker overall interhemispheric functional connectivity than the controls (including all the channels we used, covering bilateral IFC and bilateral TC). In particular, ASD showed significantly weaker interhemispheric functional connectivity (rASD = 0.112 ± 0.098, rTD = 0.318 ± 0.163, p = 0.002) in TC (HbO). The HbO values in IFC and the data for Hb in all ROIs did not show significant difference between the two groups (p > 0.017).
Fig. 2 Interhemispheric correlation in ROIs of children with autism spectrum disorder (ASD, red, n = 10) and typically developing (TD) children (blue, n = 10). Error bars are standard error of mean across participants. Children with ASD showed significantly reduced interhemispheric correlation in overall (including all the channels, p = 0.018) and temporal cortex (TC, p = 0.002) than TDs in terms of HbO.
3.2 Inter-region correlations
Table 1 summarized the average maximum correlation values between each ROI and each seed for HbO and Hb across all children in each group. With the purpose of comparing the strength of connectivity between all ROIs and finding out the discrepancy between ASD and TD children, we firstly used General Liner Model to test the effect of the seed selection (within factor: six different seeds locations), group (between factor: ASD vs. TD) and their interactions on the correlation values in terms of HbO and Hb in each ROI. Then, we used a post hoc analysis to do pair-wise comparisons (with Bonferroni correction). We believed this was an optimal method for taking the interaction of different variables into concern, improving the efficiency of statistics and preventing errors of multiple comparisons.
Table 1. Means and standard errors of means across all participants for maximum correlation of HbO and Hb, between 4 ROIs and 6 seeds.
The interaction effect of seed and group did not reach the significance (F ≤ 1.30, p ≥ 0.317), meaning that the influence of the group (ASD or TD children) on the correlation values did not significantly differ when we chose different seeds, or equivalently speaking, the influence of the seed on the correlation value was not modulated or modified by the chosen groups. Under such circumstances, we just needed to take the main effect of seed and group into concern.
Our results showed that the main effect of seeds' selection was significant across all ROIs and all types of hemoglobin (F ≥ 7.92, p ≤ 0.001). As the results were nearly the same for HbO and Hb, we only gave the average maximum correlation values for HbO in Fig. 3. Each subfigure showed how the different seeds connected to the given cortex. According to post hoc analysis (with Bonferroni correction), left IFG seed and left STG seed got stronger correlation with left IFC than all other seeds (see Fig. 3(a), p < 0.001). In Fig. 3(b), right IFG seed showed the strongest correlation with right IFC (p < 0.02). The correlation between right STG seed and right IFC was significantly stronger than correlation between right MTG seed and right IFC seed (p = 0.012). In Fig. 3(c), left STG seed and left MTG seed got equally strong correlation with left TC and both were stronger than other seeds (p < 0.001). In Fig. 3(d), right STG seed and right MTG seed got equally stronger correlation with right TC and both were stronger than other seeds (p < 0.03). These results showed strong local connectivity within all ROIs. However, in the left hemisphere, STG seed in the left hemisphere exhibited a stronger connectivity with IFC than that in the right hemisphere. Thus, the inter-region connectivity from STG to IFC only appeared in the left hemisphere (in line with Hb analysis).
Fig. 3 Inter-region correlation (HbO) between four ROIs and six seeds of children with ASD (red line, n = 10) and TDs (blue line, n = 10). Four ROIs are left interior frontal cortex (a), right inferior frontal cortex (b), left temporal cortex (c), and right temporal cortex (d). Six seeds are left inferior frontal gyrus (Channel 4), left superior temporal gyrus (Channel 9), left middle temporal gyrus (Chanel 14), right inferior frontal gyrus (Channel 26), right superior temporal gyrus (Channel 31), and right middle temporal gyrus (Channel 36). Error bars are standard error of mean across participants. Significant group difference was marked with red asterisks: * means 0.01<p<0.05, ** means p<0.01. The ASDs showed significantly weaker inter-region correlation between right TC and left STG, and weaker local correlation in right TC than controls (d).
The main effect of groups was heterogeneous. There were significant differences between groups in left IFC (HbO, Hb), in left TC (Hb) and in right TC (HbO and Hb). Considering HbO only, all the differences were in right TC (Fig. 3(d), group differences were marked with red asterisks). Considering the functional connectivity to right TC, children with ASD showed significantly lower intensity in left STG seed (p = 0.027), right STG seed (p = 0.005) and right MTG seed (p = 0.034) than TD children, suggesting that children with ASD showed reduced interhemispheric and local connectivity in right TC. Regarding Hb, significantly lower inter-region correlations for children with ASD were in the following pairs: right STG to left IFC (p = 0.001), left MTG to left TC (p = 0.039), right STG to left TC (p = 0.040), right MTG to left TC (p = 0.032), and left MTG to right TC (p = 0.01), suggesting that children with ASD showed reduced local connectivity in left TC, lower connectivity from bilateral STG to left IFC, and lower interhemispheric connectivity in TC. Converging all the results in group analysis, children with ASD showed reduced local connectivity in left temporal cortex (only for Hb) and right temporal cortex (only for HbO), reduced interhemispheric connectivity in temporal cortex (for HbO and Hb), reduced connectivity from right superior temporal gyrus to left inferior frontal cortex (only for Hb).
3.3 Correlation maps
For comparison, Fig. 4 gives correlation maps (HbO) of all the channels for a TD child and an ASD child, separating by different seed locations: Three seeds in the left hemisphere and three corresponding seeds in the right hemisphere. When the seeds were located in left hemisphere, a TD child showed both stronger local and interhemispheric correlations than an ASD child. Nearly no symmetrical pattern was observed for an ASD child when the seed was located in the left temporal cortex (Channel 9 and Channel 14). When the seeds were located at the right hemisphere, children with ASD also showed reduced interhemispheric correlation in temporal cortex, but not so obviously than the left-seeded. All TD children displayed a similar correlation pattern with strong left-and-right symmetry; however, the correlation maps for those with ASD are not so similar in patterns, but share very poor left-and-right symmetry. Correlation maps for Hb gave nearly the same patterns as for HbO.
Fig. 4 HbO correlation maps for a typically developing (TD) child and an autism spectrum disorder (ASD) child. The selected seeds (Channel 4, Channel 9, Channel 14 in the left hemisphere and Channel 26, Channel 31, Channel 36 in the right hemisphere) were marked on the maps. (a) HbO correlation maps for a TD child (interhemispheric correlation value is 0.530). (b) HbO correlation maps for an ASD child (interhemispheric correlation value is 0.036). Color bar represents the strength of the correlation with the seed.
4. Discussion
During the twenty-year development, fNIRS has been shown to be an effective tool to monitor human brain function and widely applied in many areas [43]. In this work, we used fNIRS for the first time to study cortical resting-state functional connectivity of the children with autism spectrum disorder (ASD) and typical developing (TD) ones. We firstly investigated interhemispheric correlations and then inter-region correlations of inferior frontal cortex (IFC) and temporal cortex (TC) of the brain for each group. We also statistically analyzed the difference between ASDs and TDs.
By calculating the Pearson correlation coefficient r between the time course of each channel and the symmetrical channel in the other hemisphere from the regions of interest (ROIs; overall, IFC and TC), After that, to reveal more details of functional connectivity between all regions in this study (bilateral inferior cortices, bilateral temporal cortices), we chose six seeds based on the 10-10 international system. Each seed represents an anatomy location of the brain (bilateral inferior frontal gyrus, bilateral superior temporal gyrus, and bilateral middle temporal gyrus) [40]. We calculated r between the time course of the seed and the time course of all other channels in all the ROIs. By analyzing the main effect of the seeds selection, we could see how the six seeds connected to the four ROIs. Furthermore, correlation maps were generated to visualize the pattern of connectivity.
Our results of interhemispheric correlation analysis have indicated that children with ASD showed reduced interhemispheric correlation in TC. Inter-region analysis showed strong local connectivity within all ROIs. Moreover, in the left hemisphere, STG showed stronger connectivity to IFC. By analyzing the group difference, different patterns of inter-region correlations between children with ASD and TD children have been revealed. Results varied between different hemoglobin contrasts. For HbO, children with ASD showed reduced local connectivity in right TC and lower interhemispheric connectivity between bilateral TCs. Hb analyses would give the consistent findings for reduced interhemispheric connectivity in TC of children with ASD, besides the additional finding that children with ASD showed reduced local connectivity in left TC, lower connectivity from bilateral STG to left IFC. Furthermore, TD children showed obvious local connectivity in the left hemisphere and strong left-right symmetry in maps, while ASD children did not show local connectivity and lost left-right symmetry when the seeds ware located in the left temporal cortex.
Summarizing all the results, reduced interhemispheric connectivity in temporal cortex was our main finding when we tried to reveal the cortical resting-state functional connectivity of ASD children by fNIRS. Our findings are in line with fMRI studies with adolescents, adults and toddlers with ASD [31,35,36]. Furthermore, as compared with TD children, children with ASD exhibited reduced local connectivity of bilateral temporal cortices (note that these results were revealed by analysis of HbO and Hb separately). As we have mentioned in the introductions, the temporal regions may be involved in the sensation in auditory, processing of word, and mentalizing network, where ASD children have functional and anatomical difference from TD children [14–29]. These results provide the evidence for intrinsic hemodynamic difference between ASD and TD children in temporal cortex, suggesting fNIRS is a much cheaper and yet effective technique to detect the atypical neural activity of awoken children with ASD or other developmental disorder of the brain [9]. However, in order to test the feasibility of fNIRS in diagnosis and assessment on ASD, larger sample size and more specific assessment of symptoms will be needed.
Although it was not the main purpose of the present study to reveal the feature of functional connectivity of different regions, we also found some interesting results about connectivity between IFC and TC and lateralization. For both groups, local correlation in the left TC was stronger than that in the right TC. This lateralization in left temporal cortex was in line with previous fNIRS study [44]. Furthermore, the connectivity from STG to IFC only appeared in the left hemisphere. According to the dual-stream model of speech processing, the link from STG to IFC is believed to be part of the dorsal circuit of speech processing and STG to MTG is the ventral stream [45]. Dorsal stream is strongly left-dominate, while the ventral one is largely bilaterally organized [45]. Our result on dorsal stream (from STG to LFC) was also left dominate, suggesting fNIRS as well as the RSFC method have potential to study more specific cortical organization of the brain.
Considering the different hemoglobin contrasts, HbO provided the best sensitivity to reflect the difference in interhemispheric correlation and the best convergence on the functional related region of the cortex, suggesting that this contrast could be recommended to identify the atypical neural connectivity of ASD. Hb showed good sensitivity on detecting the different pattern of inter-region connectivity between groups. In some previous study on human adults [8] and rats [46], HbT provide better maps of functional location than other two hemoglobin contrasts. However, some studies also supported the better mapping for HbO and Hb [47]. Future studies on different samples would be required to make physiologic judgments about the origins of these mapping differences.
In our study, we have used global signal removal to avoid the impact of systematic physiological noise on RSFC. Although this method was widely used in fMRI RSFC studies, there were some arguments that removal of global signal would induce anti-correlation in resting-state networks of the brain [48]. However, the anti-correlation effect, which was possibly caused by removal of global signal, impacted the default model network (task negative network) and extended dorsal attention system (task positive network, including the intraparietal sulcus and the junction of the precentral and superior frontal sulcus in each hemisphere). Those regions were not included in our study. Thus we believe the removal of global signal will not cause artifacts of results. Even so, the systematic noise should be more specifically defined and measured by auxiliary channels in further studies.
5. Conclusions
In summary, by using fNIRS to study RSFC of ASD and TD children, for the first time, we have observed reduced interhemispheric functional connectivity in children with ASD. Compared with the controls, children with ASD showed significantly lower interhemispheric correlation in temporal cortex. Children with ASD also showed significantly lower local connectivity in bilateral temporal cortices. In correlation maps (HbO), an ASD child did not present a symmetric pattern like a TD child in temporal cortex. For the three hemodynamic parameters measured by fNIRS, HbO provided the optimal sensitivity to reflect the divergence of function connectivity of different groups, suggesting that this contrast may be used to identify the atypical neural connectivity of ASD. This research suggests that fNIRS is a much cheaper (as compared with fMRI) and effective technique to detect the atypical neural activity of awoken children with autism spectrum disorder.
Acknowledgments
This work was supported by Guangdong Innovative Research Team Program (No. 201001D0104799318), the National Basic Research Program (973) of China (2011CB503700), Guangdong Science and Technology Program (2012B03180000), Macau Foundation (CUM-16), the Swedish Research Council and SOARD. We thank Prof. Jun Li in Centre for Optical and Electromagnetic Research of SCNU for helping with data analysis and imaging processing. We also thank Profs. Heyong Shen and Lan Gao of SCNU for discussion, and Zhifang Zhu, Lina Qiu, Xinge Li and Wei Cao for experiments.
References and links
1. F. F. Jöbsis, “Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters,” Science 198(4323), 1264–1267 (1977). [CrossRef] [PubMed]
2. A. Villringer and B. Chance, “Non-invasive optical spectroscopy and imaging of human brain function,” Trends Neurosci. 20(10), 435–442 (1997). [CrossRef] [PubMed]
3. A. Villringer, J. Planck, C. Hock, L. Schleinkofer, and U. Dirnagl, “Near infrared spectroscopy (NIRS): a new tool to study hemodynamic changes during activation of brain function in human adults,” Neurosci. Lett. 154(1-2), 101–104 (1993). [CrossRef] [PubMed]
4. M. Tamura, Y. Hoshi, and F. Okada, “Localized near-infrared spectroscopy and functional optical imaging of brain activity,” Philos. Trans. R. Soc. Lond. B Biol. Sci. 352(1354), 737–742 (1997). [CrossRef] [PubMed]
5. Z. Guo, F. Cai, and S. He, “Optimization for Brain Activity Monitoring with Near Infrared Light in a Four-Layered Model of the Human Head,” Prog. Electromagnetics Res. 140, 277–295 (2013). [CrossRef]
6. C.-M. Lu, Y.-J. Zhang, B. B. Biswal, Y.-F. Zang, D.-L. Peng, and C.-Z. Zhu, “Use of fNIRS to assess resting state functional connectivity,” J. Neurosci. Methods 186(2), 242–249 (2010). [CrossRef] [PubMed]
7. B. R. White, S. M. Liao, S. L. Ferradal, T. E. Inder, and J. P. Culver, “Bedside optical imaging of occipital resting-state functional connectivity in neonates,” Neuroimage 59(3), 2529–2538 (2012). [CrossRef] [PubMed]
8. R. C. Mesquita, M. A. Franceschini, and D. A. Boas, “Resting state functional connectivity of the whole head with near-infrared spectroscopy,” Biomed. Opt. Express 1(1), 324–336 (2010). [CrossRef] [PubMed]
9. R. E. Vanderwert and C. A. Nelson, “The use of near-infrared spectroscopy in the study of typical and atypical development,” Neuroimage 85(Pt 1), 264–271 (2014). [CrossRef] [PubMed]
10. B. Biswal, F. Z. Yetkin, V. M. Haughton, and J. S. Hyde, “Functional connectivity in the motor cortex of resting human brain using echo-planar MRI,” Magn. Reson. Med. 34(4), 537–541 (1995). [CrossRef] [PubMed]
11. M. E. Raichle, “Two views of brain function,” Trends Cogn. Sci. (Regul. Ed.) 14(4), 180–190 (2010). [CrossRef] [PubMed]
12. P. J. Uhlhaas and W. Singer, “Neural synchrony in brain disorders: relevance for cognitive dysfunctions and pathophysiology,” Neuron 52(1), 155–168 (2006). [CrossRef] [PubMed]
13. American Psychiatry Association, Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR® (American Psychiatric Pub, 2000).
14. E. Courchesne, R. Carper, and N. Akshoomoff, “Evidence of brain overgrowth in the first year of life in autism,” JAMA 290(3), 337–344 (2003). [CrossRef] [PubMed]
15. R. A. Carper and E. Courchesne, “Inverse correlation between frontal lobe and cerebellum sizes in children with autism,” Brain 123(4), 836–844 (2000). [CrossRef] [PubMed]
16. N. Boddaert, N. Chabane, H. Gervais, C. D. Good, M. Bourgeois, M. H. Plumet, C. Barthélémy, M. C. Mouren, E. Artiges, Y. Samson, F. Brunelle, R. S. Frackowiak, and M. Zilbovicius, “Superior temporal sulcus anatomical abnormalities in childhood autism: a voxel-based morphometry MRI study,” Neuroimage 23(1), 364–369 (2004). [CrossRef] [PubMed]
17. G. M. McAlonan, V. Cheung, C. Cheung, J. Suckling, G. Y. Lam, K. S. Tai, L. Yip, D. G. Murphy, and S. E. Chua, “Mapping the brain in autism. A voxel-based MRI study of volumetric differences and intercorrelations in autism,” Brain 128(2), 268–276 (2004). [CrossRef] [PubMed]
18. R. K. Kana, T. A. Keller, N. J. Minshew, and M. A. Just, “Inhibitory control in high-functioning autism: decreased activation and underconnectivity in inhibition networks,” Biol. Psychiatry 62(3), 198–206 (2007). [CrossRef] [PubMed]
19. M. Dapretto, M. S. Davies, J. H. Pfeifer, A. A. Scott, M. Sigman, S. Y. Bookheimer, and M. Iacoboni, “Understanding emotions in others: mirror neuron dysfunction in children with autism spectrum disorders,” Nat. Neurosci. 9(1), 28–30 (2006). [CrossRef] [PubMed]
20. G. Dawson, C. Finley, S. Phillips, and L. Galpert, “Hemispheric specialization and the language abilities of autistic children,” Child Dev. 57(6), 1440–1453 (1986). [CrossRef] [PubMed]
21. M. R. Prior and J. L. Bradshaw, “Hemisphere functioning in autistic children,” Cortex 15(1), 73–81 (1979). [CrossRef] [PubMed]
22. U. Frith and C. D. Frith, “Development and neurophysiology of mentalizing,” Philos. Trans. R. Soc. Lond. B Biol. Sci. 358(1431), 459–473 (2003). [CrossRef] [PubMed]
23. F. Castelli, C. Frith, F. Happé, and U. Frith, “Autism, Asperger syndrome and brain mechanisms for the attribution of mental states to animated shapes,” Brain 125(8), 1839–1849 (2002). [CrossRef] [PubMed]
24. L. Brothers, “The social brain: a project for integrating primate behavior and neurophysiology in a new domain,” Concepts Neurosci. 1, 27–51 (1990).
25. S. Baron-Cohen, H. A. Ring, S. Wheelwright, E. T. Bullmore, M. J. Brammer, A. Simmons, and S. C. Williams, “Social intelligence in the normal and autistic brain: an fMRI study,” Eur. J. Neurosci. 11(6), 1891–1898 (1999). [CrossRef] [PubMed]
26. H. Koshino, R. K. Kana, T. A. Keller, V. L. Cherkassky, N. J. Minshew, and M. A. Just, “fMRI investigation of working memory for faces in autism: visual coding and underconnectivity with frontal areas,” Cereb. Cortex 18(2), 289–300 (2008). [CrossRef] [PubMed]
27. M. A. Just, V. L. Cherkassky, T. A. Keller, and N. J. Minshew, “Cortical activation and synchronization during sentence comprehension in high-functioning autism: evidence of underconnectivity,” Brain 127(8), 1811–1821 (2004). [CrossRef] [PubMed]
28. G. Rizzolatti and L. Craighero, “The mirror-neuron system,” Annu. Rev. Neurosci. 27(1), 169–192 (2004). [CrossRef] [PubMed]
29. M. R. Herbert, D. A. Ziegler, N. Makris, P. A. Filipek, T. L. Kemper, J. J. Normandin, H. A. Sanders, D. N. Kennedy, and V. S. Caviness Jr., “Localization of white matter volume increase in autism and developmental language disorder,” Ann. Neurol. 55(4), 530–540 (2004). [CrossRef] [PubMed]
30. M. Murias, S. J. Webb, J. Greenson, and G. Dawson, “Resting state cortical connectivity reflected in EEG coherence in individuals with autism,” Biol. Psychiatry 62(3), 270–273 (2007). [CrossRef] [PubMed]
31. V. L. Cherkassky, R. K. Kana, T. A. Keller, and M. A. Just, “Functional connectivity in a baseline resting-state network in autism,” Neuroreport 17(16), 1687–1690 (2006). [CrossRef] [PubMed]
32. S.-J. Weng, J. L. Wiggins, S. J. Peltier, M. Carrasco, S. Risi, C. Lord, and C. S. Monk, “Alterations of resting state functional connectivity in the default network in adolescents with autism spectrum disorders,” Brain Res. 1313, 202–214 (2010). [CrossRef] [PubMed]
33. M. Assaf, K. Jagannathan, V. D. Calhoun, L. Miller, M. C. Stevens, R. Sahl, J. G. O’Boyle, R. T. Schultz, and G. D. Pearlson, “Abnormal functional connectivity of default mode sub-networks in autism spectrum disorder patients,” Neuroimage 53(1), 247–256 (2010). [CrossRef] [PubMed]
34. D. P. Kennedy and E. Courchesne, “The intrinsic functional organization of the brain is altered in autism,” Neuroimage 39(4), 1877–1885 (2008). [CrossRef] [PubMed]
35. J. S. Anderson, T. J. Druzgal, A. Froehlich, M. B. DuBray, N. Lange, A. L. Alexander, T. Abildskov, J. A. Nielsen, A. N. Cariello, J. R. Cooperrider, E. D. Bigler, and J. E. Lainhart, “Decreased interhemispheric functional connectivity in autism,” Cereb. Cortex 21(5), 1134–1146 (2011). [CrossRef] [PubMed]
36. I. Dinstein, K. Pierce, L. Eyler, S. Solso, R. Malach, M. Behrmann, and E. Courchesne, “Disrupted neural synchronization in toddlers with autism,” Neuron 70(6), 1218–1225 (2011). [CrossRef] [PubMed]
37. Y. Kawakubo, H. Kuwabara, K.-i. Watanabe, M. Minowa, T. Someya, I. Minowa, T. Kono, H. Nishida, T. Sugiyama, N. Kato, and K. Kasai, “Impaired prefrontal hemodynamic maturation in autism and unaffected siblings,” PLoS ONE 4(9), e6881 (2009). [CrossRef] [PubMed]
38. U. Chaudhary, M. Hall, A. Gutierrez, D. Messinger, G. Rey, and A. Godavarty, “Joint attention studies in normal and autistic children using NIRS,” Proc. SPIE 7883, 788348 (2011). [CrossRef]
39. Y. Kita, A. Gunji, Y. Inoue, T. Goto, K. Sakihara, M. Kaga, M. Inagaki, and T. Hosokawa, “Self-face recognition in children with autism spectrum disorders: a near-infrared spectroscopy study,” Brain Dev. 33(6), 494–503 (2011). [CrossRef] [PubMed]
40. L. Koessler, L. Maillard, A. Benhadid, J. P. Vignal, J. Felblinger, H. Vespignani, and M. Braun, “Automated cortical projection of EEG sensors: Anatomical correlation via the international 10-10 system,” Neuroimage 46(1), 64–72 (2009). [CrossRef] [PubMed]
41. P. M. Macey, K. E. Macey, R. Kumar, and R. M. Harper, “A method for removal of global effects from fMRI time series,” Neuroimage 22(1), 360–366 (2004). [CrossRef] [PubMed]
42. M. D. Fox, D. Zhang, A. Z. Snyder, and M. E. Raichle, “The global signal and observed anticorrelated resting state brain networks,” J. Neurophysiol. 101(6), 3270–3283 (2009). [CrossRef] [PubMed]
43. D. A. Boas, C. E. Elwell, M. Ferrari, and G. Taga, “Twenty years of functional near-infrared spectroscopy: introduction for the special issue,” Neuroimage 85(Pt 1), 1–5 (2014). [CrossRef] [PubMed]
44. Y.-J. Zhang, C.-M. Lu, B. B. Biswal, Y.-F. Zang, D.-L. Peng, and C.-Z. Zhu, “Detecting resting-state functional connectivity in the language system using functional near-infrared spectroscopy,”J. Biomed. Opt. 15, 047003 (2010).
45. G. Hickok and D. Poeppel, “The cortical organization of speech processing,” Nat. Rev. Neurosci. 8(5), 393–402 (2007). [CrossRef] [PubMed]
46. A. Devor, I. Ulbert, A. K. Dunn, S. N. Narayanan, S. R. Jones, M. L. Andermann, D. A. Boas, and A. M. Dale, “Coupling of the cortical hemodynamic response to cortical and thalamic neuronal activity,” Proc. Natl. Acad. Sci. U.S.A. 102(10), 3822–3827 (2005). [CrossRef] [PubMed]
47. B. R. White, A. Z. Snyder, A. L. Cohen, S. E. Petersen, M. E. Raichle, B. L. Schlaggar, and J. P. Culver, “Resting-state functional connectivity in the human brain revealed with diffuse optical tomography,” Neuroimage 47(1), 148–156 (2009). [CrossRef] [PubMed]
48. K. Murphy, R. M. Birn, D. A. Handwerker, T. B. Jones, and P. A. Bandettini, “The impact of global signal regression on resting state correlations: are anti-correlated networks introduced?” Neuroimage 44(3), 893–905 (2009). [CrossRef] [PubMed]
