1 Introduction

Autism Spectrum Disorder (ASD) is a heterogeneous neuro-developmental disorder that begins in early childhood and persists throughout life. It is characterized by social deficits, restricted interests, stereotyped behavior, and alterations in sensory processing [1]. Approximately 1/100 children are diagnosed with ASD around the world [2, 3]. For a diagnosis of High-Functioning Autism (HFA), the subject had to meet the required criteria for autism based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders and also have a full scale IQ of at least 70 on a test of intelligence [4]. They are characterized by possessing simple basic abilities in normal language use, cognition, and social skills. Due to their relatively less severe core symptoms, they are often overlooked even by medical professionals. Past studies show that ASD children benefit from theory of mind behavioral intervention tasks to promote the development of competencies [5, 6]. However, the precondition for effectively completing the theory of mind task is to have the corresponding basic ability, and HFA children are the most suitable candidates [6, 7]. Therefore, it is crucial to early identify HFA children, which may assist in enhancing rehabilitation outcomes and improving prognosis. Despite significant progress in understanding the behavioral and cognitive aspects of HFA, the specific neuropathological changes and pathogenesis mechanisms remain somewhat unclear [8].

Now a days, the development of magnetic resonance imaging (MRI) technology has explored the abnormal changes in brain structure and function in HFA patients, trying to provide a more objective basis for early diagnosis [9-11]. Indeed, functional MRI is the gold standard for in vivo imaging of the human brain, but functional near-infrared spectroscopy (fNIRS) stands out for its high portability, robustness to noise, relatively low costs, and small size, bringing functional imaging into much more realistic environments [12, 13]. fNIRS is an optical neuroimaging technique that assesses cerebral activity based on hemodynamics, which is associated with changes in the transmission of low power near-infrared light directed through the scalp and skull into the brain [14, 15]. fNIRS has the capacity to measure different types of concentration changes in hemoglobin parameters, which provides a more comprehensive description of functional connectivity (FC). Moreover, fNIRS is quiet, comfortable, and insensitive to subject motion, which is well suited for special populations such as infants and patients. Finally, the portable and cost-effective features of fNIRS make it much easier to translate FC approaches from laboratory environments to clinical applications. Accordingly, fNIRS is becoming a promising imaging modality for RSFC assessment complementary to functional MRI in spite of its limited spatial resolution [16, 17].

Most of the higher and more complex motor, cognitive, and emotional behavioral functions are thought to be found primarily in the frontal lobes [17]. An increasing number of studies have found structural abnormalities related to ASD in multiple brain regions, especially in the frontal lobe [18, 19], and research shows that defects in prefrontal cortex function are associated with core clinical symptoms of ASD [20-23]. In the past decade, resting-state fNIRS has been increasingly used to depict the typical and atypical development of cortical FC and network topological features within the age range from neonates to adolescents [24]. Regions whose blood oxygen level-dependent (BOLD) signal fluctuations show a high degree of temporal correlation are presumed to constitute a tightly coupled neural network [25]. Current results demonstrate that the activation-deactivation dichotomy routinely observed in response to attention-demanding tasks is represented intrinsically in the resting human brain, demonstrable in the absence of any overt task or behavior [24]. Brain regions that are simultaneously active during the resting state show a positive temporal correlation in their associated BOLD signals, and these brain regions together form an intrinsic functional network [26] when describing the connectivity network of the brain in a spontaneous state. It is an endogenously mediated network that is activated during the resting state and when performing social/emotional tasks, but is inhibited when performing non-social tasks that require high cognitive demands [24-26]. Therefore, we speculate that exploring the mechanisms of prefrontal changes in ASD can provide a basis for the early diagnosis and precise treatment of children with ASD. To the best of our knowledge, there have been no reported studies using fNIRS to search for brain function markers of HFA.

The purpose of this study is to use fNIRS technology to investigate the abnormal prefrontal FC in children with HFA and typically developing (TD) children.

2 Materials and Methods

2.2 fNIRS Acquisition

The experiment was conducted in a quiet and dimly lit room. Participants sat on a comfortable chair and were asked to close their eyes, stay awake, keep their bodies as still as possible, and avoid thinking about anything. This study used a fNIRS optical topographic system (employing a dual-wavelength combination of 695/830 nm, 22 channels, model ETG-ONE, manufactured by Hitachi Medical Corporation, Tokyo, Japan). Within the near-infrared spectral window, near-infrared light has strong penetrability, capable of penetrating tissues to a certain depth and reaching the cerebral cortex 20–30 mm beneath the skull. The probe was fixed to the scalp via a headcap, with a scanning frequency of 10 Hz. The system comprised 8 transmitting probes and 7 detecting probes, with a fixed distance of 30 mm between the transmitting and detecting probes. The headcap was generally rectangular, positioned in the prefrontal region, forming 22 measurement channels. Brain signal changes in the prefrontal cortex of the HFA group and TD children were measured in the resting state. To more accurately maintain the coverage area, we referred to the international 10–20 system positioning method to confirm that the lines of the red 12, blue 14, and red 17 optical fibers were located in the midline, and the probes in the next column were positioned on the line of Fp1-Fpz-Fp2 in the international electroencephalogram 10–20 positioning method, as shown in Figure 1.

According to the modified Beer–Lambert law, the concentration changes of oxyhemoglobin and deoxyhemoglobin chromophores were calculated through the degree of infrared light attenuation. When the activation of a certain brain region increases, compared with the baseline period, an increase in oxyhemoglobin concentration and a decrease in deoxyhemoglobin concentration can be observed. The start time was set according to the measurement mode of the fNIRS system. A “ding” prompt sound was emitted at the beginning of the task, and 8 min of data were recorded. The probes were positioned according to the international 10–20 system, as shown in Figure 1.

3 Results

3.2 Multi-Channel Topographic Map

We used topographic maps to visually display the brain functional activation in the area covered by the headcap. A color scale was used to mark the activation levels of different regions, where colors closer to the red end indicate higher blood oxygen metabolism levels, and colors closer to the blue end indicate lower blood oxygen metabolism levels. Based on the coordinate information of channels with significantly higher blood oxygen metabolism levels, the relevant brain regions were determined by comparing with the Transcranial Brain Atlas (TBA) [29].

The TBA is a probabilistic mapping P(L|S) from scalp space to label space. To achieve this mapping, we need: First, mathematically define the scalp space S, that is, design a coordinate system suitable for the scalp; Second, establish a probabilistic mapping from scalp space to brain space, namely solve P(B|S), while considering the influence of individual differences; Finally, combine the scalp-brain mapping P(B|S) with the brain-label mapping P(L|B) to derive the scalp-label mapping P(L|S). As a probabilistic mapping P(L|S), the TBA supports transcranial brain imaging research in two different ways. In the forward application of this mapping, we can use P(L|S) to obtain label information from scalp positions, thereby clarifying the recording or stimulation targets of the placed devices. In the reverse application, we can use P(L|S) to estimate the scalp positions that can best record or stimulate specific target brain regions.

Analysis of topographic maps for the HFA group and TD group showed that blood oxygen metabolism levels were particularly prominent in Channels 6, 7, and 11 in both groups (Figure 2A,B). Comparison of topographic maps between the two groups revealed that blood oxygen metabolism levels in Channels 6, 7, and 11 were higher in the HFA group than in the TD group. Specifically, Channel 11 was the most active in the HFA group, followed by Channel 6 and then Channel 7; in the TD group, Channel 7 was the most active, followed by Channel 11 and then Channel 6. The activation regions clearly indicate significant differences in resting-state brain function between the two groups. Combined with MNI coordinate information, the coordinates corresponding to each channel are described in Table 2. By comparing with the TBA, it can be seen that the functional areas corresponding to Channel 6, Channel 7, and Channel 11 are the right prefrontal cortex region.

TABLE 2. The coordinates corresponding to each channel.

Channel number	MNI_X	MNI_Y	MNI_Z
CH1	28.33	53.67	37.33
CH2	11	62	36
CH3	−12	60.67	37
CH4	−30.67	51.67	34.67
CH5	42.33	50.33	28.33
CH6	21	67	25
CH7	−1.67	66	24.33
CH8	−22.67	66	23.67
CH9	−41.33	50.33	25.67
CH10	32.33	63.33	14.33
CH11	13.33	73	12.67
CH12	−14	72.67	12.33
CH13	−33.67	63	13
CH14	48	54.67	2.33
CH15	21	70.33	1.33
CH16	−2.67	70.67	0.33
CH17	−23.33	70	0.67
CH18	−44	56.33	1.33
CH19	38.33	63.67	−10.67
CH20	13.33	71.33	−11.33
CH21	−13.33	70.33	−11.33
CH22	−36	62.67	−10

3.3 Visualization of Differences in Patient FC Matrices Between Groups

We presented the FC between channels through FC matrices. FC matrices serve as crucial tools for understanding how different brain regions collaborate. The degree of connectivity is represented by a color gradient, where the transition from red to blue indicates the correlation between channels from high to low. Results showed that in the HFA group (Figure 3A), high correlations were mainly observed in regions of Channels 15–19, 14–18, and 18–22. In the TD group (Figure 3B), high correlations were primarily distributed in Channels 15–19, 14–15, 14–19, 10–14, and 9–18. Notably, compared with the HFA group, the TD group exhibited stronger FC with a more concentrated distribution area.

On the other hand, through brain network connectivity maps, the color of inter-channel connections can elucidate the brain network synergy between two points. The brain network connectivity maps of the two groups showed that active channel connections in the TD group (Figure 4B) were dominated by red and yellow, with the region concentrated between channels 10–14–15–19. In contrast, the HFA group (Figure 4A) had significantly fewer red and yellow lines representing tight inter-channel connections than the TD group, with a more dispersed distribution area. This result indicates that the active regions of the brain network in the HFA group were less concentrated than those in the TD group. In addition, although channels 18–12 are in the significantly activated region in the HFA group, there is no functional connection in the TD group.

4 Discussion

This study shows that fNIRS technology can observe differences in resting-state brain FC between TD children and HFA children. It also confirms that the prefrontal lobe of HFA children is more activated at rest than that of TD children, but the distribution of brain FC in HFA children is more dispersed than that in TD children, and the number and intensity of brain FC in the HFA group are significantly lower than those in the TD group.

By comparing the HFA and TD groups, we found that there were similar active regions in both groups. Our results indicate that the blood oxygen metabolism levels in Channels 6, 7, and 11 were higher in the HFA group than in the TD group, and Channels 6, 7, and 11 in the HFA group were more active. The channels 6, 7, and 11 correspond to the right prefrontal cortex region. Joshi G et al.'s [30] findings suggest that the architecture and strength of the default mode network FC are altered in HFA. The default mode network with the medial prefrontal cortex in the HFA group showed weaker correlations in comparison with healthy controls [30]. Previous studies suggest that during the observation of some tasks, frontal cortex activation is affected in HFA children using functional MRI [31, 32]. These research findings are similar to our results. The right prefrontal cortex has been shown to be associated with prosocial behavior [33], which is closely associated with general social skills deficits, and there are significant impairments in the measurements of working memory and processing speed [18]. A study designed to investigate the neural mechanisms of change following social skills interventions shows that the social cognitive skills group showed greater increases in activity in the medial prefrontal cortex, implicated in theory of mind, relative to the comparison group for both irony comprehension and gaze processing tasks [34]. These findings show that the ability of emotion recognition and theory of mind in the HFA may be related to the right prefrontal cortex.

Moreover, in the brain network connection diagrams of the two groups, the number of connections in the HFA group is significantly lower than that in the TD group, which may be related to the impaired brain functions of patients with ASD. In order to further understand the differences in brain activities between the two groups, we analyzed the differences in connectivity between channels. It is worth noting that the stronger functional connections in children with HFA mainly occur in Channels 15–19, 14–18, and 18–22; the stronger functional connections in TD children mainly occur in Channels 15–19, 14–15, 14–19, 10–14, and 9–18. These regions with obvious differences in connectivity are all located in the medial prefrontal cortex region and the dorsolateral prefrontal cortex region [29]. Studies have shown that patients with damage to the ventromedial prefrontal cortex will exhibit social isolation and apathy [35] and show a reduction in prosocial behavior in some social decision-making games [36]. From another perspective, in the brain network connection diagrams of the two groups, the number of connections in the HFA is significantly lower than that in the TD. Obviously, the results we observed are related to the changes in the function of the prefrontal lobe of the brain. In any case, the fNIRS technology reveals the differences in brain functions between the two groups during the resting state.

Previous fNIRS studies have shown that functional deficits in the prefrontal cortex are associated with ASD. Individuals with ASD exhibit atypical social behaviors and deficits in social cognition, such as impaired theory of mind and a lack of social interest [18]. Xiao et al. found that compared with TD children, children with HFA showed lower activation in the right prefrontal cortex when performing the “go/no-go” task and the Stroop task, which also linked the differences in executive functions to the prefrontal regions [37]. Funahashi et al. measured the activation changes in the prefrontal cortex and the temporal cortex of adolescents with ASD and their matched TD adolescents during listening and ignoring tasks. The results showed that there were significant activation differences in the prefrontal cortex between adolescents with ASD and those with typical development, but no such differences were found in the auditory brain regions [38]. This indicates that the lack of awareness of sounds in children with ASD may be due to deficits in attention control rather than auditory cortex dysfunction. Specifically, attention control ability may be one of the causes of non-social difficulties in ASD. Therefore, the dysfunction of the prefrontal cortex may be a developmental consequence resulting from other early deviations.

This study still has some limitations. Firstly, our sample lacks population diversity. Conducting a multicenter study with a larger sample size and involving people from multiple nationalities would be more representative. Secondly, although we randomly selected children from the HFA group and TD group as participants, it is obvious that there is a significant difference in gender ratio in the HFA group. It has been reported that the dominance of males in the autism population is a consistent finding. Finally, we cannot rule out the possibility of underestimating the intelligence level of children with HFA. Whether the WISC is a suitable measurement tool for assessing the cognitive status of children with HFASD is still controversial and worthy of further exploration in the future. Therefore, more carefully designed prospective studies using innovative methods are needed to clarify this issue.

5 Conclusion

In conclusion, we observed significant differences in signal channels in the prefrontal cortex region between preschool children with HFA and TD children. Moreover, the multi-channel topographic map showed a high correlation with children with HFA. The research results further indicate that the default network mode of children with HFA is relatively active in the right medial prefrontal cortex region, and the medial prefrontal cortex region may potentially serve as a new therapeutic target. These preliminary research findings may provide new insights into the potential pathophysiological basis of the autistic brain.

Author Contributions

Conceptualization, W.S. and Z.Z.; Methodology, W.S., B.Z., C.J. and Z.Z.; Resources, Z.Z.; Data Collection, W.S. and Z.Z.; Formal Analysis, W.S., B.Z., C.J. and Z.Z.; Writing – Original Draft Preparation, W.S.; Writing – Review and Editing, W.S., B.Z., C.J. and Z.Z.; Visualization, B.Z., C.J. and Z.Z.; Supervision, C.J.; Funding Acquisition, Z.Z. and C.J.; All authors have read and agreed to the published version of the manuscript.

Ethics Statement

The study was approved by the Institutional Review Board of The Beibei Affiliated Hospital of Chongqing Medical University/The Ninth's People's Hospital of Chongqing (No. JY2021-4-14).

Consent

Informed consent was obtained from all subjects involved in the study.

Conflicts of Interest

The authors declare no conflicts of interest.