Brain function evaluation system for patients with disturbance of consciousness based on neural multi-mode monitoring technology

Abstract

The invention relates to a brain function evaluation system for a patient with disturbance of consciousness based on a neural multi-modal monitoring technology, and belongs to the technical field of biomedicine, and the system comprises a multi-modal data acquisition module which is used for acquiring multi-modal data of an evaluated person, the multi-modal data comprises clinical behavioral data, cerebral blood flow data, brain image data and neural electrophysiological data, the neural electrophysiological data comprises electroencephalogram and evoked potential data, and the brain image data comprises brain structure images and brain function images; the multi-modal data evaluation module is used for evaluating the clinical behavioral data, the cerebral blood flow data, the brain image data and the neural electrophysiological data to obtain an evaluation value of each modal data; and the comprehensive evaluation module is used for processing the plurality of evaluation values by using a disturbance of consciousness awakening prediction model to obtain an evaluation result. The method has the effects that the brain function of the patient with the disturbance of consciousness is evaluated by using a multi-mode neural function monitoring technology, and the evaluation accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of biomedicine, and in particular to a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology. Background Art

[0002] Impaired consciousness is the most common clinical manifestation of a variety of critical neurological diseases such as stroke, craniocerebral trauma, and hypoxic-ischemic encephalopathy. The extent of brain damage in patients with impaired consciousness, whether they can "wake up", and whether their perception can be restored are one of the primary concerns of clinicians and patients' families, and are also major challenges faced by clinicians. Currently, the assessment of the degree of impaired consciousness mainly relies on clinical scores (such as the modified coma recovery scale [CRS-R]) or single physiological signals (such as electroencephalogram or brain imaging).

[0003] However, with the development of science and technology in recent years, many emerging neurological monitoring and brain information analysis technologies have been continuously applied in clinical practice, greatly improving the accuracy of clinical diagnosis and the understanding of consciousness disorders. The concept and method of multimodal neurological function monitoring are also gradually applied to critical care medicine practice. How to use multimodal neurological function monitoring technology to evaluate the brain function of patients with consciousness disorders has become one of the urgent problems to be solved in the biomedical field. Summary of the invention

[0004] In order to use multimodal neural function monitoring technology to evaluate the brain function of patients with impaired consciousness, the present application provides a brain function evaluation system for patients with impaired consciousness based on neural multimodal monitoring technology.

[0005] The present application provides a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology, which adopts the following technical solutions:

[0006] A multimodal data acquisition module, used to acquire multimodal data of the person being evaluated, wherein the multimodal data includes clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data, wherein the neuroelectrophysiological data includes electroencephalogram and evoked potential data, and the brain imaging data includes brain structure image and brain function image;

[0007] A multimodal data evaluation module, used to evaluate the clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data to obtain an evaluation value of each modality data;

[0008] The comprehensive evaluation module is used to process the multiple evaluation values ​​using the consciousness disorder awakening prediction model to obtain an evaluation result.

[0009] Optionally, the multimodal data evaluation module includes:

[0010] A first evaluation submodule is used to perform CRS-R scoring on the clinical behavioral data to obtain a clinical behavioral data score, and use the clinical behavioral data score as a first evaluation value;

[0011] A second evaluation submodule is used to determine the lesion grade based on the brain structure image, and use the lesion grade as a second evaluation value;

[0012] A third evaluation submodule, configured to determine a brain network score based on the brain function image, and use the brain network score as a third evaluation value;

[0013] a fourth evaluation submodule, configured to determine a blood flow grade based on the cerebral blood flow data, and use the blood flow grade as a fourth evaluation value;

[0014] a fifth evaluation submodule, for determining the Synek grade, the sleep EEG structure grade, the EEG reactivity grade, the ABCD pattern grade, and the phase synchronization based on the EEG, taking the Synek grade as the fifth evaluation value, taking the sleep EEG structure grade as the sixth evaluation value, taking the EEG reactivity grade as the seventh evaluation value, taking the ABCD pattern grade as the eighth evaluation value, and taking the phase synchronization as the ninth evaluation value;

[0015] The sixth evaluation submodule is used to determine the somatosensory potential grade and the auditory potential grade based on the evoked potential data, and use the somatosensory potential grade as the tenth evaluation value and the auditory potential grade as the eleventh evaluation value.

[0016] Optionally, the consciousness disorder awakening prediction model is composed of a time series neural network, which obtains the evaluation values ​​of the patient at multiple time points, forms numerical variables through normalization processing, and inputs the numerical variables into the time series neural network for feature extraction, and inputs the extracted features into the multi-layer perception layer of the time series neural network. The multi-layer perception layer analyzes and calculates the extracted features in a preset time step sequence to obtain a prediction of the state outcome result in the next time stage.

[0017] Optionally, the first evaluation submodule is specifically used to:

[0018] CRS-R scores were performed on hearing, vision, movement, language, communication and wakefulness respectively to obtain a total CRS-R score, which was used as a clinical behavioral data score.

[0019] Optionally, the second evaluation submodule is specifically used to:

[0020] Based on the 3T MRI T1, T2 and Flair sequences, the brain structure images were evaluated from localized to extensive, unilateral or bilateral perspectives to obtain the lesion grade.

[0021] Optionally, the third evaluation submodule is specifically used to:

[0022] Brain functional images were evaluated based on resting-state functional magnetic resonance imaging of the default network, dorsal attention network, executive control network, salience network, sensory input auditory network, visual network, and sensorimotor network to obtain a brain network score.

[0023] Optionally, the fourth evaluation submodule is specifically used to:

[0024] The blood flow grade was obtained based on the evaluation of the mean blood flow velocity and pulsatility index of the bilateral middle cerebral arteries through the temporal window or eye window by transcranial Doppler.

[0025] Optionally, the fifth evaluation submodule is specifically used to:

[0026] The EEG in the non-stimulated state was evaluated by Synek grading to obtain the Synek grading;

[0027] The modified Valente score was used to grade the patients' sleep EEG and obtain the sleep EEG structure grade;

[0028] Use sound, pain, passive eye opening and other stimuli to evaluate and grade EEG reactivity, to EEG reactivity grading;

[0029] Preprocess the EEG in the non-stimulation state, divide the EEG data into multiple frequency bands based on preset rules, and use the Welch method to analyze multiple frequency bands to obtain the EEG power spectrum;

[0030] Determine the ABCD pattern classification based on the EEG power spectrum;

[0031] Based on the formula Calculate the phase synchronization, where P xy (f) is the cross spectrum between x(t) and y(t), P xx (f) and P yy (f) is the corresponding autospectrum at frequency f, Coh(f) is the frequency band phase synchronization of the frequency band, where x corresponds to the electrode data of the left hemisphere and y corresponds to the right hemisphere data.

[0032] Optionally, the sixth evaluation submodule is specifically used to:

[0033] Acquiring the waveform, amplitude and latency of each wave of the detected somatosensory potential, and determining the somatosensory potential classification based on the waveform, amplitude and latency of each wave of the somatosensory potential;

[0034] The waveform, amplitude and latency of each wave of the detected auditory potential are acquired, and the auditory potential grade is determined based on the waveform, amplitude and latency of each wave of the auditory potential.

[0035] Optionally, also include:

[0036] Data preprocessing module, used to perform layer deletion, time correction, head motion correction, image registration, spatial normalization and spatial smoothing on functional magnetic resonance data;

[0037] It is also used for EEG channel localization, useless electrode removal, filtering, downsampling, segmentation, interpolation of bad electrodes, cutting of bad segments, independent component analysis, artifact removal, extreme value removal and re-referencing.

[0038] In summary, the present application includes at least one of the following beneficial technical effects:

[0039] The present invention collects and integrates neural multimodal detection data from four aspects, namely brain structure imaging, brain function imaging, cerebral blood flow and neuroelectrophysiology, to systematically evaluate the brain function of patients with impaired consciousness, and combines neural multimodal detection technology with clinical behavioral scale evaluation to improve the accuracy of diagnosis of the degree of impaired consciousness. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a structural block diagram of a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology in an embodiment of the present application.

[0041] Figure 2 It is a structural block diagram of the sub-modules in the multimodal data evaluation module in the embodiment of the present application.

[0042] Attached figure numbers: 100, multimodal data evaluation module; 200, multimodal data evaluation module; 201, first evaluation submodule; 202, second evaluation submodule; 203, third evaluation submodule; 204, fourth evaluation submodule 204; 205, fifth evaluation submodule; 206, sixth evaluation submodule; 300, comprehensive evaluation module; 400, data processing module. DETAILED DESCRIPTION

[0043] The present application is further described in detail below in conjunction with the accompanying drawings.

[0044] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make modifications to the present embodiment without any creative contribution as needed, but such modifications are protected by the patent law as long as they are within the scope of the claims of the present application.

[0045] / or" is just a way to describe the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0046] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0047] The present application embodiment provides a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology, such as Figure 1 As shown, the brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology includes a multimodal data acquisition module 100, a multimodal data assessment module 200 and a comprehensive assessment module 300.

[0048] The multimodal data acquisition module 100 is used to acquire multimodal data of the person being evaluated, wherein the multimodal data includes clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data, wherein the neuroelectrophysiological data includes electroencephalogram and evoked potential data, and the brain imaging data includes brain structure images and brain function images.

[0049] A multimodal data evaluation module 200 is used to evaluate clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data to obtain an evaluation value of each modality data;

[0050] Among them, Figure 2 As shown, the multimodal data evaluation module 200 includes a first evaluation submodule 201 , a second evaluation submodule 202 , a third evaluation submodule 203 , a fourth evaluation submodule 204 , a fifth evaluation submodule 205 , and a sixth evaluation submodule 206 .

[0051] The first evaluation submodule 201 is used to perform CRS-R scoring on the clinical behavioral data to obtain a clinical behavioral data score, and use the clinical behavioral data score as a first evaluation value;

[0052] Specifically, CRS-R scores were performed on hearing, vision, movement, language, communication and wakefulness respectively to obtain a total CRS-R score, which was used as a clinical behavioral data score.

[0053] Table 1 is the CRS-R scoring table.

[0054] Table 1

[0055]

[0056]

[0057]

[0058] Among them, * represents the minimum state of consciousness minus; ▅ represents the minimum state of consciousness plus; ○ represents leaving the minimum state of consciousness.

[0059] In this embodiment, according to the CRS-R scoring table in Table 1, the six aspects of hearing, vision, movement, language, communication and wakefulness of the evaluated person are scored by CRS-R respectively, and the CRS-R scores of the six aspects are summarized to obtain the CRS-R total score, which is used as the clinical behavioral data score.

[0060] The second evaluation submodule 202 is used to determine the lesion grade based on the brain structure image and use the lesion grade as the second evaluation value.

[0061] Specifically, based on 3T MRI T1, T2 and Flair sequences, brain structure images are evaluated from localized or extensive, unilateral or bilateral perspectives to obtain lesion grades.

[0062] In this embodiment, the T1 sequence, T2 sequence and Flair sequence are described:

[0063] T1 sequence: mainly used to observe the anatomical structure of brain tissue, such as white matter, gray matter and cerebrospinal fluid. In T1 images, white matter usually shows a slightly low signal, gray matter shows a medium signal, and cerebrospinal fluid shows a high signal; T2 sequence: very sensitive to water molecules in the brain, often used to observe edema or inflammatory response in the lesion site, T2 signal enhancement is usually related to the increase in water content in the lesion site; Flair sequence: that is, fluid attenuated inversion recovery sequence, which can effectively suppress cerebrospinal fluid signals and make lesions adjacent to cerebrospinal fluid clearer.

[0064] Lesion extent: Assess whether the lesion is localized or widespread. Localized lesions are usually confined to a specific area, while widespread lesions may involve multiple brain areas. Lesion location: Determine whether the lesion is unilateral or bilateral. Unilateral lesions only affect one side of the brain, while bilateral lesions affect both sides of the brain at the same time.

[0065] The grading method can be level one: normal; level two: unilateral cerebral hemisphere or corpus callosum lesions; level three: bilateral cerebral hemispheres, brainstem, or diffuse swelling lesions.

[0066] The evaluation process can be:

[0067] Image acquisition: Use 3T MRI equipment to obtain the patient's T1, T2 and Flair sequence images; Image analysis: Experienced radiologists or neurologists conduct detailed analysis of the images to evaluate the extent and location of the lesions; Lesion grading: Lesions are divided into different levels based on the evaluation results.

[0068] The third evaluation submodule 203 is used to determine a brain network score based on the brain function image, and use the brain network score as a third evaluation value.

[0069] Specifically, brain functional images were evaluated based on the default network, dorsal attention network, executive control network, salience network, sensory input auditory network, visual network, and sensorimotor network of resting-state functional MRI to obtain brain network scores.

[0070] In this embodiment, brain function imaging assessment is based on the evaluation of 7 high-order brain functional networks of resting-state functional magnetic resonance imaging.

[0071] In this embodiment, it includes data preprocessing, independent component analysis, network identification and selection, seed point correlation analysis, network evaluation and result interpretation.

[0072] Data preprocessing included: preprocessing of resting-state functional MRI data using SPM12, FSL, and in-house codes running under Matlab; co-registration of three-dimensional T1-weighted images with average EPI images (echo-planar imaging) to extract tissue probability maps of white matter, cerebrospinal fluid, and gray matter; removal of confounding factors by multiple linear regression, including head motion parameters, average WM and CSF signals, and linear trends obtained with fourth-order polynomials and outlier volumes; application of 0.01-0.1 Hz bandpass filtering to focus on the frequency range associated with neural activity; normalization of three-dimensional T1 and EPI images to the MNI template for inter-group comparison; and spatial smoothing of EPI images using a 6-mm full-width at half-maximum Gaussian filter to reduce noise and improve the signal-to-noise ratio.

[0073] Independent component analysis included: independent component analysis was performed using FSL's MELODIC tool to identify functional networks in the brain. The number of components was fixed to 30, and the threshold was set to z ≥ 3 to select significant components; ICA maps of healthy controls were generated, and binary templates of 7 networks were created.

[0074] Network identification and selection included: using fslcc to calculate the cross-correlation between the template of each network and the individual components of patients with impaired consciousness, and preliminarily selecting more likely neural networks; combining the consensus of two experienced observers on the spatial map, time series shape, and power spectral density to determine the components corresponding to each network.

[0075] Seed point correlation analysis included: using the DPARSF-A tool to perform seed point correlation analysis on 7 networks, including high-order networks (such as the default mode network DMN, executive control network ECN, dorsal attention network DAN, salience network SN) and low-order networks related to sensory perception (such as auditory network AN, visual network VN, sensorimotor network SMN).

[0076] Network assessment and result interpretation included scoring of brain networks for each patient by neuroimaging experts. Each node of the ICA map and SBA map was evaluated according to the following criteria: 0 (no map or unclear map), 1 (low map specificity and unclear clusters), 2 (neuronal maps have specific clusters at several levels), and 3 (neuronal maps have highly specific clusters at many levels). For SBA maps, the threshold was changed to reduce noise when necessary. The presence of a brain network was defined as a score ≥2 for at least one node in ICA or SBA. Based on the evaluation results, the differences in network connectivity and functional activity between DOC patients and healthy controls were analyzed. These differences may help understand the pathophysiological mechanisms of patients with disorders of consciousness and provide new insights into treatment and rehabilitation.

[0077] The fourth evaluation submodule 204 is used to determine the blood flow grade based on the cerebral blood flow data, and use the blood flow grade as a fourth evaluation value.

[0078] Specifically, the mean blood flow velocity and pulsatility index of the bilateral middle cerebral arteries were evaluated through the temporal window or the ocular window based on transcranial Doppler to obtain the blood flow grade.

[0079] In this embodiment, we can classify the blood flow into different levels according to the values ​​of the mean blood flow velocity (Vm) and the pulsatility index (PI).

[0080] Level 1 (normal): Vm value is within the normal range (the normal range may vary depending on factors such as age and gender); PI value is within the normal range, reflecting stable blood flow and moderate resistance.

[0081] Level 2 (mild abnormality): The Vm value is slightly lower or slightly higher than the normal range, but does not reach a significantly abnormal level; the PI value is slightly increased or decreased, which may indicate a slight change in vascular resistance or a slight decrease in blood flow stability.

[0082] Level 3 (moderate abnormality):

[0083] A Vm value that is significantly lower or higher than the normal range may indicate insufficient or excessive cerebral blood perfusion; a significantly increased or decreased PI value reflects a significant increase or decrease in vascular resistance and a significant decrease in blood flow stability.

[0084] Level 4 (severe abnormality):

[0085] The blood flow in the anterior and posterior circulations shows oscillatory waves, sharp contraction waves or disappearance of blood flow signals.

[0086] The fifth evaluation submodule 205 is used to determine the Synek grade, the sleep EEG structure grade, the EEG reactivity grade, the ABCD pattern grade and the phase synchronization based on the electroencephalogram, and uses the Synek grade as the fifth evaluation value, the sleep EEG structure grade as the sixth evaluation value, the EEG reactivity grade as the seventh evaluation value, the ABCD pattern grade as the eighth evaluation value, and the phase synchronization as the ninth evaluation value.

[0087] Specifically, the EEG in the non-stimulated state is evaluated by Synek grading to obtain the Synek grading; the EEG in the sleeping state is scored by Valente to obtain the sleeping EEG structure grading; the EEG in the stimulated state is evaluated to obtain the EEG reactivity grading; the EEG in the non-stimulated state is preprocessed, and the EEG data is divided into multiple frequency bands based on preset rules, and the Welch method is used to analyze multiple frequency bands to obtain the EEG power spectrum; the ABCD pattern grading is determined based on the EEG power spectrum; based on the formula Calculate the phase synchronization, where P xy (f) is the cross spectrum between x(t) and y(t), P xx (f) and P yy (f) is the corresponding autospectrum at frequency f, Coh(f) is the frequency band phase synchronization of the frequency band, where x corresponds to the electrode data of the left hemisphere and y corresponds to the right hemisphere data.

[0088] In this embodiment, Synek evaluates the EEG in the non-stimulated state, including: based on the Synek classification, the EEG of the DOC patient is divided into 12 levels: level I, regular alpha rhythm, some theta waves, responsive; level II A, theta waves are dominant, normal voltage, responsive; level II B, theta waves are dominant, low voltage, unresponsive; level III A, delta waves are dominant (150 μV), comprehensive, rhythmic, responsive; level III B, spindle wave coma; level III C, delta waves are dominant, low amplitude (50 μV), irregular, unresponsive; level III D, delta waves are dominant, medium-low amplitude (100-150 μV), irregular, often unresponsive; level IVA, burst-suppression; level IVB, alpha coma; level IVC, theta coma; level IID, low voltage delta (<20 μV); level V, suppression, electrical quiet (<2 μV).

[0089] In this embodiment, the Valente-based evaluation of the EEG during sleep includes: using the modified Valente score to divide the sleep EEG of DOC patients into 5 levels: Level 1 is a monophasic pattern; Level 2 is a cyclic alternating pattern; Level 3 is a spindle-like pattern; Level 4 is a spindle wave pattern; Level 5 is stage 3 sleep and / or rapid eye movement sleep pattern.

[0090] In this embodiment, the patient's electroencephalogram under stimulation is evaluated, including: performing the following X stimulations on the patient: clapping sound stimulation, calling the name, pressing the nail bed, passive eye opening, pressing the sternum, electroacupuncture of the philtrum, and warm water stimulation. Each stimulation lasts for 5 seconds, with a stimulation interval of 60 seconds, and each stimulation is repeated 3 times. Determined EEG reactivity is defined as: clear EEG background activity changes appear at least 2 times out of 3 for any one stimulus. Possible EEG reactivity is defined as: clear EEG background activity changes appear at least 1 time out of 3 for any one stimulus. No EEG reactivity is defined as: no clear EEG background activity changes appear for all stimuli.

[0091] Table 2 shows the EEG responsiveness assessment

[0092] Table 2

[0093]

[0094] In this embodiment, MATLAB2016b is used to preprocess the EEG data, and the power spectrum and brain network connection analysis are performed. The number in the row where the test box is located represents the number of times. Each stimulus must be tested 3 times for EEG reactivity, and the EEG reactivity is divided into 2 levels: EEG reaction exists and EEG reaction does not exist. At least 2 visible EEG changes after the same stimulus are considered to be EEG reactivity; otherwise, it is considered that EEG reactivity does not exist.

[0095] 1) Preprocessing

[0096] Two qualified EEG experts manually removed significant motion artifacts and extracted artifact-free EEG segments of 5 seconds duration. First, EEG data were processed with a notch filter (50 Hz) and a bandpass filter (0.1-70 Hz) successively. Then, independent component analysis was applied to remove blink and eye movement artifacts. Finally, the signal was re-referenced to an offline common mean.

[0097] 2) Power spectrum analysis

[0098] The EEG data of each lead is divided into four different frequency bands: Delta (0.5-4Hz), Theta (4-8Hz), Alpha (8-13Hz) and Beta (13-30Hz). Higher frequency bands are excluded because they are easily affected by muscle artifacts. The signal power spectrum analysis is performed using the Welch method. First, the signal x(n) is decomposed into L data segments x of length M. i (n)(i=1,2,…,L,n=1,2,…,M). Then, the power spectrum of the i-th data segment is expressed as:

[0099]

[0100] in is the normalization coefficient, ensuring that the power spectrum obtained is an asymptotically unbiased estimate; d2(n) is the window function, w is the angular frequency; e is the base of the natural logarithm; j is the imaginary unit, i.e. j 2 =-1.

[0101] After calculating the power spectrum of each data segment, the average power spectrum of all L segments is calculated, and the Welch power spectrum estimate of the signal x(n) is obtained:

[0102]

[0103] in, is the normalization coefficient; d2(n) is the window function, w is the angular frequency; e is the base of the natural logarithm; j is the imaginary unit, i.e. j 2 =-1.

[0104] 3) EEG pattern analysis

[0105] EEG patterns were determined based on the EEG power spectrum and then classified into four categories based on the “ABCD” model, which assesses the integrity of thalamocortical circuits and the ability to retain consciousness.

[0106] Pattern A is a complete loss of thalamocortical integrity, with the EEG power spectrum limited to <4Hz; Pattern B is a narrow oscillation of layer V pyramidal cells in the theta frequency range (4-7Hz), which is caused by the depolarization of neocortical neurons and the inhibition of membrane potential; Pattern C refers to the oscillation of theta and beta frequencies due to the partial recovery of neocortical membrane potential and the coincident burst of stagnant thalamic neurons; Pattern D is the ignition pattern of normal neocortical neurons at alpha and beta frequencies. The highest level of EEG pattern was selected from the four leads C3, C4, O1 and O2.

[0107] 4) Brain network coherence analysis

[0108] The EEG network coherence is calculated by analyzing the symmetrical electrode data of the left and right hemispheres. Coherence represents the linear relationship between two signals at a specific frequency and is defined as:

[0109]

[0110] Where P xy (f) is the cross spectrum between x(t) and y(t), P xx (f) and P yy (f) is the corresponding autospectrum at frequency f, where x corresponds to the electrode data of the left hemisphere and y corresponds to the right hemisphere data.

[0111] The sixth evaluation submodule 206 is used to determine the somatosensory potential grade and the auditory potential grade based on the evoked potential data, and use the somatosensory potential grade as the tenth evaluation value and the auditory potential grade as the eleventh evaluation value.

[0112] Specifically, the waveforms, amplitudes and latency of each wave of the detected somatosensory potential are obtained, and the grade of the somatosensory potential is determined based on the waveforms, amplitudes and latency of each wave of the somatosensory potential; the waveforms, amplitudes and latency of each wave of the detected auditory potential are obtained, and the grade of the auditory potential is determined based on the waveforms, amplitudes and latency of each wave of the auditory potential.

[0113] In this embodiment, the somatosensory evoked potential test uses bilateral median nerve stimulation, and the recording electrodes are placed on C3' or C4' (1 cm posterior to C4 after C3), Erb point and Cv7 spinous process of the contralateral head, and the reference electrode is placed at the contralateral Erb point to record the waveform, amplitude and latency of each wave, that is, the waveform, amplitude and latency of each wave of the somatosensory potential.

[0114] Furthermore, the brainstem auditory evoked potential test is performed by click short sound stimulation, with the recording electrode placed subcutaneously behind one ear and the reference electrode placed at the midpoint of the forehead to record the waveform, amplitude and latency of each wave, that is, the waveform, amplitude and latency of each wave of the auditory potential.

[0115] The waveforms, amplitudes and latencies of each wave detected by the electrodes corresponding to C3 or C4, Erb point and Cv7 spinous process placed on the contralateral head are obtained, and the somatosensory potential grade is determined based on the waveforms, amplitudes and latencies of each wave; the waveforms, amplitudes and latencies of each wave detected by the electrodes placed subcutaneously behind one ear and at the midpoint of the forehead are obtained, and the auditory potential grade is determined based on the waveforms, amplitudes and latencies of each wave.

[0116] In this example, auditory evoked potentials are graded as follows:

[0117] Grade 1: Brainstem auditory evoked potentials are present bilaterally.

[0118] Grade 2: Only unilateral brainstem auditory evoked potentials are present.

[0119] Grade 3: Brainstem auditory evoked potentials are absent bilaterally.

[0120] In this embodiment, the somatosensory potential classification includes level 2, level 2 and level 3.

[0121] Grade 1: Somatosensory evoked potentials are present on both sides.

[0122] Grade 2: Only unilateral somatosensory evoked potentials are present.

[0123] Grade 3: Somatosensory evoked potentials are absent on both sides.

[0124] The comprehensive evaluation module 300 is used to process multiple evaluation values ​​using the consciousness disorder awakening prediction model to obtain an evaluation result.

[0125] Among them, the consciousness disorder awakening prediction model is composed of a time series neural network. The consciousness disorder awakening prediction model obtains the patient's evaluation values ​​at multiple time points, forms numerical variables through normalization processing, inputs the numerical variables into the time series neural network for feature extraction, and inputs the extracted features into the multi-layer perception layer of the time series neural network. The multi-layer perception layer analyzes and calculates the extracted features according to the preset time step sequence to obtain the prediction of the state outcome result in the next time stage.

[0126] In this embodiment, all evaluation values ​​are input into the consciousness disorder awakening prediction model to obtain an evaluation result, wherein the evaluation result is a data value, and the data value is a percentage, that is, the patient's awakening probability.

[0127] Among them, the multi-layer perception layer includes a fully connected input layer, a hidden layer and a regression output layer based on the ReLU activation function.

[0128] A brain function assessment system for patients with consciousness disorders based on neural multimodal monitoring technology also includes: a data preprocessing module 400, which is used to delete layers, time correction, head movement correction, image registration, spatial standardization and spatial smoothing of functional magnetic resonance data of brain function images; and is also used to perform channel positioning, removal of useless electrodes, filtering, downsampling, segmentation, interpolation of bad electrodes, cutting of bad segments, independent component analysis, removal of artifacts, removal of extreme values ​​and re-reference of electroencephalograms.

[0129] The terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus.

[0130] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of application involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned application concept. For example, the above features are replaced with (but not limited to) technical features with similar functions applied in the present application.

Claims

1. A brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology, characterized in that: include: A multimodal data acquisition module, used to acquire multimodal data of the person being evaluated, wherein the multimodal data includes clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data, wherein the neuroelectrophysiological data includes electroencephalogram and evoked potential data, and the brain imaging data includes brain structure image and brain function image; A multimodal data evaluation module, used to evaluate the clinical behavioral data, cerebral blood flow data, brain imaging data and neuroelectrophysiological data to obtain an evaluation value of each modality data; The comprehensive evaluation module is used to process the multiple evaluation values ​​using the consciousness disorder awakening prediction model to obtain an evaluation result.

2. According to claim 1, a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology is characterized in that: The multimodal data evaluation module comprises: A first evaluation submodule is used to perform CRS-R scoring on the clinical behavioral data to obtain a clinical behavioral data score, and use the clinical behavioral data score as a first evaluation value; A second evaluation submodule is used to determine the lesion grade based on the brain structure image, and use the lesion grade as a second evaluation value; A third evaluation submodule, configured to determine a brain network score based on the brain function image, and use the brain network score as a third evaluation value; a fourth evaluation submodule, configured to determine a blood flow grade based on the cerebral blood flow data, and use the blood flow grade as a fourth evaluation value; a fifth evaluation submodule, for determining the Synek grade, the sleep EEG structure grade, the EEG reactivity grade, the ABCD pattern grade, and the phase synchronization based on the EEG, taking the Synek grade as the fifth evaluation value, taking the sleep EEG structure grade as the sixth evaluation value, taking the EEG reactivity grade as the seventh evaluation value, taking the ABCD pattern grade as the eighth evaluation value, and taking the phase synchronization as the ninth evaluation value; The sixth evaluation submodule is used to determine the somatosensory potential grade and the auditory potential grade based on the evoked potential data, and use the somatosensory potential grade as the tenth evaluation value and the auditory potential grade as the eleventh evaluation value.

3. A system for evaluating brain function of patients with impaired consciousness based on neural multimodal monitoring technology according to claim 1 or 2, characterized in that: The consciousness disorder awakening prediction model is composed of a time series neural network. The consciousness disorder awakening prediction model obtains the evaluation values ​​of the patient at multiple time points, forms numerical variables through normalization processing, inputs the numerical variables into the time series neural network for feature extraction, and inputs the extracted features into the multi-layer perception layer of the time series neural network. The multi-layer perception layer analyzes and calculates the extracted features according to the preset time step sequence to obtain the prediction of the state outcome result of the next time stage.

4. According to claim 2, a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology is characterized in that: The first evaluation submodule is specifically used for: CRS-R scores were performed on hearing, vision, movement, language, communication and wakefulness respectively to obtain a total CRS-R score, which was used as a clinical behavioral data score.

5. According to claim 2, a brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology is characterized in that: The second evaluation submodule is specifically used for: Based on the 3T MRI T1, T2 and Flair sequences, the brain structure images were evaluated from localized to extensive, unilateral or bilateral perspectives to obtain the lesion grade.

6. The brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology according to claim 2, characterized in that: The third evaluation submodule is specifically used for: Brain functional images were evaluated based on resting-state functional magnetic resonance imaging of the default network, dorsal attention network, executive control network, salience network, sensory input auditory network, visual network, and sensorimotor network to obtain a brain network score.

7. The brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology according to claim 2, characterized in that: The fourth evaluation submodule is specifically used for: The blood flow grade was obtained based on the evaluation of the mean blood flow velocity and pulsatility index of the bilateral middle cerebral arteries through the temporal window or eye window by transcranial Doppler.

8. The brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology according to claim 2, characterized in that: The fifth evaluation submodule is specifically used for: The EEG in the non-stimulated state was evaluated by Synek grading to obtain the Synek grading; The modified Valente score was used to grade the patients' sleep EEG and obtain the sleep EEG structure grade; Use sound, pain, passive eye opening and other stimuli to evaluate and grade EEG reactivity, to EEG reactivity grading; Preprocess the EEG in the non-stimulation state, divide the EEG data into multiple frequency bands based on preset rules, and use the Welch method to analyze multiple frequency bands to obtain the EEG power spectrum; Determine the ABCD pattern classification based on the EEG power spectrum; Based on the formula Calculate the phase synchronization, where P xy (f) is the cross spectrum between x(t) and y(t), P xx (f) and P yy (f) is the corresponding autospectrum at frequency f, Coh(f) is the frequency band phase synchronization of the frequency band, where x corresponds to the electrode data of the left hemisphere and y corresponds to the right hemisphere data.

9. The brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology according to claim 2, characterized in that: The sixth evaluation submodule is specifically used for: Acquiring the waveform, amplitude and latency of each wave of the detected somatosensory potential, and determining the somatosensory potential classification based on the waveform, amplitude and latency of each wave of the somatosensory potential; The waveform, amplitude and latency of each wave of the detected auditory potential are acquired, and the auditory potential grade is determined based on the waveform, amplitude and latency of each wave of the auditory potential.

10. The brain function assessment system for patients with impaired consciousness based on neural multimodal monitoring technology according to claim 1, characterized in that: Also includes: Data preprocessing module, used to delete layers, perform time correction, head motion correction, image registration, spatial normalization and spatial smoothing on functional magnetic resonance imaging data; It is also used for EEG channel localization, useless electrode removal, filtering, downsampling, segmentation, interpolation of bad electrodes, cutting of bad segments, independent component analysis, artifact removal, extreme value removal, and re-referencing.

Images