CN115349857A

CN115349857A - Dynamic rehabilitation assessment method and system based on fNIRS brain function map

Info

Publication number: CN115349857A
Application number: CN202210843314.2A
Authority: CN
Inventors: 李增勇; 霍聪聪; 徐功铖; 谢晖; 张静莎; 张腾宇; 吕泽平
Original assignee: National Research Center for Rehabilitation Technical Aids
Current assignee: National Research Center for Rehabilitation Technical Aids
Priority date: 2022-07-18
Filing date: 2022-07-18
Publication date: 2022-11-18
Anticipated expiration: 2042-07-18
Also published as: CN115349857B

Abstract

The invention provides a dynamic rehabilitation assessment method and system based on an fNIRS brain function map, wherein the method comprises the following steps: selecting a brain function detection paradigm matched with the motor ability of the user from a plurality of pre-stored brain function detection paradigms based on the motor ability of the user; acquiring a cerebral blood oxygen signal which is acquired by a near-infrared cerebral function acquisition device in real time and induced and generated by a user under the matched cerebral function detection mode; calculating near-infrared multi-mode brain function characteristic parameters based on the acquired brain blood oxygen signals, wherein the multi-modes are an activation mode and a connection mode, and the parameters form multi-mode brain function characteristic combination vectors; based on the obtained multi-mode brain function feature combination vector, performing brain function feature matching classification by an integrated multi-classification method to obtain a matching classification result; and providing a visual evaluation report through a brain function state evaluation module based on the acquired matching classification result.

Description

Dynamic rehabilitation assessment method and system based on fNIRS brain function map

Technical Field

The invention relates to the technical field of brain function rehabilitation assessment, in particular to a dynamic rehabilitation assessment method and system based on an fNIRS brain function map.

Background

The recovery of motor dysfunction following stroke is often attributed to functional remodeling or reorganization of the cerebral cortex. Real-time monitoring of brain functional status is of great significance to stroke rehabilitation. The injury type, degree and position of the patient with the apoplexy have individual difference, and nerve remodeling and migration phenomena with different modes are presented in the rehabilitation process. However, due to the lack of dynamic quantitative evaluation and real-time feedback means, the stroke rehabilitation technology still has the key problems of unclear action mechanism, inaccurate action target, unclear evaluation index and the like, and the clinical rehabilitation effect is not uniform. Aiming at the root cause of cortical function damage after stroke, the clinical rehabilitation intervention scheme is still a clinical core problem how to individually select and perform layered treatment on patients to promote the patients to obtain greater rehabilitation benefit.

The development of neuroimaging techniques provides new methods for studying and guiding neuroplasticity. Common brain function detection technologies such as functional nuclear magnetic resonance imaging (fMRI) and electroencephalogram (EEG) are difficult to acquire dynamic brain function information under exercise or rehabilitation intervention due to poor anti-electromagnetic interference and anti-exercise interference capabilities. The functional near-infrared spectroscopy (fNIRS) is used as an optical non-invasive brain function imaging technology, non-invasive visual brain nerve activity information is provided by detecting the cerebral blood oxygen metabolism condition, the method has the advantages of being simple to operate, strong in anti-interference performance, good in electromagnetic compatibility and the like, the brain function of a stroke patient can be rapidly detected in a multi-scene in real time, and the method has important clinical application value in the field of nerve rehabilitation. However, the current fNIRS imaging system still has the following key problems in stroke rehabilitation: the method is lack of standard detection paradigm and imaging marks, the brain function remodeling condition of the auxiliary evaluation application of the brain function is limited to be complex, and the characteristic quantity obtained by a single brain function analysis method cannot fully represent the complex brain function network characteristics.

Therefore, a dynamic evaluation system for brain function rehabilitation, which can detect the brain function of a patient more accurately, is needed.

Disclosure of Invention

In view of the above, the embodiments of the present invention provide a dynamic rehabilitation assessment method and system based on fNIRS brain function maps, so as to eliminate or improve one or more defects existing in the prior art.

One aspect of the invention provides a dynamic rehabilitation assessment method based on a fNIRS brain function map, which comprises the following steps:

selecting a brain function detection paradigm matched with the motor ability of the user from a plurality of pre-stored brain function detection paradigms based on the motor ability of the user;

acquiring a cerebral blood oxygen signal which is acquired by a near-infrared cerebral function acquisition device in real time and induced and generated by a user under the matched cerebral function detection mode;

calculating near-infrared multi-mode brain function characteristic parameters based on the obtained brain blood oxygen signals, wherein the multi-modes are an activation mode and a connection mode, and the parameters form multi-mode brain function characteristic combination vectors;

based on the obtained multi-mode brain function feature combination vector, performing brain function feature matching classification by an integrated multi-classification method to obtain a matching classification result, wherein the matching classification result is that the motor function rehabilitation effect is good, the effect is general and the effect is poor;

and providing a visual evaluation report through a brain function state evaluation module based on the acquired matching classification result.

In some embodiments of the invention, the pre-stored plurality of brain function detection paradigms comprises a typical fNIRS brain function detection paradigms of resting state and motor facing neural loop remodeling and sensory-motor loop remodeling; the detection paradigm for performing neural loop remodeling facing motion is specifically based on a repeated extending motion task of a target-oriented hemiplegic upper limb; the detection paradigm facing sensory-motor nerve loop remodeling is a central-peripheral combined magnetoelectric stimulation technology, and the detection paradigm activates a motor-related brain region and induces proprioception to be transmitted into a central cortex.

In some embodiments of the present invention, the method includes combining network information of an activation mode and a connection mode by using a brain function multi-modal feature fusion method, wherein the activation mode describes activation distribution of various regions of the brain, and the connection mode represents information interaction process under typical brain tasks.

In some embodiments of the present invention, the method further comprises performing time-frequency conversion on the cerebral blood oxygen signal using a wavelet transform time-frequency analysis method based on the wavelet transform time-frequency analysis method having the capability of decoupling signal components and providing local phase information.

In some embodiments of the present invention, the near-infrared multimode brain function characteristic parameters include hemisphere activation laterality, hemisphere connection laterality, hemisphere autonomy coefficients, undirected weighted topology parameters, directed weighted topology parameters, and the like.

In some embodiments of the invention, the method further comprises classifying motor dysfunction including mild, moderate and severe dysfunction using cluster analysis; clustering the motion function assessment scale set into mild, moderate and severe motion dysfunction subsets by using a K-means clustering algorithm; the set of athletic performance assessment scales includes a functional exercise assessment (Fugl-Meyer, FM) scale, a grip index, and an upper limb action study ARAT scale.

In some embodiments of the invention, the method further comprises selecting a brain function response feature based on a combination of group-level statistical tests and functionally-related regression fusion analysis;

the group level statistical test is used as a characteristic screening and filtering method to extract a brain function characteristic index with statistical significance;

the function-dependent regression fusion analysis is to fuse clinical function evaluation information on the basis of primary feature screening by group-level statistical test, establish a segmented regression model, select features with statistical significance, identify an optimal feature subset related to dysfunction in a specific state, and select multi-task-multi-mode fusion features.

In some embodiments of the invention, the method further comprises dealing with the imbalance problem of each set of data using an integrated multi-classification Support Vector Machine (SVM) model based on the imbalance of the clinical data; the integrated multi-classification support vector machine SVM model adopts a radial basis kernel function RBF to perform high-dimensional mapping on near-infrared brain functional characteristics, and a grid optimization method is adopted to search an optimal punishment factor and a kernel radius parameter.

Another aspect of the present invention provides a dynamic rehabilitation assessment system based on fNIRS brain function maps, comprising: the motion detection paradigm selection module is used for selecting a brain function detection paradigm matched with the motion capability of the user from a plurality of pre-stored brain function detection paradigms according to the motion capability of the user;

the near-infrared brain function acquisition module is used for acquiring multichannel cerebral blood oxygen signals induced by a user under the matched brain function detection mode in real time;

the brain function response characteristic analysis module is used for preprocessing the collected brain blood oxygen signals through the fNIRS data collection and multi-mode characteristic calculation module and calculating near-infrared multi-mode brain function characteristic parameters to obtain brain function multi-mode response characteristic vectors;

the fNIRS brain function map module is used for constructing a fNIRS brain function characteristic and exercise rehabilitation mapping model according to an integrated multi-classification method, and inputting the obtained brain function multi-mode response characteristic vector to the module for matching classification;

and the brain function state evaluation reporting module is used for outputting the matching result obtained by the fNIRS brain function mapping module and providing a visual evaluation report.

In some embodiments of the invention, the fNIRS brain function map module comprises: the system comprises a motion function evaluation module, an fNIRS data acquisition and multi-mode feature calculation module, a multitask-multi-mode fusion feature selection module and an fNIRS brain function and motion rehabilitation mapping module;

the motion function evaluation module is used for storing the limb motion function evaluation results of the user at different time nodes, including comprehensive motion dysfunction evaluation and comprehensive motion function rehabilitation evaluation results;

the fNIRS data acquisition and multi-mode feature calculation module is used for acquiring fNIRS data under a specific detection paradigm by using a near-infrared brain function imaging device, and performing preprocessing and multi-mode brain function feature calculation;

the multitask-multimode fusion feature selection module is used for extracting multitask-multimode fNIRS brain function fusion features related to the user movement ability based on two-step feature dimensionality reduction formed by group-level statistics and function-related regression fusion analysis;

the fNIRS brain function and motion rehabilitation mapping module is used for screening fNIRS brain function fusion characteristics and rehabilitation effect classification labels by the multi-task and multi-mode fusion characteristic selection module, constructing an integrated multi-classification Support Vector Machine (SVM) model, and selecting a parameter with the highest classification accuracy as the fNIRS brain function and motion rehabilitation mapping classification model through training and parameter optimization, namely constructing the fNIRS brain function map.

In some embodiments of the invention, the system further comprises a processor and a memory, the memory having stored therein computer instructions for executing the computer instructions stored in the memory, the system implementing the steps of the method as described above when the computer instructions are executed by the processor.

In another aspect of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the method as set forth above.

According to the dynamic rehabilitation assessment method and system based on the fNIRS brain function map, disclosed by the invention, the nerve function state of the motor dysfunction population after brain injury is dynamically assessed by combining a near-infrared brain function imaging technology with a specific motion detection paradigm, so that an important iconography basis is provided for formulation and optimization of a rehabilitation scheme; the multi-mode feature fusion method can be adopted to combine activation and connection network information, and a plurality of different brain function analysis methods can be used for more abundantly describing brain function maps; the integrated multi-classification support vector machine model can be adopted to solve the limitation of unbalanced clinical data and solve the mapping prediction model bias phenomenon caused by data unbalance. The brain function specific response and remodeling characteristics under various typical task paradigms can be fused, a dynamic mapping model of near-infrared brain function indexes and motor function rehabilitation is constructed, namely a fNIRS brain function map for motor rehabilitation is constructed, and the brain function state evaluation of patients with different rehabilitation levels is realized; the technical advantages of the near-infrared brain function imaging technology of motion resistance and electromagnetic interference resistance can be utilized, a typical monitoring paradigm facing to a motor neural loop is designed, brain function characteristics related to motion functions are fully obtained, and brain stroke neural function assessment except for resting state is enriched; the method can extract the fNIRS multi-mode brain function response characteristics under the resting state and the typical detection task paradigm, establish a mapping model of the fNIRS multi-mode brain function response characteristics and motor function rehabilitation, establish a fNIRS brain function map facing motor dysfunction and realize dynamic assessment of the brain function state of a brain injury patient.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.

It will be appreciated by those skilled in the art that the objects and advantages that can be achieved with the present invention are not limited to the specific details set forth above, and that these and other objects that can be achieved with the present invention will be more clearly understood from the detailed description that follows.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principle of the invention. In the drawings:

fig. 1 is a schematic flow chart of a dynamic rehabilitation assessment method based on an fNIRS brain function map in an embodiment of the present invention.

Fig. 2 is a block diagram of a dynamic rehabilitation assessment system based on fNIRS brain function maps according to an embodiment of the present invention.

FIG. 3 is a block diagram of a brain function response feature analysis module according to an embodiment of the present invention.

FIG. 4 is a framework diagram of a functional brain atlas of fNIRS in an embodiment of the invention.

FIG. 5 is a block diagram of a multitasking-multimodal fusion feature selection module in accordance with an embodiment of the present invention.

FIG. 6 is a diagram of a model framework of an integrated multi-class support vector machine according to an embodiment of the present invention.

Fig. 7 is a flowchart illustrating the operation of a dynamic rehabilitation assessment system based on the fNIRS brain function atlas according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the following embodiments and the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.

It should be noted that, in order to avoid obscuring the present invention with unnecessary details, only the structures and/or processing steps closely related to the scheme according to the present invention are shown in the drawings, and other details not so relevant to the present invention are omitted.

It should be emphasized that the term "comprises/comprising" when used herein, is taken to specify the presence of stated features, elements, steps or components, but does not preclude the presence or addition of one or more other features, elements, steps or components.

It is also noted herein that the term "coupled," if not specifically stated, may refer herein to not only a direct connection, but also an indirect connection in which an intermediate is present.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings, the same reference numerals denote the same or similar parts, or the same or similar steps.

In order to overcome the problems that in the prior art, standard detection paradigm and imaging marks are lacked for cerebral apoplexy rehabilitation, brain function remodeling conditions are complex, a brain function analysis method is simplified, and the like, the invention provides a dynamic rehabilitation assessment method and a system based on an fNIRS brain function map, wherein the neural function state of a motor dysfunction population after brain injury is dynamically assessed by combining a near-infrared brain function imaging technology and a specific motion detection paradigm, and important imaging basis is provided for formulation and optimization of a rehabilitation scheme; the multi-mode feature fusion method can be adopted to combine activation and connection network information, and a plurality of different brain function analysis methods can be used for more abundantly describing brain function maps; the method can solve the limitation of unbalanced clinical data by adopting an integrated multi-classification support vector machine model and solve the mapping prediction model bias phenomenon caused by unbalanced data. The brain function specific response and remodeling characteristics under various typical task paradigms can be fused, and a dynamic mapping model of near-infrared brain function indexes and motor function rehabilitation is constructed; the technical advantages of the near-infrared brain function imaging technology of motion resistance and electromagnetic interference resistance can be utilized, a typical monitoring paradigm facing to a motor neural loop is designed, brain function characteristics related to motion functions are fully obtained, and brain stroke neural function assessment except for resting state is enriched; the fNIRS multi-mode brain function response characteristics under a resting state and a typical detection task paradigm can be extracted, a mapping model of the fNIRS multi-mode brain function response characteristics and motor function rehabilitation is established, a fNIRS brain function map facing motor dysfunction is established, and dynamic assessment of brain function states of brain injury patients is achieved. The dynamic evaluation of the brain function rehabilitation is more real-time, more comprehensive, more continuous, more objective, more accurate and more targeted, so that the evaluation of the brain function state of the stroke patient can be aimed at, the clinical function evaluation and the treatment response detection are assisted, and the treatment target is provided for the personalized intervention.

Fig. 1 is a schematic flow chart of a dynamic rehabilitation assessment method based on a fNIRS brain function map according to an embodiment of the present invention, as shown in fig. 1, the method of the embodiment includes the following steps:

step S110, selecting a brain function detection pattern matching with the user 'S exercise ability from a plurality of pre-stored brain function detection patterns based on the user' S exercise ability.

In this step, the pre-stored plurality of brain function detection paradigms include a typical fNIRS brain function detection paradigms of resting state and motor-oriented execution neural circuit remodeling and sensory-motor neural circuit remodeling; the detection paradigm facing the movement execution neural loop remodeling is specifically based on a target guidance hemiplegia upper limb repeated extension movement task; the detection paradigm facing sensory-motor nerve loop remodeling is a central-peripheral combined magnetoelectric stimulation technology, and the detection paradigm activates a motor-related brain region and induces proprioception to be transmitted into a central cortex.

In this embodiment, for example, a typical fNIRS brain function detection model for performing a loop remodeling and a sensory-motor loop remodeling for a motion is designed, specifically, a motor-performed cortical reorganization response task model for inducing a stroke user to perform a motion based on a target-oriented hemiplegic upper limb repeated extension motion model; based on the central combined peripheral magnetoelectric stimulation technology, the motor-related brain area is activated and proprioception is induced to be transmitted into the central cortex task paradigm. The exercise execution task paradigm requires that a participant has certain exercise capacity and can participate in the paradigm under the condition of assistance or initiative; the sensory-motor task paradigm is completed by combining central magnetic stimulation and peripheral electrical stimulation without requiring a participant to perform specific activities; therefore, for users without any athletic ability, the possibility of magnetic and electric taboo is eliminated, and the resting state and the sensorimotor task paradigm are preferentially selected; for users with exercise ability, the possibility of magnetic and electric taboo is eliminated, and a resting state and an exercise execution task paradigm are selected preferentially; if the possibility of magnetic and electric taboo is eliminated, a resting state, a motion execution task paradigm and a sensory and motor task paradigm are preferably selected.

And step S120, acquiring a brain blood oxygen signal induced by the user under the matched brain function detection mode in real time by the near-infrared brain function acquisition device.

In the step, based on a brain function detection model matched with the motor ability of the user, a near-red brain function acquisition device is adopted to acquire a brain blood oxygen signal induced by the user under the matched brain function detection model in real time, wherein the brain blood oxygen signal comprises an oxygenated hemoglobin concentration delta [ oxy-Hb ] signal, a deoxygenated hemoglobin concentration delta [ deoxy-Hb ] signal and a total oxygenated hemoglobin concentration delta [ total-Hb ] signal.

In an embodiment of the invention, a multichannel near-infrared brain function imaging system is used to acquire cerebral blood oxygen signals of a stroke patient, for example, and the near-infrared brain function imaging system comprises 17 transmitters and 16 receivers, and 52 fNIRS acquisition channels are formed. According to the international 10-20 system arrangement, the prefrontal lobe, the motor area, the sensory motor and the occipital cortex of the left hemisphere and the right hemisphere are respectively covered; the distance between the light source and the detector is 3cm, and the sampling frequency is 10Hz. Each user finishes the acquisition of the fNIRS data according to the matched brain function detection paradigm, the focused frequency band of the near-infrared brain blood oxygen signal is 0.009-0.01 Hz, the lowest frequency is 0.009Hz, namely the period of the lowest frequency is 112s, so the minimum measurement time is 560s, the fluctuation of 40s before the data acquisition starts is considered, namely the validity of near-infrared brain function indexes in low-frequency cycles at least containing more than 5 in one measurement is ensured, the total acquisition time is ensured to be 600s, and therefore the acquisition time of each group of tasks is not less than 10 minutes.

Step S130, calculating near-infrared multi-mode brain function characteristic parameters based on the acquired cerebral blood oxygen signals, wherein the multi-modes are an activation mode and a connection mode, and the parameters form multi-mode brain function characteristic combination vectors.

In this step, the cerebral blood oxygen signal is input into the cerebral function response characteristic analysis module based on the collected cerebral blood oxygen signal, and the collected data is preprocessed and the near-infrared multi-mode cerebral function parameter is calculated by the fNIRS data collection and multi-mode characteristic calculation module in the cerebral function response characteristic analysis module, as shown in fig. 3.

In the embodiment of the invention, the activation distribution condition of each area of the brain can be described based on the activation mode, and the connection mode among different brain areas can express the information interaction process under the brain typical task, so the activation and connection network information is combined by adopting a brain function multi-mode feature fusion method to comprehensively construct a brain function map. Establishing indexes such as brain activation, brain function connection, brain effect connection and the like based on a time-frequency analysis method of wavelet transformation, and constructing a near-infrared brain function analysis model; on the basis, indexes such as hemisphere activation lateral deviation, hemisphere connection lateral deviation, hemisphere autonomous coefficients, undirected weighted topological parameters and directed weighted topological parameters are further calculated, and near-infrared brain function network characteristics are comprehensively analyzed. The method is characterized in that signals need to be preprocessed before calculating near-infrared multimode brain function parameters, and the specific method comprises the following steps:

(1) Primary visual inspection and signal-to-noise ratio inspection: and checking the signal-to-noise ratio of the signal according to light intensity data acquired by the two-wavelength near infrared light, and removing a noise channel. In order to estimate the signal-to-noise ratio of the data channel, the signal-to-noise ratio of the channel data is estimated by calculating the relative coefficient of variation (CV,%) for the original signals at 760nm and 850nm, which is expressed as follows:

CV＝σ/μ×100％

where μ is the data mean of the signal and σ is the standard deviation of the signal. And calculating the CV value of each channel in the whole task paradigm acquisition process, and eliminating the measurement channels with the CV values of any wavelength signals larger than 15%.

(2) Band-pass filtering: and performing band-pass filtering (0.005 Hz-2 Hz) on the original light intensity signal by using a zero-phase sixth-order Butterworth filter to remove high-frequency and ultra-low-frequency interference components.

(3) And (3) concentration conversion: according to the modified Beer-Lambert law, namely the basic law of light absorption, the brain blood oxygen concentration signals are converted according to the detected light intensity change, wherein the brain blood oxygen concentration signals comprise oxygenated hemoglobin concentration delta [ oxy-Hb ] signals, deoxygenated hemoglobin concentration delta [ deoxy-Hb ] signals and total oxygenated hemoglobin concentration delta [ total-Hb ] signals.

(4) Removing motion artifacts: and eliminating obvious abnormal points in the signal by adopting a moving average method, and then removing the motion artifact by adopting a method based on moving standard deviation and cubic spline interpolation.

(5) Removing physiological interference: scalp signals acquired by short-distance data acquisition channels are utilized, regression analysis is carried out on blood oxygen signals of each channel to remove scalp interference noise, and physiological interference in fNIRS measurement, including heartbeat, respiration signals, blood pressure signals and the like, is reduced by combining with principal component analysis (ICA). All ICA decomposed signal components are individually visually examined to determine components that may be associated with noise and artifacts. The component of interest is visually identified according to the criterion that the brain function signal should have a significant low frequency (in the range of 0.01-0.08 Hz) oscillation spectrum, indicating that the acquired signal contains brain functional hemodynamic response information.

(6) Wavelet transform time-frequency analysis: the time-frequency analysis method based on wavelet transform has the capability of decoupling signal components and providing local phase information, is suitable for analyzing biological oscillation signals, and in the embodiment of the invention, for example, complex Morlet wavelet is selected as a wavelet basis function to perform time-frequency conversion on cerebral blood oxygen signals, and the brain blood oxygen signals are subjected to specific frequency f and time t _n The complex wavelet coefficient obtained by wavelet transform is defined as:

the generated wavelet coefficient is represented as a complex number on a time-frequency plane, and the absolute amplitude W of each frequency and time can be defined _i (f,t _n ) And instantaneous phase information

Wherein i represents the near-red cerebral blood oxygen signal of a certain channel, a _i (f,t _n ) Is the real part of a complex number, b _i (f,t _n ) Is the imaginary part of the complex number; the amplitude and phase information can be used to represent the relationship between brain activation and different oscillation signals, respectively. On the basis, indexes such as hemisphere activation lateral deviation, hemisphere connection lateral deviation, hemisphere autonomous coefficients, undirected weighted topological parameters and directed weighted topological parameters are further calculated, near-infrared brain function network characteristics are comprehensively analyzed, and brain function multi-mode response characteristic vectors are obtained, and the specific analysis method is as follows:

1) Brain activation response and hemispheric laterality

Based on wavelet transformation, the amplitude information of the cerebral blood oxygen signal at a certain frequency and time point is defined as W _i (f,t _n )：

The Wavelet Amplitude (WA) is defined as the average result of a time-frequency signal in a time domain, can be expressed as a power index, reflects the oscillation amplitude of different frequency components of an original signal, and can be used for representing the cortical activity strength; WA index for each fNIRS channel is calculated within the frequency band of interest to represent the hemodynamic oscillation amplitude of the covered brain region.

In addition, in order to observe inter-hemisphere balance of stroke user activation, activation hemisphere lateral deviation index (LI) WAs further calculated based on WA results _WA ):

Wherein WA _ipsi Is the WA value of the fnIRS channel of the affected hemisphere; WA (water in oil) _contra WA index for fNIRS channels of the healthy lateral hemisphere; LI (lithium ion) material _WA The magnitude of the values is between-1 and 1, -1 indicates complete activation of the healthy lateral hemisphere, and 1 indicates complete activation of the affected lateral hemisphere.

2) Brain functional connectivity analysis

The dynamic phase information can be used to study the relationship between different oscillations. In an embodiment of the invention, a brain Functional Connectivity (FC) is described using a wavelet phase coherence indicator (WPCO) that quantifies the tendency of the phase difference between two signals to remain constant at a particular frequency. WPCO may be defined as:

wherein

Representing the instantaneous phase difference of the two fNIRS channel signals. In the embodiment of the invention, the WPCO value between every two fNIRS channels is calculated, amplitude adaptive Fourier transform substitute signals are adopted to check whether the calculated WPCO value is effective or not, and after the check of the substitute signals, functional connection matrixes corresponding to the number of the fNIRS channels are generated.

In addition, in order to observe the balance between hemispheres of the stroke user functional connection network, a connection hemisphere lateral deviation index (LI) is further calculated based on the WPCO result _WPCO )：

N _ipsi-i Represents the number of functional linkages to the affected lateral hemispheric channel i; t is _ipsi-i Represents all possible functional connections of the affected hemispheric channel i; n is a radical of _contra-j Represents the number of functional connections to the healthy lateral hemisphere channel j; t is _contral-j Representing all possible function connection quantity of the healthy side hemispherical channel j, wherein the channel i and the channel j are symmetrically distributed in a hemisphere; LI (lithium ion) material _WPCO The magnitude of the values is between-1 and 1, -1 indicates a fully healthy hemispheric junction and 1 indicates a fully affected hemispheric junction.

In addition, in order to observe the difference between intra-hemispheric connection and inter-hemispheric connection of stroke user functional connection network, a connection hemisphere autonomy coefficient (AI) is further calculated based on WPCO result _WPCO )：

N _i The number of channels which are functionally connected with the channel i in the hemisphere at the same side is represented; n is a radical of _c Representing the number of channels in the contralateral hemisphere that are functionally connected to channel i; t is a unit of _i And T _c Respectively representing the total number of channels of the same side hemisphere and the opposite side hemisphere; the greater the AI value is, the stronger the connection strength between the passage and the hemisphere is, and the comparison of the AI values of bilateral hemisphere can be used to reflect the autonomous condition of cerebrum hemisphere, so the methodThe method is based on the comparison of functional connection between hemispheres and hemispheres, reflects the imbalance of functional connection between hemispheres and hemispheres, and indirectly reflects the lateralization of the cortex by comparing the sizes of the left hemisphere and the right hemisphere AI.

3) Brain effect junction analysis

In the embodiment of the present invention, for example, a coupling phase oscillation model between two signals is constructed in a frequency band of interest by using dynamic phase information after wavelet transform:

w _i (t) is a parameter of the natural frequency, [ xi ] _i (t) is white Gaussian noise, and the phases of the two oscillators are phi _i And phi _j Function q of _i Representing a coupling equation; and calculating the optimal parameters for describing the model by utilizing dynamic Bayesian inference, and further acquiring and describing the coupling relationship between the two oscillation signals, namely Effect Connection (EC). On the basis, an amplitude adaptive Fourier transform method is adopted to check whether the calculated effect connection value is effective or not; after the test of the substitution signal, the corresponding effect connection matrix is generated on the number of the fNIRS channels.

On the basis of the above, the connection lateral deviation index LI can be also obtained _EC And hemispherical autonomy factor AI _EC The two metrics based on the effect connection network are calculated.

4) Brain function network index calculation

In the embodiment of the invention, the interesting brain areas are divided according to the coverage positions of the fNIRS channels, and the undirected weighted network and the directed weighted network characteristic parameters (G) are respectively calculated based on the functional connection and effect connection results, wherein the undirected weighted network characteristic parameters comprise global characteristic topology parameters (small world network, global efficiency, fuji club and hierarchy), and the directed weighted network characteristic parameters comprise node characteristic topology parameters (clustering coefficient, shortest path, node efficiency, local efficiency of nodes, centrality of point degree and centrality of intermedium).

In conclusion, the fNIRS multimode brain function feature combination vector V in each state is calculated:

V＝[WA，LI _WA ，FC，LI _FC ，AI _FC ，EC，LI _EC ，AI _EC ，G]

and step S140, based on the obtained multi-mode brain function feature combination vector, performing brain function feature matching classification by an integrated multi-classification method to obtain a matching classification result, wherein the matching classification result is that the motor function rehabilitation effect is good, the effect is general and the effect is poor.

In this step, the fNIRS multimode brain function feature combination vector V calculated in step S130 is input into the fNIRS brain function atlas module constructed by using the integrated multi-classification method and related to motor rehabilitation, so as to perform matching classification of brain function features and determine the neurological function status of the user according to the matching classification result; the fNIRS brain function map module comprises a motion function evaluation module, an fNIRS data acquisition and multi-mode feature calculation module, a multi-task and multi-mode fusion feature selection module and an fNIRS brain function and motion rehabilitation mapping module; as shown in fig. 4, the fNIRS brain function atlas module uses the motor function assessment results to group the motor dysfunction of the user and calculate a comprehensive motor assessment score; grouping the motor function rehabilitation effects of the users according to the motor function evaluation results of the front and back times in the rehabilitation process and calculating a comprehensive motor rehabilitation score; acquiring cerebral blood oxygen signals of a stroke user under various typical task paradigms based on a near-infrared cerebral function imaging device, performing cerebral function multi-mode feature analysis, and realizing near-infrared cerebral function feature and dysfunction index fusion feature selection by using group-level statistics and function regression; the multitask-multimode fNIRS brain function fusion index is input into the integrated multi-classification classifier to realize classification mapping of the rehabilitation effect of the stroke user, wherein the classification mapping comprises good motor function rehabilitation effect, general effect and poor effect.

In an embodiment of the present invention, the exercise function evaluation module is configured to record basic information of a user, where the basic information includes: age, sex, type of injury (ischemic/hemorrhagic), location of lesion. Evaluating motor function of a brain injury patient using an evaluation scale comprising a grip index representing a relative grip of a patient's diseased hand relative to a healthy hand calculated as mean grip (sick side)/mean grip (healthy side) multiplied by 100, a Fugl-Meyer (FM) scale, and an upper limb movement study (ARAT) scale; the FM scale is used to assess the ability of a patient to move individual joints under synergistic effects, including assessment of motion control and strength; the ARAT scale is used to assess complex movements of the upper limbs in daily life, including grasping, pinching, and gross movement of the shoulders and elbows while flexed and extended. The module stores the limb movement function evaluation results of the user at different time nodes, and can calculate the comprehensive movement function disorder evaluation and the comprehensive movement function rehabilitation evaluation results.

Aiming at the comprehensive motor function rehabilitation evaluation, a K-means clustering (K-means) algorithm is adopted to divide a motor function evaluation table set (comprising an FM (frequency modulation) table, a grip index and an ARAT (auto-regressive artificial neural network) into three categories, namely mild, moderate and severe motor dysfunction subsets, and a Principal Component Analysis (PCA) method is further adopted to construct a comprehensive motor evaluation index M in each functional dysfunction subset as a preamble process of a subsequent feature fusion module based on a clinical motor function evaluation result (comprising the FM table, the grip index and the ARAT table). For each subclass, PCA with FM scale, ARAT scale and grip index as input variables yields a one-factor solution, and the number of principal components is determined on the basis of the principle that the cumulative variance contribution is not less than 90%. In the embodiment of the present invention, the factor value of the first principal component derived from principal component analysis using FM scale, ARAT and grip index as input variables is defined as the composite motor function score M.

For comprehensive motor function rehabilitation evaluation, relative difference values of various evaluation indexes (including FM (frequency) scale, grip strength index and ARAT (auto-responsive assessment) scale) at different time points are calculated firstly, namely

d(x,y)＝(x-y)/(1+y)

Wherein d represents the evaluation index difference, x represents the evaluation score of the node at the previous time, and y represents the evaluation score of the node at the later time. After z-standardization processing is carried out on each evaluation index difference value, a K-means clustering (K-means) algorithm is adopted to divide a z-standardized motion function evaluation difference value set (comprising an FM (frequency modulation) scale difference value, a grip index difference value and an ARAT (auto-regressive actuator) scale difference value) into three categories, namely good recovery, general recovery and poor recovery subsets, which are used as a preorder process of a subsequent classification mapping task module. In addition, each evaluation index difference value after the z-standardization processing is input into the PCA, and a comprehensive exercise rehabilitation index R is constructed. And determining the number of the main components according to the principle that the cumulative variance contribution rate is not lower than 90%. And (3) carrying out weighted summation on the principal component scores according to the proportion of the variance contribution rate of each principal component in the accumulated variance contribution rate of the extracted principal component:

PCA with FM scale, ARAT scale and grip index as input variables yields a composite motor rehabilitation score index R, where negative values do not represent a deterioration in motor performance, but rather a relatively lesser improvement in motor performance compared to the entire group.

In the embodiment of the present invention, in combination with fig. 4, the fNIRS data acquisition and multi-mode feature calculation module is used to acquire fNIRS data in a specific detection paradigm by using a near-infrared brain function imaging device, and perform preprocessing and multi-mode brain function feature calculation. The specific detection paradigm for the fNIRS data collection and multi-mode feature computation module includes a resting state and a typical fNIRS brain function detection paradigm for performing loop remodeling and sensorimotor loop remodeling with facing motion; the detection paradigm for performing the functional reconstruction of the neural loop facing the movement is a repeated extending movement task of the target-oriented hemiplegic upper limb; the detection paradigm facing sensory-motor nerve loop function reconstruction is based on a central-peripheral combined magnetoelectric stimulation technology to activate a motor-related brain region and induce a proprioceptive task.

In the embodiment of the invention, the multitask-multimode fusion feature selection module is used for realizing the multitask-multimode fNIRS brain function fusion feature extraction related to the movement function of the user based on two-step feature dimension reduction formed by group-level statistics and function-related regression. In order to fully describe the brain function state, the system designs a plurality of typical task paradigms and calculates a fNIRS multi-mode brain function characteristic combination vector V under the plurality of typical task paradigms from an activation mode and a connection mode, so that the formed brain function characteristic indexes have large dimension and large quantity, in this case, if all the brain function characteristics are used as the characteristics in a classifier, the overfitting problem is caused, a large amount of redundant information exists, the classification performance is not ideal, and the brain function network related to a specific dysfunction is generally concentrated on a small part of all possible indexes. In order to reduce the redundancy of index features, the feature selection is realized by combining group-level statistical test and function-related regression fusion analysis, and the distinguishing capability of the feature subset is ensured. The specific steps are shown in fig. 5:

based on the near-infrared multi-mode brain function feature vector, preliminarily selecting a patient task-specific near-infrared brain function response feature with statistical significance by using group-level statistical examination, aiming at a resting state fNIRS brain function response feature, and performing feature selection by using significance difference between a patient and a healthy control group by using the system; the system utilizes the significant difference between the patient's task state and resting state for the fNIRS brain function response characteristics under both motor executive and sensorimotor tasks for characteristic selection. And multi-mode near-infrared brain function characteristics under multiple tasks are preliminarily screened through the inter-group significance test. The group-level statistical test in the system does not use the label information of the samples to select the features, and reduces the bias of classification results.

On the basis of primary screening of the group-level statistical test on the features, clinical function evaluation information is fused, a regression model is established, the features with statistical significance are selected, the optimal feature subset related to the dysfunction in a specific state is identified, and multi-task-multi-mode fusion feature selection is realized. The motion function evaluation-based module provides a motion function classification label for the module through a K-means algorithm and calculates a comprehensive motion evaluation score in each motion function obstacle subclass based on a PCA algorithm. With reference to fig. 5, a multivariate piecewise regression model is respectively established according to the task-specific near-infrared brain function response characteristics and the motor function comprehensive evaluation scores, the comprehensive motor function scores of the motor function disorder subclasses are selected as segmentation nodes, the comprehensive motor scores are used as dependent variables, the near-infrared brain function characteristics are used as independent variables, the mapping relation analysis between the fNIRS brain function indexes of the stroke patients and the motor function disorders in the resting state is realized, and the resting state fNIRS brain function maps of the stroke patients with mild, moderate and severe motor function disorders are established; analyzing a mapping relation between fNIRS brain function response indexes of stroke patients induced by movement execution tasks and movement dysfunction, and constructing fNIRS brain function maps related to task execution of stroke mild, moderate and severe movement dysfunction patients; and (3) analyzing the mapping relation between the fNIRS brain function response indexes of the stroke patients induced by the sensory-motor tasks and the motor dysfunction, and constructing fNIRS brain function maps related to the sensory-motor tasks of the stroke patients with mild, moderate and severe motor dysfunction. In the regression process, the least square sum of the difference between the given function value and the model predicted value is used as a loss function, and the final fitting parameters are calculated by using a least square method and a gradient descent method. The module is used for performing regression fusion analysis on specific response indexes of brain functions under specific tasks and dyskinesia assessment, comprehensively screening multitask multimode near-infrared brain function specific response characteristics which are obviously related to the brain function of a stroke patient and the motor function, and constructing the fNIRS brain function maps of the stroke patient with different dyskinesia degrees under a multitask state. In addition, the module further realizes the reduction of the dimension of the feature data, the filtration and the extraction of the fusion feature vector of the fNIRS brain function index and the clinical motion assessment, and provides the near-infrared brain function features which are obviously related to the motion function for the near-infrared index and stroke rehabilitation mapping module.

In the embodiment of the invention, the fNIRS brain function and motor rehabilitation mapping module is used for inputting the multitask-multimode fNIRS brain function fusion index and the comprehensive motor rehabilitation index into the integrated classifier to realize the classification mapping of the fNIRS brain function and rehabilitation effect. According to the multitask fNIRS brain function index screened out by the multitask-multimode fusion characteristic selection module and clinical evaluation fusion characteristics, the rehabilitation effect evaluation module provides sample labels for the module, wherein the sample labels comprise three types of good recovery, general recovery and poor recovery; the clinical data are often unbalanced, so that the recovery rate of stroke patients is unevenTherefore, an integrated multi-classification Support Vector Machine (SVM) model is designed, as shown in figure 6. The integrated SVM model mainly comprises N multi-classification SVM models, wherein the numerical value of N is determined according to the situation; each SVM model will input a balanced sub-training set containing the same number of patients with good recovery, general recovery, and poor recovery. The sub-training set generation principle is that according to the number of the minority samples, the sample amount consistent with the number of the minority samples is randomly selected from the majority samples; each sub-training set formed in the way is different, but a few classes of samples are fully utilized, and each sub-training set is guaranteed to be balanced. Constructing on each multi-classification SVM model

And 3 classifiers, where k =3 in the embodiment of the present invention, are constructed, and the value of k is merely an example, and the present invention is not limited thereto. The model adopts a radial basis kernel function (RBF) to carry out high-dimensional mapping on near-infrared brain functional characteristics, and adopts a grid optimization method to search an optimal penalty factor C and a kernel radius sigma parameter, and the specific process is as follows:

assuming that the effective range distribution of the penalty factor C and the kernel radius sigma is a and b, and the search step is k, the coordinate system (C) is obtained _i ,σ _i ) C in _i ＝C*i/k,σ _i = σ i/k, for different C's, according to cross validation calculation procedure _i ,σ _i And traversing combination is carried out, the classification accuracy of the sample is calculated, and after traversing is finished, the classifier parameter with the maximum cross validation classification accuracy is selected to establish a classifier model. And determining the classification result of each multi-classification SVM through the weight value according to a voting method. The integrated SVM model combines a plurality of SVM by adopting a bagging majority voting method to form a strong learner.

The method comprises the steps of forming a data set consisting of brain function indexes and function rehabilitation labels after the fNIRS data set and the clinical evaluation set of a patient with brain injury in a multi-task state are collected clinically and processed, using the data set as the input of an integrated learning device for training and parameter optimization, and selecting the parameter with the highest classification accuracy as an fNIRS brain function and motor rehabilitation mapping model, namely constructing a near-infrared brain function map for brain function dynamic evaluation and rehabilitation effect prediction.

In the embodiment of the invention, the fNIRS brain function map is a data set formed by brain function indexes and function rehabilitation labels after the acquired fNIRS data set and clinical evaluation set in a brain injury user specific detection paradigm are processed, the data set is used as the input of an integrated learning device for training and parameter optimization, the parameter with the highest classification accuracy is selected as the fNIRS brain function and motor rehabilitation mapping classification model, and a fNIRS brain function map knowledge base for motor rehabilitation is constructed.

And step S150, providing a visual evaluation report through a brain function state evaluation module based on the acquired matching classification result.

In this step, based on the fNIRS brain function map knowledge base for motor rehabilitation constructed in step S140, brain function response characteristics of the stroke patient in a specific detection paradigm are input into the fNIRS brain function map module for classification and matching, the brain function state of the patient can be judged according to the result of the classification and matching, and a brain function state evaluation result is output and a visual evaluation report is provided to be fed back to the patient and medical care personnel.

In accordance with the above method, the present invention further provides a system, as shown in fig. 2, including: the brain function evaluation system comprises a motion detection paradigm selection module, a near-infrared brain function acquisition module, a brain function response characteristic analysis module, a fNIRS brain function atlas module and a brain function state evaluation report module; the motion detection paradigm selection module is used for selecting a specific brain function detection paradigm according to the motion function of a user; the near-infrared brain function acquisition module is used for acquiring multichannel cerebral blood oxygen signals of a user in a specific detection paradigm in real time; the brain function response characteristic analysis module is used for preprocessing the collected brain blood oxygen signals under the specific detection paradigm, calculating near-infrared multi-mode brain function characteristic parameters and constructing brain function response characteristic vectors induced by the specific detection paradigm; the fNIRS brain function map module is used for constructing a mapping model of the fNIRS brain function features and the motor rehabilitation according to an integrated multi-classification method, namely a fNIRS brain function map knowledge base related to the motor rehabilitation; inputting brain function response characteristics induced by the user specific detection paradigm obtained by the brain function response characteristic analysis module into the module for matching and classification, and judging the neural function state of the user; and the brain function state evaluation report module is used for outputting the matching result obtained by the fNIRS brain function atlas module, providing a visual evaluation report and feeding back the visual evaluation report to the user and medical staff.

In an embodiment of the present invention, the system further includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is configured to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, as shown in fig. 7, the following steps are performed:

and S1, removing the possibility of magnetic and electric taboo according to the motor function condition of the brain injury patient, and selecting a detection paradigm in a motion detection paradigm selection module.

Step S2: the brain blood oxygen signals of the user under a specific detection paradigm are collected through a near-infrared brain function collecting device, and each state is not less than 10 minutes.

And step S3: and inputting the collected cerebral blood oxygen signals under the specific detection paradigm into a cerebral function response characteristic analysis module, and calculating near-infrared multi-mode cerebral function parameters to obtain cerebral function multi-mode response characteristics induced by the specific detection paradigm.

And step S4: training the constructed integrated multi-classifier by taking multi-task-multi-mode fNIRS brain function characteristics and clinical rehabilitation effect classification labels as data sets, storing, training, adjusting parameters and testing to form an optimal detection model, namely forming a near-infrared brain function map for cerebral apoplexy motor function rehabilitation; and inputting the brain function response characteristics induced by the specific detection paradigm into a fNIRS brain function map knowledge base for matching and classification, and judging the nerve function state of the user.

Step S5: and matching and outputting the fNIRS brain function map to obtain a result for visual report, realizing the brain function state evaluation of the stroke patient, feeding the brain function state evaluation back to the user and medical personnel, assisting in clinical function evaluation and treatment response detection, and providing a treatment target for personalized intervention.

The invention provides a dynamic rehabilitation assessment method and a dynamic rehabilitation assessment system based on a fNIRS brain function map, wherein the neural function state of a motor dysfunction population after brain injury is dynamically assessed by combining a near-infrared brain function imaging technology with a specific motion detection paradigm, and an important imaging basis is provided for formulation and optimization of a rehabilitation scheme; the multi-mode feature fusion method can be adopted to combine activation and connection network information, and a plurality of different brain function analysis methods can be used for more abundantly describing brain function maps; the integrated multi-classification support vector machine model can be adopted to solve the limitation of unbalanced clinical data and solve the mapping prediction model bias phenomenon caused by data unbalance. The brain function specific response and remodeling characteristics under various typical task paradigms can be fused, and a dynamic mapping model of near-infrared brain function indexes and motor function rehabilitation is constructed; the technical advantages of the near-infrared brain function imaging technology of motion resistance and electromagnetic interference resistance can be utilized, a typical monitoring paradigm facing to a motor neural loop is designed, brain function characteristics related to motion functions are fully obtained, and brain stroke neural function assessment except for resting state is enriched; the fNIRS multi-mode brain function response characteristics under a resting state and a typical detection task paradigm can be extracted, a mapping model of the fNIRS multi-mode brain function response characteristics and motor function rehabilitation is established, a fNIRS brain function map facing motor dysfunction is established, and dynamic assessment of brain function states of brain injury patients is achieved. The dynamic evaluation of the brain function rehabilitation is more real-time, more comprehensive, more continuous, more objective, more accurate and more targeted, so that the evaluation of the brain function state of the stroke patient can be aimed at, the clinical function evaluation and the treatment response detection are assisted, and the treatment target is provided for the personalized intervention. The method is simple to operate, and can realize the evaluation of the brain function states of patients with different rehabilitation levels.

Embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the foregoing steps of the edge computing server deployment method. The computer readable storage medium may be a tangible storage medium such as Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

Those of ordinary skill in the art will appreciate that the various illustrative components, systems, and methods described in connection with the embodiments disclosed herein may be implemented as hardware, software, or combinations of both. Whether this is done in hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the invention are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link.

It is to be understood that the invention is not limited to the specific arrangements and instrumentality described above and shown in the drawings. A detailed description of known methods is omitted herein for the sake of brevity. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present invention are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications and additions or change the order between the steps after comprehending the spirit of the present invention.

Features that are described and/or illustrated with respect to one embodiment may be used in the same way or in a similar way in one or more other embodiments and/or in combination with or instead of the features of the other embodiments in the present invention.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made to the embodiment of the present invention by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic rehabilitation assessment method based on an fNIRS brain function map, which is characterized by comprising the following steps:

selecting a brain function detection paradigm matched with the user's athletic ability from a plurality of pre-stored brain function detection paradigms based on the user's athletic ability;

calculating near-infrared multi-mode brain function characteristic parameters based on the acquired brain blood oxygen signals, wherein the multi-modes are an activation mode and a connection mode, and the parameters form multi-mode brain function characteristic combination vectors;

2. The method of claim 1, wherein the pre-stored plurality of brain function detection paradigms includes a typical fNIRS brain function detection paradigms for resting and motor-oriented execution of neural circuit remodeling and sensory-motor neural circuit remodeling; the detection paradigm facing the movement execution neural loop remodeling is specifically based on a target guidance hemiplegia upper limb repeated extension movement task; the detection paradigm facing sensory-motor nerve loop remodeling is a central-peripheral combined magnetoelectric stimulation technology, and the detection paradigm activates a motor-related brain region and induces proprioception to be transmitted into a central cortex.

3. The method according to claim 1, wherein the method comprises combining network information of an activation mode and a connection mode by using a brain function multi-mode feature fusion method, wherein the activation mode is used for describing activation distribution of all areas of the brain, and the connection mode is used for representing information interaction processes under typical tasks of the brain.

4. The method of claim 1, further comprising performing time-frequency transformation on the cerebral blood oxygen signal using a wavelet transform time-frequency analysis method based on the ability of the wavelet transform time-frequency analysis method to decouple signal components and provide local phase information.

5. The method according to claim 1, wherein the near-infrared multi-mode brain function characteristic parameters comprise hemisphere activation laterality, hemisphere connection laterality, hemisphere autonomy coefficients, undirected weighted topology parameters, directed weighted topology parameters, and the like.

6. The method of claim 1, further comprising classifying motor dysfunctions including mild, moderate and severe dysfunctions using cluster analysis; clustering the motion function assessment scale set into mild, moderate and severe motion dysfunction subsets by using a K-means clustering algorithm; the set of athletic performance assessment scales includes a functional exercise assessment (Fugl-Meyer, FM) scale, a grip index, and an upper limb action study ARAT scale.

7. The method of claim 1, further comprising selecting a brain functional response feature based on a combination of group-level statistical tests and functionally-related regression fusion analysis;

the group-level statistical test is used as a characteristic screening and filtering method to extract a brain function characteristic index with statistical significance;

8. The method of claim 1, further comprising addressing the imbalance problem for each set of data with an integrated multi-classification Support Vector Machine (SVM) model based on imbalance of clinical data; the integrated multi-classification support vector machine SVM model adopts a radial basis kernel function RBF to carry out high-dimensional mapping on near-infrared brain function characteristics, and a grid optimization method is adopted to search an optimal punishment factor and a kernel radius parameter.

9. A dynamic rehabilitation assessment system based on fNIRS brain function maps, the system comprising:

the motion detection paradigm selection module is used for selecting a brain function detection paradigm matched with the motion capability of the user from a plurality of pre-stored brain function detection paradigms according to the motion capability of the user;

the near-infrared brain function acquisition module is used for acquiring multichannel brain blood oxygen signals induced by a user under the matched brain function detection mode in real time;

the fNIRS brain function map module is used for constructing a fNIRS brain function characteristic and motor rehabilitation mapping model according to an integrated multi-classification method, and inputting the obtained brain function multi-mode response characteristic vector to the module for matching classification;

10. The system of claim 9, wherein the fNIRS brain function map module comprises: the system comprises a motion function evaluation module, an fNIRS data acquisition and multi-mode feature calculation module, a multi-task and multi-mode fusion feature selection module and an fNIRS brain function and motion rehabilitation mapping module;

the motion function evaluation module is used for storing the limb motion function evaluation results of the user at different time nodes, and the results comprise comprehensive motion dysfunction evaluation and comprehensive motion function rehabilitation evaluation results;

11. A dynamic rehabilitation assessment system based on a fNIRS brain function atlas, the system further comprising a processor and a memory, wherein the memory has stored therein computer instructions for executing the computer instructions stored in the memory, the system implementing the steps of the method according to any one of claims 1 to 8 when the computer instructions are executed by the processor.

12. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of a method according to any one of claims 1 to 8.