Intelligent automatic pain analysis system

Abstract

The invention provides an intelligent automatic pain analysis system, which is characterized by comprising a hand wearable device, a wrist wearable device, a foot wearable device, a finger EDA skin electric signal measuring device, a finger EDA skin electric signal measuring device, a finger EDA skin electric signal measuring device, a finger EDA skin electric signal measuring device, a finger EDA skin electric signal measuring device, a finger EDA skin electric signal measuring device and a finger EDA skin electric signal measuring device, and the finger EDA skin electric signal measuring device is connected with the wrist wearable device to collect hand skin electric signals. Signals generated by the hand wearable device and the foot wearable device are transmitted to the wrist wearable device in a wireless mode, and finger skin electric signals measured by the wrist wearable device and the skin electric signal measuring device are transmitted to the remote server. And predicting the pain level by the remote server by using a pre-trained intelligent model based on a heterogeneous signal time domain attention mechanism. On the basis of data monitoring of the heart rate, the blood pressure, the hand and foot touch pressure and the galvanic skin, a remote server is used for carrying out an intelligent model based on a heterogeneity attention signal time domain attention mechanism to obtain automatic analysis of the pain level, and human body detection of pain can be used for drug adverse clinical response research outside a hospital without abortion.

Description

Technical Field

[0001] The invention relates to an automatic pain analysis system, in particular to an intelligent automatic pain analysis system, and belongs to the field of intelligent medical systems. Background Art

[0002] Existing medicine uses mice for experiments to study the effects of drugs on pain. However, the results of studies from animals to humans are often not linear. Therefore, it is most accurate to directly collect human pain data to study the stimulation of drug clinical trials. Therefore, it is necessary to combine the data foundation of pain signals with the analysis of the model to obtain accurate drug evaluation.

[0003] On the other hand, it is also extremely important to collect data from life outside the hospital. This allows volunteers to conduct large-scale experiments without leaving their jobs, thereby improving the efficiency of the experiments. Therefore, how to use flexible and convenient data collection devices is also the key to the technology.

[0004] Finally, since pain is a subjective feeling of a person and can be described in words, the mechanism behind it is the abnormality of physiological parameters, so it can be described as abnormal parameters-pain description. The abnormal parameters are caused by factors such as medication, so pain can be described as a process of drug-parameter-pain description in a time series, which is essentially a process of natural language generation. Therefore, if the attention mechanism is considered, it can replace people and automatically analyze the level description of pain in real time. Summary of the invention

[0005] Based on the above problems in the prior art, the present invention will consider the following technical points: first, consider a wearable device for collecting heart rate, blood pressure, and finger surface pressure data of hands and feet; second, use an intelligent model with an attention mechanism based on the time interval of heterogeneous signals to analyze pain data, so as to train the pain level based on the temporal relationship between the four types of heterogeneous signals of hand, wrist, foot, and skin electricity.

[0006] Based on the above considerations, the present invention provides an intelligent automatic pain analysis system, including three wearable devices at the hand, wrist, and foot, which are respectively used for data collection of hand touch pressure, heart rate and blood pressure, and foot touch pressure over time function, a finger EDA skin electrical signal measurement device, which is used to connect with the wrist wearable device to collect hand skin electrical signals, and a remote server, wherein: The signals generated by the wearable devices on the hands and feet are wirelessly transmitted to the wearable device on the wrist, and the wearable device on the wrist sends the finger electrocutaneous signals measured by the electrocutaneous signal measuring device together to the remote server. The remote server predicts the pain level using a pre-trained intelligent model based on the time domain attention mechanism of heterogeneous signals.

[0007] Optionally, the hand wearable device includes a plurality of hand touch pressure sensors arranged on a flexible substrate, and a first integrated circuit connected to all the hand touch pressure sensors arranged on the flexible substrate; the foot wearable device includes a flexible foot cover, the flexible foot cover includes a base pad provided with a plurality of foot touch pressure sensors and a second integrated circuit connected to all the foot touch pressure sensors; the wrist wearable device includes a host and a wristband connected to the host; the EDA electrodermal signal measuring device includes a plurality of electrodermal acquisition sheets, which are connected to a detachable electrically connected interface on the host via a signal line; wherein the first integrated circuit and the second integrated circuit both include a data processor and a data wireless transmitter for communicating with the host or directly sending data to the remote server.

[0008] The pre-training method of the intelligent model based on the temporal attention mechanism of heterogeneous signals includes the following steps: S1 constructs the foot touch pressure, hand touch pressure, SCL (skin electrical signal level), and heart rate time change function to obtain all SCL peaks; S2 For each SCL peak, divide the time span of the peak into n time units , , . . . , , constructing a long short-term memory model with each time unit as a unit ,Will Microelement correspondence And the detected heart rate Input the first unit, input the next unit through the intermediate transmission layer, and then The detected heart rate corresponding to the microelement Input the second unit, and then input the next unit through the intermediate transmission layer, and so on, we get Multiple output values, input first In the function, multiple outputs are obtained , forming the output matrix ; At the same time, the blood pressure measured in each time unit , , . . . , The matrix formed With the output matrix Perform matrix multiplication to get the attention value ,in The heart rate matrix formed by the heart rates of n time units; S3 For each SCL peak, follow the steps of S2 based on the same n time elements. , , . . . , Establishing a short-term memory model of hand touch and pressure , also using the second The function gets multiple outputs, forming a matrix , and the matrix formed by foot contact pressure Matrix multiplication to form motion values , is the hand touch pressure matrix; S4 Calculation of pain level , repeat steps S2-S3 to obtain the pain level of all SCL peaks, thus completing the construction of the model. , it is during exercise that the pain level cannot be judged. , for mood swings, , for mild pain, , moderate pain, , very painful, The pain threshold assessed by all volunteers; S5 recruits volunteers to wear wearable devices on their hands, wrists, and feet, and finger EDA electrodermal signal measurement devices to take drugs. The remote server collects data for training the model according to steps S2-S4.

[0009] It should be understood that by introducing the hand and foot touch pressure training, the movement value obtained excludes the skin electrical fluctuations caused by factors such as movement and emotions, and filters out relatively pure contributions belonging to pain.

[0010] Optionally, the hand wearable device is not attached to the palm surface through the hydrogel, and the wrist wearable device is a smart watch.

[0011] Beneficial Effects Through sensors and intelligent devices on the hands, feet, and wrists, data monitoring of heart rate, blood pressure, hand and foot touch pressure, and skin electricity is achieved, and the data is sent to a remote server for automatic analysis of pain levels based on an intelligent model of attention mechanism in the time domain of heterogeneous attention signals. This enables human body detection of pain to be used in the study of adverse clinical drug reactions outside the hospital without leaving the workplace. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of the composition of a wearable intelligent automatic pain analysis system described in an embodiment of the present invention, Figure 2 Function graph of foot touch pressure, hand touch pressure, SCL, and heart rate signals changing with time. Figure 3 Algorithm logic diagram of the pre-training method of the intelligent model based on the temporal attention mechanism of heterogeneous signals. DETAILED DESCRIPTION

[0013] like Figure 1A wearable intelligent automatic pain analysis system is presented, including a flexible substrate with hydrogel attached to the palm surface, a main body with a touch screen, and a smart watch and foot cover with wristbands connected to both sides of the main body, which are used to respectively analyze the hand touch pressure, heart rate and blood pressure, and foot touch pressure over time (e.g. Figure 2 As shown in the figure, a finger EDA skin electrical signal measuring device is used to connect with a wrist wearable device to collect hand skin electrical signals, and a remote server.

[0014] Figure 1 A hand pressure sensor A is arranged on the flexible base of the mesohydrogel attached to the palm surface, one for each of the thumb, index finger and middle finger of the right hand, and two for the palm. The set includes a bottom pad and a top surface. The bottom pad is provided with a foot pressure sensor B, one for the heel, one for the front end of the foot, and one for the big toe. Each hand pressure sensor A and foot pressure sensor is connected to a first integrated circuit at the heel and a second integrated circuit arranged on the top surface through a flexible wire. Both integrated circuits have data processing and wireless transmitters, which are respectively used to process the collected hand pressure and foot pressure data and transmit them to a remote server.

[0015] The finger EDA skin electrical signal measurement device has three skin electrical acquisition pieces, which are set on the thumb, index finger and palm of the left hand. They are connected to the detachable electrical connection interface on the host through a signal line, and the host uploads the skin electrical signal to the remote server. The remote server uses a pre-trained intelligent model based on the attention mechanism of heterogeneous signal time intervals to input the received data and predict the pain level.

[0016] Figure 2 The function graphs of foot touch pressure, hand touch pressure, SCL, and heart rate signals changing with time are given, including two stages when the subjects take the medicine and after taking the medicine. For convenience, the time period between taking the medicine and the onset of pain is cut out for ease of explanation. In the figure, after the pain occurs, the heart rate begins to speed up, which is manifested as a shortened cycle, and three corresponding main peaks appear on the SCL (indicated by the downward arrow). The hand touch pressure function graph shows multiple peaks caused by disassembling the packaging when taking the medicine, as well as a stable pressure signal after holding the cup.

[0017] The foot touch pressure function graph shows two pressure reductions caused by lifting the legs due to pain, indicating that the subject moved his body due to pain. However, since there are fewer peaks in the hand touch pressure and foot touch pressure, the factor of physical exercise can be ruled out. Therefore, from the overall analysis of heterogeneous signals, to achieve intelligent automatic analysis and perception of pain, it is necessary to analyze the attention mechanism of each signal.

[0018] The pre-training method of the intelligent model based on the temporal attention mechanism of heterogeneous signals includes the following steps: S1 constructs the foot touch pressure, hand touch pressure, SCL skin electrodermal signal level, and heart rate time variation function to obtain all SCL peaks (e.g. Figure 2 ); S2 For each SCL peak, divide the time span of the peak into n time units , , . . . , , constructing a long short-term memory model with each time unit as a unit ,Will Microelement correspondence And the detected heart rate Input the first unit, input the next unit through the intermediate transmission layer, and then The detected heart rate corresponding to the microelement Input the second unit, and then input the next unit through the intermediate transmission layer, and so on, we get Multiple output values, input first In the function, multiple outputs are obtained , forming the output matrix ; At the same time, the blood pressure measured in each time unit , , . . . , The matrix formed With the output matrix Perform matrix multiplication to get the attention value ,in The heart rate matrix formed by the heart rates of n time units; S3 For each SCL peak, follow the steps of S2 based on the same n time elements. , , . . . , Establishing a short-term memory model of hand touch and pressure , also using the second Function gets multiple outputs , forming a matrix , and the matrix formed by foot contact pressure Matrix multiplication to form motion values , is the hand touch pressure matrix; S4 Calculation of pain level , repeat steps S2-S3 to obtain the pain level of all SCL peaks, thus completing the construction of the model. , it is during exercise that the pain level cannot be judged. , for mood swings, , for mild pain, , moderate pain, , very painful, The pain threshold assessed by all volunteers; S5 recruits volunteers to wear wearable devices on their hands, wrists, and feet, and finger EDA electrodermal signal measurement devices to take drugs. The remote server collects data for training the model according to steps S2-S4.

Claims

1. An intelligent automatic pain analysis system, characterized in that: It includes three wearable devices, namely, a hand wearable device, a wrist wearable device, and a foot wearable device, which are respectively used for data collection of hand touch pressure, heart rate and blood pressure, and foot touch pressure over time function, a finger EDA skin electrical signal measurement device, which is used to connect with the wrist wearable device to collect hand skin electrical signals, and a remote server, wherein, The signals generated by the wearable devices on the hands and feet are wirelessly transmitted to the wearable device on the wrist, and the wearable device on the wrist sends the finger electrocutaneous signals measured by the electrocutaneous signal measuring device together to the remote server. The remote server predicts the pain level using a pre-trained intelligent model based on the time domain attention mechanism of heterogeneous signals.

2. The analysis system according to claim 1, characterized in that The hand wearable device includes a plurality of hand touch pressure sensors arranged on a flexible substrate, and a first integrated circuit connected to all the hand touch pressure sensors arranged on the flexible substrate; the foot wearable device includes a flexible foot cover, the flexible foot cover includes a base pad provided with a plurality of foot touch pressure sensors and a second integrated circuit connected to all the foot touch pressure sensors; the wrist wearable device includes a host and a wristband connected to the host; the EDA skin electrodermal signal measuring device includes a plurality of skin electrodermal acquisition sheets, which are connected to a detachable electrically connected interface on the host via a signal line; wherein the first integrated circuit and the second integrated circuit both include a data processor and a data wireless transmitter for communicating with the host or directly sending data to the remote server.

3. The method according to claim 2, characterized in that The pre-training method of the intelligent model based on the temporal attention mechanism of heterogeneous signals includes the following steps: S1 constructs the time-varying functions of foot touch pressure, hand touch pressure, skin electrodermal signal level (SCL), and heart rate, and obtains all SCL peaks; S2 For each SCL peak, divide the time span of the peak into n time units , , . . . , , constructing a long short-term memory model with each time unit as a unit ,Will Microelement correspondence And the detected heart rate Input the first unit, input the next unit through the intermediate transmission layer, and then The detected heart rate corresponding to the microelement Input the second unit, and then input the next unit through the intermediate transmission layer, and so on, we get Multiple output values, input first In the function, multiple outputs are obtained , forming the output matrix ; At the same time, the blood pressure measured in each time unit , , . . . , The matrix formed With the output matrix Perform matrix multiplication to get the attention value ,in The heart rate matrix formed by the heart rates of n time units; S3 For each SCL peak, follow the steps of S2 based on the same n time elements. , , . . . , Establishing a short-term memory model of hand touch and pressure , also using the second Function gets multiple outputs , forming a matrix , and the matrix formed by foot contact pressure Matrix multiplication to form motion values , is the hand touch pressure matrix; S4 Calculation of pain level , repeat steps S2-S3 to obtain the pain level of all SCL peaks, thus completing the construction of the model. , it is during exercise that the pain level cannot be judged. , for mood swings, , for mild pain, , moderate pain, , very painful, The pain threshold assessed by all volunteers; S5 recruits volunteers to wear wearable devices on their hands, wrists, and feet, and finger EDA electrodermal signal measurement devices to take drugs. The remote server collects data for training the model according to steps S2-S4.

4. The method according to claim 3, characterized in that: The hand wearable device is not attached to the palm surface through the hydrogel, and the wrist wearable device is a smart watch.

Images