A method and system for autism assessment using near-infrared spectroscopy and eye tracking

Abstract

This invention discloses a method and system for autism assessment using near-infrared spectral imaging and eye-tracking, relating to the field of autism risk assessment technology. The key technical points include the following steps: collecting the age, developmental level, and past medical history of suspected autism patients to generate autism patient profiles; establishing personal information files; collecting near-infrared spectral imaging data and eye-tracking data of the patients; screening abnormal correlation data of potential autism pathological features to obtain correlation data to be assessed; extracting the functional coupling characteristics, temporal synchronization, and intensity attenuation patterns of abnormal correlations in the correlation data to be assessed; and determining whether it is necessary to perform scheme fusion processing on the correlation data to be assessed from different testing schemes based on the functional coupling characteristics and temporal synchronization. The effect is that it effectively captures the specific pathological characteristics of patients, avoids the problem of a uniform testing paradigm masking individual differences, and further improves the accuracy and specificity of the assessment.

Description

Technical Field

[0001] This invention relates to the field of autism risk assessment technology, and more specifically, to an autism assessment method and system based on near-infrared spectral imaging and eye-tracking. Background Technology

[0002] Autism, a neurodevelopmental disorder, is characterized by impaired social communication, repetitive and stereotyped behaviors, and restricted interests. Onset typically occurs in infancy, and failure to identify and intervene early can have long-term impacts on a patient's social development. Current clinical assessment of autism relies on traditional methods centered on behavioral scales. However, these methods have significant limitations: firstly, assessment results are highly dependent on the subjective experience of healthcare professionals and are easily influenced by factors such as the patient's testing status and environmental interference, leading to low consistency among assessors; secondly, behavioral scales only capture outward manifestations and cannot explain the pathological mechanisms of autism from a neurophysiological perspective, making it difficult to distinguish the physiological roots behind behavioral abnormalities or identify risks before very early behavioral symptoms fully manifest.

[0003] Furthermore, existing autism assessment protocols based on multimodal data fusion often remain at the level of simply stitching together data after parallel acquisition, without establishing a correlation analysis model between neurophysiological and behavioral signals. For example, some protocols only record cerebral blood oxygenation data and eye movement data simultaneously, without exploring the functional coupling, temporal synchronization, and other correlation features between the two. This results in the data value not being fully utilized, and the sensitivity and specificity of the assessment failing to meet clinical needs. At the same time, due to significant individual differences in developmental levels and symptom presentation among different patients, existing protocols lack personalized testing and data processing mechanisms, often using a uniform testing paradigm to collect data. This leads to the masking of specific pathological characteristics in some patients, further reducing the accuracy of the assessment. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for autism assessment using near-infrared spectral imaging and eye-tracking.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for autism assessment using near-infrared spectral imaging and eye-tracking, comprising the following steps: The process involves collecting data on the age, developmental level, and medical history of suspected autism patients to generate autism patient profiles, establishing personal information files, collecting near-infrared spectral imaging data and eye-tracking data of patients, and screening abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated. Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extract the feature quantification parameters of abnormal associations from the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Based on the feature quantification parameters and intensity decay law of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results. The diagnostic assessment results are compared with the diagnostic criteria and standards for autism, and the final autism assessment report for the patient is generated.

[0006] Preferably, the following steps are taken: collecting the age, developmental level, and past medical history of suspected autism patients to establish personal information files; collecting near-infrared spectral imaging data and eye-tracking data of the patients; and screening abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated. The age, developmental level, and past medical history of suspected autism patients are collected to construct personal information files. Individualized testing plans are matched based on the personal information files. The individualized testing plans include social cognition tests, emotion recognition tests, and attention allocation tests. The brain function activation data and eye movement behavior pattern data of near-infrared spectral imaging data and eye movement tracking data of patients in the individualized testing protocol are collected to generate an individualized assessment dataset; Based on the individualized assessment dataset, abnormal correlation data reflecting potential pathological characteristics of autism are extracted and marked as correlation data to be assessed.

[0007] Preferably, based on functional coupling characteristics and time synchronization, it is determined whether it is necessary to perform scheme fusion processing on the correlation data to be evaluated for different test schemes, and the corresponding correlation data to be evaluated is marked as core evaluation correlation data. Specifically, this includes the following steps: If the functional coupling characteristics of the same type of anomalies are in the evaluation data of a single test plan, then the evaluation data of that test plan is directly marked as the core evaluation data. If the functional coupling characteristics of the same type of anomaly are scattered across the evaluation data of at least two test schemes, then the evaluation data of each scheme is segmented and extracted according to the test scheme type to obtain scheme data fragments. Based on the time synchronization of abnormal correlations, characteristic synchronization markers in the data segments of each scheme are identified. The time of each scheme data segment is calibrated according to the characteristic synchronization markers to obtain calibration scheme data. The data of each calibration scheme are then fused to form core evaluation correlation data.

[0008] Preferably, calibration scheme data is obtained by performing time calibration on the data segments of each scheme according to the feature synchronization mark, and the data of each calibration scheme are fused to form core evaluation correlation data, specifically including the following steps: Calculate the time difference of feature synchronization labels in at least two data segments; If the time difference is greater than the preset time difference threshold, the corresponding scheme data segment is marked as the scheme data to be calibrated. Based on the feature synchronization mark, the deviation of the scheme data to be calibrated is corrected to obtain the calibration scheme data. The calibration scheme data is then fused to form the core evaluation correlation data. If the time difference is less than or equal to the preset time difference threshold, the data fragments of each scheme will be directly fused to form core evaluation associated data based on the correspondence of the feature synchronization tags.

[0009] Preferably, based on the characteristic quantification parameters and intensity decay laws of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results, specifically including the following steps: Based on the feature quantification parameters and intensity decay law of abnormal associations, it is determined whether there is a cascading effect of abnormal associations in the core assessment association data. If there is no cascading effect in the abnormal associations in the core assessment data, the reliability of the first diagnosis of autism based on the abnormal association type and feature quantification parameters, and the corresponding core diagnostic dimensions are used to form the first diagnostic judgment result. By comparing the feature quantification parameters of each anomaly association with the association parameters and the diagnostic reliability table, the basic reliability and initial diagnostic dimension corresponding to each anomaly association are obtained. Based on the confidence level of the basic reliability, the initial diagnostic dimension is divided into high-confidence diagnostic dimension and low-confidence diagnostic dimension; Calculate the overall diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension, and combine it with the corresponding diagnostic dimension to form the first diagnostic judgment result; If there is a cascading effect of abnormal associations in the core assessment data, the abnormal associations are calculated based on the cascading pattern and feature quantification parameters to form a second diagnostic judgment result for the reliability of the second diagnosis of autism and the corresponding extended diagnostic dimensions. The diagnostic assessment result is obtained by integrating the results of the first and second diagnostic assessments.

[0010] Preferably, based on the characteristic quantification parameters and intensity decay law of abnormal associations, it is determined whether there is a cascading effect of abnormal associations in the core assessment association data, specifically including the following steps: If there is only a single functional type of abnormal association in the core assessment data, it is determined that the abnormal association does not have a cascading effect. If there are at least two types of abnormal associations in the core assessment data, calculate the parametric correlation index and intensity decay synergy value of the abnormal associations of different functional types. The parameter correlation index is compared with a preset correlation threshold, and the intensity attenuation synergy value is compared with a preset attenuation synergy threshold. If the parameter correlation index is lower than the preset correlation threshold and the intensity attenuation synergy value is lower than the preset attenuation synergy threshold, then it is determined that the abnormal correlation does not have a cascading effect. If the parameter correlation index is higher than or equal to the preset correlation threshold, or the intensity attenuation synergy value is higher than or equal to the preset attenuation synergy threshold, then the abnormal correlation is determined to have a cascading effect.

[0011] Preferably, the comprehensive diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension is calculated, and the first diagnostic judgment result is formed by combining the corresponding diagnostic dimensions. This specifically includes the following steps: Assign a first weight value to the high-confidence diagnostic dimension and a second weight value to the low-confidence diagnostic dimension, wherein the first weight value is greater than the second weight value; The high-confidence weighted total reliability is obtained by summing the products of the basic reliability of the high-confidence diagnostic dimension and the first weight value. The low-confidence weighted total reliability is obtained by summing the products of the base reliability of the low-confidence diagnostic dimension and the second weight value. The overall diagnostic reliability is obtained by adding the high-confidence weighted total reliability and the low-confidence weighted total reliability. The first diagnostic judgment result is formed by combining the high-confidence diagnostic dimension and the low-confidence diagnostic dimension.

[0012] Preferably, the abnormal associations are calculated based on the cascading patterns and feature quantification parameters to form a second diagnostic judgment result for the reliability of the autism diagnosis and the corresponding extended diagnostic dimensions. This specifically includes the following steps: Identify cascading patterns of abnormal associations, including positive cascading patterns and negative cascading patterns; If the abnormal association presents a positive cascading pattern, the reliability of the cascading diagnosis is calculated based on the feature quantification parameters and positive cascading coefficient of the abnormal association of each functional type, the core extended diagnostic dimension corresponding to the positive cascading is determined, and a second diagnostic judgment result is formed. If the abnormal association presents a negative cascading pattern, the reliability of cascading cancellation is calculated based on the feature quantification parameters of the abnormal association of each functional type and the negative cascading coefficient, the secondary extended diagnostic dimension corresponding to the negative cascading is determined, and a second diagnostic judgment result is formed.

[0013] A near-infrared spectral imaging and eye-tracking autism assessment system includes: Information collection module: Collects the age, developmental level and past medical history of suspected autism patients to generate suspected autism patient profiles, establishes personal information files, collects near-infrared spectral imaging data and eye-tracking data of patients, and screens abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated; Extraction Module 1: Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extraction Module 2: Extract the feature quantification parameters of abnormal associations in the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Judgment Module: Based on the feature quantification parameters and intensity decay law of abnormal associations, the module judges the cascading effect of abnormal associations in the core assessment association data and generates diagnostic assessment results. Output module: Compares the diagnostic assessment results with the diagnostic criteria and standards for autism, and outputs the patient's final autism assessment report.

[0014] Compared with existing technologies, this invention has the following beneficial effects: By integrating neurophysiological data from near-infrared spectral imaging with behavioral data from eye-tracking, an assessment system centered on objective quantitative indicators of functional coupling characteristics and temporal synchronization is constructed. This reduces interference caused by environment and testing conditions, improves the consistency of assessment results among different assessors, and enhances the objectivity and reliability of the assessment. By mining abnormal correlation features between near-infrared spectral imaging data and eye-tracking data, a correlation system between neurophysiology and external behavior is established. This elucidates the neurophysiological roots of behavioral abnormalities and enables risk identification when behavioral symptoms are not fully manifested, improving the accuracy of identification in young suspected cases. Through the fusion judgment mechanism of functional coupling characteristics and temporal synchronization, and the quantitative analysis of cascading effects, the correlation value of multimodal data is fully explored, improving the sensitivity and specificity of the assessment and meeting the clinical demand for high-precision assessment. By collecting patient age and developmental level information to match personalized testing plans and performing differentiated data processing for abnormal correlations of different functional types, the specific pathological characteristics of patients are effectively captured, avoiding the problem of uniform testing paradigms masking individual differences, and further improving the accuracy and targeting of the assessment. Attached Figure Description

[0015] Figure 1 This is a schematic diagram illustrating the steps of an autism assessment method based on near-infrared spectral imaging and eye tracking, as provided in an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the steps in obtaining diagnostic assessment results in an autism assessment method based on near-infrared spectral imaging and eye tracking, as provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of a near-infrared spectral imaging and eye-tracking autism assessment system provided in an embodiment of the present invention. Detailed Implementation

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0018] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0019] Reference Figures 1-3 As shown.

[0020] This embodiment further illustrates the autism assessment method and system based on near-infrared spectral imaging and eye tracking proposed in this invention.

[0021] A method for autism assessment using near-infrared spectral imaging and eye-tracking, comprising the following steps: The process involves collecting data on the age, developmental level, and medical history of suspected autism patients to generate autism patient profiles, establishing personal information files, collecting near-infrared spectral imaging data and eye-tracking data of patients, and screening abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated. Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extract the feature quantification parameters of abnormal associations from the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Based on the feature quantification parameters and intensity decay law of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results. The diagnostic assessment results are compared with the diagnostic criteria and standards for autism, and the final autism assessment report for the patient is generated.

[0022] The diagnostic criteria for autism refer to the currently used clinical diagnostic standards, such as the diagnostic entries for autism spectrum disorders in the DSM-5 Diagnostic and Statistical Manual of Mental Disorders and the ICD-11 International Classification of Diseases. These standards clarify the core characteristics of autism, such as social communication impairment, repetitive and stereotyped behaviors, and severity classification.

[0023] During the comparison process, the system will compare the quantitative characteristics corresponding to the preliminary diagnostic assessment results (such as the duration of abnormal associations corresponding to the severe effects of cascade effects and the overlap rate of activation areas) with the core characteristics in the diagnostic criteria one by one. For example, if the preliminary assessment is highly suspected of autism, it is necessary to verify the abnormal neurophysiological and behavioral associations corresponding to the result, such as whether the low overlap rate of prefrontal activation and eye movement fixation under social stimulation matches the core diagnostic item of persistent deficit in social interaction and communication in DSM-5. At the same time, the matching degree is calculated. If the matching degree reaches 80% or above, it means that the preliminary assessment result is highly consistent with the core characteristics of the authoritative standard, and the assessment result can be directly confirmed. If the matching degree is less than 80%, it is necessary to review the patient's personal information file and combine it with their developmental level, such as the degree of social developmental delay assessed by the Gesell scale, past medical history, such as family history of neurodevelopmental disorders, to correct the preliminary assessment result and avoid misjudgment due to local bias of multimodal data.

[0024] After matching and correction, the system integrates all information and outputs the final autism assessment report. The autism assessment report not only includes the patient's basic information such as age and developmental level, but also presents multimodal data characteristics from near-infrared spectroscopy imaging and eye tracking, such as the duration of abnormal associations, the value of activation area overlap rate, the analysis process of cascade effect, and the matching results with authoritative diagnostic standards. This allows clinicians to see objective quantitative data and understand the basis of the assessment conclusions, thus providing a comprehensive and reliable reference for the development of subsequent intervention plans.

[0025] The process involves collecting information on the age, developmental level, and medical history of suspected autistic patients to establish personal information files. Near-infrared spectral imaging data and eye-tracking data are also collected. Abnormal correlation data related to potential pathological features of autism are then screened to obtain correlation data to be evaluated. This process includes the following steps: The age, developmental level, and past medical history of suspected autism patients are collected to construct personal information files. Individualized testing plans are matched based on the personal information files. The individualized testing plans include social cognition tests, emotion recognition tests, and attention allocation tests. The brain function activation data and eye movement behavior pattern data of near-infrared spectral imaging data and eye movement tracking data of patients in the individualized testing protocol are collected to generate an individualized assessment dataset; Based on the individualized assessment dataset, abnormal correlation data reflecting potential pathological characteristics of autism are extracted and marked as correlation data to be assessed.

[0026] We collect basic information on suspected autism patients, including their age, developmental level, and medical history. This information is used to create a personal profile for each patient. Because each patient's situation is different, we need to match a personalized testing plan based on their profile. This personalized testing plan includes social cognition testing, emotion recognition testing, and attention allocation testing. For example, if the patient is a young child with a developmental level indicating weak social interaction skills, we will focus on social cognition testing, showing them materials involving facial interactions. If the patient shows signs of delayed emotional perception, we will add emotion recognition testing, displaying materials showing facial expressions of different emotions.

[0027] During the patient's participation in the personalized testing program, brain function activation data and eye movement behavior pattern data are collected simultaneously. Brain function activation data is obtained through existing near-infrared spectroscopy imaging technology. For example, when the patient takes a social cognition test, changes in blood oxygen concentration in the prefrontal cortex, superior temporal gyrus, and other brain regions related to social function are monitored to reflect the brain's activation state. Eye movement behavior pattern data is recorded through existing eye-tracking technology. For example, the patient's fixation point position, fixation duration, and saccade trajectory information when viewing emotional expression materials. The brain function activation data and eye movement behavior pattern data are integrated together to form a personalized assessment dataset for the patient.

[0028] Based on personalized assessment datasets, abnormal correlation data that reflect potential pathological characteristics of autism were screened and marked as correlation data to be evaluated. Specifically, the correlation between brain function activation and eye movement behavior in normal individuals under the same testing protocol was compared. If the patient's data deviated from the normal pattern, it was considered abnormal correlation data. For example, in social cognition tests, normal individuals tend to focus more on the eye area when viewing facial images, and the corresponding superior temporal gyrus shows higher activation. However, suspected autistic patients may avoid focusing on the eye area, and the activation level of the superior temporal gyrus is significantly lower. This mismatch between brain activation and eye movement behavior was extracted as correlation data to be evaluated.

[0029] Based on functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the correlation data to be evaluated for different test schemes, and mark the corresponding correlation data to be evaluated as core evaluation correlation data. The specific steps include: If the functional coupling characteristics of the same type of anomalies are in the evaluation data of a single test plan, then the evaluation data of that test plan is directly marked as the core evaluation data. If the functional coupling characteristics of the same type of anomaly are scattered across the evaluation data of at least two test schemes, then the evaluation data of each scheme is segmented and extracted according to the test scheme type to obtain scheme data fragments. Based on the time synchronization of abnormal correlations, characteristic synchronization markers in the data segments of each scheme are identified. The time of each scheme data segment is calibrated according to the characteristic synchronization markers to obtain calibration scheme data. The data of each calibration scheme are then fused to form core evaluation correlation data.

[0030] Functional coupling characteristics refer to the mutual modulation strength between brain functional signals from near-infrared spectroscopy imaging and behavioral signals from eye-tracking. For example, in social cognition tests, the degree of influence of brain region activation signals on eye-tracking fixation signals is a manifestation of functional coupling characteristics. Temporal synchronization refers to the key characteristics of these two types of signals, such as the degree of time matching of signal peaks. The smaller the time difference between the peak of brain region activation and the change in eye-tracking fixation focus, the higher the temporal synchronization.

[0031] First, determine the distribution of functional coupling characteristics associated with the same type of abnormality. If the functional coupling characteristics corresponding to the abnormal association only appear in the data to be evaluated in a single test scheme, such as the abnormality of low association between brain region activation and eye movement fixation, which is only reflected in the data to be evaluated in the social cognition test, then the data to be evaluated under that test scheme is directly marked as the core evaluation data.

[0032] If the functional coupling characteristics of similar abnormal associations are scattered across the data to be evaluated in at least two test schemes, such as the low association between brain region activation and eye-tracking fixation appearing simultaneously in the data to be evaluated in both social cognition and emotion recognition tests, then it is necessary to first extract the data to be evaluated for each scheme according to the type of test scheme to obtain the corresponding scheme data fragments. For example, extract the time period data corresponding to the abnormal association from the social cognition test data as the social cognition scheme data fragment; extract the corresponding time period data from the emotion recognition test data as the emotion recognition scheme data fragment.

[0033] Based on the temporal synchronization of anomalous correlations, feature synchronization markers are identified in the data segments of each scheme. Specifically, firstly, the time anchor points of brain functional signals and eye-movement behavioral signals in different scheme data segments are found. For example, the time point of peak brain region activation in the social cognition scheme data segment and the time point of peak activation in the same brain region in the emotion recognition scheme data segment can both serve as feature synchronization markers. Then, based on these feature synchronization markers, the time dimensions of each scheme data segment are time-calibrated to ensure consistency, resulting in calibrated scheme data.

[0034] Finally, these calibrated scheme data are merged to form core evaluation correlation data. For example, the calibrated social cognition scheme data and emotion recognition scheme data are concatenated along the time dimension, while retaining the functional coupling characteristics of the two types of data. The resulting integrated data is the core evaluation correlation data, which can be used for subsequent feature extraction and evaluation analysis.

[0035] Based on the feature synchronization markers, time calibration is performed on the data segments of each scheme to obtain calibration scheme data. The data of each calibration scheme are then fused to form core evaluation correlation data, which specifically includes the following steps: Calculate the time difference of feature synchronization labels in at least two data segments; If the time difference is greater than the preset time difference threshold, the corresponding scheme data segment is marked as the scheme data to be calibrated. Based on the feature synchronization mark, the deviation of the scheme data to be calibrated is corrected to obtain the calibration scheme data. The calibration scheme data is then fused to form the core evaluation correlation data. If the time difference is less than or equal to the preset time difference threshold, the data fragments of each scheme will be directly fused to form core evaluation associated data based on the correspondence of the feature synchronization tags.

[0036] Identifying key features for synchronization is crucial for understanding brain function and eye movement signals across different data segments. Examples include peak brain activation and the time points corresponding to eye focus switching. The time difference between synchronization markers of the same type of feature should be calculated across at least two data segments. For instance, taking the peak brain activation time point T1 from the social cognition test data segment and the peak brain activation time point T2 from the emotion recognition test data segment, the time difference ΔT = |T1 - T2| can be calculated.

[0037] The time difference ΔT is compared with a preset time difference threshold T0. The preset time difference threshold T0 is determined based on the fluctuation range of the characteristic synchronous marking time of normal people under multiple testing schemes, and is usually set to 0.5 seconds.

[0038] If the calculated time difference ΔT is greater than the preset time difference threshold T0, it indicates a significant deviation in the time dimension between the two data segments, requiring data calibration. In this case, the data segment with the larger time dimension deviation is marked as the data segment to be calibrated. Then, deviation correction is applied based on the feature synchronization marker: for example, if the feature synchronization marker time of the data segment to be calibrated is 0.8 seconds later than the target time, all timestamps of that data segment are uniformly subtracted by 0.8 seconds. After correction, the calibrated data is obtained. Subsequently, the calibrated data is concatenated with other data segments along the time dimension to form the core evaluation correlation data.

[0039] If the time difference ΔT is less than or equal to the preset time difference threshold T0, it indicates that the temporal dimension matching degree of the data segments from different schemes is high, and no additional calibration is required. At this time, the data segments of each scheme are directly integrated in chronological order based on the correspondence of the feature synchronization tags. For example, the brain function signals and eye movement behavior signals corresponding to the timestamps in the data segments of the social cognition test and emotion recognition test are spliced ​​together one by one to form the core assessment correlation data.

[0040] For example, the feature synchronization marking time for the social cognition test is 10 seconds after the test starts, and the feature synchronization marking time for the emotion recognition test is 10.3 seconds after the test starts. The preset time difference threshold is 0.5 seconds. In this case, ΔT = 0.3 seconds ≤ 0.5 seconds, and the data from the two schemes are directly fused according to their timestamps. If the feature synchronization marking time for the emotion recognition test is 10.6 seconds, and ΔT = 0.6 seconds > 0.5 seconds, then the timestamps of the data segments from the emotion recognition test are uniformly reduced by 0.6 seconds, calibrated, and then fused with the data from the social cognition test.

[0041] Based on the characteristic quantification parameters and intensity decay laws of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results, specifically including the following steps: Based on the feature quantification parameters and intensity decay law of abnormal associations, it is determined whether there is a cascading effect of abnormal associations in the core assessment association data. If there is no cascading effect in the abnormal associations in the core assessment data, the reliability of the first diagnosis of autism based on the abnormal association type and feature quantification parameters, and the corresponding core diagnostic dimensions are used to form the first diagnostic judgment result. By comparing the feature quantification parameters of each anomaly association with the association parameters and the diagnostic reliability table, the basic reliability and initial diagnostic dimension corresponding to each anomaly association are obtained. Based on the confidence level of the basic reliability, the initial diagnostic dimension is divided into high-confidence diagnostic dimension and low-confidence diagnostic dimension; Calculate the overall diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension, and combine it with the corresponding diagnostic dimension to form the first diagnostic judgment result; If there is a cascading effect of abnormal associations in the core assessment data, the abnormal associations are calculated based on the cascading pattern and feature quantification parameters to form a second diagnostic judgment result for the reliability of the second diagnosis of autism and the corresponding extended diagnostic dimensions. The diagnostic assessment result is obtained by integrating the results of the first and second diagnostic assessments.

[0042] The core assessment focuses on the cascading effect in correlated data. A cascading effect refers to a chain reaction where one abnormal correlation triggers abnormalities in other correlations. For example, low correlation between brain region activation and eye fixation (the first abnormality) can further lead to accelerated signal intensity decay during emotion recognition (the second abnormality). This chain reaction is called the cascading effect. The basis for determining the existence of this effect is the combination of feature quantification parameters: correlation duration, activation region overlap rate, intensity change slope, and intensity decay pattern. If the parameter changes of multiple abnormal correlations show a sequential and interconnected trend, a cascading effect is determined to exist; otherwise, it does not exist.

[0043] If no cascading effect is determined, firstly, the feature quantification parameters of each abnormal association are compared with the preset association parameters and diagnostic reliability table. The diagnostic reliability table is established based on a large amount of sample data from normal people and autistic patients. For example, the basic reliability corresponding to an association duration of more than 30 seconds is 0.7, and the basic reliability corresponding to an activation region overlap rate of less than 20% is 0.8. At the same time, the corresponding initial diagnostic dimensions are matched, such as the former corresponding to social interaction deficits and the latter corresponding to cognitive processing abnormalities. Thus, the basic reliability and initial diagnostic dimensions of each abnormal association are obtained.

[0044] Then, based on the confidence level of the basic reliability, the initial diagnostic dimension is usually classified as a high-confidence diagnostic dimension if the basic reliability is ≥0.8, and if it is <0.8, it is classified as a low-confidence diagnostic dimension.

[0045] Next, the overall diagnostic reliability is calculated using the formula: Overall Reliability = (Mean Reliability of High Confidence Dimensions × 0.7) + (Mean Reliability of Low Confidence Dimensions × 0.3). The weights are set according to the confidence level of the dimensions, and then combined with the corresponding diagnostic dimensions to form the first diagnostic judgment result.

[0046] If a cascading effect is determined, the cascading pattern of abnormal associations is first identified. For example, brain region activation, low eye movement fixation association, accelerated signal intensity attenuation, and interruption of emotion recognition association constitute a cascading pattern. Then, the reliability of the second diagnosis is calculated by combining feature quantification parameters. The base reliability of each abnormal association in the cascading pattern will be superimposed in a chain order. For example, if the base reliability of the first abnormality is 0.7 and the second is 0.8, then the reliability of the second diagnosis = 0.7 + (0.8 × 0.5) = 1.1. After normalization to the 0-1 interval, it is taken as 0.9. At the same time, the diagnostic dimensions are expanded accordingly. For example, the cascading pattern corresponds to multi-dimensional defects in social, cognitive, and emotional aspects, and finally, the second diagnostic judgment result is formed.

[0047] Finally, the first and second diagnostic results are fused. If only one result exists, that result is used directly. If both results exist, the result with higher overall reliability is taken as the main conclusion, while the dimensional information of the other result is supplemented to obtain a complete diagnostic assessment result. For example, in the patient's core assessment correlation data, there is only an anomaly of activation region overlap rate less than 20%, and no chain reaction, indicating no cascading effect. After comparing with the correlation parameter table, its basic reliability is 0.85, and the initial diagnostic dimension is social interaction deficit. Because the reliability is ≥0.8, it is classified as a high-confidence diagnostic dimension, with an overall reliability of 0.85, forming a first diagnostic result of social interaction deficit with a diagnostic reliability of 0.85. If the patient also has a cascading pattern of low activation region overlap rate and increased absolute value of intensity change slope, the second diagnostic reliability of the corresponding cascading pattern is 0.92, and the extended diagnostic dimension is a combined social and cognitive deficit. Then, the final diagnostic assessment result after fusion is a combined social and cognitive deficit with an overall diagnostic reliability of 0.92.

[0048] Based on the feature quantification parameters and intensity decay patterns of anomalous associations, the determination of whether anomalous associations in the core assessment data have a cascading effect includes the following steps: If there is only a single functional type of abnormal association in the core assessment data, it is determined that the abnormal association does not have a cascading effect. If there are at least two types of abnormal associations in the core assessment data, calculate the parametric correlation index and intensity decay synergy value of the abnormal associations of different functional types. The parameter correlation index is compared with a preset correlation threshold, and the intensity attenuation synergy value is compared with a preset attenuation synergy threshold. If the parameter correlation index is lower than the preset correlation threshold and the intensity attenuation synergy value is lower than the preset attenuation synergy threshold, then it is determined that the abnormal correlation does not have a cascading effect. If the parameter correlation index is higher than or equal to the preset correlation threshold, or the intensity attenuation synergy value is higher than or equal to the preset attenuation synergy threshold, then the abnormal correlation is determined to have a cascading effect.

[0049] Abnormal associations of functional types refer to abnormal association types classified according to functional dimensions. For example, abnormal associations of social cognition correspond to abnormal associations between brain function and eye movement behavior under social stimuli, while abnormal associations of emotion recognition correspond to abnormal signal associations in emotion material tests. Different functional types represent different pathological dimensions.

[0050] First, determine the number of abnormal associations in the core assessment data. If there are only abnormal associations of a single functional type, such as only social cognition-related abnormalities and no abnormal associations of other functional dimensions, then it is directly determined that there is no cascading effect of abnormal associations. Since the cascading effect requires the chain reaction between different associations, a single functional type cannot form a chain reaction.

[0051] If there are abnormal associations between at least two functional types, it is necessary to further calculate the parameter correlation index and intensity attenuation synergy value.

[0052] The parameter correlation index measures the degree of correlation between quantitative feature parameters that correlate anomalies of different functional types. The formula is: Parameter Correlation Index = (Correlation coefficient of association duration for different anomalies + Correlation coefficient of activation region overlap rate + Correlation coefficient of intensity change slope) / 3. For example, the correlation coefficient between the association duration of social cognition anomalies and the association duration of emotion recognition anomalies is 0.8, the correlation coefficient of activation region overlap rate is 0.7, and the correlation coefficient of intensity change slope is 0.6. Therefore, the parameter correlation index = (0.8 + 0.7 + 0.6) / 3 = 0.7.

[0053] The intensity decay synergy value measures the degree of synchronization in the intensity decay patterns of abnormal associations of different functional types. The calculation formula is: Intensity Decay Synergy Value = 1 - |decay coefficient of abnormal association A - decay coefficient of abnormal association B| / (decay coefficient of abnormal association A + decay coefficient of abnormal association B). For example, the decay coefficient for social cognition abnormalities is 0.5, and the decay coefficient for emotion recognition abnormalities is 0.6. Therefore, the intensity decay synergy value = 1 - |0.5 - 0.6| / (0.5 + 0.6) ≈ 0.91.

[0054] After calculating the parameter correlation index and intensity attenuation synergy value, it is necessary to make a double judgment with the preset threshold. The preset correlation threshold is usually set to 0.6, and the preset attenuation synergy threshold is usually set to 0.8.

[0055] If the parameter correlation index is below 0.6 and the intensity decay synergy value is below 0.8, it indicates low correlation and asynchronous decay patterns between abnormal associations of different functional types, and no cascading effect is determined. If the parameter correlation index is above or equal to 0.6, or the intensity decay synergy value is above or equal to 0.8, it indicates correlation or synchronous decay between abnormal associations of different functional types, and a cascading effect is determined. For example, if a patient's core assessment correlation data shows two abnormal associations—social cognition and emotion recognition—and the calculated parameter correlation index is 0.7 > 0.6, and the intensity decay synergy value is 0.85 > 0.8, then a cascading effect is determined for these two abnormal associations. If another patient's two indicators are 0.5 and 0.75 respectively, both below the threshold, then no cascading effect is determined.

[0056] The comprehensive diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension is calculated, and the first diagnostic judgment result is formed by combining the corresponding diagnostic dimensions. The specific steps include: Assign a first weight value to the high-confidence diagnostic dimension and a second weight value to the low-confidence diagnostic dimension, wherein the first weight value is greater than the second weight value; The high-confidence weighted total reliability is obtained by summing the products of the basic reliability of the high-confidence diagnostic dimension and the first weight value. The low-confidence weighted total reliability is obtained by summing the products of the base reliability of the low-confidence diagnostic dimension and the second weight value. The overall diagnostic reliability is obtained by adding the high-confidence weighted total reliability and the low-confidence weighted total reliability. The first diagnostic judgment result is formed by combining the high-confidence diagnostic dimension and the low-confidence diagnostic dimension.

[0057] High-confidence diagnostic dimensions are those whose baseline reliability reaches a preset high-confidence standard. The assessment data of these dimensions are more stable and have a higher degree of matching with the pathological characteristics of autism. Low-confidence diagnostic dimensions, on the other hand, are those whose baseline reliability does not reach the high-confidence standard, and their assessment data are relatively less stable.

[0058] Before calculating the overall diagnostic reliability, weight values ​​need to be assigned to the high-confidence diagnostic dimension and the low-confidence diagnostic dimension. Since the high-confidence diagnostic dimension has higher reference value, the first weight value assigned to it will be greater than the second weight value of the low-confidence diagnostic dimension. A common weight setting is a first weight value of 0.7, a second weight value of 0.3, and a total weight of 1, so as to ensure that the calculation results are within a reasonable range.

[0059] To calculate the weighted total reliability, first, calculate the high-confidence weighted total reliability by multiplying the baseline reliability of each high-confidence diagnostic dimension by the first weight value, and then summing all the products. For example, if a patient has two high-confidence diagnostic dimensions with baseline reliability of 0.85 and 0.9 respectively, then the high-confidence weighted total reliability = (0.85 × 0.7) + (0.9 × 0.7) = 1.225.

[0060] Then, calculate the low-confidence weighted total reliability by multiplying the baseline reliability of each low-confidence diagnostic dimension by the second weight value and summing the results. For example, if the patient has one low-confidence diagnostic dimension with a baseline reliability of 0.75, then the low-confidence weighted total reliability = 0.75 × 0.3 = 0.225.

[0061] Finally, the high-confidence weighted total reliability is added to the low-confidence weighted total reliability to obtain the comprehensive diagnostic reliability. Taking the above example, the comprehensive diagnostic reliability = 1.225 + 0.225 = 1.45.

[0062] After the calculation is completed, the pathological directions corresponding to the high-confidence and low-confidence diagnostic dimensions are combined. For example, the high-confidence dimension corresponds to social interaction deficits, and the low-confidence dimension corresponds to attention allocation abnormalities. This forms the initial diagnostic judgment. For instance, a comprehensive diagnostic reliability of 1.45 corresponds to suspected manifestations of both social interaction deficits and attention allocation abnormalities. For example, a patient has two high-confidence diagnostic dimensions with baseline reliability of 0.8 and 0.82 respectively, and a first weight value of 0.7; and one low-confidence diagnostic dimension with a baseline reliability of 0.7 and a second weight value of 0.3. Then, the high-confidence weighted total reliability = (0.8 × 0.7) + (0.82 × 0.7) = 1.134, the low-confidence weighted total reliability = 0.7 × 0.3 = 0.21, and the comprehensive diagnostic reliability = 1.134 + 0.21 = 1.344. Combining these with the corresponding diagnostic dimensions forms the initial diagnostic judgment.

[0063] Based on the cascading patterns and feature quantification parameters of abnormal associations, abnormal associations are calculated to form a second diagnostic determination result for the reliability of the autism diagnosis and the corresponding extended diagnostic dimensions. The specific steps include: Identify cascading patterns of abnormal associations, including positive cascading patterns and negative cascading patterns; If the abnormal association presents a positive cascading pattern, the reliability of the cascading diagnosis is calculated based on the feature quantification parameters of the abnormal association of each functional type and the positive cascading number, the core extended diagnostic dimension corresponding to the positive cascading is determined, and a second diagnostic judgment result is formed; among them, the association duration and activation region overlap rate in the feature quantification parameters of the abnormal association of each functional type are extracted, and the single-function diagnostic reliability of the abnormal association of each functional type is calculated. Based on the positive cascade connection number and the single-function diagnostic reliability quantity associated with anomalies of each functional type, the cascade diagnostic reliability calculation formula is as follows: , where R diag For the reliability of cascaded diagnostics, D i p represents the reliability quantity of the i-th single-function diagnosis, that is, the reliability measure of a single diagnostic function. i The cascade gain reliability value is represented by the i-th positive cascade correlation coefficient G, and the cascade diagnostic reliability is obtained. The cascade gain reliability value is calculated by combining the parameter correlation index and the intensity attenuation synergy value. If the abnormal association presents a negative cascading pattern, the reliability of cascading cancellation is calculated based on the feature quantification parameters of the abnormal association of each functional type and the negative cascading coefficient. The secondary extended diagnostic dimension corresponding to the negative cascading is determined to form the second diagnostic judgment result. In this process, the slope of intensity change and the duration of association in the feature quantification parameters of the abnormal association of each functional type are obtained to determine the negative weight coefficient of the abnormal association of each functional type. Based on the negative cascading correlation coefficient, negative weighting coefficient, and the basic reliability of abnormal associations for each functional type, the reliability calculation formula through cascading cancellation is as follows: R cascade For cascaded cancellation reliability, R i w represents the i-th basic reliability. i c represents the i-th negative weight coefficient. i Let represent the negative cascade correlation coefficient, and then obtain the cascade cancellation reliability.

[0064] First, we identify the cascading patterns of abnormal associations and divide them into positive and negative cascading patterns. Positive cascading patterns refer to the fact that abnormal associations of the preceding functional type amplify the pathological manifestations of abnormal associations of the following functional type. For example, abnormalities in social cognition amplify the degree of abnormalities in emotion recognition. Negative cascading patterns refer to the fact that abnormal associations of the preceding functional type weaken the pathological manifestations of abnormal associations of the following functional type. For example, abnormalities in attention allocation reduce the significance of abnormalities in social interaction.

[0065] If the abnormal associations exhibit a positive cascading pattern, the reliability of the cascading diagnosis needs to be calculated step by step, and the core extended diagnostic dimensions need to be determined. First, extract the association duration and activation region overlap rate from the feature quantification parameters of abnormal associations for each functional type, and calculate the single-function diagnostic reliability. The single-function diagnostic reliability is the diagnostic reliability corresponding to an abnormal association of a single functional type, and is usually determined by the association duration and activation region overlap rate. For example, the association duration of social cognition abnormalities is 40 seconds, and the activation region overlap rate is 15%. Referring to the association parameter table, its single-function diagnostic reliability is 0.8; the association duration of emotion recognition abnormalities is 35 seconds, and the activation region overlap rate is 18%, corresponding to a single-function diagnostic reliability of 0.75.

[0066] Next, the positive cascade connection number is determined. This coefficient measures the degree to which the associations of different functional types of anomalies reinforce each other in the positive cascade. For example, the positive cascade connection number for social cognitive anomalies to emotion recognition anomalies is 0.6, indicating that the former will increase the reliability of the latter by 60%. At the same time, the cascade gain reliability value G is calculated by combining the parameter correlation index and the intensity decay synergy value. The formula is G = (parameter correlation index + intensity decay synergy value) / 2. For example, if the parameter correlation index is 0.7 and the intensity decay synergy value is 0.85, then G = (0.7 + 0.85) / 2 = 0.775.

[0067] Next, the results are calculated using the cascade diagnostic reliability calculation formula: D i p is the reliability quantity of the i-th single-function diagnostic. i It is the number of the i-th positive cascade. For example, if there are two abnormal associations of different functional types, D1=0.8, p1=0.6, and D2=0.75, p2=0.5, the calculated reliability of the cascade diagnosis is 1.63.

[0068] Finally, the core extended diagnostic dimensions corresponding to the positive cascade are determined, such as the joint deficiency of social cognition and emotion recognition. Combined with the cascade diagnostic reliability of 1.63, a second diagnostic judgment result is formed.

[0069] If the abnormal associations exhibit a negative cascading pattern, the reliability of cascading cancellation needs to be calculated step-by-step, and the secondary extended diagnostic dimensions need to be determined. First, the slope of intensity change and the duration of association in the feature quantification parameters of abnormal associations for each functional type should be obtained to determine the negative weight coefficient. The negative weight coefficient measures the degree to which an abnormal association of a single functional type is weakened in the negative cascading. For example, if the absolute value of the slope of intensity change for an attention allocation anomaly is 0.9 and the duration of association is 25 seconds, the corresponding negative weight coefficient is 0.4; if the absolute value of the slope of intensity change for a social interaction anomaly is 0.8 and the duration of association is 30 seconds, the corresponding negative weight coefficient is 0.3.

[0070] Then, the negative cascade number is determined. This coefficient measures the degree to which the associations of different functional types of anomalies weaken each other in the negative cascade. For example, the negative cascade number of attention allocation anomalies to social interaction anomalies is 0.5, which means that the former will reduce the reliability of the latter by 50%.

[0071] Next, the results are calculated using the cascaded cancellation reliability calculation formula: , where R i It is the i-th basic reliability, w i It is the i-th negative weight coefficient, c iIt is the i-th negative cascade correlation number. For example, if the baseline reliability of patient attention allocation abnormalities is R1=0.8, w1=0.4, c1=0.5; and the baseline reliability of social interaction abnormalities is R2=0.7, w2=0.3, c2=0.4, then the cascade cancellation reliability R cascade It is 0.244.

[0072] Finally, secondary extended diagnostic dimensions corresponding to the negative cascade are determined, such as the weakening association between attention allocation abnormalities and social interaction deficits. Combined with the cascade cancellation reliability of 0.244, a second diagnostic judgment is formed. For example, if a patient presents a positive cascade pattern, including abnormalities in both social cognition and emotion recognition, with D1=0.8, p1=0.6, D2=0.75, p2=0.5, and G=0.775, the cascade diagnostic reliability R is calculated. diag =1.63, the core extended diagnostic dimension is the joint deficiency of social cognition and emotion recognition, and the final second diagnostic judgment result is a comprehensive cascade diagnostic reliability of 1.63, corresponding to the suspected manifestation of the joint deficiency of social cognition and emotion recognition.

[0073] A near-infrared spectral imaging and eye-tracking autism assessment system includes: Information collection module: Collects the age, developmental level and past medical history of suspected autism patients to generate suspected autism patient profiles, establishes personal information files, collects near-infrared spectral imaging data and eye-tracking data of patients, and screens abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated; Extraction Module 1: Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extraction Module 2: Extract the feature quantification parameters of abnormal associations in the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Judgment Module: Based on the feature quantification parameters and intensity decay law of abnormal associations, the module judges the cascading effect of abnormal associations in the core assessment association data and generates diagnostic assessment results. Output module: Compares the diagnostic assessment results with the diagnostic criteria and standards for autism, and outputs the patient's final autism assessment report.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for autism assessment using near-infrared spectral imaging and eye-tracking, characterized in that, The method includes the following steps: The process involves collecting data on the age, developmental level, and medical history of suspected autism patients to generate autism patient profiles, establishing personal information files, collecting near-infrared spectral imaging data and eye-tracking data of patients, and screening abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated. Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extract the feature quantification parameters of abnormal associations from the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Based on the feature quantification parameters and intensity decay law of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results. The diagnostic assessment results are compared with the diagnostic criteria and standards for autism, and the final autism assessment report for the patient is generated.

2. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 1, characterized in that, The process involves collecting information on the age, developmental level, and medical history of suspected autistic patients to establish personal information files. Near-infrared spectral imaging data and eye-tracking data are also collected. Abnormal correlation data related to potential pathological features of autism are then screened to obtain correlation data to be evaluated. This process includes the following steps: The age, developmental level, and past medical history of suspected autism patients are collected to construct personal information files. Individualized testing plans are matched based on the personal information files. The individualized testing plans include social cognition tests, emotion recognition tests, and attention allocation tests. The brain function activation data and eye movement behavior pattern data of near-infrared spectral imaging data and eye movement tracking data of patients in the individualized testing protocol are collected to generate an individualized assessment dataset; Based on the individualized assessment dataset, abnormal correlation data reflecting potential pathological characteristics of autism are extracted and marked as correlation data to be assessed.

3. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 2, characterized in that, Based on functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the correlation data to be evaluated for different test schemes, and mark the corresponding correlation data to be evaluated as core evaluation correlation data. The specific steps include: If the functional coupling characteristics of the same type of anomalies are in the evaluation data of a single test plan, then the evaluation data of that test plan is directly marked as the core evaluation data. If the functional coupling characteristics of the same type of anomaly are scattered across the evaluation data of at least two test schemes, then the evaluation data of each scheme is segmented and extracted according to the test scheme type to obtain scheme data fragments. Based on the time synchronization of abnormal correlations, characteristic synchronization markers in the data segments of each scheme are identified. The time of each scheme data segment is calibrated according to the characteristic synchronization markers to obtain calibration scheme data. The data of each calibration scheme are then fused to form core evaluation correlation data.

4. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 3, characterized in that, Based on the feature synchronization markers, time calibration is performed on the data segments of each scheme to obtain calibration scheme data. The data of each calibration scheme are then fused to form core evaluation correlation data, which specifically includes the following steps: Calculate the time difference of feature synchronization labels in at least two data segments; If the time difference is greater than the preset time difference threshold, the corresponding scheme data segment is marked as the scheme data to be calibrated. Based on the feature synchronization mark, the deviation of the scheme data to be calibrated is corrected to obtain the calibration scheme data. The calibration scheme data is then fused to form the core evaluation correlation data. If the time difference is less than or equal to the preset time difference threshold, the data fragments of each scheme will be directly fused to form core evaluation associated data based on the correspondence of the feature synchronization tags.

5. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 4, characterized in that, Based on the characteristic quantification parameters and intensity decay laws of abnormal associations, the cascading effect of abnormal associations in the core assessment association data is judged to generate diagnostic assessment results, specifically including the following steps: Based on the feature quantification parameters and intensity decay law of abnormal associations, it is determined whether there is a cascading effect of abnormal associations in the core assessment association data. If there is no cascading effect in the abnormal associations in the core assessment data, the reliability of the first diagnosis of autism based on the abnormal association type and feature quantification parameters, and the corresponding core diagnostic dimensions are used to form the first diagnostic judgment result. By comparing the feature quantification parameters of each anomaly association with the association parameters and the diagnostic reliability table, the basic reliability and initial diagnostic dimension corresponding to each anomaly association are obtained. Based on the confidence level of the basic reliability, the initial diagnostic dimension is divided into high-confidence diagnostic dimension and low-confidence diagnostic dimension; Calculate the overall diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension, and combine it with the corresponding diagnostic dimension to form the first diagnostic judgment result; If there is a cascading effect of abnormal associations in the core assessment data, the abnormal associations are calculated based on the cascading pattern and feature quantification parameters to form a second diagnostic judgment result for the reliability of the second diagnosis of autism and the corresponding extended diagnostic dimensions. The diagnostic assessment result is obtained by integrating the results of the first and second diagnostic assessments.

6. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 5, characterized in that, Based on the feature quantification parameters and intensity decay patterns of anomalous associations, the determination of whether anomalous associations in the core assessment data have a cascading effect includes the following steps: If there is only a single functional type of abnormal association in the core assessment data, it is determined that the abnormal association does not have a cascading effect. If there are at least two types of abnormal associations in the core assessment data, calculate the parametric correlation index and intensity decay synergy value of the abnormal associations of different functional types. The parameter correlation index is compared with a preset correlation threshold, and the intensity attenuation synergy value is compared with a preset attenuation synergy threshold. If the parameter correlation index is lower than the preset correlation threshold and the intensity attenuation synergy value is lower than the preset attenuation synergy threshold, then it is determined that the abnormal correlation does not have a cascading effect. If the parameter correlation index is higher than or equal to the preset correlation threshold, or the intensity attenuation synergy value is higher than or equal to the preset attenuation synergy threshold, then the abnormal correlation is determined to have a cascading effect.

7. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 5, characterized in that, The comprehensive diagnostic reliability of the high-confidence diagnostic dimension and the low-confidence diagnostic dimension is calculated, and the first diagnostic judgment result is formed by combining the corresponding diagnostic dimensions. The specific steps include: Assign a first weight value to the high-confidence diagnostic dimension and a second weight value to the low-confidence diagnostic dimension, wherein the first weight value is greater than the second weight value; The high-confidence weighted total reliability is obtained by summing the products of the basic reliability of the high-confidence diagnostic dimension and the first weight value. The low-confidence weighted total reliability is obtained by summing the products of the base reliability of the low-confidence diagnostic dimension and the second weight value. The overall diagnostic reliability is obtained by adding the high-confidence weighted total reliability and the low-confidence weighted total reliability. The first diagnostic judgment result is formed by combining the high-confidence diagnostic dimension and the low-confidence diagnostic dimension.

8. The autism assessment method based on near-infrared spectral imaging and eye-tracking according to claim 5, characterized in that, Based on the cascading patterns and feature quantification parameters of abnormal associations, abnormal associations are calculated to form a second diagnostic determination result for the reliability of the autism diagnosis and the corresponding extended diagnostic dimensions. The specific steps include: Identify cascading patterns of abnormal associations, including positive cascading patterns and negative cascading patterns; If the abnormal association presents a positive cascading pattern, the reliability of the cascading diagnosis is calculated based on the feature quantification parameters and positive cascading coefficient of the abnormal association of each functional type, the core extended diagnostic dimension corresponding to the positive cascading is determined, and a second diagnostic judgment result is formed. If the abnormal association presents a negative cascading pattern, the reliability of cascading cancellation is calculated based on the feature quantification parameters of the abnormal association of each functional type and the negative cascading coefficient, the secondary extended diagnostic dimension corresponding to the negative cascading is determined, and a second diagnostic judgment result is formed.

9. An autism assessment system based on near-infrared spectral imaging and eye-tracking, applied to the autism assessment method based on near-infrared spectral imaging and eye-tracking as described in claims 1-8, characterized in that, include: Information collection module: Collects the age, developmental level and past medical history of suspected autism patients to generate suspected autism patient profiles, establishes personal information files, collects near-infrared spectral imaging data and eye-tracking data of patients, and screens abnormal correlation data of potential pathological features of autism to obtain correlation data to be evaluated; Extraction Module 1: Extract the functional coupling characteristics, time synchronization, and intensity decay patterns of abnormal associations in the data to be evaluated. Based on the functional coupling characteristics and time synchronization, determine whether it is necessary to perform scheme fusion processing on the data to be evaluated from different test schemes, and mark the corresponding data to be evaluated as core evaluation data. Extraction Module 2: Extract the feature quantification parameters of abnormal associations in the core evaluation association data. The feature quantification parameters include association duration, activation region overlap rate, and intensity change slope. Judgment Module: Based on the feature quantification parameters and intensity decay law of abnormal associations, the module judges the cascading effect of abnormal associations in the core assessment association data and generates diagnostic assessment results. Output module: Compares the diagnostic assessment results with the diagnostic criteria and standards for autism, and outputs the patient's final autism assessment report.

Images