Multimodal biofeedback driven dynamic fractionation intervention device for chronic pain

Abstract

The present application relates to the technical field of intelligent sensing and control equipment, in particular to a chronic pain dynamic grading intervention device driven by multi-modal biological feedback, comprising a collection module, a reference module, a correction module, a grading module and a control module, collecting heart beat, skin electricity, muscle electricity, limb micro-motion and time data, forming physiological signal time sequence data, establishing grading threshold, transmitting upper limit of hysteresis and safety release threshold, generating pain load grade through hysteresis correction, and outputting stimulation control instruction combined with last period record. In the present application, through multi-source physiological feedback, resting reference, time sequence correction, continuous connection of dynamic grading and tolerance margin constraint, the pain load judgment is synchronized into the degree of physiological abnormality, the signal reaches hysteresis and the individual bearing state, the grading lag and stimulation deviation caused by fixed threshold triggering are reduced, and the real-time, adaptability, closed-loop and safety of intervention control are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent sensing and control equipment technology, and in particular to a multimodal biofeedback-driven dynamic grading intervention device for chronic pain. Background Technology

[0002] The field of intelligent sensing and control equipment technology involves computer devices that collect human body state data and form interactive control links through sensors, processors, memory, and control execution components. Among these, a traditional multimodal biofeedback-driven dynamic grading intervention device for chronic pain refers to a device composed of wearable sensors, data acquisition circuits, a processor, a display component, and a stimulation output component. The sensors collect physiological feedback signals and transmit them to the processor via the acquisition circuit. The processor reads the grading thresholds and intervention parameters from the memory according to a preset program and sends corresponding control commands to the display component or stimulation output component.

[0003] Traditional chronic pain grading intervention devices typically rely on wearable sensors to collect single or small amounts of physiological feedback signals. These signals are then transmitted to a processor via a collection circuit, and output display or stimulation commands according to stored thresholds and fixed intervention parameters. The operation process tends to focus on static grading and preset triggers, making it difficult to synchronously reflect the combined effects of heart rate fluctuations, skin conductance changes, continuous activity of electromyography, limb micro-movements, and communication delays on pain load assessment. The stimulation output is not sufficiently connected with the tolerance state of the previous cycle, which can easily lead to problems such as grading lag, intervention intensity deviating from actual tolerance, and unclear safety control margins. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a multimodal biofeedback-driven dynamic grading intervention device for chronic pain.

[0005] On the one hand, a multimodal biofeedback-driven dynamic grading intervention device for chronic pain is provided, including: The acquisition module is used to collect heart rate fluctuation values, skin conductance response rise slope values, muscle electrical activity duration amplitude values, limb micro-movement amplitude values, sampling time recording, and communication arrival time to generate physiological signal time series data. The baseline module, connected to the acquisition module, is used to generate physiological grading thresholds, transmission hysteresis upper limits, and safe release thresholds based on the initial resting sampling data, and to receive the execution record of the previous intervention cycle. A correction module, connected to the acquisition module and the reference module, is used to calculate the transmission hysteresis value based on the physiological signal time series data, compare the transmission hysteresis value with the transmission hysteresis upper limit, and generate a time series correction signal set. A grading module, connected to the correction module, is used to compare the physiological values ​​in the time-series correction signal set with the physiological grading threshold, and generate a pain load level based on the number of over-threshold items and the over-threshold amplitude after channel baseline normalization. The control module, connected to the grading module and the benchmark module, is used to generate a stimulus tolerance margin from the previous intervention cycle execution record, compare the stimulus tolerance margin with the safe release threshold, and output stimulus control commands according to the pain load level.

[0006] As a further embodiment of the present invention, the acquisition module includes a time reference submodule, a multi-channel acquisition submodule, and a timing encapsulation submodule; The time reference submodule receives the sampling time record and writes the same sampling window number to the heartbeat interval sampling channel, skin electric field sampling channel, muscle electric field sampling channel and limb micro-motion sampling channel to generate channel time markers; The multi-channel acquisition submodule acquires the heart rate fluctuation value, the skin conductance response rise slope value, the muscle electrical activity duration amplitude value, and the limb micro-movement amplitude value according to the channel time stamp, and writes the communication arrival time into the corresponding sampling window to obtain the channel acquisition record; The timing encapsulation submodule associates the channel acquisition records with the sampling window number, writes a missing marker at the missing channel position, and outputs the physiological signal timing data.

[0007] As a further aspect of the present invention, the time stamp writing process of the time reference submodule includes: Obtain the sampling interval between two adjacent sampling time records, and compare the sampling interval with a preset sampling interval range. When the sampling interval is within the preset sampling interval range, write the current sampling time record into the corresponding channel acquisition record. When the sampling interval exceeds the preset sampling interval range, mark the current channel acquisition record as a record to be corrected. The timing encapsulation submodule establishes a time-stamped difference record according to the sampling time record and communication arrival time of the record to be corrected, and associates the time-stamped difference record with the record to be corrected and writes it into the physiological signal timing data.

[0008] As a further embodiment of the present invention, the reference module includes a resting screening submodule, a threshold generation submodule, and a recording and receiving submodule; The resting screening submodule removes sampling segments from the initial resting sampling data whose communication arrival time exceeds the preset resting lateness boundary relative to the sampling time, and removes sampling segments that do not simultaneously have heart rate interval fluctuation values, skin conductance response rise slope values, muscle electrical activity duration amplitude values, and limb micro-movement amplitude values. It extracts the effective resting values ​​according to the physiological sampling channels and generates the resting channel baseline. The threshold generation submodule determines the heart rate interval fluctuation threshold, skin conductance slope threshold, muscle conductance amplitude threshold, and limb micro-movement amplitude threshold based on the resting channel baseline. It also extracts the sampling time record and communication arrival time of each sampling segment from the initial resting sampling data retained by the resting screening submodule and generates a set of resting hysteresis values ​​according to the lag relationship between the communication arrival time and the sampling time record. The set of resting hysteresis values ​​is sorted according to the hysteresis magnitude, and the hysteresis value corresponding to the preset hysteresis quantile position is selected as the upper limit of transmission hysteresis, so that the upper limit of transmission hysteresis is objectively determined by the actual communication hysteresis distribution in the resting state. The physiological grading threshold is generated based on the cardiac interval fluctuation threshold, the skin electrophysiology slope threshold, the muscle electrical amplitude threshold, the limb micro-movement amplitude threshold, and the transmission hysteresis upper limit. The record receiving submodule receives the execution record of the previous intervention cycle and extracts the stimulus duration, stimulus intensity change and termination state from the execution record of the previous intervention cycle to generate an intervention baseline record.

[0009] As a further aspect of the present invention, the threshold generation process of the threshold generation submodule includes: The resting center value and resting fluctuation range of each physiological sampling channel in the resting channel baseline are obtained. The resting center value and the resting fluctuation range are jointly determined as the threshold generation basis for the corresponding physiological sampling channel. The threshold index relationship is established according to the channel order of heartbeat interval, skin conductance response, muscle electrical activity and limb micro-movement. The threshold generation submodule writes the threshold index relationship and the physiological grading threshold into the threshold configuration table, and provides the threshold configuration table to the grading module for judging the number of over-threshold items and the over-threshold amplitude after normalization according to the resting fluctuation range.

[0010] As a further embodiment of the present invention, the correction module includes a hysteresis calculation submodule, an upper limit comparison submodule, and a signal correction submodule; The hysteresis calculation submodule acquires the sampling time record and communication arrival time in the physiological signal time series data, and determines the transmission hysteresis value according to the lag relationship between the communication arrival time and the sampling time record within the same sampling window. The upper limit comparison submodule compares the transmission hysteresis value with the transmission hysteresis upper limit, and generates a normal timing state, a correctable hysteresis state, or an over-limit hysteresis state based on the comparison result. The signal correction submodule writes correction flags to the physiological values ​​in the corresponding sampling window according to the normal timing state, the correctable hysteresis state, or the excessive hysteresis state, and generates the timing correction signal set.

[0011] As a further aspect of the present invention, the correction mark writing process of the signal correction submodule includes: When the comparison result is the normal time series state, the correlation between the physiological values ​​and the sampling time records within the corresponding sampling window is maintained, and a reliable marker is written to generate a reliable time series signal. When the comparison result is the correctable hysteresis state, the corresponding physiological value is mapped to the adjacent sampling window and written into the correction mark according to the delay direction between the communication arrival time and the sampling time record, thereby generating a correction timing signal. When the comparison result is the over-limit hysteresis state, the corresponding physiological value is retained and written into the isolation mark, an isolation timing signal is generated, and the reliable timing signal, the correction timing signal and the isolation timing signal are collected into the timing correction signal set.

[0012] As a further embodiment of the present invention, the grading module includes a threshold reading submodule, an over-threshold determination submodule, and a grade generation submodule; The threshold reading submodule acquires the time-series correction signal set and the physiological grading threshold, and establishes a threshold matching record according to the channel correspondence of heartbeat interval, skin conductance response, muscle electrical activity and limb micro-movement; The threshold determination submodule compares each physiological value in the time-series correction signal set with the corresponding threshold in the threshold matching record to determine the number of threshold items and the threshold amplitude of each threshold item. The physiological values ​​with confidence and correction labels in the time-series correction signal set are compared with the physiological grading threshold. Physiological values ​​with isolation or missing labels are not used as the basis for generating the current pain load level. The level generation submodule generates the pain load level based on the number of overthreshold items and the overthreshold amplitude of each overthreshold item.

[0013] As a further aspect of the present invention, the level generation process of the level generation submodule includes: The number of out-of-threshold items and the out-of-threshold amplitude corresponding to each out-of-threshold item are obtained. The maximum out-of-threshold amplitude is determined from the out-of-threshold amplitudes corresponding to each out-of-threshold item. Candidate pain load levels are determined according to a preset range of item numbers, and the candidate pain load levels are verified according to a preset range of amplitudes. When the maximum overthreshold amplitude reaches the amplitude range corresponding to the candidate pain load level, the candidate pain load level is maintained; when the maximum overthreshold amplitude does not reach the amplitude range corresponding to the candidate pain load level, the candidate pain load level is reduced, and the pain load level is generated.

[0014] As a further embodiment of the present invention, the control module includes a recording and parsing submodule, a margin generation submodule, and an instruction output submodule. The recording and parsing submodule obtains the execution record of the previous intervention cycle and extracts the stimulus intensity record, stimulus duration record, and stop triggering state from the execution record of the previous intervention cycle to generate an execution boundary record. The margin generation submodule determines the stimulus intensity boundary and stimulus duration boundary that can be released in the current intervention cycle based on the execution boundary record, and determines the stimulus intensity boundary and the stimulus duration boundary together as the stimulus tolerance margin. The stimulation tolerance margin includes an intensity margin for limiting the upper limit of stimulation release intensity in the current intervention cycle and a duration margin for limiting the upper limit of stimulation release duration in the current intervention cycle. The intensity margin is used to compare with the safe intensity threshold in the safe release threshold, and the duration margin is used to compare with the safe duration threshold in the safe release threshold. The instruction output submodule compares the intensity margin with the safe intensity threshold and the duration margin with the safe duration threshold. When the intensity margin meets the safe intensity threshold and the duration margin meets the safe duration threshold, it outputs a stimulus control instruction that allows release in conjunction with the pain load level. When either the intensity margin or the duration margin fails to meet the corresponding safety threshold, a stimulus control command to limit release and stop release is output.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, a continuous feedback link is constructed around the fluctuation of heart rate interval, the rising slope of skin conductance response, the sustained amplitude of muscle electrical activity, the amplitude of limb micro-movement, the sampling time recording, and the communication arrival time. An individualized threshold is established based on the resting state, and transmission hysteresis boundaries and safe release conditions are established. This allows physiological data to undergo temporal consistency correction before entering the grading judgment. Pain load is characterized by the number and amplitude of multiple overthresholds. Stimulus output is combined with the execution record of the previous intervention cycle to form a tolerance margin judgment. A closed-loop correlation is formed between the grading result and the stimulus intensity, which is constrained by hysteresis and safe release. This reduces misjudgment and delay caused by static threshold triggering, and makes the intervention rhythm more in line with the real-time pain load and individual tolerance. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the multimodal biofeedback-driven dynamic grading intervention for chronic pain according to the present invention. Figure 2 This is a schematic diagram of the structural relationship of the dynamic hierarchical intervention device of the present invention; Figure 3 This is a schematic diagram comparing the states before and after timing correction in this invention; Figure 4 This is a schematic diagram illustrating the pain load level generation effect of the present invention; Figure 5 This is a schematic diagram illustrating the effect of the stimulus control safety constraint of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[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] Example

[0021] Please see Figures 1 to 5This embodiment provides a multimodal biofeedback-driven dynamic grading intervention device for chronic pain. In the actual operation of the device, the patient undergoes continuous physiological monitoring and is graded according to changes in pain load. Changes in heart rate interval, skin conductance response, muscle electrical activity and limb micro-movements are simultaneously entered into the same pain load judgment link. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain includes a data acquisition module, a reference module, a correction module, a grading module and a control module.

[0022] During chronic pain intervention, issues arise such as inconsistent sampling windows for different physiological signals, delayed communication arrival, and constraints imposed by the previous intervention cycle on the release of current stimuli. This embodiment addresses these issues by using an acquisition module to generate time-series physiological signal data with sampling time records and communication arrival times. A baseline module generates grading thresholds, hysteresis boundaries, and safe release boundaries. A calibration module then marks the reliability of the time-series data. A grading module generates pain load levels based on a unified threshold index relationship. Finally, a control module, in conjunction with the previous intervention cycle, executes recorded stimulus control commands, creating a continuous data flow loop encompassing acquisition, calibration, grading, and control.

[0023] Data Acquisition Module: The data acquisition module receives data on heart rate fluctuations, skin conductance slope, muscle electrical activity amplitude, limb micro-movement amplitude, sampling time, and communication arrival time from subjects undergoing chronic pain intervention. It organizes these physiological values ​​into a time-series physiological signal data within the same sampling window. Heart rate fluctuations are records of fluctuations between adjacent heartbeats, reflecting rhythmic changes related to the autonomic nervous system. Skin conductance slope records the rate of change during the rising phase of skin conductance, reflecting sympathetic arousal. Muscle electrical activity amplitude records the amplitude during sustained contraction or tension, reflecting pain-related muscle tension. Limb micro-movement amplitude records the amplitude generated by subtle limb movements, reflecting postural adjustments or involuntary movements triggered by pain. Among them, sampling time record refers to the time stamp formed when the physiological signal is sampled, and communication arrival time refers to the time stamp formed when the corresponding sampling record arrives at the subsequent processing link. Physiological signal time series data refers to a data set organized according to the sampling window number, which carries physiological values ​​of each channel, sampling time record, communication arrival time, channel time stamp, gap marker and time stamp difference record, and its output is used by the reference module and the calibration module; The acquisition module includes a time reference submodule, a multi-channel acquisition submodule, and a timing encapsulation submodule. The time reference submodule receives sampling time records and writes the same sampling window number to the cardiac interval sampling channel, skin conductance sampling channel, muscle conductance sampling channel, and limb micro-motion sampling channel, generating a channel time stamp. The channel time stamp is a record field indicating that each physiological sampling channel belongs to the same sampling window; it carries the sampling window number and the corresponding sampling time record, and serves as the basis for subsequent channel association. The multi-channel acquisition submodule acquires the values ​​of heart rate interval fluctuation, skin conductance response slope, muscle electrical activity duration amplitude, and limb micro-movement amplitude based on the channel time stamp, and writes the communication arrival time into the corresponding sampling window to obtain the channel acquisition record. The channel acquisition record refers to the record item formed within each physiological sampling channel, which includes the channel type, physiological value, sampling time record, communication arrival time, and sampling window number, and is associated by the timing encapsulation submodule according to the sampling window number. The timing encapsulation submodule associates the channel acquisition records according to the sampling window number, writes a missing marker for the missing channel position, and outputs the physiological signal timing data. A missing marker is a status field written when a physiological value for a certain channel is not obtained in the corresponding sampling window. It is used to indicate that subsequent correction and grading processes should not directly use this position as a normal complete value. For sampling windows with missing markers, the timing encapsulation submodule retains the already acquired channel acquisition records and outputs the missing status along with the sampling window number, enabling subsequent processing to identify data integrity boundaries. The time reference submodule's time stamp writing process involves acquiring the sampling interval between two adjacent sampling time records and comparing this interval with a preset sampling interval range. The preset sampling interval range refers to the allowable sampling rhythm boundary determined by the device during initial configuration. It originates from the sampling rule configuration of each physiological sampling channel by the acquisition module and serves as a configurable field for determining the continuity of sampling time. When the sampling interval is within the preset range, the time reference submodule writes the current sampling time record into the corresponding channel's acquisition record; when the sampling interval exceeds the preset range, the time reference submodule marks the current channel's acquisition record as a record to be corrected. Among them, the record to be calibrated refers to the channel acquisition record where the sampling time record is inconsistent with the preset sampling rhythm and the hysteresis state needs to be further determined in the calibration module. The timing encapsulation submodule establishes a time-stamped difference record according to the sampling time record and the communication arrival time of the record to be calibrated, and associates the time-stamped difference record with the record to be calibrated and writes it into the physiological signal timing data. The time-stamped difference record is a status field used to describe the sequential relationship between the sampling time record and the communication arrival time. It does not express the degree of lag in the form of a formula, but forms a record based on the relationship of sampling time first and communication arrival second and the corresponding sampling window, which is used by the calibration module to generate a set of timing calibration signals; In this embodiment, the acquisition module unifies each physiological sampling channel under the sampling window number and simultaneously retains the sampling time record and communication arrival time. This enables the subsequent correction module to identify normal records, correctable records, and records that need to be isolated, thereby avoiding the direct mixing of different physiological signals in pain load grading due to inconsistent arrival times.

[0024] Baseline module: The baseline module refers to the component that receives initial resting sampling data and the execution record of the previous intervention cycle, and generates physiological grading thresholds, transmission hysteresis upper limits, safe release thresholds, and intervention baseline records. Initial resting sampling data refers to physiological signal records collected before the intervention begins or when the current operation phase reaches a stable resting state. It carries the physiological values ​​of each physiological sampling channel in the resting state, sampling time records, and communication arrival times. Physiological grading thresholds refer to the set of thresholds used in the grading module to determine whether the physiological values ​​of each channel exceed the resting baseline boundary. These include the heart rate interval fluctuation threshold, skin conductance slope threshold, muscle electrical amplitude threshold, and limb micro-movement amplitude threshold. The transmission hysteresis upper limit is a boundary field used to determine whether the delay in communication arrival time relative to the sampling time record still allows for correction. It originates from the hysteresis state of communication arrival time in the initial resting sampling data. The safe release threshold is a safety boundary field used to constrain the release of stimulus in the current intervention cycle. It, together with the execution record of the previous intervention cycle, is used by the control module to compare the stimulus tolerance margin. The execution record of the previous intervention cycle refers to the stimulus control record completed before the current intervention cycle, which includes the stimulus duration, stimulus intensity change, stimulus intensity record, stimulus duration record, and stop triggering status. The baseline module includes a resting interval screening submodule, a threshold generation submodule, and a recording and receiving submodule. The resting interval screening submodule removes sampling segments from the initial resting interval sampling data whose communication arrival time relative to the sampling time exceeds a preset resting interval lateness boundary. It also removes sampling segments that do not simultaneously possess values ​​for heart rate interval fluctuation, skin conductance response slope, muscle electrical activity duration amplitude, and limb micro-movement amplitude. The preset resting interval lateness boundary is a configurable boundary used to determine whether the communication arrival of a resting interval sampling segment meets the baseline generation requirements; it originates from the sampling and communication status configuration during device initialization. The retained sampling segments are used to extract valid resting interval values ​​according to the physiological sampling channels, generating resting interval channel baselines. Among them, the valid resting value refers to the resting physiological value that simultaneously meets the requirements of channel integrity, identifiable sampling time, communication arrival not exceeding the preset resting lateness boundary, and not being marked with a gap. The resting channel baseline refers to the baseline record formed separately for each physiological sampling channel, which carries the resting center value, resting fluctuation range, and channel correspondence. The resting center value is used to represent the reference position of the corresponding channel in the resting state, and the resting fluctuation range is used to represent the natural variation boundary that the channel is allowed to appear in the resting state. Both serve as the basis for threshold generation. The threshold generation submodule determines the thresholds for cardiac interval fluctuation, skin conductance slope, muscle electrical amplitude, and limb micro-movement amplitude based on the resting channel baseline. It also determines the upper limit of transmission hysteresis based on the communication arrival time in the initial resting sampling data, generating physiological grading thresholds. The threshold generation process does not use publicly available formulas. Specifically, the submodule reads the resting center value and resting fluctuation range of each channel, using the resting center value as the reference position for the corresponding channel and the resting fluctuation range as the allowable resting variation boundary. It then determines the threshold boundaries for grading based on the physiological significance of the channel. For the cardiac interval fluctuation channel, the threshold boundaries reflect the degree to which the rhythm fluctuation deviates from the resting state; for the skin conductance response channel, the threshold boundaries reflect the degree of enhancement of the rising change relative to the resting state; for the muscle electrical activity channel, the threshold boundaries reflect the degree of tension of the sustained amplitude relative to the resting state; and for the limb micro-movement channel, the threshold boundaries reflect the activity level of the micro-movement amplitude relative to the resting state. The threshold generation submodule establishes a threshold index relationship according to the channel order of heartbeat interval, skin conductance response, muscle electrical activity, and limb micro-movement. The threshold index relationship refers to a lookup table field that permanently associates each physiological sampling channel with its corresponding threshold boundary, preventing the grading module from cross-referencing channels with different physical meanings during comparison. The threshold generation submodule writes the threshold index relationship and physiological grading thresholds into a threshold configuration table, and provides this table to the grading module for determining the number and magnitude of threshold exceedances. The threshold configuration table contains configuration records for channel names, threshold sources, threshold boundary states, channel order, and subsequent call positions. It does not contain data instances and is updated synchronously with the resting channel baseline after initializing the resting sampling data update. The recording receiving submodule receives the execution record of the previous intervention cycle and extracts the stimulus duration, stimulus intensity change, and stop trigger state from the previous intervention cycle execution record to generate an intervention baseline record. The intervention baseline record refers to the data record used by the control module to parse the stimulus release boundary of the previous cycle; it carries the duration state, intensity adjustment state, and whether the stimulus released in the previous cycle ended due to a stop condition. When the device is in the first intervention cycle and there is no execution record from the previous intervention cycle, the recording receiving submodule generates the intervention baseline record using initial resting sampling data and an initial execution state without a stop trigger. After completing the first intervention cycle, it switches to a stable operating rule using the execution record of the previous intervention cycle as input. In this embodiment, the resting screening submodule excludes late and incomplete resting sampling segments, so that the threshold generation submodule generates a threshold configuration table based only on valid resting values ​​that meet the temporal integrity requirements. This provides the grading module with a stable benchmark corresponding to each physiological channel when making subsequent over-threshold judgments.

[0025] Calibration module: The correction module is a component that calculates transmission hysteresis values ​​based on physiological signal time-series data, compares these values ​​with a transmission hysteresis upper limit, and generates a time-series correction signal set. The transmission hysteresis value refers to the time difference between the communication arrival time and the recorded sampling time within the same sampling window. This time difference is expressed in milliseconds, seconds, or a time unit consistent with the sampling period. When the communication arrival time is later than the recorded sampling time, the time difference characterizes the data transmission delay degree of the corresponding sampling window; when the communication arrival time is not later than the recorded sampling time, the transmission hysteresis value is written as a normal hysteresis value. The transmission hysteresis value is calculated by the hysteresis calculation submodule based on the recorded sampling time and communication arrival time within the same sampling window, and is compared using the same time unit or the same hysteresis level as the transmission hysteresis upper limit. The time-series correction signal set is a data set formed by aggregating reliable time-series signals, correction time-series signals, and isolation time-series signals. It carries physiological values, sampling window numbers, transmission hysteresis values, correction markers, and subsequent hierarchical call status. The correction module includes a hysteresis calculation submodule, an upper limit comparison submodule, and a signal correction submodule. The hysteresis calculation submodule acquires the sampling time record and communication arrival time from the physiological signal time-series data, and determines the transmission hysteresis value according to the lag relationship between the communication arrival time and the sampling time record within the same sampling window. For channel positions marked with a gap, the hysteresis calculation submodule retains the gap state and does not include it as a complete hysteresis record in the generation of reliable time-series signals; for channel acquisition records marked with a record to be corrected, the hysteresis calculation submodule treats them as objects to be included in the upper limit comparison. The upper limit comparison submodule compares the transmission hysteresis value with the transmission hysteresis upper limit and generates a normal timing state, a correctable hysteresis state, or an over-limit hysteresis state based on the comparison result. A normal timing state means that the delay of the communication arrival time relative to the sampling time record is such that the original sampling window association can be directly maintained. A correctable hysteresis state means that the delay relationship has affected the original sampling window association, but it can still be mapped to adjacent sampling windows based on the delay direction between the communication arrival time and the sampling time record. An over-limit hysteresis state means that the delay relationship exceeds the transmission hysteresis upper limit and cannot be directly used as a reliable input for the current classification judgment. The signal correction submodule writes correction flags to the physiological values ​​within the corresponding sampling window based on the normal timing state, correctable hysteresis state, or excessive hysteresis state, generating a timing correction signal set. The correction flag is a field indicating whether the physiological value is in a reliable, corrected, or isolated state within the current sampling window; it is read by the grading module and used to determine the data range that enters the threshold judgment. When the comparison result is in a normal time series state, the signal correction submodule maintains the correlation between the physiological values ​​and the sampling time records within the corresponding sampling window, writes a trustworthy flag, and generates a trustworthy time series signal. A trustworthy time series signal refers to a signal record in which the physiological values ​​and the sampling window number maintain their original correlation and meet the hysteresis boundary requirements; it directly enters the threshold matching processing of the grading module. When the comparison result indicates a correctable hysteresis state, the signal correction submodule maps the corresponding physiological value to an adjacent sampling window and writes a correction mark based on the lag direction between the communication arrival time and the sampling time record, generating a corrected timing signal. Mapping to an adjacent sampling window means assigning the physiological value to a sampling window more consistent with its arrival state according to the hysteresis direction, rather than changing the physiological value itself. When the corrected timing signal enters the grading module, the grading module reads its correction mark and uses it as an input that can participate in threshold comparison but whose correction source must be retained. When the comparison result indicates an excessive hysteresis state, the signal correction submodule retains the corresponding physiological value and writes it into an isolation marker, generating an isolated time-series signal. The isolated time-series signal refers to a signal record that retains the original physiological value and sampling window affiliation but does not directly enter the pain load level generation; it is used to maintain the traceability of the sampling link. The signal correction submodule aggregates the reliable time-series signal, the corrected time-series signal, and the isolated time-series signal into a time-series correction signal set and outputs it to the grading module. In this embodiment, the correction module distinguishes between reliable, correctable, and isolated states before grading, enabling the grading module to perform threshold matching only on physiological values ​​that meet the time boundary or have been corrected by adjacent windows. This ensures that communication arrival delays do not directly disrupt the pain load level generation link.

[0026] Hierarchical module: The grading module receives a set of timing correction signals and physiological grading thresholds, compares physiological values ​​with the corresponding thresholds, and generates a pain load level based on the number of items exceeding the threshold and the magnitude of the exceedance. The number of items exceeding the threshold refers to the number of items in each physiological sampling channel within the same grading judgment window that are determined to exceed the corresponding threshold boundary. The magnitude of the exceedance refers to the deviation of a physiological value from the corresponding threshold boundary; it is not expressed in formulaic form but is entered into the grading generation submodule as a textual state such as not exceeding, approaching the grading boundary, reaching the magnitude range corresponding to the candidate level, or not reaching the magnitude range corresponding to the candidate level. The grading module includes a threshold reading submodule, an over-threshold determination submodule, and a grade generation submodule. The threshold reading submodule acquires the time-series correction signal set and physiological grading thresholds, and establishes threshold matching records according to the channel correspondences of heartbeat interval, skin conductance response, muscle electrical activity, and limb micro-movements. The threshold matching record is a record that associates the channel type, sampling window number, correction marker, and corresponding threshold boundary in the threshold configuration table within the time-series correction signal set; its output is used by the over-threshold determination submodule. The threshold determination submodule compares each physiological value in the time-series corrected signal set with the corresponding threshold in the threshold matching record to determine the number of threshold-exceeding items and the threshold-exceeding magnitude of each item. For physiological values ​​with a confidence marker, the threshold determination submodule compares them according to their original sampling window; for physiological values ​​with a correction marker, the threshold determination submodule compares them according to the corrected sampling window and retains the correction source status in the threshold-exceeding result; for physiological values ​​with an isolation marker or a missing marker, the threshold determination submodule does not use them as a direct basis for judging the current pain load level and writes the corresponding channel status into the grading process record. The pain load generation submodule generates pain load levels based on the number of overthreshold items and the overthreshold amplitude of each item. The pain load level refers to a pain load state field formed by multiple physiological sampling channels, used by the control module to select stimulus control commands. The level generation submodule first reads the number of overthreshold items in the same judgment window and determines candidate pain load levels according to a preset item number range. The preset item number range refers to a level mapping rule established based on the number of channels participating in the overthreshold; it originates from the level configuration field in the threshold configuration table and is not expressed using data instances. The grade generation submodule then determines the maximum threshold amplitude from the threshold amplitudes corresponding to each threshold item and verifies the candidate pain load level according to the preset amplitude range. The maximum threshold amplitude refers to the threshold state that deviates most significantly from the corresponding threshold boundary within the current judgment window. The preset amplitude range refers to the amplitude state boundary corresponding to the candidate pain load level, which is derived from the grade verification rules in the threshold configuration table. When the maximum threshold amplitude reaches the amplitude range corresponding to the candidate pain load level, the grade generation submodule maintains the candidate pain load level; when the maximum threshold amplitude does not reach the amplitude range corresponding to the candidate pain load level, the grade generation submodule reduces the candidate pain load level and generates a pain load level. During the startup phase, when the reliable time-series signals and correction time-series signals available for grading within the current judgment window are insufficient to cover all physiological sampling channels, the grading generation submodule generates a grading process record based on the obtained available channel status, and treats missing or isolated channels as constraint states that do not improve the candidate grading level. Once subsequent judgment windows have a complete set of available time-series correction signals, the grading generation submodule switches to a stable operating rule that uses all available channels to jointly determine the candidate pain load level. In this embodiment, the grading module first performs channel matching according to the threshold index relationship, then determines the candidate pain load level according to the number of items exceeding the threshold, and verifies the candidate level with the maximum threshold exceedance. This ensures that the pain load level is constrained by both the multi-channel participation status and the deviation degree of a single channel, thereby avoiding the direct triggering of mismatched intervention control based solely on a single channel abnormality or solely on changes in the number of items.

[0027] Control module: The control module receives the pain load level and the execution record of the previous intervention cycle, generates a stimulus tolerance margin, compares the stimulus tolerance margin with the safe release threshold, and outputs a stimulus control command. The stimulus tolerance margin is a control constraint field formed by the allowed stimulus intensity and duration boundaries for the current intervention cycle, determined by the stimulus intensity record, stimulus duration record, and stop-trigger status from the previous intervention cycle's execution record. The stimulus control command is the control output used to indicate the stimulus release status for the current intervention cycle, including whether the stimulus is released, the release intensity boundary, the release duration boundary, and the restriction status after stop-trigger. The control module includes a recording and parsing submodule, a margin generation submodule, and an instruction output submodule. The recording and parsing submodule acquires the execution record of the previous intervention cycle and extracts the stimulus intensity record, stimulus duration record, and stop trigger state from it to generate an execution boundary record. The execution boundary record describes the stimulus release boundary and the reason for cessation in the previous intervention cycle and is output to the margin generation submodule. If a stop trigger state exists in the previous intervention cycle's execution record, the recording and parsing submodule uses this stop trigger state as the priority criterion for limiting stimulus release in the current cycle; if no stop trigger state exists, the recording and parsing submodule uses the stimulus intensity record and stimulus duration record from the previous cycle as inputs for generating the stimulus tolerance margin. The reserve generation submodule determines the permissible stimulus intensity and duration boundaries for the current intervention cycle based on the execution boundary records, and jointly defines the stimulus tolerance reserve as the stimulus tolerance reserve. The stimulus intensity boundary refers to the boundary state that the stimulus release intensity is allowed to reach within the current intervention cycle, and the stimulus duration boundary refers to the boundary state that the stimulus release duration is allowed to reach within the current intervention cycle. When generating the stimulus tolerance reserve, the reserve generation submodule prioritizes the stop triggering state, uses the stimulus intensity change as the basis for intensity boundary adjustment, and uses the stimulus duration as the basis for duration boundary adjustment, ensuring that the current intervention cycle does not directly release the stimulus outside the execution state of the previous intervention cycle. The instruction output submodule compares the stimulus tolerance margin with the safe release threshold and outputs stimulus control instructions based on the pain load level. The safe release threshold is a boundary field used to limit whether the stimulus tolerance margin is allowed to enter the release state; it originates from the baseline module and is used in the control judgment along with the intervention baseline record. When the stimulus tolerance margin meets the release conditions corresponding to the safe release threshold, the instruction output submodule determines the release level of the stimulus control instruction based on the pain load level; when the stimulus tolerance margin does not meet the release conditions corresponding to the safe release threshold, the instruction output submodule outputs a stimulus control instruction to restrict or stop release, and associates the reason for the restriction with the execution boundary record. When the pain load level is low, the instruction output submodule outputs control instructions matching the low load state and constrains stimulus release by the stimulus tolerance margin. When the pain load level is intermediate, the instruction output submodule outputs stimulus control instructions matching the load state within the range allowed by the safe release threshold. When the pain load level is high, the instruction output submodule still first checks the comparison result between the stimulus tolerance margin and the safe release threshold, and only outputs the corresponding stimulus control instruction if the release boundary is met. Thus, the pain load level is used to determine the intervention requirement, and the stimulus tolerance margin and safe release threshold are used to limit the actual release boundary. When the execution record of the previous intervention cycle is missing, the stop triggering state cannot be identified, or the stimulus intensity and duration records are incomplete, the record parsing submodule writes the corresponding record into the abnormal state field of the execution boundary record, and the margin generation submodule generates the stimulus tolerance margin using the boundary state of restricted release. When the instruction output submodule reads this abnormal state field, it does not directly increase the stimulus release according to the pain load level, but outputs a restricted stimulus control instruction, and waits for the execution record of the next intervention cycle to be restored to complete before entering the stable control rule; In this embodiment, the pain load level and the execution record of the previous intervention cycle are incorporated into the stimulation control instruction generation process through the control module. This ensures that the stimulation control instruction not only responds to the current multimodal physiological load, but is also constrained by the stimulation tolerance margin and the safe release threshold, thereby making the intervention output consistent with the current pain load state, the execution boundary of the previous cycle, and the safe release conditions.

[0028] Based on the above device structure, the physiological signal time-series data output by the acquisition module enters the baseline module and the correction module. The physiological grading threshold, transmission hysteresis upper limit, and safe release threshold output by the baseline module enter the grading module, the correction module, and the control module, respectively. The time-series correction signal set output by the correction module enters the grading module. The pain load level output by the grading module enters the control module. The control module outputs stimulation control commands and forms the execution record of the previous intervention cycle that can be received by the baseline module for the next intervention cycle. Through this data flow, the multimodal biofeedback-driven chronic pain dynamic grading intervention device forms a closed loop between acquisition, baseline generation, time-series correction, pain load grading, and stimulation control, ensuring the continuity of the input sources, output destinations, and subsequent call relationships of each component.

[0029] 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 multimodal biofeedback-driven dynamic grading intervention device for chronic pain, characterized in that, include: The acquisition module is used to collect heart rate fluctuation values, skin conductance response rise slope values, muscle electrical activity duration amplitude values, limb micro-movement amplitude values, sampling time recording, and communication arrival time to generate physiological signal time series data. The baseline module, connected to the acquisition module, is used to generate physiological grading thresholds, transmission hysteresis upper limits, and safe release thresholds based on the initial resting sampling data, and to receive the execution record of the previous intervention cycle. A correction module, connected to the acquisition module and the reference module, is used to calculate the transmission hysteresis value based on the physiological signal time series data, compare the transmission hysteresis value with the transmission hysteresis upper limit, and generate a time series correction signal set. A grading module, connected to the correction module, is used to compare the physiological values ​​in the time-series correction signal set with the physiological grading threshold, and generate a pain load level based on the number of over-threshold items and the over-threshold amplitude after channel baseline normalization. The control module, connected to the grading module and the benchmark module, is used to generate a stimulus tolerance margin from the previous intervention cycle execution record, compare the stimulus tolerance margin with the safe release threshold, and output stimulus control commands according to the pain load level.

2. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 1, characterized in that, The acquisition module includes a time reference submodule, a multi-channel acquisition submodule, and a timing encapsulation submodule; The time reference submodule receives the sampling time record and writes the same sampling window number to the heartbeat interval sampling channel, skin electric field sampling channel, muscle electric field sampling channel and limb micro-motion sampling channel to generate channel time markers; The multi-channel acquisition submodule acquires the heart rate fluctuation value, the skin conductance response rise slope value, the muscle electrical activity duration amplitude value, and the limb micro-movement amplitude value according to the channel time stamp, and writes the communication arrival time into the corresponding sampling window to obtain the channel acquisition record; The timing encapsulation submodule associates the channel acquisition records with the sampling window number, writes a missing marker at the missing channel position, and outputs the physiological signal timing data.

3. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 2, characterized in that, The time base module's time stamp writing process includes: Obtain the sampling interval between two adjacent sampling time records, and compare the sampling interval with a preset sampling interval range. When the sampling interval is within the preset sampling interval range, write the current sampling time record into the corresponding channel acquisition record. When the sampling interval exceeds the preset sampling interval range, mark the current channel acquisition record as a record to be corrected. The timing encapsulation submodule establishes a time-stamped difference record according to the sampling time record and communication arrival time of the record to be corrected, and associates the time-stamped difference record with the record to be corrected and writes it into the physiological signal timing data.

4. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 1, characterized in that, The benchmark module includes a resting screening submodule, a threshold generation submodule, and a record receiving submodule; The resting screening submodule removes sampling segments from the initial resting sampling data whose communication arrival time exceeds the preset resting lateness boundary relative to the sampling time, and removes sampling segments that do not simultaneously have heart rate interval fluctuation values, skin conductance response rise slope values, muscle electrical activity duration amplitude values, and limb micro-movement amplitude values. It extracts the effective resting values ​​according to the physiological sampling channels and generates the resting channel baseline. The threshold generation submodule determines the heart rate interval fluctuation threshold, skin conductance slope threshold, muscle conductance amplitude threshold, and limb micro-movement amplitude threshold based on the resting channel baseline. It also extracts the sampling time record and communication arrival time of each sampling segment from the initial resting sampling data retained by the resting screening submodule and generates a set of resting hysteresis values ​​according to the lag relationship between the communication arrival time and the sampling time record. The set of resting hysteresis values ​​is sorted according to the hysteresis magnitude, and the hysteresis value corresponding to the preset hysteresis quantile position is selected as the upper limit of transmission hysteresis, so that the upper limit of transmission hysteresis is objectively determined by the actual communication hysteresis distribution in the resting state. The physiological grading threshold is generated based on the cardiac interval fluctuation threshold, the skin electrophysiology slope threshold, the muscle electrical amplitude threshold, the limb micro-movement amplitude threshold, and the transmission hysteresis upper limit. The record receiving submodule receives the execution record of the previous intervention cycle and extracts the stimulus duration, stimulus intensity change and termination state from the execution record of the previous intervention cycle to generate an intervention baseline record.

5. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 4, characterized in that, The threshold generation process of the threshold generation submodule includes: The resting center value and resting fluctuation range of each physiological sampling channel in the resting channel baseline are obtained. The resting center value and the resting fluctuation range are jointly determined as the threshold generation basis for the corresponding physiological sampling channel. The threshold index relationship is established according to the channel order of heartbeat interval, skin conductance response, muscle electrical activity and limb micro-movement. The threshold generation submodule writes the threshold index relationship and the physiological grading threshold into the threshold configuration table, and provides the threshold configuration table to the grading module for judging the number of over-threshold items and the over-threshold amplitude after normalization according to the resting fluctuation range.

6. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 1, characterized in that, The correction module includes a hysteresis calculation submodule, an upper limit comparison submodule, and a signal correction submodule; The hysteresis calculation submodule acquires the sampling time record and communication arrival time in the physiological signal time series data, and determines the transmission hysteresis value according to the lag relationship between the communication arrival time and the sampling time record within the same sampling window. The upper limit comparison submodule compares the transmission hysteresis value with the transmission hysteresis upper limit, and generates a normal timing state, a correctable hysteresis state, or an over-limit hysteresis state based on the comparison result. The signal correction submodule writes correction flags to the physiological values ​​in the corresponding sampling window according to the normal timing state, the correctable hysteresis state, or the excessive hysteresis state, and generates the timing correction signal set.

7. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 6, characterized in that, The correction mark writing process of the signal correction submodule includes: When the comparison result is the normal time series state, the correlation between the physiological values ​​and the sampling time records within the corresponding sampling window is maintained, and a reliable marker is written to generate a reliable time series signal. When the comparison result is the correctable hysteresis state, the corresponding physiological value is mapped to the adjacent sampling window and written into the correction mark according to the delay direction between the communication arrival time and the sampling time record, thereby generating a correction timing signal. When the comparison result is the over-limit hysteresis state, the corresponding physiological value is retained and written into the isolation mark, an isolation timing signal is generated, and the reliable timing signal, the correction timing signal and the isolation timing signal are collected into the timing correction signal set.

8. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 1, characterized in that, The grading module includes a threshold reading submodule, an over-threshold determination submodule, and a grading generation submodule; The threshold reading submodule acquires the time-series correction signal set and the physiological grading threshold, and establishes a threshold matching record according to the channel correspondence of heartbeat interval, skin conductance response, muscle electrical activity and limb micro-movement; The threshold determination submodule compares each physiological value in the time-series correction signal set with the corresponding threshold in the threshold matching record to determine the number of threshold items and the threshold amplitude of each threshold item. The physiological values ​​with confidence and correction labels in the time-series correction signal set are compared with the physiological grading threshold. Physiological values ​​with isolation or missing labels are not used as the basis for generating the current pain load level. The level generation submodule generates the pain load level based on the number of overthreshold items and the overthreshold amplitude of each overthreshold item.

9. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 8, characterized in that, The level generation process of the level generation submodule includes: The number of out-of-threshold items and the out-of-threshold amplitude corresponding to each out-of-threshold item are obtained. The maximum out-of-threshold amplitude is determined from the out-of-threshold amplitudes corresponding to each out-of-threshold item. Candidate pain load levels are determined according to a preset range of item numbers, and the candidate pain load levels are verified according to a preset range of amplitudes. When the maximum overthreshold amplitude reaches the amplitude range corresponding to the candidate pain load level, the candidate pain load level is maintained; when the maximum overthreshold amplitude does not reach the amplitude range corresponding to the candidate pain load level, the candidate pain load level is reduced, and the pain load level is generated.

10. The multimodal biofeedback-driven dynamic grading intervention device for chronic pain according to claim 1, characterized in that, The control module includes a recording and parsing submodule, a margin generation submodule, and an instruction output submodule. The recording and parsing submodule obtains the execution record of the previous intervention cycle and extracts the stimulus intensity record, stimulus duration record, and stop triggering state from the execution record of the previous intervention cycle to generate an execution boundary record. The margin generation submodule determines the stimulus intensity boundary and stimulus duration boundary that can be released in the current intervention cycle based on the execution boundary record, and determines the stimulus intensity boundary and the stimulus duration boundary together as the stimulus tolerance margin. The stimulation tolerance margin includes an intensity margin for limiting the upper limit of stimulation release intensity in the current intervention cycle and a duration margin for limiting the upper limit of stimulation release duration in the current intervention cycle. The intensity margin is used to compare with the safe intensity threshold in the safe release threshold, and the duration margin is used to compare with the safe duration threshold in the safe release threshold. The instruction output submodule compares the intensity margin with the safe intensity threshold and the duration margin with the safe duration threshold. When the intensity margin meets the safe intensity threshold and the duration margin meets the safe duration threshold, it outputs a stimulus control instruction that allows release in conjunction with the pain load level. When either the intensity margin or the duration margin fails to meet the corresponding safety threshold, a stimulus control command to limit release and stop release is output.

Images