Methods and systems for predicting the efficacy of stem cell therapy for spinal cord injury

Abstract

This invention relates to the field of therapeutic efficacy prediction technology, specifically to a method and system for predicting the therapeutic efficacy of stem cell therapy for spinal cord injury. The method includes the following steps: acquiring multi-parameter metabolic sequences and constructing an activation rhythm atlas; identifying synchronous segments of neural repair; extracting the dominant metabolic pathways and evaluating electrical signal responses; generating alignment layers and predicting therapeutic trends. In this invention, by introducing the synchronous identification of trends in cross-temporal metabolism and electrophysiological indicators, dynamic mapping of stem cell activation rhythms is achieved. This is linked to changes in neural axon morphology and cerebrospinal fluid metabolic parameters, clarifying the distribution of key response points throughout the treatment cycle. A causal path is established based on the sequence of metabolic mutations and neural repair events. The evolutionary stages dominated by metabolic driving forces are extracted. The coordination trend between therapeutic response and intervention pathways is marked on the timeline. The number and location of trend convergence, regression, and interruption regions are identified, achieving coordinated prediction of repair progress and therapeutic trends.

Description

Technical Field

[0001] This invention relates to the field of efficacy prediction technology, and in particular to a method and system for predicting the efficacy of stem cell therapy for spinal cord injury. Background Technology

[0002] The field of efficacy prediction technology involves the prospective assessment of the outcomes of medical interventions and is an important component of healthcare management and clinical decision support. Core aspects of this technology include building predictive models based on biomedical information, estimating patient treatment responses using specific rules or training sets, and providing risk assessment and grading support before or during treatment. Systematically, this field encompasses multiple steps, including data source structuring (e.g., electronic medical records, laboratory tests, imaging information), feature extraction and normalization, model building (e.g., generating prediction rules through case review analysis), algorithm training and validation, and output and visualization of predicted values. It is widely applied in medical activities such as personalized treatment pathway recommendations, rehabilitation prognosis assessment, and treatment course management adjustments. In particular, the efficacy prediction method for stem cell therapy for spinal cord injury refers to a method that predicts efficacy based on patient basic information and clinical manifestations, using linear regression models or manual scoring mechanisms. This type of method mainly relies on medical history records, neurological function scoring scales, and expert experience-based statistical summaries of postoperative rehabilitation indicators. It predicts efficacy by setting static scoring standards or empirical formulas, and uses multi-factor manual weighted analysis, static comparison of physiological parameters, or time-point neural recovery trend methods to infer treatment effects.

[0003] Existing technologies rely solely on static scoring standards or linear estimation methods to predict efficacy before or during treatment, lacking the tracking and quantification of multidimensional physiological dynamic changes in patients. Especially when metabolic parameters and neural structure recovery are asynchronous, traditional methods struggle to identify potential neural remodeling pathways guided by metabolic mutations, resulting in predictions limited to single-factor assessments at specific time points. This fails to reflect the synergistic changes at different stages of treatment. In the early stages of intervention, when metabolism is activated before neuromorphological changes are apparent, traditional scoring mechanisms underestimate the progress of efficacy, affecting the suitability of individualized intervention strategies and the timeliness of feedback regulation. Summary of the Invention

[0004] To address the shortcomings of existing technologies in tracking and quantifying the multidimensional dynamic changes in patients' physiological processes, especially in cases where metabolic parameters and neural structure recovery are asynchronous, traditional methods struggle to identify potential neural remodeling pathways guided by metabolic mutations. This results in predictions limited to single-factor assessments at specific time points, failing to reflect the synergistic trends across different stages of treatment. Furthermore, in the early stages of intervention, when metabolism is activated before neuromorphological changes are apparent, traditional scoring mechanisms underestimate efficacy progress, impacting the suitability of individualized intervention strategies and the timeliness of feedback regulation. Therefore, this invention provides a method and system for predicting the efficacy of stem cell therapy for spinal cord injury. The technical solution is as follows:

[0005] On the one hand, a method for predicting the efficacy of stem cell therapy for spinal cord injury is provided, which includes: S1: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy; extract continuous phases with consistent change direction; record the time coverage of continuous phases in the complete sequence; and obtain the activation rhythm mapping map. S2: Based on the stem cell activation initiation time period in the activation rhythm mapping map, obtain the concentration change sequence of lactic acid, glutamate, and myelin fragments in the patient's cerebrospinal fluid during the time period, summarize the distribution of response points in the treatment cycle, and obtain a list of synchronous segments for neural repair. S3: Based on the time segments identified in the list of neural repair synchronization segments, trace the location of the first mutation behavior in the changes in cerebrospinal fluid metabolism, and obtain the set of metabolic driving pathway sequences according to the time sequence between the mutation point and the neural extension inflection point. S4: Call the number of the active mutation segment in the time period of the metabolic drive pathway sequence set, the corresponding electrical signal response amplitude sequence of the matching segment, evaluate the active release frequency and morphological response density of the implanted cells, and generate a repair progress alignment layer.

[0006] As a further embodiment of the present invention, the activation rhythm mapping map includes change stage label number, matching start time index, and stage occurrence frequency distribution; the list of neural repair synchronization segments includes response time node set, trend synchronization matching rate, and periodic distribution position index; the metabolic driving path sequence set includes mutation start coordinate, path evolution direction encoding, and metabolic leader identifier sequence number; and the repair progress alignment layer includes active response frequency band index, electrical signal amplitude mapping intensity, and release morphology density index.

[0007] As a further aspect of the present invention, the step of obtaining the activation rhythm mapping map specifically includes: S101: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy. Determine the same direction of the three sets of sequences, extract continuous time periods with the same direction of change, record the time index interval, and generate time intervals of the same direction stage. S102: Based on the three values ​​of each time period in the direction-consistent phase time interval, extract and encode the phase label combination composed of triples, summarize the number of times the phase label combination appears in the intervention cycle, and generate the phase label combination frequency value. S103: Call the frequency value of the stage label combination, and perform horizontal matching of the first complete occurrence time period of the stage label combination based on the patient's treatment start time. Compare the time start time corresponding to the label combination with the treatment start time in sequence. Record the location time point for the stage combination that meets the matching conditions. Use the label combination as the horizontal axis and the location time point as the vertical axis to generate an activation rhythm mapping map.

[0008] As a further aspect of the present invention, the step of obtaining the list of neural repair synchronization segments specifically includes: S201: Call the stem cell activation start time period indicated by the activation rhythm mapping map, obtain the lactate concentration sequence, glutamate concentration sequence and myelin fragment concentration sequence in the patient's cerebrospinal fluid by using the time interval as the index, and sort the three concentration data in order of time to generate the cerebrospinal fluid biochemical sequence group. S202: Based on the time range covered by the cerebrospinal fluid biochemical sequence group, the neural axon extension length sequence and axon branch number sequence within the time period are linked and called, and the two values ​​are arranged into a bivariate synchronous sequence in chronological order. The neural structural feature value is calculated, the structural remodeling change trajectory is identified, and a neural structural feature sequence is generated. The neural structural feature values ​​are calculated using the following formula: ; in, Representing the Neural structural feature values ​​within a time period Representing the The number of sampling points within the time period Representing the Time period Number of axonal branches at each time point Representing the Time period The length of nerve axon extension at each time point Representing the Time period The glucose concentration values ​​in the cerebrospinal fluid at each time point Representing the The average glucose concentration in cerebrospinal fluid at sampling points within the time period. Representing the The average value of the sampling points within the time period; S203: Call the time points in the neural structure feature sequence, perform point-to-point comparison operation based on the time position, identify the time points where the direction of concentration change is consistent with the direction of neural axis mutation, record multiple types of positions as a set of response points, and generate a group of points with consistent response trends. S204: Based on the time index of the points in the consistent response trend point group within the intervention period, summarize the continuous distribution segments of the response points in chronological order, and merge the cases where the time interval between adjacent response points is less than a fixed interval, mark them as continuous response segments, and generate a list of neural repair synchronization segments.

[0009] As a further aspect of the present invention, the step of obtaining the metabolic drive pathway sequence set specifically includes: S301: Based on the start and end indexes of time segments in the list of neural repair synchronization segments, trace back the sequences of three metabolic parameters—lactic acid, glutamate, and myelin fragments—in the cerebrospinal fluid within the time period, identify the inflection point where the parameters first show a change in value, record the time index of the mutation point and identify the parameter source to which it belongs, and generate a list of metabolic mutation time points. S302: Based on the position of the time points in the list of metabolic mutation time points, take the time point of the first change in the curve of the corresponding nerve axon extension length or axon branch number as the inflection point of the nerve change, construct a time path structure with metabolic change first and nerve change second between the two in chronological order, and connect the time nodes in the structure according to the combination index method to generate a sequence of leading paths. S303: The path sequence formed by the leading path is within the continuous evolutionary segment in terms of time span, and the metabolic change time index is earlier than the neural change time index. The path structures that meet the conditions are numbered and recorded, and the number of recurrences in the entire intervention period is counted. The path structure is paired with the frequency and labeled to generate a set of metabolic-driven path sequences.

[0010] As a further aspect of the present invention, the step of obtaining the repair progress alignment layer specifically includes: S401: Call the time period corresponding to the path structure with marked frequency in the metabolic drive path sequence set, extract the number information of the active mutation segment, match the data of the same time period in the electrical signal response amplitude sequence according to the segment time index, and organize them according to the segment number to generate an electrical signal response amplitude index table. S402: Based on the amplitude values ​​in each segment sequence in the electrical signal response amplitude index table, extract the periodic change frequency between the peak and trough of the electrical signal sequence in the segment as the active release frequency of the implanted cells, calculate the periodic change difference frequency value, count the total number of inflection points of the electrical signal per unit time, pair the two types of values ​​according to the segment number and align them on the time axis to generate a repair progress alignment layer.

[0011] As a further aspect of the present invention, the frequency value of the periodic variation difference is expressed by the formula: ; in, Representing the The frequency difference of periodic changes in segments Representing the The number of cycles extracted from segments. Representing the The first section of the segment The peak electrical signal amplitude corresponding to each cycle Representing the The first section of the segment The amplitude of the low-valley electrical signal corresponding to each cycle Representing the The first section of the segment The absolute difference between the mean values ​​of the preceding and following periods. It is a constant.

[0012] As a further aspect of the present invention, the method further includes step S5: S5: Based on the timeline covered by the repair progress alignment layer, call the position span of the pre-intervention, treatment, and expected post-intervention stages within the layer, retrieve the number and distribution of trend convergence areas, trend interruption areas, and trend reversal areas in the alignment layer, identify the direction of change in the degree of cooperation between the repair response and the intervention path, and obtain the prediction results of the change in efficacy. The predicted results of changes in therapeutic efficacy include the trend of changes in compatibility, the number of response segment types, and the time axis alignment distribution map.

[0013] As a further aspect of the present invention, the step of obtaining the predicted results of changes in therapeutic efficacy specifically includes: S501: Based on the complete timeline covered by the repair progress alignment layer, call the time position span corresponding to the pre-intervention, treatment and expected post-intervention stages within the layer, divide the time range of the layer according to the stage label, index and organize the time periods of the three stages, and generate a set of treatment stage time intervals. S502: Based on the layer trend within the multi-stage segment of the treatment phase time interval, retrieve the change trajectory of electrical signal frequency and morphological response density in the time dimension, record the trend convergence area with consistent trend direction and continuous amplitude within the multi-stage, mark the start and end time and quantity of multiple types of areas, and generate the layer trend distribution structure. S503: Based on the number and distribution time of the three types of regions in the differentiated stage in the layer trend distribution structure, compare the proportional changes between the trend convergence area and the interruption and withdrawal area, identify the direction of the change in the degree of cooperation between the intervention path and the repair response over time, summarize the cooperation characteristics under multiple stages according to the comparison relationship between the trend structure and the stage time, and obtain the prediction results of the change in efficacy.

[0014] On the other hand, the stem cell therapy efficacy prediction system for spinal cord injury is used to execute the above-mentioned stem cell therapy efficacy prediction method for spinal cord injury, and the system includes: The metabolic stage identification module acquires the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence. It calculates the change direction of the three sequences at adjacent time points in chronological order, calls the change direction value for joint judgment, and filters out continuous stages in which the three indicators change in the same time period with the same change direction to obtain the activation rhythm mapping map. The activation initiation response module calls the activation initiation time period identified in the activation rhythm mapping map, obtains the concentration sequence of lactate, glutamate and myelin fragments in cerebrospinal fluid, and calls the synchronous axon extension length and branch value to compare the changing trends and obtain a list of synchronous neural repair segments. The metabolic pathway deduction module traces the mutation points and neural extension inflection points in the cerebrospinal fluid metabolic changes based on the time points of the synchronous segments in the list of neural repair synchronous segments, calculates the time sequence, and filters the leading segments of metabolic mutations to obtain a set of metabolic driving pathway sequences. The electrical signal pairing module calls the metabolic driving pathway sequence set to obtain the corresponding electrical signal response amplitude and release morphology density values, and obtains the repair progress alignment layer. The efficacy trend prediction module aligns the timeline of the positioning layer with the repair progress, calls up multi-stage time intervals, and statistically analyzes the number and distribution of trend convergence segments, interruption segments, and regression segments to obtain the efficacy change prediction results.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By introducing the synchronous identification of trends in cross-temporal metabolic and electrophysiological indicators, dynamic mapping of stem cell activation rhythms is achieved. This is linked to changes in neural axon morphology and cerebrospinal fluid metabolic parameters, clarifying the distribution of key response points in the treatment cycle. A causal path is established based on the sequence of metabolic mutations and neural repair events, extracting the evolutionary stage dominated by metabolic driving forces. Cell activity and release density are characterized by combining electrical signal response characteristics. The coordination trend of therapeutic response and intervention path is marked on the time axis, identifying the number and location of trend convergence, regression, and interruption regions. This enables the synergistic prediction of repair progress and therapeutic trend. The achieved results include the temporal coupling identification of synchronous physiological changes and structural repair characteristics, the evolutionary labeling of metabolic-dominant patterns within the disease course, the alignment analysis of intervention rhythm and response efficiency, and the accurate deduction of therapeutic trend changes. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the workflow of the present invention; Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] Please see Figure 1 This invention provides a method for predicting the efficacy of stem cell therapy for spinal cord injury. The processing flow of this method may include the following steps: S1: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy. Arrange the three sequences in chronological order, extract continuous stages with consistent change directions, record the time coverage of continuous stages in the complete sequence, form stage label combinations, count the frequency of stage combinations in the intervention period, call the patient's treatment start time point to locate and compare the stage label combinations, and obtain the activation rhythm mapping map. S2: Based on the stem cell activation initiation time in the activation rhythm mapping map, obtain the concentration change sequence of lactic acid, glutamate and myelin fragments in the patient's cerebrospinal fluid during the time period, link and call the synchronous nerve axon extension length and axon branch number sequence, identify response points with consistent change trends in the two sets of data, summarize the distribution of response points in the treatment cycle, and obtain a list of synchronous nerve repair segments. S3: Based on the time segments identified in the list of synchronous neural repair segments, trace the location of the first mutation behavior in the changes in cerebrospinal fluid metabolism, establish a path sequence with metabolic changes as the leader according to the time sequence between the mutation point and the neural extension inflection point, screen the time periods in continuous evolution segments where changes in metabolic parameters take precedence over neural changes, and mark the frequency of occurrence of the path to obtain a set of metabolic-driven path sequences. S4: Call the number of the active mutation segment in the time period of the metabolic drive pathway sequence, match the electrical signal response amplitude sequence of the corresponding segment, evaluate the active release frequency and morphological response density of the implanted cells, and generate a repair progress alignment layer. S5: Based on the timeline covered by the alignment layer of the repair progress, call the position span of the pre-intervention, treatment, and expected post-intervention stages in the layer, retrieve the number and distribution of trend convergence areas, trend interruption areas, and trend reversal areas in the alignment layer, identify the direction of change in the degree of cooperation between the repair response and the intervention path, and obtain the prediction results of the change in efficacy. The activation rhythm mapping map includes the change stage label number, matching start time index, and stage occurrence frequency distribution. The list of neural repair synchronization segments includes the response time node set, trend synchronization matching rate, and periodic distribution location index. The metabolic driving path sequence set includes mutation start coordinates, path evolution direction encoding, and metabolic leader identifier sequence number. The repair progress alignment layer includes the active response frequency band index, electrical signal amplitude mapping intensity, and release morphology density index. The efficacy change prediction results include the cooperation change trend, the number of response segment types, and the time axis alignment distribution map.

[0021] The specific steps for obtaining the activation rhythm mapping map are as follows: S101: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy. Determine the same direction of the three sets of sequences, extract continuous time periods with the same direction of change, record the time index interval, and generate time intervals of the same direction stage. Multiple sets of data need to be collected during the monitoring phase after stem cell transplantation or injection. This process can be combined with the physiological parameter collection function set in the hospital ICU monitoring system. The program is set to automatically record the patient's blood glucose level, blood lactate concentration, and mitochondrial membrane potential changes every 5 minutes. The monitoring device starts recording continuous index data from the first minute after the patient is injected with stem cells. Each sequence data is simultaneously marked with a timestamp to generate a structured three-dimensional time series. The collected sequence data is cleaned and standardized to remove occasional jumps and noise. The three data corresponding to each time segment are compared to identify whether the direction of change is consistent. If the glucose concentration gradually decreases, the lactate release gradually decreases, and the mitochondrial membrane potential also decreases within a certain time period, it can be regarded as the direction of change of the indexes during that period. The start and end time points of recording this stage are regarded as a stage with consistent direction. For example, if a patient's three physiological indicators decrease simultaneously from the 10th minute to the 40th minute, this is determined to be a stage with consistent direction. The 10th minute to the 40th minute is used as the index range of this stage interval to prepare for subsequent stage feature extraction and generate the time interval of the stage with consistent direction.

[0022] S102: Based on the three values ​​of each time period in the direction-consistent phase time interval, extract and encode the phase label combination composed of triples, summarize the number of times the phase label combination appears in the intervention period, and generate the phase label combination frequency value. The three physiological data points at each time point are classified and coded using a mapping method. Continuous values ​​are grouped into several preset intervals to form label combinations. The rate of glucose decline is set into three categories: low, medium, and high. Lactate release and membrane potential changes are also categorized using the same standard. Standard labels are formed through this segmented coding. The combination of low glucose, medium lactate, and high membrane potential can be labeled as LMH. This label corresponds to the data status at each specific time point. A set of continuous label combinations is generated based on the time series. Three time points within a 5-minute interval are labeled as LMH, LMH, and MMH. A sliding window mechanism is used to select recurring label combinations as stage label combinations, and their frequency of occurrence throughout the entire treatment cycle is counted. If a certain LMH combination appears three times consecutively in the entire monitoring sequence, its frequency is recorded as 3. This frequency is saved as part of the stage combination feature vector. In practice, doctors can observe the repeatability of a certain type of metabolic rhythm based on this frequency value to determine whether it is a stable manifestation type. The high-frequency MMH label indicates the regularity of metabolic fluctuations, and the stage label combination frequency value is generated.

[0023] S103: Call the frequency value of the stage label combination, and perform horizontal matching of the first complete occurrence time of the stage label combination based on the patient's treatment start time. Compare the time start time corresponding to the label combination with the treatment start time in sequence. Record the location time point for the stage combination that meets the matching conditions. Use the label combination as the horizontal axis and the location time point as the vertical axis to generate an activation rhythm mapping map. Using the patient's treatment start time as the baseline time point, the time intervals for the first complete appearance of various tag combinations are recorded and located. For example, if a patient starts stem cell therapy at minute 0, and the first complete appearance of the LMH combination is detected between minutes 25 and 35, then minute 25 can be taken as the activation time point for that combination. If this time point meets the consistency conditions of combination matching in duration and direction, it will be recorded as the location time corresponding to the tag combination. Simultaneously, a mapping relationship map is constructed with the tag combination name on the horizontal axis and the first appearance time point on the vertical axis. The map lists combination tags such as LMH, MML, and HLH on the horizontal axis, and the vertical axis represents the number of minutes of the first activation of each tag type during the patient's treatment cycle, facilitating a direct observation of the activation sequence of various metabolic modes. This process is performed independently for different patients, forming personalized rhythm maps. In practice, doctors can use these maps to assess the initial response time distribution of patients to stem cell intervention. If a certain tag combination, such as MML, appears before minute 30 in multiple patients, an activation rhythm mapping map is generated.

[0024] The specific steps for obtaining the list of neural repair synchronization segments are as follows: S201: Call the stem cell activation initiation time period marked by the activation rhythm mapping map, obtain the lactate concentration sequence, glutamate concentration sequence and myelin fragment concentration sequence in the patient's cerebrospinal fluid by indexing the time interval, and sort the three concentration data in chronological order to generate the cerebrospinal fluid biochemical sequence set. The activation combination time range listed in the retrieval map was used as a time index for extracting cerebrospinal fluid (CSF) biochemical data. CSF samples were collected from patients within the corresponding time periods, and the concentrations of lactate, glutamate, and myelin fragments were measured sequentially. The sampling frequency was set to once every 10 minutes. Automated detection platforms such as liquid chromatography-mass spectrometry were used to measure the values ​​of the three types of indicators. The concentration values ​​of each type were numbered sequentially according to time to form a standardized concentration time series. To maintain data consistency, the units of the raw values ​​were standardized and the influence of extreme values ​​was removed. The LMH activation period for patients was set from the 30th to the 80th minute. The corresponding lactate concentrations (e.g., 2.3, 2.1, 1.9 mmol / L), glutamate concentrations (e.g., 6.5, 6.2, 5.9 μmol / L), and myelin fragment concentrations (e.g., 1.2, 1.3, 1.1 μg / mL) in this interval were automatically read. The three indicators were arranged sequentially according to time to generate a CSF biochemical sequence set.

[0025] S202: Based on the time range covered by the cerebrospinal fluid biochemical sequence group, the neural axon extension length sequence and axon branch number sequence within the time period are linked and called, and the two values ​​are arranged into a bivariate synchronous sequence in chronological order. The neural structural feature value is calculated, the structural remodeling change trajectory is identified, and a neural structural feature sequence is generated. Neural structural feature values ​​are calculated using the following formula: ; in, Representing the Neural structural feature values ​​within a time period Representing the The number of sampling points within the time period Representing the Time period Number of axonal branches at each time point Representing the Time period The length of nerve axon extension at each time point Representing the Time period The glucose concentration values ​​in the cerebrospinal fluid at each time point Representing the The average glucose concentration in cerebrospinal fluid at sampling points within the time period. Representing the The average value of the sampling points within the time period; The calculation logic of the formula is as follows: by weighting the number of nerve axon branches and the corresponding axon extension length collected within a selected time period with the square root, and normalizing the fluctuation of glucose concentration in cerebrospinal fluid, the comprehensive expression value of each group of samples is calculated, and the absolute value of the deviation between the average value and the sample expression mean is taken as the structural feature index of that time period; this logic integrates the dynamic relationship between the three time series variables to measure the stability of structural parameters under multi-source physiological signal interference; Neural structural feature values ​​are used to characterize the consistency and fluctuation trend of neuronal structural performance over a certain period of time. They are a quantitative indicator of the synchronicity of axonal morphological remodeling. The closer the value is to zero, the more uniform the neuromorphological structural performance at each time point is, and the more stable the remodeling activity is. Meaning of parameters and derivation of formulas: In the analysis of neural structural remodeling, the first The time period was set from April 1, 2025 to April 30, 2025. A total of 5 sets of synchronization data were collected during this period. , Parameter description and data source: : Number of axonal branches, identified by laser confocal imaging technology and automatic image segmentation and extraction algorithms, representing the total number of branches in a neuron image, with the data unit being the number of branches; Axon extension length: The three-dimensional geometric path length of each axon is calculated using a three-dimensional reconstruction algorithm. The data unit is micrometers, and the measurement tool is a high-resolution time-series microscopic imaging device. The glucose concentration in cerebrospinal fluid was analyzed by high performance liquid chromatography (HPLC) of the glucose content in each sample. The data unit is mmol / L. Calculate all samples from the above 5 samplings. The arithmetic mean; : Calculate the average of the values ​​of the expressions within the formula, i.e. As a balancing or baseline term, its main purpose is to offset the shift caused by overall background fluctuations over different time periods, reflecting the background level of structural features in the absence of sudden changes. Table 1: Sample Data Collection Table t 1 18 125 3.4 2 21 132 3.9 3 17 120 4 4 20 118 3.5 5 19 129 3.7 The calculation steps are as follows: calculate : ; Calculate the value of each expression. : ; ; ; ; ; calculate : ; Calculate the average of the overall expression: ; Calculate neural structural feature values: ; The results show that the neural structural feature value is 0 in the i-th time period, indicating that the expression of axonal branches is relatively stable. There is no obvious distribution deviation among the sampling points under the influence of cerebrospinal fluid glucose concentration. The neural structural feature sequence is highly consistent in this time period. The calculation result is directly used as the neural structural feature value in the step to generate the neural structural feature sequence.

[0026] S203: Call the time points in the neural structure feature sequence, perform point-to-point comparison based on the time position, identify the time points where the direction of concentration change is consistent with the direction of neural axis mutation, record multiple types of positions as a set of response points, and generate a group of points with consistent response trends. For each feature point containing a time index, a point-to-point comparison operation is performed. The direction of change of the three cerebrospinal fluid concentrations at the corresponding time point is matched with the direction of change of axon extension and branch number. If the two types of change directions are consistent, both are set to increase or both are set to decrease, then the time point is recorded as a consistent response trend point. The time index that meets this condition is added to the response point set. If, during the activation cycle, a decrease in lactate and glutamate concentrations is detected at minutes 35, 45, and 55, and axon length and branch number are also decreasing, then the three biochemical indicators and two neural structural parameters have consistent change directions, generating a consistent response trend point group.

[0027] S204: Based on the time index of the points in the consistent response trend point group within the intervention period, summarize the continuous distribution location segments of the response points in chronological order, and merge the cases where the time interval between adjacent response points is less than a fixed interval, mark them as continuous response segments, and generate a list of neural repair synchronous segments. The locations throughout the entire intervention period are summarized sequentially, and the distribution characteristics between consecutive time points are identified. If the time interval between two response points is less than the set maximum allowable interval threshold (less than 20 minutes), these two response points are classified as the same response segment. This process is then continuously expanded to adjacent response points for merging, generating multiple sets of continuous response time period markers. These response segments are called continuous response segments. During the output process, parameters such as start and end time index, number of points included, and average response intensity are marked for each segment to construct a standardized list of synchronous neural repair segments. If a patient records multiple response points within 30 to 70 minutes after intervention, with an interval of no more than 15 minutes between them, this segment is identified as a continuous response segment after merging and marked as "Segment 1". The time index 30-70 and the number of response points are recorded as 5. This information is used by doctors to assess the time range of synchronous neural repair after stem cell activation and to generate a list of synchronous neural repair segments.

[0028] The specific steps for obtaining the metabolic drive pathway sequence set are as follows: S301: Based on the start and end indexes of time segments in the list of synchronous neural repair segments, trace back the sequences of three metabolic parameters in cerebrospinal fluid—lactate, glutamate, and myelin fragments—within the time period, identify the inflection point where the parameters first show a change in value, record the time index of the mutation point and identify the parameter source to which it belongs, and generate a list of metabolic mutation time points. Using the corresponding start and end time indices as the data extraction window, sequences of three metabolic parameters—lactate, glutamate, and myelin fragments—were extracted from the patient's cerebrospinal fluid sample within this time period. The numerical changes of each metabolic parameter within this segment were continuously monitored. Starting from the beginning of the sequence, the values ​​of adjacent time points were compared to check for abrupt differences. A sudden and significant increase or decrease after several consecutive stable changes was recorded as an inflection point for that parameter. Upon detecting an inflection point, its time index was immediately recorded, along with the corresponding metabolic type and source. The identified mutation points formed a list of metabolic mutation time points. For example, if a patient's lactate concentration significantly decreased starting at 38 minutes, glutamate sharply increased at 42 minutes, and myelin fragments suddenly increased at 50 minutes within the 45-minute timeframe, these three time points and their corresponding parameters would be recorded. Each item in the list includes fields such as time index, metabolic parameter type, and direction of change, ensuring clear identification of the pre-relationship between metabolic activity and neural repair activity, thus generating a list of metabolic mutation time points.

[0029] S302: Based on the position of the time points in the list of metabolic mutation time points, take the time point of the first change in the curve of the corresponding nerve axon extension length or axon branch number as the inflection point of the neural change, construct a time path structure between the two in chronological order, with metabolic change preceding neural change, and connect the time nodes in the structure according to the combined index method to generate a sequence of leading paths. After finding the corresponding time point, the first response time point in the neural structural parameters, that is, the time point at which the change trend is detected in the axon extension length sequence or axon branch number sequence, is used to compare the onset time of neural changes with the time of metabolic mutation to determine whether metabolism precedes neural structural response. A time path structure between the two is constructed, with the metabolic mutation point as the path start point and the neural change point as the path end point. If the lactate concentration mutates at 35 minutes and the axon length first increases at 48 minutes, a path structure is formed: lactate-35→axon-48. The time index range between the two time points is recorded in the sequence structure. Each path is identified according to the metabolic source and neural structural type and assigned an independent number to generate a leader path that constitutes a sequence.

[0030] S303: Construct a path sequence through a leader path where the time span is within a continuous evolutionary segment and the metabolic change time index is earlier than the neural change time index. Number and record the path structures that meet the conditions, and count the number of recurrences throughout the entire intervention period. Pair the path structure with the frequency and label it to generate a set of metabolic-driven path sequences. Each pathway structure is evaluated sequentially to determine if it forms a complete chain of change within a continuous evolutionary segment, while simultaneously requiring that the metabolic mutation time index precedes the neural change time index. If both conditions are met, the pathway is identified as a valid metabolic-driven pathway. Qualified pathway structures are numbered and recorded in a table. Each record includes information such as the type of metabolic parameter, time span, and corresponding neural structural change type. The number of times the pathway structure recurs within the complete intervention period is then counted. If the "lactic acid drop → axonal elongation" pathway is detected in all 5 different stages, it is counted as a recurrence count of 5. The pathway structure and frequency are combined and labeled. If the pathway "LAC38→EXT52" appears 5 times, a record item "LAC38→EXT52×5" is formed. The pathway structure and frequency are paired to generate a set of metabolic-driven pathway sequences.

[0031] The specific steps for obtaining the alignment layer of the repair progress are as follows: S401: Call the time period corresponding to the path structure with labeled frequency in the metabolic drive path sequence set, extract the number information of the active mutation segment, match the data of the same time period in the electrical signal response amplitude sequence according to the segment time index, and organize them according to the segment number to generate an electrical signal response amplitude index table. Based on this time range, the specific segment numbers appearing in the intervention period are found, and records marked as active mutation segments are filtered out. The segment number is used as the primary key to call the corresponding electrical signal response amplitude sequence recorded in the electrophysiological monitoring. All electrical signal amplitude data within the segment start and end time index range are matched, and the acquired data are classified and organized according to the segment number. The start and end times and amplitude sequence values ​​are listed one by one according to the segment number, and timestamp information is attached to ensure that each amplitude data is strictly aligned with its corresponding metabolic driving pathway. A certain metabolic pathway "GLU42→EXT55" is marked as segment 3, with a time index of 42 to 55 minutes. That is, all amplitude change data within 42 to 55 minutes are extracted from the electrical signal records and classified under segment 3. This index table can serve as the basis for subsequent analysis of the relationship between neural discharge response and cell signal release, ensuring that each metabolic driving pathway has a clear corresponding electrical signal record at the electrophysiological level, which facilitates researchers to conduct multidimensional comparison and trend overlay analysis and generate an electrical signal response amplitude index table.

[0032] S402: Based on the amplitude values ​​in each segment sequence in the electrical signal response amplitude index table, extract the periodic change frequency between the peak and trough of the electrical signal sequence in the segment as the active release frequency of the implanted cells, calculate the periodic change difference frequency value, count the total number of inflection points of the electrical signal per unit time, pair the two types of values ​​according to the segment number and align them on the time axis to generate a repair progress alignment layer. The frequency value of the periodic variation difference is calculated using the formula: ; in, Representing the The frequency difference of periodic changes in segments Representing the The number of cycles extracted from segments. Representing the The first section of the segment The peak electrical signal amplitude corresponding to each cycle Representing the The first section of the segment The amplitude of the low-valley electrical signal corresponding to each cycle Representing the The first section of the segment The absolute difference between the mean values ​​of the preceding and following periods. It is a constant; Formula calculation logic: By extracting the peak and trough values ​​of the electrical signal waveform within each segment, the amplitude difference between them is calculated as the numerator. At the same time, the difference between the mean peak and trough values ​​of adjacent cycles is introduced to construct a difference adjustment factor. After merging, it is used for the square root normalization of the denominator. The overall calculation process iterates through the cycle samples in each segment, standardizes the amplitude difference of each cycle, takes the absolute value, and then obtains the representative value of the cycle frequency difference of the segment through a weighted average method. This formula also integrates the fluctuation amplitude, the trend of adjacent cycle changes, and the influence of noise interference to ensure the comparability of signal change characteristics in different segments within each cycle and improve the response capability to the cycle activity of electrical signals under non-stationary conditions. The frequency value of the period variation reflects the normalized fluctuation intensity of the peak-valley potential difference amplitude in each electrical signal period within a segment, and is used to measure the significance of the amplitude change during the period within a segment. Meaning of parameters and derivation of formulas: The electrical signal data acquisition process is based on an implantable microelectrode array with a sampling frequency of 1000Hz per second. For each segment, the time interval is 0.5 seconds, which is equivalent to 500 sampling points. The local periodic features are obtained by using a data window sliding method. In the segment a=3, the periodic parameters obtained from b=1 to b=4 are as follows: , ; , ; , ; , ; The peak and trough voltages mentioned above are determined by signal peak detection. The maximum and minimum points are identified when the first derivative of the waveform is zero. The average between the peak and trough forms a complete cycle. Periodic mean difference parameter It is calculated by the average of the peaks and troughs of the preceding and following periods, and the quantification method is as follows: ; Since the first cycle has no preceding cycle and the fourth cycle has no following cycle, no edge cycle is introduced. The average difference is directly recorded as 0.5 μV. Actual calculations are performed for b=2 and b=3: ; ; Offset constant The value is set to 0.1 μV, based on the lower limit standard for signal detection noise. In the sampled floor noise measurements, the minimum average interference voltage is between 0.08 and 0.12 μV; therefore, it is set to the median value. Substitute the above data into each item and perform the calculations: ; ; ; ; Sum the above four terms and divide by : ; The results show that the frequency difference of the periodic variation of the amplitude of the unit periodic electrical signal between peaks and valleys in the third segment is 2.068. The higher the value, the stronger the amplitude fluctuation within the period. As a characteristic value of the active repair frequency, it is used to match the segment number of the inflection point counting result and draw it in alignment with the time axis for use in the output of repair progress alignment layer.

[0033] The specific steps for obtaining the predictive results of changes in treatment efficacy are as follows: S501: Based on the complete timeline covered by the repair progress alignment layer, call the time position span corresponding to the pre-intervention, treatment and expected post-intervention stages within the layer, divide the time range of the layer according to the stage label, index and organize the time periods of the three stages, and generate a set of treatment stage time intervals. Extract the corresponding time periods of each stage—before intervention, during treatment, and after expected intervention—into the layer. Divide the timeline according to the intervention start time, treatment duration, and the doctor-set end time of the observation period, labeling them as Stage A (before intervention), Stage B (during treatment), and Stage C (after expected intervention). Set several pre-start times, such as the 20th minute, the treatment period (20 to 80 minutes), and the expected observation period (80 to 120 minutes). Then, organize the corresponding index ranges into Segment A (20 minutes), Segment B (80 minutes), and Segment C (80 to 120 minutes), and distinguish them in the layer using color or stage labels. The organized time segments are uniformly constructed into a treatment stage time interval set. This interval set serves as the basic framework for subsequent layer trend extraction and compliance analysis, ensuring that the data points on the layer accurately match the stages of the treatment process in time, thus generating the treatment stage time interval set.

[0034] S502: Based on the layer trend within the multi-stage segment of the treatment phase time interval, retrieve the change trajectory of electrical signal frequency and morphological response density in the time dimension, record the trend convergence area with consistent trend direction and continuous amplitude within the multi-stage, mark the start and end time and quantity of multiple types of areas, and generate the layer trend distribution structure. The repair progress of each stage segment is retrieved by performing trend retrieval on the layer curves. The time change trajectories of the electrical signal release frequency curve and the morphological response density curve in each time period are extracted. The segments where the two layer curves have the same trend direction and change continuously in the same stage are identified. The judgment criteria are set as the trend direction remaining the same in three or more consecutive time points. If the frequency and density curves continue to rise in the 30th to 50th minute of stage B, or continue to fall in stage C, then the time period is determined as the trend convergence zone. The start and end time index, stage, number of data points and curve continuation direction of each identified convergence zone are recorded. Multiple regions are numbered and classified according to stage to generate the layer trend distribution structure.

[0035] S503: Based on the number and distribution time of the three types of regions in the differentiated stage in the layer trend distribution structure, compare the proportional changes between the trend convergence area and the interruption and withdrawal area, identify the direction of the change in the degree of cooperation between the intervention path and the repair response over time, summarize the cooperation characteristics under multiple stages according to the comparison relationship between the trend structure and the stage time, and obtain the predictive results of the change in efficacy. Based on the number, start and end positions, and total duration of trend convergence zones identified in the three stages of the layer trend distribution structure, a comparative analysis is performed with the discontinuous zones (i.e., interruption zones) and reversal zones (i.e., trend reversal zones) that appear during the same period. By statistically analyzing the proportional relationship of various segments on the time axis, the direction of change in the temporal coordination between metabolic driving pathways and neural repair signals during the treatment process is determined. If the proportion of convergence zones exceeds 70% in treatment stage B, while the proportion of reversal zones increases in stage C, it indicates that the signal coordination decreases in the later stage of treatment. Based on the changing trend, coordination characteristics under multiple stages are summarized, such as "initial synchronous enhancement, mid-term high coupling, and late-term weakened synergy". Combined with the length proportion of each stage interval, a coordination characteristic structure table is output, and the efficacy change prediction result is generated based on the evolution trend of coordination.

[0036] Please see Figure 2 A predictive system for the efficacy of stem cell therapy for spinal cord injury, comprising: The metabolic stage identification module acquires the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence. It calculates the change direction of the three sequences at adjacent time points in chronological order, calls the change direction value for joint judgment, and filters out continuous stages in which the three indicators change in the same time period with the same change direction to obtain the activation rhythm mapping map. The activation initiation response module calls the activation initiation time period marked in the activation rhythm mapping map, obtains the concentration sequence of lactate, glutamate and myelin fragments in cerebrospinal fluid, and calls the synchronous axon extension length and branch value in conjunction to compare the changing trends and obtain a list of synchronous neural repair segments. The metabolic pathway deduction module traces the mutation points and neural extension inflection points in the cerebrospinal fluid metabolic changes based on the time points of the synchronous segments in the list of synchronous segments for neural repair, calculates the time sequence, and filters the leading segments of metabolic mutations to obtain a set of metabolic driving pathway sequences. The electrical signal pairing module calls the metabolic driving path sequence set to obtain the corresponding electrical signal response amplitude and release morphology density values, and obtains the repair progress alignment layer. The efficacy trend prediction module aligns the timeline of the layer with the repair progress, calls up multiple time intervals, and counts and distributes the number and distribution of trend convergence segments, interruption segments, and regression segments to obtain the efficacy change prediction results.

[0037] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for predicting the efficacy of stem cell therapy for spinal cord injury, characterized in that, Includes the following steps: S1: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy; extract continuous phases with consistent change direction; record the time coverage of continuous phases in the complete sequence; and obtain the activation rhythm mapping map. S2: Based on the stem cell activation initiation time period in the activation rhythm mapping atlas, obtain the concentration change sequence of lactate, glutamate, and myelin fragments in the patient's cerebrospinal fluid during the time period, summarize the distribution location of response points in the treatment cycle, and obtain a list of neural repair synchronization segments. The specific steps for obtaining this list are as follows: S201: Call the stem cell activation start time period indicated by the activation rhythm mapping map, obtain the lactate concentration sequence, glutamate concentration sequence and myelin fragment concentration sequence in the patient's cerebrospinal fluid by using the time interval as the index, and sort the three concentration data in order of time to generate the cerebrospinal fluid biochemical sequence group. S202: Based on the time range covered by the cerebrospinal fluid biochemical sequence group, the neural axon extension length sequence and axon branch number sequence within the time period are linked and called, and the two values ​​are arranged into a bivariate synchronous sequence in chronological order. The neural structural feature value is calculated, the structural remodeling change trajectory is identified, and a neural structural feature sequence is generated. The neural structural feature values ​​are calculated using the following formula: ； in, Representing the Neural structural feature values ​​within a time period Representing the The number of sampling points within the time period Representing the Time period Number of axonal branches at each time point Representing the Time period The length of nerve axon extension at each time point Representing the Time period The glucose concentration values ​​in the cerebrospinal fluid at each time point Representing the The average glucose concentration in cerebrospinal fluid at sampling points within the time period. Representing the The average value of the sampling points within the time period; S203: Call the time points in the neural structure feature sequence, perform point-to-point comparison operation based on the time position, identify the time points where the direction of concentration change is consistent with the direction of neural axis mutation, record multiple types of positions as a set of response points, and generate a group of points with consistent response trends. S204: Based on the time index of the points in the group of points with consistent response trends within the intervention period, summarize the continuous distribution segments of response points in chronological order, and merge the cases where the time interval between adjacent response points is less than a fixed interval, mark them as continuous response segments, and generate a list of neural repair synchronous segments. S3: Based on the time segments identified in the list of neural repair synchronization segments, trace the location of the first mutation behavior in the changes in cerebrospinal fluid metabolism, and obtain the set of metabolic driving pathway sequences according to the time sequence between the mutation point and the neural extension inflection point. S4: The active mutation segments within the time period of the metabolic drive pathway sequence are retrieved, and the corresponding electrical signal response amplitude sequences of the matched segments are used to evaluate the active release frequency and morphological response density of implanted cells, generating a repair progress alignment layer. The specific steps are as follows: S401: Call the time period corresponding to the path structure with marked frequency in the metabolic drive path sequence set, extract the number information of the active mutation segment, match the data of the same time period in the electrical signal response amplitude sequence according to the segment time index, and organize them according to the segment number to generate an electrical signal response amplitude index table. S402: Based on the amplitude values ​​in each segment sequence in the electrical signal response amplitude index table, extract the periodic change frequency between the peak and trough of the electrical signal sequence in the segment as the active release frequency of the implanted cells, calculate the periodic change difference frequency value, count the total number of inflection points of the electrical signal per unit time, pair the two types of values ​​according to the segment number and align them on the time axis to generate a repair progress alignment layer. The frequency value of the periodic variation difference is expressed by the formula: ； in, Representing the The frequency difference of periodic changes in segments Representing the The number of cycles extracted from segments. Representing the The first section of the segment The peak electrical signal amplitude corresponding to each cycle Representing the The first section of the segment The amplitude of the low-valley electrical signal corresponding to each cycle Representing the The first section of the segment The absolute difference between the mean values ​​of the preceding and following periods. It is a constant.

2. The method for predicting the efficacy of stem cell therapy for spinal cord injury according to claim 1, characterized in that, The activation rhythm mapping map includes change stage label numbers, matching start time index, and stage occurrence frequency distribution. The list of neural repair synchronization segments includes response time node set, trend synchronization matching rate, and periodic distribution location index. The metabolic driving path sequence set includes mutation start coordinates, path evolution direction encoding, and metabolic leader identifier sequence number. The repair progress alignment layer includes active response frequency band index, electrical signal amplitude mapping intensity, and release morphology density index.

3. The method for predicting the efficacy of stem cell therapy for spinal cord injury according to claim 1, characterized in that, The specific steps for obtaining the activation rhythm mapping map are as follows: S101: Obtain the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence of patients with spinal cord injury treated with stem cell therapy. Determine the same direction of the three sets of sequences, extract continuous time periods with the same direction of change, record the time index interval, and generate time intervals of the same direction stage. S102: Based on the three values ​​of each time period in the direction-consistent phase time interval, extract and encode the phase label combination composed of triples, summarize the number of times the phase label combination appears in the intervention cycle, and generate the phase label combination frequency value. S103: Call the frequency value of the stage label combination, and perform horizontal matching of the first complete occurrence time period of the stage label combination based on the patient's treatment start time. Compare the time start time corresponding to the label combination with the treatment start time in sequence. Record the location time point for the stage combination that meets the matching conditions. Use the label combination as the horizontal axis and the location time point as the vertical axis to generate an activation rhythm mapping map.

4. The method for predicting the efficacy of stem cell therapy for spinal cord injury according to claim 1, characterized in that, The specific steps for obtaining the metabolic drive pathway sequence set are as follows: S301: Based on the start and end indexes of time segments in the list of neural repair synchronization segments, trace back the sequences of three metabolic parameters—lactic acid, glutamate, and myelin fragments—in the cerebrospinal fluid within the time period, identify the inflection point where the parameters first show a change in value, record the time index of the mutation point and identify the parameter source to which it belongs, and generate a list of metabolic mutation time points. S302: Based on the position of the time points in the list of metabolic mutation time points, take the time point of the first change in the curve of the corresponding nerve axon extension length or axon branch number as the inflection point of the nerve change, construct a time path structure with metabolic change first and nerve change second between the two in chronological order, and connect the time nodes in the structure according to the combination index method to generate a sequence of leading paths. S303: The path sequence formed by the leading path is within the continuous evolutionary segment in terms of time span, and the metabolic change time index is earlier than the neural change time index. The path structures that meet the conditions are numbered and recorded, and the number of recurrences in the entire intervention period is counted. The path structure is paired with the frequency and labeled to generate a set of metabolic-driven path sequences.

5. The method for predicting the efficacy of stem cell therapy for spinal cord injury according to claim 1, characterized in that, The method further includes step S5: S5: Based on the timeline covered by the repair progress alignment layer, call the position span of the pre-intervention, treatment, and expected post-intervention stages within the layer, retrieve the number and distribution of trend convergence areas, trend interruption areas, and trend reversal areas in the alignment layer, identify the direction of change in the degree of cooperation between the repair response and the intervention path, and obtain the prediction results of the change in efficacy. The predicted results of changes in therapeutic efficacy include the trend of changes in compatibility, the number of response segment types, and the time axis alignment distribution map.

6. The method for predicting the efficacy of stem cell therapy for spinal cord injury according to claim 5, characterized in that, The specific steps for obtaining the predicted results of changes in therapeutic efficacy are as follows: S501: Based on the complete timeline covered by the repair progress alignment layer, call the time position span corresponding to the pre-intervention, treatment and expected post-intervention stages within the layer, divide the time range of the layer according to the stage label, index and organize the time periods of the three stages, and generate a set of treatment stage time intervals. S502: Based on the layer trend within the multi-stage segment of the treatment phase time interval, retrieve the change trajectory of electrical signal frequency and morphological response density in the time dimension, record the trend convergence area with consistent trend direction and continuous amplitude within the multi-stage, mark the start and end time and quantity of multiple types of areas, and generate the layer trend distribution structure. S503: Based on the number and distribution time of the three types of regions in the differentiated stage in the layer trend distribution structure, compare the proportional changes between the trend convergence area and the interruption and withdrawal area, identify the direction of the change in the degree of cooperation between the intervention path and the repair response over time, summarize the cooperation characteristics under multiple stages according to the comparison relationship between the trend structure and the stage time, and obtain the prediction results of the change in efficacy.

7. A system for predicting the efficacy of stem cell therapy for spinal cord injury, characterized in that, The system is used to implement the method for predicting the efficacy of stem cell therapy for spinal cord injury as described in any one of claims 1-6, the system comprising: The metabolic stage identification module acquires the glucose dissipation rate sequence, lactate release rate sequence, and mitochondrial membrane potential change sequence. It calculates the change direction of the three sequences at adjacent time points in chronological order, calls the change direction value for joint judgment, and filters out continuous stages in which the three indicators change in the same time period with the same change direction to obtain the activation rhythm mapping map. The activation initiation response module calls the activation initiation time period identified in the activation rhythm mapping map, obtains the concentration sequence of lactate, glutamate and myelin fragments in cerebrospinal fluid, and calls the synchronous axon extension length and branch value to compare the changing trends and obtain a list of synchronous neural repair segments. The metabolic pathway deduction module traces the mutation points and neural extension inflection points in the cerebrospinal fluid metabolic changes based on the time points of the synchronous segments in the list of neural repair synchronous segments, calculates the time sequence, and filters the leading segments of metabolic mutations to obtain a set of metabolic driving pathway sequences. The electrical signal pairing module calls the metabolic driving pathway sequence set to obtain the corresponding electrical signal response amplitude and release morphology density values, and obtains the repair progress alignment layer. The efficacy trend prediction module aligns the timeline of the positioning layer with the repair progress, calls up multi-stage time intervals, and statistically analyzes the number and distribution of trend convergence segments, interruption segments, and regression segments to obtain the efficacy change prediction results.

Images