Stroke Detection Sensor
Data is collected through multiple wearable sensors and the probability of stroke is determined in the data processing device. Combined with countdown alarm and user test verification, the problem of frequent false alarms in the prior art is solved, and the accuracy and reliability of stroke detection is improved.
Patent Information
- Application Number
- CN201980085334.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-11-12
- Filing Date
- 2019-12-19
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2039-12-19
AI Technical Summary
The existing stroke detection system is difficult to generate reliable stroke detection signals without triggering false alarms, and it is inconvenient to handle false alarms when false alarms occur.
A plurality of wearable sensors are used to collect motion data, determine the first probability in the first time period of the stroke and the second probability in the second time period through the data processing device, generate a stroke detection signal, and verify it through the countdown alarm and user test. Signal accuracy reduces false alarms.
It improves the accuracy and reliability of stroke detection, reduces the occurrence of false alarms, and ensures that false alarms can be handled appropriately when false alarms do occur, avoiding inconvenience to users.
Smart Images

Figure CN113226176B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to apparatus and methods for detecting a stroke in a patient. Background Art
[0002] A stroke is a medical condition in which poor blood flow to the brain causes cell death. There are two main causes of a stroke. An ischemic stroke is caused by insufficient blood flow to multiple parts of the brain, usually due to a blockage in an artery that supplies blood to the brain. Blood normally carries oxygen and nutrients to the brain. Once oxygen and nutrients are cut off by the blockage, brain cells cannot make enough energy and will eventually stop working. If the blockage is not cleared, brain cells eventually die. A hemorrhagic stroke is caused by bleeding in the brain. The bleeding is usually caused by damaged blood vessels leaking blood. A hemorrhagic stroke can also be caused by a ruptured brain aneurysm. In both cases, blood spreads into the surrounding brain tissue, causing increased pressure that restricts the functioning of brain cells and ultimately damages brain tissue.
[0003] In both types of stroke, because brain cells stop functioning normally, the result is a change in brain function. This change can be seen through physical symptoms, such as the inability to move or feel one side of the body, problems communicating, and vision loss. These physical symptoms tend to appear more or less immediately after the stroke begins.
[0004] Historically, strokes have been difficult to treat. Although a patient's symptoms may be recognized and diagnosed as a stroke, there are limited treatment options available. When a stroke is identified as an ischemic stroke, the patient is given a clot-dissolving drug (such as tissue plasminogen activator) intravenously in the hope that the drug will reach the clot and dissolve it enough to allow blood flow to resume through the affected artery. This treatment needs to be successfully administered within a few hours of the onset of the stroke to ensure that damage to brain tissue is limited. Some studies have shown that the window of time to get the best effect from clot-dissolving drugs is three hours from the first signs of a stroke.
[0005] If a stroke is not successfully treated, damage to brain tissue is inevitable and the only recourse is to provide care and rehabilitation training to the patient.
[0006] Recently, ischemic strokes have been successfully treated via an endovascular procedure called "mechanical thrombectomy," in which a blood clot is removed by delivering a clot-retrieval device to the site of the blocked blood vessel in the brain. The device secures the clot and pulls it out of the vessel when the device is removed. Similarly, hemorrhagic strokes can also be treated via an endovascular procedure by delivering a metal clip to a damaged blood vessel or ruptured aneurysm. The clip is secured to restrict blood flow and further prevent blood from seeping into surrounding tissue. As with clot-dissolving drugs, this procedure can significantly reduce the damage if performed within a few hours of the first signs of a stroke.
[0007] Therefore, as treatment options for ischemic stroke improve, the importance of identifying and characterizing stroke symptoms increases. For patients at increased risk for stroke, a method is needed to provide continuous monitoring to rapidly diagnose and report stroke to healthcare providers.
[0008] Known techniques utilize the use of sensors worn on the patient to detect stroke in the patient early. Villar.J R.'s "A Hybrid Intelligent Identification System for Early Detection of Stroke" describes the use of a wearable device that generates an alarm and automatically connects to an e-health service when a stroke is detected. The above method utilizes two wearable devices to monitor motion data and uses a genetic fuzzy finite state machine and time series (TS) analysis to determine stroke.
[0009] US patent application 2018153477 discloses a device for monitoring a patient for stroke via a plurality of sensors for determining "physiological signals". Physiological signals may include heart rate signals, atrial rate signals, heart rate variability signals, blood pressure signals, blood pressure variability signals, heart sound signals, etc.
[0010] U.S. Patent 7,981,058 discloses a device for monitoring a patient using a low-cost dual-axis motion sensor. A first sensor captures objective acceleration data, and a second dual-axis sensor captures subjective acceleration data at least relative to the first accelerometer. The acceleration data is then used to determine nonlinear parameters and generate at least two levels of motor function information.
[0011] US application US2017281054 discloses a device for monitoring a patient for stroke via a plurality of motion sensors located at one or more anatomical locations on the patient's body and configured to detect a plurality of motion parameters corresponding to the motion of a part of the patient's body. The motion of the patient's body part is then based on a plurality of predetermined motion determination features.
[0012] US application US2015157252 discloses a device for monitoring a patient via multiple sensors. A technique for switching between sensors to determine which sensor provides the best indication of the patient's physiological signal state is described.
[0013] However, problems associated with the above-described systems include problems with generating a reliable signal indicating that a patient has had a stroke without triggering too many false alarms. What is needed is a system that can generate a stroke detection signal with minimal false alarms, and in the event that false alarms do occur, the system can handle them appropriately without causing too much inconvenience to the user. Summary of the invention
[0014] It is an object of the present invention to at least partially overcome one or more of the above-mentioned limitations of the prior art.
[0015] One or more of these objects, as well as further objects that may emerge from the description below, are at least partially achieved by: a method for data processing, a computer-readable medium, an apparatus for data processing, and a device according to the independent claim, embodiments of the device being defined by the dependent claims.
[0016] One embodiment of the present disclosure describes a stroke detection device, which includes: a data processing device, which includes a processor; at least one wearable sensor, which is constructed to be worn by a user and is configured to transmit the user's motion data to the data processing device, the data processing device is configured to process the motion data to determine a first probability of the user having a stroke in a first time period and a second probability of the user having a stroke in a second time period, and generate a stroke detection signal based on at least the first probability and the second probability.
[0017] Another embodiment of the present disclosure describes a method for generating a stroke detection signal, the method comprising: using at least one wearable sensor device to generate motion data of a user, the wearable sensor device being configured to be worn by the user, transmitting the motion data to a processing device, processing the motion data at the processing device to determine a first probability that the user has had a stroke in a first time period and a second probability that the user has had a stroke in a second time period, and generating a stroke detection signal based on at least the first probability and the second probability. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Embodiments of the present invention will be described in more detail with reference to the accompanying schematic drawings. Figure 1 is a schematic illustration showing a patient wearing a multi-sensor stroke detection device. Figure 2 A perspective view showing a stroke detection device according to an embodiment of the present application. Figure 3is a schematic diagram of a stroke detection device according to an embodiment of the present application. Figure 4a and Figure 4b is a process flow diagram showing the execution flow of the wearable sensor and the control device respectively. Figure 5a and Figure 5b is a sequence diagram showing a data transmission order between a first wearable sensor device and a second wearable sensor device and a control device. Figure 6 1 is a process flow diagram showing the execution flow of the control device 30 . Figure 7 2 is a process flow diagram showing an alternative execution flow of the control device 30 . Figures 8a to 8c is a diagram of exemplary motion sensor signals for a stroke detection device. Fig. 9 1 is a processing flowchart showing an alternative execution flow of the control device 30 . Fig.10 The time it takes for an embodiment of the system to detect a stroke given a particular configuration is shown. DETAILED DESCRIPTION
[0019] The present invention relates to a wearable sensor and a technique for providing monitoring and detection of a patient's condition, and reporting the patient's condition where necessary. The same reference numerals are used throughout this specification to refer to corresponding elements.
[0020] Figure 1 is a schematic diagram showing a patient 10 equipped with a stroke detection device 100, which includes a plurality of wearable sensors 20. Figure 1 In the embodiment, the wearable sensors 20a, 20b are attached to the patient's wrists. In other embodiments, the wearable sensors can be worn on the ankles as an alternative to the wrists or in combination with the wrists. Other locations where the wearable sensors can be worn include shoes, headwear, and attached to clothing such as pants, and upper body clothing.
[0021] Figure 2 An embodiment of a wearable sensor 20 is shown. In one embodiment, the wearable sensor 20 includes a band 21 and a sensor body 22, the band 21 being configured to fix the wearable sensor 20 to the patient, and the sensor body 22 accommodating a processing board 25. The processing board may include a power supply 26, a data processing device 27, and a sensor assembly 28. The sensor assembly 28 may include any suitable sensor component configured to measure the inclination, position, orientation, and / or acceleration of the portion of the patient's body to which the wearable sensor 20 is attached. The sensor assembly 28 may include a piezoelectric component, a piezoresistive component, and / or a capacitive component to convert mechanical motion into an electrical signal. Piezoelectric ceramics (e.g., lead zirconate titanate) or single crystals (e.g., quartz, tourmaline) may be used. Preferably, a capacitive accelerometer is used because the capacitive accelerometer has superior performance in the low frequency range.
[0022] The data processing device 27 can be implemented by special software (or firmware) running on one or more general or special computing devices (such as hardware processors). Each "element" or "device" of such a computing device refers to the conceptual equivalent of a method step; there is not always a one-to-one correspondence between an element / device and a specific hardware or software program. A piece of hardware sometimes includes different devices / elements. For example, a processing unit acts as an element / device when executing one instruction, but acts as another element / device when executing another instruction. In addition, an element / device can be implemented by one instruction in some cases, but can be implemented by multiple instructions in some other cases. Such a software-controlled computing device may include one or more processing units, such as a CPU ("central processing unit"), a DSP ("digital signal processor"), an ASIC ("application-specific integrated circuit"), discrete analog and / or digital components, or some other programmable logic devices such as an FPGA ("field programmable gate array"). The data processing device 27 may also include a system memory and a system bus, which couples various system components including the system memory to the processing unit. The system bus may be any of a variety of types of bus structures, including a memory bus or a memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. The system memory may include computer storage media in the form of volatile and / or nonvolatile memory, such as read only memory (ROM), random access memory (RAM), and flash memory. The application-specific software may be stored in the system memory, or on other removable / non-removable volatile / non-volatile computer storage media included in or accessible to the computing device, such as magnetic media, optical media, flash memory cards, digital tapes, solid-state RAM, solid-state ROM, etc. The application-specific software may be provided to the data processing device 27 on any suitable computer-readable medium, including recording media and read-only memory.
[0023] The data processing device 27 includes one or more communication interfaces and one or more data acquisition devices, the communication interface is, for example, a serial interface, a USB interface, a wireless interface, etc., and the data acquisition device is, for example, an A / D converter. In one example, the data processing device 27 may include a transmitter component, which is configured to send sensor data received from the sensor assembly 28 and processed by the data processing device 27 and / or the A / D converter through one or more communication interfaces. In one embodiment, via A WLAN or WiFi transceiver provides a communication interface, and the processed sensor data is sent to the control device 30 (as described below). Alternatively, the processed sensor data may be sent to one or more remote devices via a GSM, LTE or similar mobile communication interface.
[0024] The power source 26 may include a battery, a kinetic energy source or other power source suitable for the wearable device. The power source 26 is arranged to supply power energy to the data processing device 27 and the sensor assembly 28 of the processing board 25.
[0025] The wearable sensor 20 may further include a fastening component 23 configured to be secured together with a corresponding component 24 to enable the device to be secured to a limb of the patient 10. In one embodiment, the fastening component 23 includes a sensor configured to determine whether the band 21 of the wearable sensor 20 is in an "open" configuration or in a "secured" configuration. Figure 2 An example of an “open” configuration of the strap is shown in , where the wearable sensor 20 is not secured to anything. Figure 1 An example of a "closed" configuration is shown in FIG, wherein the wearable sensor 20 is secured to the patient. Figure 3 , an example of a "closed" configuration is shown in which the fastening component 23 has been fastened to the corresponding component 24 to arrange the band 21 of the wearable sensor 20 into a fixed loop. In one embodiment, the sensor of the fastening component 23 is electrically connected to the processing board 25 so that the data processing device 27 can determine the configuration of the band 21 of the wearable sensor 20.
[0026] The wearable sensor 20 may also include a feedback mechanism including at least one of a vibration motor or an equivalent physical actuator, a light source (e.g., an LED), or a sound source (e.g., a speaker). The feedback mechanism may be configured to generate one of a vibration prompt, a visual prompt, or an auditory prompt that can attract the attention of a healthy patient.
[0027] Figure 3 An exemplary wireless network according to an embodiment of the present application is shown. Figure 3 In the embodiment, two wearable sensors 20, shown as 20a and 20b, are worn by a patient (not shown), and the corresponding band 21 of each wearable sensor is in a "closed" configuration. Each wearable sensor 20a, 20b collects sensor data from the patient, processes the data, and transmits the data to a control device 30 via a wireless networking interface 34. Figure 4a The process is described in more detail. The control device 30 may be a mobile phone device, such as an Apple TM , Apple mobile phone TM , Apple tablet TM , Apple WatchTM , Android TM Device, Wearable Operating System TM The control device 30 receives data from the wearable sensor, processes the data, determines the patient's condition, and performs reporting processing when appropriate. Figure 4b Describe the process in more detail.
[0028] Figure 4a An embodiment of a processing flow for the data processing device 27 is shown. Figure 4a The process flow shown in can be executed in a continuous loop or periodically. In the case of periodically executing the process flow, a power saving mode can be adopted between executions to minimize battery use.
[0029] In step 410, the data processing device 27 receives the sensor output from the sensor assembly 28. In the case where the sensor data is generated by, for example, a piezoelectric component, the output from the piezoelectric component may require preprocessing to ensure that the desired analog signal is generated. An example of a preprocessing step may be to reduce high frequency noise generated by the piezoelectric component from the analog signal. The preprocessing step 420 may occur on the sensor assembly 28 or the data processing device 27.
[0030] After the pre-processing step 420 , a conversion of the signal from an analog signal to a digital signal is performed in a step 430 . This analog-to-digital conversion can take place at the sensor assembly 28 or at the data processing device 27 .
[0031] The digital signal is then processed in step 440 to reduce noise and timestamp the sensor reading. Step 440 may include converting the acceleration vector generated by the accelerometer to the norm of the acceleration vector. That is, the acceleration vector is converted to a strictly positive length in a single direction. This provides a number of advantages, including reduced storage space for storing vector data and invariance to accelerometer orientation. Other filters are contemplated in conjunction with the above steps or independently to ensure that the filtered acceleration vector signal is invariant to gravity or orientation of the sensor. In a preferred embodiment, any of the aforementioned filtering steps are performed locally on the wearable sensor.
[0032] Step 440 may further include applying a low pass filter in conjunction with the above step or independently to remove the acceleration vector on the accelerometer due to gravity. This can be achieved by removing slow or constant acceleration vectors from the difference of the acceleration vector. Advantageously, this enables the removal of noise caused by gravity.
[0033] The key understanding of the present disclosure is the need to capture the signal that most likely contains an indication of a stroke that occurred in the user's brain. In view of the difficulty in detecting other biosignals that may indicate that this situation has occurred, the present disclosure focuses on detecting signals from the body's central nervous system that indicate changes in brain operation. In an embodiment where the sensor for capturing the signal includes an accelerometer, the signal based on the acceleration vector captures the movement of the body, and therefore, it is necessary to filter the signals caused by human body movement that are not the direct result of the signal from the brain. For example, these signals are only the indirect result of the electrical signal reaching the muscle. Examples of movements that can be filtered include movements that are only the indirect result of signals from the nervous system. For example, when walking, electrical signals may stimulate the arms to swing forward and backward to ensure balance. However, the downward swing of the arms while walking may be the result of gravity and body mechanics, rather than the stimulation of any muscles. Therefore, where possible, these movements should be identified and removed from the acceleration vectors generated by one or more accelerometers.
[0034] Therefore, signals that should be filtered to provide the clearest signal of CNS activity may include: passive limb mechanics, other biosignals (e.g., heart function, tremors), or other involuntary muscle movements. Environmental noise (e.g., bus engines)
[0035] The sensor data is then formatted in step 445 according to the defined data structure for transmission to the control device 30. Finally, in step 450, the wearable sensor 20 transmits the formatted sensor data to the control device 30 using the wireless network interface 34.
[0036] Figure 4b An embodiment of a process flow for controlling the device 30 is shown. Figure 4b The process flow shown in can be executed in a continuous loop or periodically. In the case of periodically executing the process flow, a power saving mode can be adopted between executions to minimize battery use.
[0037] In step 460, the mobile device receives the formatted sensor data from the wearable sensor 20. Then, in step 470, the received sensor data is merged with the existing data previously received from the wearable sensor 20, and any other wearable sensors that transmit sensor data to the control device 30. In one embodiment, the data is stored in a local database that is stored on the control device 30. In one embodiment, the system implements "chunking" of the formatted sensor data, which includes dividing the data into blocks, each block including a header indicating some parameters (e.g., timestamp, size, etc. for the recorded signal data). This enables each data block to resynchronize the clock of the mobile device with the wearable sensor.
[0038] In step 475, data analysis is performed on the sensor data by the control device 30. The patient condition, such as a stroke condition, is then determined based on the data analysis 475. This includes determining that the probability of the current patient condition has exceeded a critical threshold based on the sensor data. Fig. 9 The process of step 475 and step 480 is described in more detail. When the patient's condition is determined, the reporting process 490 is triggered. Figure 6 and Figure 7 The reporting process 490 is described. In step 650, once the reporting process 490 is completed and the patient has been unable to cancel the reporting of the patient's condition, the control device 30 contacts the network point 35 via the network interface 34 to request emergency services to handle the patient's condition. In one embodiment, the mobile device can communicate at least one of the patient's condition, the GPS location of the mobile device, the patient ID, the patient's medical history, recent sensor data reports, etc.
[0039] It should be noted that step 475 and subsequent steps may be performed as direct subsequent steps of step 460. Alternatively, step 475 and subsequent steps may be performed in independent cycles that are independently triggered by, for example, interrupts at periodic times.
[0040] In this description, multiple wearable sensors are used to collect data. Figure 5a A process for collecting data from more than one wearable sensor at a control device 30 is shown. In steps 510 and 520, sensor data from a wearable sensor 20a positioned on the left side of the body and a wearable sensor 20b positioned on the right side of the body are collected. In step 530, the sensor data from the wearable sensor 20b is transmitted to the wearable sensor 20b via the wireless interface 34. In step 540, once the sensor data from the wearable sensor 20a is received at the wearable sensor 20b, the sensor data from the wearable sensor 20a is combined with the sensor data from the wearable sensor 20b (previously collected in step 520) and transmitted to the control device 30.
[0041] exist Figure 5a In an alternative embodiment to the embodiment shown in Figure 5bAn alternative flow for collecting data from more than one wearable sensor at the control device 30 is shown. In steps 510 and 520, sensor data from the wearable sensor 20a positioned on the left side of the body and the wearable sensor 20b positioned on the right side of the body are collected. In step 550, the control device 30 commands the wearable sensor 20a to send the sensor data collected by the wearable sensor 20a to the control device 30 via the wireless network interface 34. In step 560, the wearable sensor 20a sends the collected data to the control device 30 via the wireless network interface 34. In step 570, the control device 30 commands the wearable sensor 20b to send the sensor data collected by the wearable sensor 20b to the control device 30 via the wireless network interface 34. In step 580, the wearable sensor 20b sends the collected data to the control device 30 via the wireless network interface 34.
[0042] Figure 6 It shows that Figure 4b 1. An embodiment of a reporting process 490 is shown. Once a threshold of probability of a current patient condition is exceeded, the described embodiments provide a process for testing the patient to determine if the patient is actually suffering from symptoms of the condition, or if the sensor data is simply causing a "false positive". The reporting process 490 includes the following steps:
[0043] In step 610, when the probability of the patient's current condition has been exceeded, the control device 30 begins the reporting process. In an embodiment of the present invention, during the testing process, one or more user tests will be used to determine the user's physical and mental state. The terms "user" and "patient" can be used interchangeably.
[0044] In one embodiment, the test process includes a countdown alarm 620 presented to the user on the display of the control device 30. The countdown alarm 620 may include a simple countdown alarm showing a short countdown before which the control device 30 will move to step 650 to alert the medical service department. The user has the option to cancel the countdown at any time. If the user fails to respond to the countdown alarm, the control device 30 will move directly to step 650 to alert the medical service department. In the case where the patient cancels the countdown by performing the required cancellation task (e.g., pressing a cancel button), the patient is presented with the option of resetting the reporting process. Alternatively, the user may be presented with the option of moving directly to step 650 to alert the medical service department when the patient feels that there is still a problem. In one embodiment, the user cannot cancel the countdown alarm until all user tests are successfully completed. In one embodiment, a two-part countdown is used. Countdown 1 is a short countdown (e.g., less than 60 seconds) and can be canceled by the user. Countdown 2 is a long countdown (e.g., greater than 60 seconds) that can only be canceled by successfully completing all user tests.
[0045] In one embodiment, the testing process includes a basic user test 630 presented to the user on a display of the control device 30. The basic user test may include one or more physical tests. The physical test may include requesting the patient to move one or more of his or her limbs in a prescribed manner. For example, move one hand or one leg, move the opposite hand or leg sequentially or simultaneously, clap the hands, stand up, jump, raise one or more knees to the patient's chest, etc. The physical test is intended to perform one or more functions. In the event that the motion sensor data indicates that movement of one or more limbs or one side of the body appears to be restricted, the physical test may be intended to provide data that can be used to reduce the likelihood of any form of paralysis that the patient is suffering from. For example, stand up or wave each hand in sequence. Alternatively, in the event that the motion sensor data indicates that there is an overall asymmetry in the patient's movement, the physical test may be intended to provide data that can be used to provide clear data about the symmetry of the patient's current movement. For example, in the event that the patient wears a wearable sensor on each wrist, clapping the hands should provide clearly symmetrical signals from each wearable sensor as long as both of the patient's arms have a full range of motion. In Figure 6In an embodiment, if the user fails the basic user test 630, the control device 30 presents further tests to the patient. In an alternative embodiment, if the patient fails the basic user test 630, the control device 30 will move directly to step 650 to alert the medical service department. In the event that the patient passes the basic user test 630, the patient is presented with the option of resetting the reporting process. Alternatively, the user may be presented with the option of continuing with further tests or moving directly to step 650 to alert the medical service department if the patient still feels that there is a problem. It should be understood that the user tests described in this description may be presented to the user in any combination or order that is considered to produce the most reliable results.
[0046] In one embodiment, the testing process includes an advanced user test 640 presented to the user on the display of the control device 30. The advanced user test may include one or more intelligence tests. The intelligence test includes asking the patient to complete one or more intelligence tests that are intended to prove that the user can think clearly and rationally, such as mental arithmetic tests, verbal reasoning tests, facial or overall image recognition tests, vocalization tests, etc. As with the basic user test 630, if the user fails the advanced user test 640, the control device 30 may present further steps to the patient before moving to step 650. In an alternative embodiment, if the patient fails the advanced user test 640, the control device 30 will move directly to step 650 to alert the medical service department. In the case where the patient passes the advanced user test 640, the patient is presented with the option of resetting the reporting process. Alternatively, the user may be presented with the option of continuing with further tests, or the option of moving directly to step 650 to alert the medical service department if the patient feels that there is still a problem.
[0047] Figure 7 Shown for Figure 6 An alternative user testing sequence to the user testing sequence shown. Figure 7 In the example embodiment, step 720 includes a countdown alarm to provide the patient with an opportunity to cancel the escalation if he or she still believes that there is no medical condition. If no response is received, the alarm will be immediately escalated. In one embodiment, the countdown continues until all subsequent tests are passed. After the countdown alarm, step 730 includes a basic user test, which includes asking the user to clap his or her hands three times. Finally, step 740 includes an advanced user test as described above. Figure 7 In an embodiment, the advanced user test is a mental arithmetic test to assess the patient's intellectual ability.
[0048] In one embodiment, the user test may include only vibration prompts. This may identify situations where the user is unable to feel the vibrations due to a sustained stroke. In one embodiment, a flashing light is used as a control device for the vibration test.
[0049] Figures 8a to 8c An exemplary set of sensor data is shown from the present device in which wearable sensors 20a and 20b include accelerometers that produce motion sensor data. Figure 8a Exemplary motion sensor data generated by a wearable sensor 20a secured to the left limb of a patient's body over a one hour period is shown. Figure 8b The wearable sensor 20b fixed on the right limb of the patient's body is shown in FIG. Figure 8a Example motion sensor data generated during the same time period as the time period of Figure 8c In the Figure 8a Motion sensor data and Figure 8b In the case where the wearable sensors 20a and 20b are symmetrically fixed to the patient's body, the difference between the motion sensor data from the wearable sensor 20a and the wearable sensor 20b, respectively, can show the low-frequency motion asymmetry between the limbs on the left side of the body to which the wearable sensors are respectively fixed and the limbs on the left side of the body. Figure 8a , the asymmetry between the patient's left and right limbs is shown to increase over time.
[0050] Fig. 9 Shows Figure 4b An extended embodiment of the embodiment of . Fig. 9 In the embodiment of the present invention, the data analysis step 475 is divided into a plurality of data analysis steps 475a, 475b, etc., which can be performed continuously or in parallel. In the present embodiment, for each of the plurality of data analysis steps, the data stored on the control device 30 can be analyzed in a different manner.
[0051] Fig.10The results of an embodiment using multiple data analysis steps 475 are shown, and the data analysis step 475 is configured to analyze motion sensor data in different time periods (shown by lines 1010 and 1020). In such an embodiment, analyzing the motion sensor data in a short time period (e.g., line 1010) enables the possibility of a stroke to be discovered quickly and relatively quickly after the stroke occurs. However, in order to minimize false alarms, the analysis in a short time period requires a narrow detection threshold. For example, the analysis in a shorter time period is more likely to have a certain degree of noise that may trigger a false alarm. Therefore, the range of acceptable positive values needs to be narrower, and a high threshold should be met before a stroke detection signal is generated. However, when the device is used in a longer time period, the high threshold may result in a lower detection rate. To compensate for this, one embodiment provides the use of a data analysis step that runs in parallel with the short time analysis, which analyzes the motion sensor data in a longer time period (e.g., line 1010). Analyzing the motion sensor data in a longer time period enables a better detection rate, but inevitably leads to a longer delay before a stroke can be detected. In one example, analyzing motion sensor data over a one hour period can achieve a very high detection rate of almost 100%, but introduces a one hour delay in detection which can make a significant difference to the outlook of a stroke victim. A fused approach (shown as line 1030) can gain the benefits of all approaches.
[0052] In one embodiment, the stroke detection device can operate in multiple modes, changing the fidelity, data rate, resolution or dimensionality of the motion sensor data stored on the device. In an embodiment including two operating modes, when the stroke detection device determines that the probability of a stroke is below a first threshold, the stroke detection device can be configured to operate in a first mode. In the first mode, the motion sensor data is stored with a first fidelity. That is, the sampling rate, sampling size, bit rate, dimensionality or equivalent for storing the motion sensor data is set to a first value. When the stroke detection device determines that the risk of being in a stroke has risen to exceed the first threshold, the stroke detection device can be configured to operate in a second mode. In the second mode, the motion sensor data is stored with a second fidelity higher than the first fidelity. That is, the sampling rate, sampling size, bit rate, dimensionality or equivalent for storing the motion sensor data is set to a second value, which is higher than the first value. This enables more detailed and concise monitoring of the patient's movement during a time period in which the patient's risk of stroke is determined to be high, while enabling improved storage and / or reduced power consumption of the device when the patient's risk of stroke is determined to be low.
[0053] In another embodiment, the stroke detection device can be configured to operate in the first mode and / or the second mode (as described above) and the third mode. When the stroke detection device is monitoring a specific movement of the user, the stroke detection device can be configured to operate in the third mode. For example, in the case where the stroke detection device has prompted the user to perform a specific movement or gesture to pass the test. In the third mode, the motion sensor data is stored with a third fidelity higher than the first fidelity and / or the second fidelity. That is, the sampling rate, sampling size, bit rate, dimension or equivalent for storing the motion sensor data is set to a third value, which is higher than the first value and / or the second value. The third mode enables the stroke detection device to accurately capture the user's movement within a short period of time with a fidelity sufficient to capture a specific gesture. In some embodiments, the first mode and / or the second mode stores the motion sensor data only in a single dimension, while the third mode stores the motion sensor data in multiple dimensions. This can more clearly determine specific patient movements.
[0054] In one embodiment, the above-described device is used in a rehabilitation context rather than in a stroke detection context. By indirectly monitoring the patient's nervous system signals in the same manner as described above for detecting a stroke, the patient's recovery from the stroke can be tracked and rehabilitation guidance can be provided. Stroke rehabilitation strives to enable the patient to recover as much function as possible that was impaired by the stroke and to help the stroke victim gain as much functional independence as possible. Stroke rehabilitation is critical to the ultimate functional outcome after a stroke and the costs associated with long-term disability after a stroke. In one configuration, the device is configured to: 1) monitor and quantify rehabilitation progress over time. 2) provide active support to the stroke victim throughout the rehabilitation process.
[0055] The process of monitoring and quantifying rehabilitation progress can include reporting the correlation between accelerometer data from the wearable sensor (i.e., the patient's motion data) and the physical therapist's evaluation of arm function before and after rehabilitation. The accelerometer data from the wearable sensor, as well as the corresponding related motion data, are presented in the application for use by the patient and the professional rehabilitation team.
[0056] The process of providing active support to a stroke victim throughout the rehabilitation process can include generating exercise instructions for the patient based on accelerometer data from a wearable sensor. In one embodiment, the difficulty, duration, and timing of the exercise instructions depend on the patient's movement data. That is, in the event that the patient's physical movement appears to be limited over a period of time, contrary to rehabilitation recommendations, the system is configured to command the user to perform more regular and intensive exercise exercises.
Claims
1. A stroke detection device, comprising: A data processing device, the data processing device comprising a processor; at least one wearable sensor, the at least one wearable sensor being configured to be worn by a user and configured to transmit the user's motion data to the data processing device, the data processing device being configured to receiving the motion data, processing the motion data to determine a probability that the user has suffered a stroke, wherein the data processing device is configured to store the motion data with fidelity according to the probability of the user having a stroke, The fidelity of the stored motion data includes at least one of a sampling rate, a sampling size, a bit rate, and a dimension of the motion data. Wherein, when it is determined that the probability of the user having a stroke is lower than a first threshold, the stroke detection device is configured to operating in a first mode in which the motion data is stored at a first fidelity and wherein, in an event that the probability of the user having a stroke is determined to be above the first threshold, The stroke detection device is configured to Operating in a second mode, in which the motion data is stored at a second fidelity, wherein the second fidelity is higher than the first fidelity.
2. The stroke detection device according to claim 1, wherein: The processing device includes a user interface and further includes generating one or more user tests at the user interface, the one or more user tests configured to determine a third probability that the user has had a stroke.
3. The stroke detection device according to claim 2, wherein: The user test comprises at least one of: a reaction time test, an arithmetic test, a reading and / or comprehension test, a memory test, and a user response test.
4. The stroke detection device according to claim 3, wherein: The stroke detection device is configured to generate one or more user tests when a probability that the current user has had a stroke has been exceeded.
5. The stroke detection device according to claim 4, wherein: The stroke detection device is configured to generate a stroke reporting signal if the user fails the one or more user tests.
6. The stroke detection device according to claim 5, wherein: An automatic emergency service request is generated according to the stroke reporting signal.
7. The stroke detection device according to claim 1, in, The stroke detection device is configured to operate in the first mode in which the processed motion data is restricted to a single dimension, and Wherein, the stroke detection device is configured to operate in the second mode, in which the processed motion data is multi-dimensional.
8. The stroke detection device according to claim 1, wherein: The stroke detection device is configured to operate in a third mode when the stroke detection device is monitoring a specific movement of a user.
9. The stroke detection device according to claim 8, wherein: The stroke detection device is configured to operate in the third mode if the stroke detection device has prompted a user to perform the specific movement to pass the test.
10. The stroke detection device according to claim 8 or 9, wherein: In a third mode, the motion sensor data is stored at a third fidelity, wherein the third fidelity is higher than the first fidelity and the second fidelity.
11. A method for detecting a stroke using the stroke detection device according to any one of claims 1 to 10.
12. A method for generating a stroke detection signal, the method comprising: generating motion data of a user using at least one wearable sensor device, the at least one wearable sensor device being configured to be worn by the user, transmitting the motion data to a processing device, receiving the motion data at the processing device, Processing the motion data at a processing device to determine a probability of a user having a stroke, the method further comprising: storing the motion data with fidelity according to the user's stroke probability, wherein the fidelity of the stored motion data includes at least one of a sampling rate, a sampling size, a bit rate, and a dimension of the motion data, storing the motion data at a first fidelity when the probability of the user having a stroke is determined to be below a first threshold, storing the motion data at a second fidelity when the probability of the user having a stroke is determined to be above the first threshold, The second fidelity is higher than the first fidelity.
Citation Information
Patent Citations
Systems and methods of limb-based accelerometer assessments of neurological disorders
US20150157252A1
Systems and methods of tracking patient movement
US20170281054A1
Multi-sensor stroke detection
US20180153477A1
Intelligent wearable monitor systems and methods
US7981058B2
A device and method for detecting an epileptic event
CN101583311A
