Children twitch disorder detection and evaluation system based on task normal form

Abstract

The invention discloses a child twitch disorder detection and evaluation system based on a task normal form, and belongs to the technical field of medical detection. The system comprises a task stimulation presentation and evaluation device which is used for executing a standardized Go / No-Go cognitive task and synchronously collecting behavior indexes of a subject; the near-infrared brain function imaging device is used for synchronously monitoring the blood oxygen concentration change of the key brain area of the prefrontal lobe of the tested brain during the execution period of the task; the data processing module is used for preprocessing the brain function signals and extracting physiological indexes from the brain function signals; and the result interpretation module is used for performing fusion analysis on the behavior indexes and the physiological indexes based on a quantitative interpretation model pre-established through clustering analysis, and automatically outputting a specific cognitive function subtype interpretation result. According to the method, objective and quantitative evaluation and typing of the core cognitive function defect of the twitch disorder are realized through multi-dimensional information fusion and standardized analysis of behavioral performance and cranial nerve activity.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of medical detection and neuroimaging technology, and particularly relates to a task paradigm-based detection and evaluation system for childhood tic disorders, and more particularly to a system combining Go / No-Go cognitive task paradigm and functional near-infrared spectroscopy technology for objectively evaluating attention and cognitive inhibition function defects of children with tic disorders. BACKGROUND

[0002] Tic disorders are common neurodevelopmental disorders in childhood, mainly manifested as involuntary, repetitive muscle tics or vocalizations. A large number of studies have shown that in addition to motor symptoms, children with tic disorders often have significant attention and cognitive inhibition function defects, especially in conflict monitoring and impulse control. The impairment of these core cognitive functions seriously affects their learning, socialization and quality of life.

[0003] Currently, clinical evaluation relies mainly on subjective methods such as doctor observation and parent questionnaire, and lacks objective and quantitative biological indicators. Go / No-Go task, as a classic cognitive paradigm, can effectively evaluate inhibition control and attention, but simple behavioral indicators (such as accuracy and reaction time) cannot directly reflect the neural activity state of the brain, making it difficult to distinguish different neural mechanisms (such as insufficient activation of the prefrontal cortex or excessive compensatory activation) behind the behavior.

[0004] Electroencephalogram technology can be used for brain function monitoring, but its spatial resolution is limited, and the signal is easily disturbed by the electromyographic artifacts generated by the involuntary movements of tic disorder patients, resulting in distorted data. Functional near-infrared spectroscopy technology has the advantages of strong anti-movement interference ability and good portability, and is suitable for children groups, but existing technologies are mostly limited to simple description of brain activation, lack of comprehensive analysis models that deeply integrate brain activation indicators and behavioral indicators, and lack of standardized quantitative interpretation standards, making it impossible to automatically and objectively interpret specific clinical subtypes from multi-dimensional data.

[0005] Therefore, there is an urgent need in the clinic for a detection and evaluation system that can integrate behavioral performance and neural activity of key brain regions, has objective and quantitative standards and automatic analysis capabilities, to improve the accuracy and reliability of tic disorder evaluation. SUMMARY

[0006] To solve the problem of strong subjectivity and difficulty in quantifying core pathological mechanisms in traditional evaluation methods (such as scales and behavior observation) for tic disorders, the present application provides a task paradigm-based detection and evaluation system for childhood tic disorders, which significantly improves the objectivity and accuracy of evaluation by combining classic cognitive neuroscience paradigms with objective brain function imaging technology.

[0007] According to an aspect of the present disclosure, a task paradigm-based detection and evaluation system for childhood tic disorders is provided, comprising: a task stimulus presentation device for evaluating the attention and cognitive inhibition ability of a subject through standardized cognitive tasks and recording the behavioral responses thereof; a near-infrared brain function imaging device for covering the core brain regions related to cognitive inhibition function through a multi-channel probe array and for synchronously monitoring the concentration changes of oxyhemoglobin and deoxyhemoglobin in the core brain regions during the execution of the cognitive tasks by using a dual-wavelength near-infrared spectroscopy technology; a data processing module connected to the task stimulus presentation device and the near-infrared brain function imaging device, respectively, for acquiring behavioral indicators from the task stimulus presentation device and pre-processing the raw light intensity signals collected by the near-infrared brain function imaging device to extract physiological indicators; a result interpretation module for taking the physiological indicators and behavioral indicators as inputs and outputting the subtype interpretation results of childhood tic disorders through standardized interpretation rules based on clustering analysis modeling.

[0008] As a further technical solution, the task stimulus presentation device executes a Go / No-Go cognitive paradigm by presenting Go task stimuli and No-Go task stimuli and records and quantifies the behavioral indicators of the subject, including the Go task accuracy, No-Go task accuracy and Go task reaction time.

[0009] As a further technical solution, the physiological indicators include the integral value, mean value and peak value extracted from the blood oxygen concentration-time curve.

[0010] As a further technical solution, the steps of executing interpretation by the result interpretation module include: determining the behavioral indicators and physiological indicators as "small", "medium" or "large" levels according to predetermined grade thresholds, respectively; based on the grade determination, outputting the subtype interpretation results according to a predetermined logic flow.

[0011] As a further technical solution, the determination threshold of the accuracy in the behavioral indicators is: the small level of Go task accuracy corresponds to a value ≤ 0.92, the medium level corresponds to 0.92 < value < 0.98, and the large level corresponds to a value ≥ 0.98; the small level of No-Go task accuracy corresponds to a value ≤ 0.92, the medium level corresponds to 0.92 < value < 0.98, and the large level corresponds to a value ≥ 0.98.

[0012] As a further technical solution, the determination threshold of the reaction time in the behavioral indicators is: The small level of the Go task reaction time corresponds to a value of 280 milliseconds or less, the medium level corresponds to a value of 280 milliseconds < value < 380 milliseconds, and the large level corresponds to a value of 380 milliseconds or more.

[0013] As a further technical solution, the determination threshold of the physiological indicators is: The small level of the integral value corresponds to a value of 0 or less, the medium level corresponds to a value of 0 < value < 6, and the large level corresponds to a value of 6 or more. The small level of the mean value corresponds to a value of 0 or less, the medium level corresponds to a value of 0 < value < 0.03, and the large level corresponds to a value of 0.03 or more. The small level of the peak value corresponds to a value of 0 or less, the medium level corresponds to a value of 0 < value < 0.08, and the large level corresponds to a value of 0.08 or more.

[0014] As a further technical solution, the predetermined logical flow includes: Step S1, task sensitivity screening: if the Go task accuracy and the No-Go task accuracy are both "small" levels and the Go task reaction time is "small" level, it is determined to be "task-insensitive type", and the process is terminated. Step S2, behavior performance determination: if it is not determined to be "task-insensitive type", when the Go task accuracy is "large" level, the No-Go task accuracy is "large" level, and the Go task reaction time is "medium" or "small" level, it is determined that the behavior performance is "high performance", otherwise it is determined to be "low performance"; Step S3, brain activation intensity determination: when the integral value is "large" level, and the peak value is "medium" or "large" level, and the mean value is "medium" or "large" level, it is determined that the brain activation is "high activation", otherwise it is determined to be "low activation"; Step S4, subtype comprehensive output: the determination results of behavior performance and brain activation intensity are combined to output low activation-high performance type, high activation-high performance type, low activation-low performance type, or high activation-low performance type.

[0015] As a further technical solution, the interpretation rule based on clustering analysis modeling in the result interpretation module is established by the following steps: A multi-dimensional feature vector containing standardized fNIRS brain function indicators, behavior indicators, and clinical scale data is constructed; A large amount of sample data is clustered and analyzed using the K-means clustering algorithm to determine the optimal cluster number and the feature center of each category; Combined with clinical diagnosis verification, each cluster category is assigned a corresponding subtype interpretation label and clinical significance.

[0016] As a further technical solution, it further comprises a visual report generation module for automatically generating and outputting a structured assessment report containing fNIRS waveform graph, brain activation topography, quantitative index table and clinical interpretation conclusion.

[0017] Compared with the prior art, the beneficial effects of the present application are that: 1. Data reliability and strong anti-interference capability: The fNIRS technology which is not sensitive to electromyography is adopted, which fundamentally avoids the signal pollution problem caused by the tic symptoms, ensures that high-quality brain function data can be obtained under the real pathological state, and lays a solid foundation for accurate evaluation.

[0018] 2. Standardized and objective interpretation is achieved: The quantitative interpretation framework integrating behavior and brain activation multi-dimensional indicators is created, and through clear numerical threshold and automatic decision logic, the experience-dependent fuzzy judgment is converted into repeatable objective classification, which greatly improves the consistency and comparability of detection results of different institutions and different time points, and promotes the standardization of diagnosis and treatment.

[0019] 3. Deepening of evaluation dimension, revealing neural mechanism: The limitation of single behavior evaluation is broken through, and through the "behavior-brain activation" two-dimensional model, individuals with similar surface behaviors but different internal neural mechanisms (such as "inefficient compensation" and "resource deficiency") can be distinguished. This provides an unprecedented neurocognitive perspective for the etiological exploration, subtype identification and personalized intervention of tic disorders.

[0020] 4. Humanized process guarantees data quality: According to the characteristics of children, a "practice-official" two-stage task and a simple guidance process are designed, which effectively reduces the anxiety and operation errors of children, ensures their full understanding and participation in the task, and thus guarantees the effectiveness and reliability of the collected data from the source.

[0021] 5. Clinical practicability and high efficiency are prominent: The whole system is highly automated, from data acquisition, processing, analysis to the generation of structured reports containing waveform graphs, topography, data tables and interpretation conclusions, without complex manual intervention throughout the process, greatly shortening the evaluation period, improving the clinical work efficiency, and easy to promote the use in medical institutions at all levels. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0023] Figure 1A detection process schematic diagram of a child tic disorder detection and evaluation system based on a task paradigm is provided for the embodiment of the present application.

[0024] Figure 2 A schematic diagram of a task stimulus presentation device is provided for the embodiment of the present application.

[0025] Figure 3 A result interpretation logic schematic diagram is provided for the embodiment of the present application.

[0026] Figure 4 A high activation low performance type schematic diagram is provided for the embodiment of the present application.

[0027] Figure 5 A high activation high performance type schematic diagram is provided for the embodiment of the present application.

[0028] Figure 6 A low activation low performance type schematic diagram is provided for the embodiment of the present application.

[0029] Figure 7 A low activation high performance type schematic diagram is provided for the embodiment of the present application.

[0030] Figure 8 A task insensitivity type schematic diagram is provided for the embodiment of the present application. DETAILED DESCRIPTION

[0031] The present application provides a child tic disorder detection and evaluation system based on a Go / No-Go task paradigm, which aims to provide objective basis for the auxiliary detection of tic disorders (Tic Disorders), the evaluation of core cognitive functions and the evaluation of rehabilitation efficacy. The system mainly comprises a task stimulus presentation device, a near-infrared brain function imaging device, a data processing module and a result interpretation module.

[0032] The core technology of the present application is that the system presents a Go / No-Go cognitive task of cartoon animal stimulation, requires the subject to quickly respond when the “Go” stimulus appears, and suppresses the response when the “No-Go” stimulus appears, so as to detect the attention and cognitive inhibition ability of the subject. While the subject performs the task and records the behavior indicators (such as reaction time and accuracy), the near-infrared brain function imaging device synchronously collects the blood oxygen concentration change signals of the key brain areas of the subject, especially the dorsolateral prefrontal cortex (DLPFC) and the frontal pole (FPA) cortex. Subsequently, the data processing module performs standardization preprocessing and feature extraction on the collected brain function signals to generate quantitative brain activation indicators (i.e. physiological indicators). Finally, the result interpretation module integrates the behavior indicators and the physiological indicators, and outputs a comprehensive evaluation report on the cognitive inhibition and attention function of the subject according to the built-in interpretation model.

[0033] The present application combines the classical cognitive neuroscience paradigm with objective brain function imaging technology, solves the problem of strong subjectivity and difficulty in quantifying the core pathological mechanism of the traditional evaluation method (such as scale, behavior observation) of tic disorders, and significantly improves the objectivity and accuracy of the evaluation. Therefore, the system has clear and wide application value: Clinical auxiliary diagnosis: can be used as an objective biological marker to provide auxiliary basis for clinical diagnosis of tic disorders in children and adolescents, and to quantitatively evaluate the severity of cognitive inhibition impairment.

[0034] High-risk population screening: can be widely used in hospitals, schools and medical centers, etc. to quickly screen the attention and impulse control ability of children, and early identify individuals at risk of tic disorders and related comorbidities.

[0035] Rehabilitation and treatment evaluation: can be used to quantitatively evaluate the improvement effect of different treatment schemes such as drug treatment and behavior intervention (such as habit reversal training) on the cognitive inhibition function of patients. By tracking the changes of brain function and behavior performance, reliable quantitative indicators are provided for developing personalized treatment schemes and dynamically adjusting intervention intensity.

[0036] The technical solutions of the embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0037] The present application provides a kind of children's tic disorder detection and evaluation system based on task paradigm, as shown in Figure 1 It mainly includes task stimulus presentation device, near-infrared brain function imaging device, data processing module and result interpretation module. The detection and evaluation process of the system is as shown in Figure 1 As follows: The task stimulus presentation device aims to accurately evaluate the attention and cognitive inhibition ability of the subject through standardized cognitive tasks.

[0038] Task design: the classic Go / No-Go paradigm is adopted, including Go task (requiring the subject to perform rapid key response to all appearing stimuli) and No-Go task (requiring the subject to only key target stimuli and inhibit response to non-target stimuli). The task flow is set to 3 Go task modules and 2 No-Go task modules, each module lasts for 27 seconds, and the total detection time is 135 seconds, to comprehensively evaluate the conflict monitoring, impulse control and selective and sustained attention ability of the subject under different conditions. The task design is as shown in Figure 2 .

[0039] Procedure control: To ensure the validity of data, the detection is divided into two stages: "practice task" and "formal check". The experimenter ensures that the subjects (especially children) fully understand the rules through standardized instructions, task demonstration and practice, and then enters the formal check procedure.

[0040] Behavior data collection: The system records and quantifies the behavior indicators of the subjects in real time through key interaction, mainly including reaction time, Go task accuracy and No-Go task accuracy, providing objective behavioral basis for subsequent comprehensive interpretation.

[0041] The detection principle of the near-infrared brain function imaging device is as follows. The module is used to monitor the neural activity of the key brain regions of the subjects in real time when performing the above cognitive tasks.

[0042] The system uses 690nm and 830nm dual-wavelength near-infrared spectroscopy technology to calculate the concentration changes of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) in the cerebral cortex in real time according to the modified Beer-Lambert law (MBLL). The selection of these two wavelengths is based on their unique optical properties for oxygenated hemoglobin and deoxygenated hemoglobin: the 690nm wavelength has high sensitivity to deoxygenated hemoglobin, while the 830nm wavelength is more sensitive to oxygenated hemoglobin. By transmitting light of these two wavelengths through the cerebral cortex and measuring the light intensity attenuation, the dynamic changes of the concentrations of the two types of hemoglobin can be effectively distinguished and quantified, thereby reflecting the activity level of the neural activity in the brain region.

[0043] The calculation method from optical density to hemoglobin concentration is as follows: The system uses the modified Beer-Lambert law (MBLL) and the optical density (OD) calculation formula to convert the measured light intensity change into the relative concentration change of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR). This method is the core of quantifying the changes in cerebral hemodynamics based on near-infrared spectroscopy technology.

[0044] Optical density (OD) is a physical quantity that measures the degree of attenuation of light penetrating through a specific medium, and its calculation formula is as follows: , Where: is the incident light intensity of the light source. is the exit light intensity received by the detector.

[0045] In practical applications, the relative change amount of optical density is of interest. Therefore, applying the above formula to the light intensity data at time t and time 0 can obtain the change amount of optical density : , The modified Beer-Lambert law establishes the relationship between the change of optical density and the change of concentration of the light-absorbing substance: , wherein: is the change of optical density. is the molar extinction coefficient of the light-absorbing substance, which is the light absorption capacity of the substance at a specific wavelength. is the change of concentration of the light-absorbing substance. L is the straight-line distance between the light source and the detector. DPF (Differential Pathlength Factor) is the differential path factor, which is used to correct the actual optical path of light in highly scattering biological tissues (such as brain tissue). G is the total light intensity attenuation that is independent of the change of hemoglobin concentration, for example, caused by head movement or probe pressure change.

[0046] In a short time, assuming that G remains unchanged, at time t and time 0, a system of two linear equations can be established by the change of optical density at two different wavelengths , and to solve the change of concentration of

[0047] Specifically, the construction and solution of the system of two linear equations are as follows: using 690 nm and 830 nm wavelengths, the following equation system can be obtained: At 690 nm wavelength: , At 830 nm wavelength: , wherein: and are the changes of optical density measured at 690 nm and 830 nm wavelengths, respectively. and are the molar extinction coefficients of oxyhemoglobin and deoxyhemoglobin at the corresponding wavelengths, respectively. These coefficients are known physical constants. and are variables to be solved, representing the relative concentration changes of oxyhemoglobin and deoxyhemoglobin.

[0048] By solving the above equation system, the changes of concentration of oxyhemoglobin (HbO) and deoxyhemoglobin (HbR) during task execution can be calculated, and thus the neural activity of the corresponding region of the cerebral cortex can be reflected.

[0049] In terms of probe placement and brain region selection, the system covers the core brain regions closely related to cognitive inhibition function through a multi-channel probe array (light source-detector distance of 2-3 cm, at least 8 groups of detection pairs). The main focus is on the dorsolateral prefrontal cortex (DLPFC) and the frontal pole (FP).

[0050] Dorsolateral prefrontal cortex (DLPFC): This brain region is considered a key node of high-level cognitive function and plays a core role in conflict detection and response inhibition. In the Go / No-Go task, it is responsible for identifying the conflict brought by the "No-Go" signal and initiating the inhibition program to prevent the individual from performing the preset "Go" response. Therefore, the activation pattern of this region is closely related to the cognitive inhibition deficits of children with Tourette syndrome.

[0051] Frontal pole (FP): As the most anterior part of the prefrontal cortex, the frontal pole is mainly involved in goal maintenance and behavior monitoring. It continuously monitors the individual's behavior performance during task execution to ensure that it meets the task goals and alerts when errors occur (such as responding to "No-Go" signals), providing information for subsequent corrective behavior.

[0052] By monitoring the blood oxygen activity of these two key brain regions, the system can quantify the cognitive inhibition ability of children with Tourette syndrome, providing objective neurophysiological indicators for the diagnosis and evaluation of the disease.

[0053] This module synchronously collects blood oxygen signals at a sampling rate of 5-10 Hz, and through flexible adaptive headbands and automatic calibration technology, it ensures the comfort of wearing and the quality of signals, maximally reduces the interference caused by children's head movements and other factors, and guarantees the reliability of data.

[0054] The data processing module deeply processes the original light intensity signals collected by the near-infrared brain function imaging device, aiming to extract effective neural activity features and quantify the cognitive inhibition function of children under the Go / No-Go task paradigm. The data processing module includes a signal preprocessing unit and a feature extraction unit.

[0055] The signal preprocessing unit converts the collected raw light intensity signals into optical density (OD) data and removes motion artifacts and physiological noise to ensure the accuracy of subsequent analysis.

[0056] a) Light intensity data conversion and motion artifact correction. First, the raw light intensity data collected and are converted into optical density (OD) data. The converted optical density data often contain motion artifacts caused by head movement. The system uses a combination of wavelet filtering and strip interpolation methods to identify and correct motion artifacts.

[0057] Wavelet Filtering: This method is based on wavelet transform to decompose the original optical attenuation signal into low-frequency and high-frequency components at different scales. Hemodynamic changes usually correspond to low-frequency components, while motion artifacts and high-frequency noise mainly correspond to high-frequency components. By thresholding the high-frequency wavelet coefficients, artifacts can be effectively removed. The mathematical representation of signal decomposition and reconstruction is as follows: Original optical attenuation signal can be decomposed into: , where: : original optical attenuation signal, i.e., optical density signal varying with time; : scaling function at the Jth layer, characterizing the low-frequency trend of the signal; : wavelet basis function at the jth layer, characterizing high-frequency details at different scales; : approximation coefficient, reflecting low-frequency components, mainly corresponding to hemodynamic changes; : detail coefficient, reflecting high-frequency components, including motion artifacts and high-frequency noise.

[0058] To eliminate artifacts, the system sets a threshold for abnormal detail coefficients , and obtains the corrected detail coefficients : , The corrected detail coefficients and the original low-frequency coefficients are used for inverse wavelet transform to obtain the corrected signal : , where is the signal after wavelet filtering.

[0059] Based on the results of wavelet filtering, the spline interpolation method is used to detect artifacts by quantifying signal fluctuations, and a smooth spline curve is used to replace the artifact data segment. The processing flow is as follows: Artifact detection: sliding time window is used on the optical density signal to calculate the standard deviation in the window. When exceeds the preset threshold T, the segment corresponding to the window is determined to have motion artifacts.

[0060] , where is the standard deviation of the ith sliding window, n is the window length, For the sample points within the window.

[0061] Spline interpolation: Smooth replacement of detected artifact segment data points with a cubic spline interpolation function : , where are the polynomial coefficients of the cubic spline interpolation.

[0062] Signal reconstruction: Final artifact-corrected signal The original signal is replaced by the smoothed spline function in the artifact segment , while the original data is kept outside the artifact.

[0063] b) Filter denoising. After artifact correction, the optical attenuation signal is band-pass filtered at 0.01-0.08 Hz to remove heartbeat (~1 Hz), respiration (~0.3 Hz) and other high-frequency noise, while eliminating low-frequency baseline drift, thus preserving the effective frequency band related to hemodynamic responses, improving signal quality and enhancing the reliability of subsequent analysis. The filter method used is a Butterworth type IIR digital band-pass filter. Its continuous domain transfer function is: , where n is the filter order, is the cutoff angular frequency. After discretization by bilinear transformation, the time-domain difference equation is obtained: , where : the discrete sequence of the original optical attenuation signal; : the filtered optical attenuation signal; : the filter coefficients calculated according to the set cutoff frequency .

[0064] This filtering step ensures that only the optical attenuation components within the 0.01-0.08 Hz range related to cerebral hemodynamics are preserved, effectively suppressing physiological noise and environmental interference.

[0065] c) Concentration conversion. The filtered optical density data are converted into the concentration changes of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) using the Modified Beer-Lambert Law (MBLL).

[0066] The feature extraction unit is responsible for extracting key physiological and behavioral feature indicators reflecting children's cognitive inhibition function from the preprocessed blood oxygen concentration data.

[0067] a) Task data block. Take the 13 seconds before the start of the Go / No-Go task as the task baseline, and the 27 seconds after the start of the task as an analysis block, i.e. the analysis time range is -13 seconds to 27 seconds.

[0068] b) Physiological index extraction. Take 13 seconds before the start of the task as the baseline and 27 seconds after the start of the task as the analysis block. For the blood oxygen concentration change curve of each target brain area (such as left / right dorsolateral prefrontal cortex, left / right frontal pole), the following physiological indexes are calculated: Integral value: Calculate the positive area under the concentration-time curve during the 0-27 second task period, reflecting the overall activation intensity of the brain area.

[0069] Mean value: Calculate the mean value of the blood oxygen concentration change during the task period, reflecting the sustained activation level.

[0070] Peak value: Calculate the maximum value of the blood oxygen concentration change during the task period (relative to the baseline), reflecting the maximum transient activation degree.

[0071] Preferably, the slope, time center of gravity and other indicators can also be calculated as supplements.

[0072] c) Behavioral index extraction. The system also synchronously records and calculates the following behavioral indexes to comprehensively evaluate the cognitive inhibition ability of children: Go task accuracy: The proportion of correct responses to Go task stimuli.

[0073] No-Go task accuracy: The proportion of correct inhibition responses to No-Go task stimuli.

[0074] Reaction time: In the Go task, the time from the appearance of the stimulus to the effective response.

[0075] The comprehensive analysis of these physiological and behavioral indexes provides comprehensive, multi-dimensional objective data support for the detection and evaluation of children's tic disorders.

[0076] The result interpretation and report module takes multi-source data (near-infrared brain function data, behavior data, clinical scale data) as input, realizes objective interpretation and visual report generation of children's tic disorders through standardized data processing, clustering analysis modeling and clinical correlation verification, and the core technology includes data foundation construction, clustering analysis model, subtype interpretation rules and report output four parts. Specifically, it includes: Data foundation and preprocessing specification: Database construction. 6000 cases of Go / No-Go task data were collected from multiple children's hospitals, including 2000 cases of children with TD and ADHD, and 4000 cases of healthy or other developmental disease children.

[0077] The database contains three types of core data: fNIRS brain function indicators: After preprocessing, the concentration change parameters of hemoglobin (HbO) and deoxyhemoglobin (HbR) in left / right dorsolateral prefrontal cortex (DLPFC) and left / right frontal pole (FP) include: integral value: the area under the concentration-time curve during the task, reflecting the overall activation intensity of the brain region; peak value: the maximum concentration difference from baseline during the task, representing the maximum transient activation level; mean value (Δ ): the average concentration difference from baseline during the task, representing the sustained activation level; Behavioral indicators: Go task accuracy (ACC ), No-Go task accuracy (ACC ), Go task reaction time (RT ); Clinical scale data: Conners Children's Behavior Rating Scale (screening, quickly identify potential ADHD / TD risk), ADHD diagnosis scale (diagnosis, quantify the severity of core symptoms) total score and each dimension score.

[0078] Data preprocessing, including: Data cleaning: remove fNIRS signal SNR <3, head motion artifact (CV coefficient of variation) >25%, finally retain 5650 valid samples; Feature standardization: to eliminate the dimensional differences of different indicators, all features are standardized by Z-score, the formula is as follows: , where, is the original value of the jth feature of the ith subject, is the sample mean of the jth feature, is the sample standard deviation of the jth feature, is the standardized feature value.

[0079] Cluster analysis model construction and verification, including: (1) Feature space construction. The standardized fNIRS indicators (12: 4 brain regions × 2 hemoglobins × 3 types of parameters), behavioral indicators (4), and scale data (8, including Conners total score, ADHD diagnosis scale 3 dimension scores, etc.) are fused to construct a 24-dimensional feature vector, represented as: , where, is the comprehensive feature vector of the ith subject, corresponding to 24 standardized features.

[0080] (2) K-means clustering algorithm implementation. K-means algorithm is used to cluster 5650 valid samples, and the core steps are as follows: Determination of the number of clusters: the silhouette coefficient of different clustering numbers k (1~10) is calculated by the "elbow method", and the silhouette coefficient is maximum (0.77) when k=4, so the optimal number of clusters is determined to be 4; Algorithm iteration process: Step 1: randomly initialize 4 cluster centers where, is the feature center of the kth class; Step 2: calculate the Euclidean distance between each sample and each cluster center , the formula is as follows: , Step 3: assign to the class corresponding to the nearest cluster center; Step 4: update the cluster center: where, is the number of samples in the kth class; Step 5: repeat steps 2~4 until the cluster center change or the number of iterations reaches 1000 times, and the algorithm converges.

[0081] (3) Verification of clustering results.

[0082] Multi-method consistency verification: simultaneously use hierarchical clustering (Ward linkage method) and DBSCAN (density clustering) for comparison, and adjust the Rand index (ARI) to quantify the consistency of different clustering results, the formula is as follows: , where a is the number of samples that are classified as A by both methods, b is the number of samples that are classified as A by method 1 and not A by method 2, c is the number of samples that are classified as not A by method 1 and A by method 2, and d is the number of samples that are classified as not A by both methods. The results show that the ARI of K-means and hierarchical clustering is 0.87, and the ARI of DBSCAN is 0.83. A "sample-sample" consistency heat map is drawn (the horizontal and vertical axes are samples, and the color depth represents the number of times two samples are classified as the same class by the same method), and the high consistency area (color depth > 85%) accounts for 89%, proving that the clustering results are stable.

[0083] Bootstrap resampling: Bootstrap resampling is used to verify the stratified sampling of the original samples to evaluate the stability and generalization ability of the clustering algorithm. The specific process is as follows: first, stratified sampling is performed according to the proportion of TD / ADHD patients and non-cases in the original samples (about 1:2) to ensure that the proportion of cases and non-cases in each resampling is consistent with the overall; then 500 samples (about 167 cases and 333 non-cases) are randomly selected from the total sample with replacement, repeated 1000 times to generate 1000 Bootstrap sample sets. For each resampling sample set, calculate the clustering purity (Purity) and adjusted Rand index (ARI): , where N = 500 is the number of sampled samples, is the number of samples in the kth cluster that belong to the lth true class. Through resampling with replacement, the robustness of the clustering results on different sample subsets can be fully evaluated, the influence of accidental sample bias on the results can be reduced, and the generalization ability of different subject groups can be quantitatively verified. Compared with single fixed sampling, the Bootstrap method can generate a large number of sample combinations, providing higher confidence support for the stability, repeatability and reliability of the clustering results for clinical interpretation.

[0084] Clustering result analysis and interpretation criteria, including: The clustering results divide the samples into four categories, three of which can correspond to "small", "medium" and "large" severity, respectively, and the other is a mixed or normal group. In order to facilitate clinical application and interpretation, the present application establishes clear interpretation criteria for each clustering result, which is based on the comprehensive performance of behavioral indicators and brain activation indicators. There are significant differences between different categories, and the interpretation criteria are shown in Table 1 as follows: Table 1 Interpretation criteria .

[0085] The interpretation criteria are based on the selection of significant dividing points between different categories after clustering analysis of a large amount of data, and are optimized in combination with clinical practice needs to provide quantitative evaluation basis for doctors.

[0086] Based on the K-means clustering results (4 categories) and clinical diagnosis verification, combined with the characteristics of fNIRS brain activation indicators and behavior indicators, the subtype interpretation rules are established, as shown in the following table: Figure 3 The integral value of the brain activation physiological indicator is taken to perform interpretation.

[0087] Based on the above grade intervals, the evaluation module performs classification interpretation according to the following logical steps: First step: Task sensitivity screening (priority determination) The system first determines whether the subject belongs to the "task insensitive type". When all the following conditions are met, it is determined to be "task insensitive type", and the subsequent classification process is terminated: Go task accuracy is in the "small" interval; and, NoGo task accuracy is in the "small" interval; and, Go task reaction time is in the "small" interval.

[0088] Second step: Behavior performance determination If the subject is not determined to be "task insensitive type", the system further evaluates the behavior performance level: High performance: When the "Go task accuracy is in the 'large' interval, and the NoGo task accuracy is in the 'large' interval, and the Go task reaction time is in the'medium' or'small' interval" condition is met, it is determined to be high performance.

[0089] Low performance: All other cases that do not meet the above "high performance" condition are determined to be low performance.

[0090] Third step: Brain activation intensity determination The system independently evaluates the dorsolateral prefrontal cortex brain oxygen activation state of the subject: High activation: When the "integral value is in the 'large' interval, and the peak value is in the'medium' or 'large' interval, and the mean value is in the'medium' or 'large' interval" condition is met, it is determined to be high activation.

[0091] Low activation: All other cases that do not meet the above "high activation" condition are determined to be low activation.

[0092] Fourth step: Subtype comprehensive output The system combines the determination results of "behavior performance" and "brain activation intensity" to output the final four types of evaluation subtypes: Low activation-high performance type (i.e. behavior performance is "high performance" and brain activation is "low activation"); High activation-high performance type (i.e. behavior performance is "high performance" and brain activation is "high activation"); Low activation-low performance (i.e., behavioral performance as "low performance" and brain activation as "low activation"); High activation-low performance (i.e., behavioral performance as "low performance" and brain activation as "high activation").

[0093] The subtype interpretation rules are associated with the following clinical correlations, where the effective sample subtypes (4 categories, a total of 5652 cases): 1. High activation-high performance (2205 cases, accounting for 39.0%) Clinical significance: This mode indicates that the individual can efficiently mobilize the prefrontal neural resources of the brain to complete cognitive inhibition and attention tasks, and is usually seen in individuals with healthy cognitive function, with a lower risk of developing tic disorders (TD) or attention deficit hyperactivity disorder (ADHD).

[0094] 2. Low activation-high performance (1642 cases, accounting for 29.0%) Clinical significance: This mode reveals that the individual uses an efficient or "energy-saving" neural strategy to complete the task, with lower cognitive resource consumption, and also has a lower risk of TD / ADHD.

[0095] 3. High activation-low performance (348 cases, accounting for 6.1%) Clinical significance: This mode indicates that there is a cognitive compensation disorder, i.e., the brain invests excessive neural resources but still fails to effectively complete the task, indicating a potential defect in cognitive control function. This subtype may be a specific subtype of TD / ADHD.

[0096] 4. Low activation-low performance (1457 cases, accounting for 25.8%) Clinical significance: This mode indicates that the individual may have a problem with insufficient mobilization of cognitive resources, resulting in the brain's failure to effectively participate in the task, thereby affecting the normal functioning of cognitive inhibition and attention. This subtype suggests a higher likelihood of developing TD or ADHD.

[0097] 5. Invalid sample subtype (task-insensitive, 347 cases, accounting for 5.8%) Clinical significance: This subtype data cannot reflect the true cognitive function status of the individual, usually due to the subject's non-compliance, excessive head movement, and other artifacts, and needs to be retested. The likelihood of developing TD / ADHD in this subtype is also higher.

[0098] After the results are interpreted, a comprehensive report is generated. After completing data collection and processing, the system can automatically generate a structured and visualized comprehensive evaluation report. This report integrates physiological indicators, behavioral indicators, and clinical interpretation, providing objective and quantitative neural function evidence for the auxiliary diagnosis and rehabilitation assessment of TD.

[0099] The evaluation report includes the following core content: 1. Visualization of Results: fNIRS Waveform Graph: This graph clearly shows the concentration-time curves of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) in the core brain regions during the task. It includes the dynamic changes of HbO and HbR concentrations in the left and right dorsolateral prefrontal cortex (DLPFC) and frontal pole (FP). The baseline and task periods are marked in the graph, which intuitively reflects the activation patterns of the brain in different task stages.

[0100] Brain Activation Topography: This graph presents the brain activation status of each brain region under the Go / No-Go task paradigm through color coding. The graph uses HbO mean as an indicator, with red representing high activation areas and blue representing low activation areas, thus providing clinicians with intuitive information on the distribution of neural function activation.

[0101] 2. Quantitative Index Table: This table lists the quantitative indicators and behavioral indicators of the core brain regions, providing data support for objective evaluation.

[0102] Physiological Indicators: These include the HbO mean, HbO integral, HbO slope, and HbO barycenter of the left and right DLPFC and FP. These indicators comprehensively reflect the activation intensity, duration, and peak position of the cerebral cortex.

[0103] Behavioral Indicators: These include the accuracy ( ) and reaction time ( ) of the Go task, as well as the accuracy ( ) of the No-Go task. These indicators are used to evaluate the subject's performance, executive control, and inhibition ability in the task.

[0104] 3. Clinical Interpretation Conclusion: Based on the above physiological and behavioral indicators, combined with the pre-trained classification model, the system automatically generates explanatory comments, providing clinicians with auxiliary diagnostic suggestions and quantitative rehabilitation evaluation basis. These conclusions help doctors quickly understand the neural function characteristics of the subject, so as to develop more accurate treatment plans.

[0105] The five subtypes defined by the system have the following report characteristics and clinical implications: High Activation-High Performance Type: As shown in Figure 5 , the report shows excellent behavioral indicators (high accuracy, fast reaction time), and significant activation in key brain regions. This pattern usually indicates efficient cognitive function, commonly seen in healthy individuals.

[0106] Low Activation-High Performance Type: As shown in Figure 7 , the subject's performance is also good, but the brain activation topography shows low activation. This may indicate that the individual has adopted an efficient "neural energy-saving" strategy.

[0107] High activation-low performance: as shown in Figure 4 the report, it shows that the brain invested significant neural resources (high activation), but the behavioral performance was poor (e.g. high No-Go error rate). This pattern strongly suggests cognitive compensation imbalance, which is an important risk subtype of tic disorders or ADHD, and provides a key clue for clinical differentiation.

[0108] Low activation-low performance: as shown in Figure 6 the report, it shows poor behavioral performance and insufficient brain activation. This pattern suggests difficulty in mobilizing cognitive resources, which is related to typical executive function deficits.

[0109] Task-insensitive: as shown in Figure 8 the report, it usually shows abnormal behavioral patterns (e.g. extremely fast but random responses) and / or irregular brain signals. This result suggests that the test data is unreliable and needs to be retested.

[0110] Through the above visualization report, the clinician can quickly and intuitively obtain objective neuroimaging evidence of the patient's cognitive inhibition function and its classification, thereby assisting in diagnosis, intervention decision-making and efficacy evaluation, beyond subjective scales.

[0111] In summary, the key point of the present application is: for the first time, an idea and method of deeply fusing and jointly analyzing fNIRS brain activation indicators (such as the integral, mean, peak value of HbO, HbR) and Go / No-Go task behavior indicators (such as reaction time, accuracy) are proposed and protected, and on this basis, a quantitative standard system containing specific numerical thresholds (such as accuracy with 0.92 and 0.98 as boundaries) is constructed, as well as an automatic classification logic (such as a four-step interpretation process) that maps multi-dimensional indicator classification results to a specific cognitive function state. This innovative system relies on the integrated scheme of Go / No-Go task and fNIRS synchronous monitoring, realizes the synchronous evaluation of behavioral performance and brain activation in one test, and its output contains waveform graph, topographic map and quantitative report standard structure, with wide clinical applicability and expandability for all age groups, thereby providing a complete solution for the objective detection, accurate classification and evaluation of tic disorders.

[0112] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent substitutions for part or all of the technical features; and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the technical solutions of the embodiments of the present application.

Claims

1. A task-based system for detecting and assessing childhood tic disorders, characterized in that, include: The task stimulus presentation device is used to assess subjects’ attention and cognitive inhibition abilities and record their behavioral responses through standardized cognitive tasks. The near-infrared brain functional imaging device covers the core brain regions related to cognitive inhibition function through a multi-channel probe array. It is used to simultaneously monitor the concentration changes of oxyhemoglobin and deoxyhemoglobin in the core brain regions during the execution of the cognitive task using dual-wavelength near-infrared spectroscopy. The data processing module is connected to the task stimulus presentation device and the near-infrared brain functional imaging device, respectively, and is used to obtain behavioral indicators from the task stimulus presentation device and preprocess the raw light intensity signal collected by the near-infrared brain functional imaging device to extract physiological indicators. The result interpretation module is used to take the physiological and behavioral indicators as input and output the subtype interpretation results of childhood tic disorders through standardized interpretation rules based on cluster analysis modeling.

2. The task-paradigm-based detection and assessment system for childhood tic disorders according to claim 1, characterized in that, The task stimulus presentation device executes the Go / No-Go cognitive paradigm by presenting Go task stimuli and No-Go task stimuli, and records and quantifies the participants' behavioral indicators, including Go task accuracy, No-Go task accuracy, and Go task reaction time.

3. The task-paradigm-based detection and assessment system for childhood tic disorders according to claim 1, characterized in that, The physiological indicators include the integral value, mean value, and peak value extracted from the blood oxygen concentration-time curve.

4. The task-paradigm-based detection and assessment system for childhood tic disorders according to claim 1, characterized in that, The result interpretation module performs the following interpretation steps: The behavioral and physiological indicators are respectively classified as "small", "medium" or "large" based on predetermined level thresholds. Based on the level determination, the subtype interpretation result is output according to the predetermined logical process.

5. A task-paradigm-based detection and assessment system for childhood tic disorders according to claim 2 or 4, characterized in that, The threshold for judging accuracy in the aforementioned behavioral indicators is: Go task accuracy is categorized as follows: minor level corresponds to a value ≤ 0.92, medium level corresponds to 0.92 < value < 0.98, and major level corresponds to a value ≥ 0.

98. For No-Go task accuracy, the lowest level corresponds to a value ≤ 0.92, the middle level corresponds to 0.92 < value < 0.98, and the highest level corresponds to a value ≥ 0.

98.

6. A task-paradigm-based detection and assessment system for childhood tic disorders according to claim 2 or 4, characterized in that, The threshold for determining reaction time in the behavioral indicators is: Go task response time is ≤280 milliseconds for the lowest level, 280 milliseconds < value < 380 milliseconds for the medium level, and ≥380 milliseconds for the highest level.

7. A task-paradigm-based detection and assessment system for childhood tic disorders according to claim 3 or 4, characterized in that, The threshold for determining the physiological indicator is: The lowest level of the points corresponds to a value ≤ 0, the middle level corresponds to 0 < value < 6, and the highest level corresponds to a value ≥ 6. The mean is as follows: small level corresponds to a value ≤ 0, medium level corresponds to 0 < value < 0.03, and large level corresponds to a value ≥ 0.03; The smaller level of the peak value corresponds to a value ≤ 0, the medium level corresponds to 0 < value < 0.08, and the larger level corresponds to a value ≥ 0.

08.

8. A task-paradigm-based detection and assessment system for childhood tic disorders according to claim 2, 3, or 4, characterized in that, The predetermined logical flow includes: Step S1, Task Sensitivity Screening: If the Go task accuracy rate and the No-Go task accuracy rate are both at the "low" level and the Go task response time is at the "low" level, then it is determined to be "task insensitive" and the process is terminated; Step S2, Behavioral Performance Judgment: If not judged as "task-insensitive", then when the Go task accuracy rate is "high", the No-Go task accuracy rate is "high", and the Go task response time is "medium" or "low", the behavioral performance is judged as "high performance"; otherwise, it is judged as "low performance". Step S3, Brain activation intensity determination: When the integral value is "large", the peak value is "medium" or "large", and the mean value is "medium" or "large", the brain activation is determined to be "high activation"; otherwise, it is determined to be "low activation". Step S4, Subtype Comprehensive Output: Combine the results of behavioral performance and brain activation intensity assessment to output low activation-high phenotype, high activation-high phenotype, low activation-low phenotype, or high activation-low phenotype.

9. The task-paradigm-based detection and assessment system for childhood tic disorders according to claim 1, characterized in that, The interpretation rules based on cluster analysis modeling in the result interpretation module are established through the following steps: Construct a multidimensional feature vector containing standardized fNIRS brain function indicators, behavioral indicators, and clinical scale data; The K-means clustering algorithm was used to perform cluster analysis on a large amount of sample data to determine the optimal number of clusters and the feature centers of each category; Based on clinical diagnostic verification, each cluster category is assigned a corresponding subtype interpretation label and clinical significance.

10. The task-paradigm-based detection and assessment system for childhood tic disorders according to claim 1, characterized in that, It also includes a visualization report generation module, which automatically generates and outputs a structured assessment report containing fNIRS waveforms, brain activation topography, quantitative indicator tables, and clinical interpretation conclusions.

Images