Non-contact mouse sleep monitoring, analysis methods and systems

Abstract

The application discloses a kind of non-contact mouse sleep monitoring, analysis method and system thereof, wherein the method comprises: real-time recording piezoelectric signal and infrared camera video;Through piezoelectric signal online analysis;Through the piezoelectric signal and infrared camera video are combined and are analyzed offline, judge mouse state.The application simultaneously uses infrared camera and piezoelectric signal to monitor mouse state, can realize the non-contact measurement of mouse physiological signal, combines online analysis and offline analysis, respectively, to infrared camera video and piezoelectric signal are analyzed and handled to obtain the sleep-wake state of mouse current brain, so that user can preview analysis result in recording process and have the potential of developing closed-loop control, and correct online analysis result after recording ends, finally, the accuracy of analysis result is judged by checking analysis, ensure that more accurate judgment result is further obtained after recording ends.

Description

Technical Field

[0001] This invention relates to the field of sleep monitoring, and more particularly to a non-contact method and system for monitoring and analyzing mouse sleep. Background Technology

[0002] In recent years, with the increasing prevalence of insomnia among younger people, sleep problems have become a global health concern, and sleep-related research has become a hot topic in life sciences, medicine, and other research fields. The sleep function and regulatory mechanisms of mice are extremely similar to those of humans; therefore, they are widely used in sleep research. Accurate and detailed analysis of the state of mice during the sleep-wake cycle is crucial for related research.

[0003] Currently, the commonly used analytical method is electrophysiological analysis combining electroencephalography (EEG) and electromyography (EMG). This involves collecting electrophysiological signals from EEG and EMG, analyzing their power spectrum and amplitude, and then determining the sleep-wake state. This type of technology requires implanting devices, such as cranial nails or electrode wires, into the skull and muscles of mice for physiological signal acquisition and electrical stimulation. For example, invention patent CN117158982A discloses a dual-mode detection and regulation micro-nano electrode array for sleep-wake brain regions, a detection system, and a fabrication method. The implantable micro-nano electrode array consists of a silicon substrate layer, a metal conductive layer, and a silicon nitride / silicon oxide insulating layer. This method requires implanting multiple devices in the brain and abdomen of experimental animals to achieve high-precision localization and implantation of multiple brain regions. However, implanting devices into laboratory animals can easily cause surgical complications such as inflammation, affecting signal acquisition and the survival time of the animals after surgery. This is especially true when using small mice as laboratory animals, as the size and weight of the implanted device have a significant impact on the spontaneous behavior of the mice. On the other hand, the data from such devices is usually transmitted via wires, requiring a transmission line to be drawn from the electrodes fixed in the mouse's head to the main body of the device, which greatly restricts the mouse's movement. As we all know, sleep is a behavior that is extremely sensitive to both internal and external environmental disturbances. Developing a non-invasive sleep monitoring system that supports the free movement of mice is of great significance for the study of sleep behavior.

[0004] Existing non-contact mouse sleep monitoring devices and methods, such as the invention patent with publication number CN116982571A, provide a non-contact mouse movement and sleep monitoring device and method, including a rectangular cage, an infrared matrix, and a base. The infrared matrix is ​​rectangular and mounted on top of the base, and the rectangular cage is mounted on top of the infrared matrix. The monitoring method involves placing a mouse within the device and positioning it inside the infrared matrix; using the center of all infrared grid points obscured by the mouse as a marker point; and determining that the mouse is asleep when the marker point remains stationary within a preset time. However, this monitoring method only determines the mouse's activity level and thus its sleep state through the infrared matrix. On the one hand, this monitoring method is highly prone to errors during experiments and does not provide other monitoring devices to verify the monitoring results, resulting in extremely low accuracy. On the other hand, this monitoring method can only determine whether the mouse is at rest based on the amplitude of its activity, and cannot further determine the specific sleep state. Furthermore, it does not provide detailed analysis methods for the collected monitoring data. Summary of the Invention

[0005] To address the above problems, this invention provides a non-contact method and system for monitoring and analyzing mouse sleep.

[0006] This invention is achieved through the following technical solution:

[0007] A non-contact method for monitoring and analyzing mouse sleep includes the following steps:

[0008] S1: Real-time monitoring of mice, recording piezoelectric signals and infrared video;

[0009] S2: Online analysis of piezoelectric signals to determine the mouse's condition;

[0010] S21: Set a minimum resolution time t1 for each interval to make a judgment, and take the time t2 before each judgment as the sampling period, and preprocess the piezoelectric signal data within the sampling period.

[0011] S22: Perform threshold analysis on the amplitude of the preprocessed data to determine whether the mouse is in a resting state. If yes, proceed to step S23; otherwise, determine that the mouse is in a non-resting state.

[0012] S23: Further analyze the power spectrum of the autocorrelation curve of the piezoelectric signal data and determine whether the mouse is in REM sleep or NREM state;

[0013] S3: Offline analysis using piezoelectric signals and infrared video recordings to determine the mouse's condition;

[0014] S31: Set a minimum resolution time t1 for each interval to make a judgment, and take the time t2 before each judgment as the sampling period, and align the time axis of the piezoelectric signal and the infrared camera video within the sampling period;

[0015] S32: Traverse the infrared photography frames within the sampling period and calculate the frame difference. Set a third threshold and determine whether the frame difference is greater than the third threshold. If yes, determine that the mouse is in a non-resting state; otherwise, determine that the mouse is in a resting state and proceed to step S33.

[0016] S33: Preprocess the piezoelectric signal data during the sampling period;

[0017] S34: Perform threshold analysis on the amplitude of the preprocessed piezoelectric signal data, and determine again whether the mouse is in a resting state. If yes, proceed to step S35; otherwise, determine that the mouse is in a non-resting state.

[0018] S35: Further analyze the power spectrum of the autocorrelation curve of the piezoelectric signal data and determine whether the mouse is in REM sleep or NREM state;

[0019] S36: After completing the judgment of the entire time series, traverse the analysis results and take the state with the highest frequency in the current 1 second and the time period before and after t4 seconds as the current state of the mouse. If the time before or after is less than t4 seconds, take the data start or end point.

[0020] Preferably, in steps S22 and S34, a first threshold is set, and threshold analysis is performed on the preprocessed data using the first threshold. If the amplitude of the piezoelectric signal data is greater than the first threshold, it is determined that the mouse is in a non-resting state; otherwise, it is determined that the mouse is in a resting state.

[0021] Preferably, in steps S23 and S35, a second threshold is set. If the frequency component with the highest power in the power spectrum of the autocorrelation curve of the piezoelectric signal data is less than the second threshold, it can be determined that the mouse is in NREM sleep state; otherwise, it can be determined that the mouse is in REM sleep state.

[0022] Preferably, step S31 includes:

[0023] The initial recording time of the piezoelectric signal and the recording time of each piezoelectric signal are obtained through the piezoelectric signal.

[0024] The initial shooting time of the infrared camera and the shooting time of each frame are obtained through infrared camera video.

[0025] By comparing the initial recording time of the piezoelectric signal and the initial shooting time of the infrared camera, the portion of either signal that precedes the other is removed, thus aligning the start times of the two signals.

[0026] The number of corresponding piezoelectric signals recorded during the capture period of each frame image is obtained, and the corresponding piezoelectric signal is downsampled to make the sampling number of the two signals uniform.

[0027] By removing the portion of either signal that lags behind the other, the end times of the two signals are aligned, thereby aligning the timelines of the piezoelectric signal and the infrared camera video.

[0028] Preferably, step S2 further includes:

[0029] S24: Set an interval t3 as an inspection period, and determine whether the frequency of the transition between the resting state and the non-resting state of the mouse in each inspection period is greater than the fourth threshold. If so, the minimum resolution time t1 needs to be increased and step S2 needs to be repeated. Otherwise, step S2 does not need to be repeated.

[0030] Preferably, step S3 further includes:

[0031] S37: Set an interval of time t3 as an inspection period, and determine whether the frame difference of the image frames in each inspection period exceeds the fifth threshold. If so, the minimum resolution time t1 needs to be increased and steps S2 and S3 need to be repeated. Otherwise, steps S2 and S3 do not need to be repeated.

[0032] A non-contact mouse sleep monitoring and analysis system, including:

[0033] The data acquisition unit includes an infrared camera, a monitoring box, a piezoelectric signal acquisition device, and an Arduino board. The infrared camera includes an infrared camera for capturing infrared video. The monitoring box includes a monitoring body for providing a space for monitoring the mouse's condition. The piezoelectric signal acquisition device includes a piezoelectric sensor for acquiring piezoelectric signals and a preprocessing mechanism for preprocessing the piezoelectric signals. The Arduino board is signal-connected to the preprocessing mechanism to receive the preprocessed piezoelectric signals and send them to the signal processing module at an adjustable frequency.

[0034] Signal processing unit: includes an online signal processing computer and an offline signal processing computer; the online signal processing computer is connected to the Arduino board signal and performs online analysis using piezoelectric signals to determine the mouse's state; the offline signal processing computer is simultaneously connected to an infrared camera and the Arduino board signal and performs offline analysis using piezoelectric signals and infrared camera video to determine the mouse's state.

[0035] Storage unit: Stores the non-contact mouse sleep monitoring and analysis method described above.

[0036] Preferably, the monitoring body includes an outer casing and a piezoelectric sensor disposed inside the outer casing. The piezoelectric sensor is a PVDF piezoelectric sensor, and the thin-film operating mode of the PVDF piezoelectric sensor is d31.

[0037] Preferably, the preprocessing mechanism includes a low-pass filter module, a signal amplification module, an amplitude modulation module, and a voltage follower module. The low-pass filter module, the signal amplification module, and the voltage follower module operate based on a CA3140E operational amplifier, and the amplitude modulation module operates based on an NE5532P operational amplifier. The effectiveness of the module has been verified through simulation and debugging. The modules within the preprocessing mechanism are connected by wires.

[0038] Preferably, the signal processing unit performs real-time data reading and processing based on the Matlab program and provides a judgment on the sleep-wake state of the mouse. The reading frequency of the piezoelectric signal is 30Hz, the effective data frequency range is 0.5Hz-10Hz, and the infrared imaging acquisition frequency is 3Hz.

[0039] The beneficial effects of the technical solution of this invention are mainly reflected in:

[0040] 1. This system simultaneously uses infrared imaging and piezoelectric signals to monitor the state of mice, enabling non-contact measurement of physiological signals and avoiding surgical complications, thus ensuring the survival time of the mice. Furthermore, the infrared imaging and piezoelectric signals are transmitted wirelessly to collect mouse state data, which does not restrict the spontaneous behavior of the mice and reduces interference from the internal and external environment. It allows for real-time assessment of the sleep-wake state of mice without invasive procedures, while supporting their free movement, thereby ensuring the accuracy of the monitoring results.

[0041] 2. In the online analysis, a preliminary judgment is made using piezoelectric signals. Subsequently, offline analysis is used in conjunction with infrared camera video to further analyze and process the piezoelectric signals to obtain the current sleep-wake state of the mouse's brain. The results are then analyzed and statistically analyzed, allowing users to preview the analysis results during the recording process and explore the potential for developing closed-loop regulation. Finally, the accuracy of the analysis results is checked and judged, and the online analysis results are corrected after the recording ends to ensure that more accurate judgment results are obtained after the recording is completed.

[0042] 3. Through real-time monitoring and analysis, users can obtain real-time mouse sleep-wake state analysis results online. Because it provides more accurate data monitoring of mouse status (including judging and analyzing whether the mouse is in a resting state, NREM sleep state, and REM sleep state), after recording, users can selectively record data and perform statistical analysis, including total wakefulness time, total sleep time, total NREM sleep time, total REM sleep time, distribution time of each state, duration of each sleep, and other information, providing data support for the diversity and accuracy of monitoring results.

[0043] 4. The hardware materials for the non-contact mouse sleep monitoring and analysis system are readily available, inexpensive, and simple to manufacture and operate, requiring no specialized equipment. The PVDF piezoelectric sensor measures pressure and vibration; its piezoelectric sensing characteristics are determined by its structure, forming a unified whole that ensures continuous and stable data measurement. The piezoelectric sensing material exhibits extremely high sensitivity when stress is applied perpendicular to the material's molecular polarization direction. Furthermore, the piezoelectric sensing material is lightweight, flexible, foldable, waterproof, and portable. The preprocessing mechanism employs a suspension design along the long axis, incorporating low-pass filtering, amplification, amplitude modulation, and voltage follower circuits to ensure accurate data acquisition. The Arduino board used in this system has excellent parallel data transmission and reception capabilities, and the acquired piezoelectric signal frequency is relatively low. Simultaneously, the system supports parallel sleep monitoring of multiple cages of mice, increasing experimental throughput and shortening experimental time. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the monitoring box in this invention from a first-view perspective;

[0045] Figure 2 This is a schematic diagram of the monitoring box in this invention from a second perspective;

[0046] Figure 3 This is a schematic diagram of a non-contact mouse sleep monitoring and analysis system;

[0047] Figure 4 This is a structural block diagram of the pretreatment mechanism in a non-contact mouse sleep monitoring and analysis system;

[0048] Figure 5 This is a process flow diagram of the online analysis steps in a non-contact mouse sleep monitoring and analysis method;

[0049] Figure 6 This is a flowchart of the offline analysis step in a preferred embodiment of a non-contact mouse sleep monitoring and analysis method. Detailed Implementation

[0050] To make the objectives, advantages, and features of the present invention clearer and more detailed, the following non-limiting description of preferred embodiments will be illustrated and explained. These embodiments are merely typical examples of applying the technical solutions of the present invention; all technical solutions formed by equivalent substitutions or equivalent transformations fall within the scope of protection claimed by the present invention.

[0051] This invention discloses a non-contact method for monitoring and analyzing mouse sleep, comprising the following steps.

[0052] S1: Real-time monitoring of mice, recording piezoelectric signals and infrared video;

[0053] In one embodiment, a transparent monitoring box provides mice with a completely transparent activity space and sufficient water and food. An infrared camera takes unobstructed infrared photographs of the transparent monitoring box. The monitoring box includes a monitoring body 1, which includes a PVDF piezoelectric sensor and an outer casing. The PVDF piezoelectric sensor is used to collect and record piezoelectric signals.

[0054] like Figure 5 As shown, step S2 is the online analysis step, which uses piezoelectric signals to perform online analysis to determine the mouse's condition. Specifically, it includes the following steps.

[0055] S21: Set a minimum resolution time t1 for each interval to make a judgment, and take the time t2 before each judgment as the sampling period, and preprocess the piezoelectric signal data within the sampling period.

[0056] In a preferred embodiment, the minimum resolution time t1 is 1 second and the t2 is 3 seconds. Therefore, in this embodiment, the mouse is judged to be asleep or awake every 1 second. The current second and the previous 2 seconds (3 seconds in total) are selected as the sampling period. The piezoelectric signal data within the sampling period is preprocessed. The preprocessing of the piezoelectric signal data includes correcting baseline drift and further filtering.

[0057] S22: Perform threshold analysis on the amplitude of the preprocessed data to determine whether the mouse is in a resting state. If yes, proceed to step S23; otherwise, determine that the mouse is in a non-resting state.

[0058] In some embodiments, in step S22, a first threshold is set, and threshold analysis is performed on the preprocessed data using the first threshold. If the amplitude of the piezoelectric signal data is greater than the first threshold, it is determined that the mouse is in a non-resting state; otherwise, it is determined that the mouse is in a resting state.

[0059] S23: Further analyze the power spectrum of the autocorrelation curve of the piezoelectric signal data and determine whether the mouse is in REM sleep or NREM state;

[0060] In some embodiments, in step S23, a second threshold is set. If the frequency component with the highest power in the power spectrum of the autocorrelation curve of the piezoelectric signal data is less than the second threshold, it can be determined that the mouse is in NREM (non-rapid eye movement sleep) sleep state; otherwise, it can be determined that the mouse is in REM (rapid eye movement sleep) sleep state. This allows for the determination of the stage of the mouse's sleep state, thereby improving the precision of mouse sleep state monitoring and analysis.

[0061] In other embodiments, the judgment can also be made based on other feature information of the autocorrelation curve and its power spectrum curve. Explicit features can be used, or implicit features can be extracted using machine learning and deep neural networks. In a preferred embodiment, the value of the second threshold can be optimized in offline analysis. The mouse's motion-resting state is further determined based on the video frame difference, and this state value is used as a label. Supervised learning training is performed on the data to obtain the optimal second threshold. The optimization method of the second threshold is not the focus of this invention and will not be described in detail here.

[0062] In some embodiments, step S2 further includes S24: setting an interval t3 as an examination period, determining whether the frequency of transitions between the resting and non-resting states of the mouse within each examination period is greater than a fourth threshold. If so, the accuracy of the mouse state judgment within the current examination period is determined to be low, and the minimum resolution time t1 needs to be increased before repeating step S2. Otherwise, the accuracy of the mouse state judgment within the current examination period is determined to be high. In a preferred embodiment, t3 is set to 3 seconds, and the fourth threshold is set to 4. That is, an examination period is set every 3 seconds, and the frequency of transitions between the resting and non-resting states of the mouse within each examination period is determined to be less than 4 times. If the frequency of transitions does not exceed 4 times, the accuracy of the mouse state judgment within the current examination period is determined to be high, and no further online analysis is required. If the frequency of transitions exceeds 4 times, it indicates that the mouse's sleep-wake state transitions are more frequent within the examination period, and the minimum resolution time t1 needs to be increased before repeating step S2 to ensure the accuracy and reliability of the analysis results.

[0063] like Figure 6 As shown, step S3 is the offline analysis step, in which:

[0064] S3: Offline analysis using piezoelectric signals and infrared video cameras is used to determine the mouse's condition.

[0065] S31: Set a minimum resolution time t1 for each interval to make a judgment, and take the time t2 before each judgment as the sampling period, and align the time axis of the piezoelectric signal and the infrared camera video within the sampling period;

[0066] In one embodiment, the piezoelectric signal should at least record the initial recording time of the piezoelectric signal and the recording time of each piezoelectric signal, and the infrared video should at least record the initial shooting time of the infrared camera and the shooting time of each frame. Step S31 specifically includes the following steps:

[0067] The initial recording time of the piezoelectric signal and the recording time of each piezoelectric signal are obtained through the piezoelectric signal.

[0068] The initial shooting time of the infrared camera and the shooting time of each frame are obtained through infrared camera video.

[0069] By comparing the initial recording time of the piezoelectric signal and the initial shooting time of the infrared camera, the portion of either signal that precedes the other is removed, thus aligning the start times of the two signals.

[0070] The number of corresponding piezoelectric signals recorded during the capture period of each frame image is obtained, and the corresponding piezoelectric signal is downsampled to make the sampling number of the two signals uniform.

[0071] By removing the portion of either signal that lags behind the other, the end times of the two signals are aligned, thereby aligning the timelines of the piezoelectric signal and the infrared camera video.

[0072] The phrase "downsampling the corresponding piezoelectric signal to make the number of samples of the two signals uniform" refers to downsampling the piezoelectric signal within the sampling interval of each frame of the infrared video signal based on the sampling interval of each frame of the video signal, and ensuring that the same number of piezoelectric signals are sampled uniformly within the sampling interval of each frame of the video signal, so that the piezoelectric signal sampling and the video signal sampling have the same distribution.

[0073] Since the video recording frame rate strictly follows the set frame rate, but there is a certain error in the time consumed by each frame during actual shooting, if the theoretical time is selected during data processing, the processing result can correspond to the video recording. If the actual time is selected during data processing, the processing result will match the actual time consumed. According to calculations, the error between the actual time consumed and the theoretical time consumed is within 5%, and the choice can be made according to actual needs.

[0074] like Figure 6 As shown, in a preferred embodiment, the minimum resolution time t1 is 1 second and t2 is 3 seconds, the same as step S21 above. Therefore, in this embodiment, the mouse is determined to be sleep-wake once every 1 second, and the current second and the previous 2 seconds (3 seconds in total) are selected as the sampling period. The piezoelectric signal and infrared video are acquired synchronously during the sampling period. Then, the time axis of the piezoelectric signal and the infrared video are aligned.

[0075] S32: Traverse the infrared imaging frames within the sampling period and calculate the frame difference. Set a third threshold and determine whether the frame difference is greater than the third threshold. If yes, determine that the mouse is in a non-resting state; otherwise, determine that the mouse is in a resting state and proceed to step S33. Specifically, when the frame difference within the sampling period is greater than the third threshold, determine that the mouse is in a non-resting state; otherwise, determine that it is in a resting state. In the resting state, it is necessary to proceed to step S33 to further analyze the piezoelectric signal of that time period.

[0076] S33: Preprocess the piezoelectric signal data during the sampling period; the preprocessing includes correcting baseline drift and further filtering.

[0077] S34: Perform threshold analysis on the amplitude of the preprocessed piezoelectric signal data, and determine again whether the mouse is in a resting state. If yes, proceed to step S35; otherwise, determine that the mouse is in a non-resting state.

[0078] In one embodiment, in step S34, a first threshold is set, and threshold analysis is performed on the preprocessed data using the first threshold. If the amplitude of the piezoelectric signal data is greater than the first threshold, the recorded piezoelectric signal amplitude is large, and the mouse is determined to be in a non-resting state; otherwise, the mouse is determined to be in a resting state. This step can eliminate interference from non-resting states with small activity amplitudes.

[0079] S35: Further analyze the power spectrum of the autocorrelation curve of the piezoelectric signal data and determine whether the mouse is in REM sleep or NREM state.

[0080] In one embodiment, in step S35, a second threshold is set. If the frequency component with the highest power in the power spectrum of the autocorrelation curve of the piezoelectric signal data is less than the second threshold, it can be determined that the mouse is in NREM sleep state; otherwise, it can be determined that the mouse is in REM sleep state.

[0081] In other embodiments, the judgment can also be made based on other feature information of the autocorrelation curve and its power spectrum curve. Explicit features can be used, or implicit features can be extracted using machine learning and deep neural networks. In a preferred embodiment, the motion-resting state of the mouse is determined based on the video frame difference. This state value is used as a label, and supervised learning training is performed on the data to obtain the optimal threshold. It is worth noting that the second threshold is susceptible to environmental factors and electromagnetic noise interference. Therefore, it needs to be optimized in each experiment. The optimization method of the second threshold is not the focus of this invention and will not be described in detail here.

[0082] S36: After completing the judgment of the entire time series, iterate through the analysis results and take the state with the highest frequency in the time period of the current 1 second and the time period t4 seconds before and after as the current state of the mouse. If the time before or after is less than t4 seconds, take the data up to the beginning or end point. Specifically, after completing the sleep-wake judgment of the entire time series, the judgment result is corrected through step S36. By further analyzing the sampled data in a relatively wide time period before and after the current time period, the accuracy of the current result is ensured, and errors in the analysis result due to local data deviation are avoided. Figure 6 As shown, in a preferred embodiment, t4 is 15 seconds.

[0083] In some embodiments, step S3 further includes step S37:

[0084] S37: Set each time interval t3 as an inspection period. Determine whether the frame difference of the image frames within each inspection period exceeds the fifth threshold. If the frame difference of the image frames within the current inspection period is greater than the fifth threshold, then the accuracy of the mouse state judgment within the current inspection period is determined to be low, and the minimum resolution time t1 needs to be increased before repeating steps S2 and S3. Otherwise, the accuracy of the mouse state judgment within the current inspection period is determined to be high, and steps S2 and S3 do not need to be repeated. Figure 6 As shown, in a preferred embodiment, t3 is 3 seconds.

[0085] This invention also discloses a non-contact mouse sleep monitoring and analysis system, comprising:

[0086] The data acquisition unit includes an infrared camera, a monitoring box, a piezoelectric signal acquisition device, and an Arduino board. The infrared camera includes an infrared camera for capturing infrared video. The monitoring box includes a monitoring body 1 for providing space for monitoring the mouse's condition. The monitoring box includes a fully transparent acrylic box body 6 and a lid 7, with the monitoring body 1 located inside the box body 6 and lid 7. The piezoelectric signal acquisition device includes a piezoelectric sensor for acquiring piezoelectric signals and a preprocessing mechanism for preprocessing the piezoelectric signals. The Arduino board is signal-connected to the preprocessing mechanism to receive the preprocessed piezoelectric signals and transmit them to the signal processing module at an adjustable frequency.

[0087] Signal processing unit: includes an online signal processing computer and an offline signal processing computer; the online signal processing computer is connected to the Arduino board signal and performs online analysis using piezoelectric signals to determine the mouse's state; the offline signal processing computer is simultaneously connected to an infrared camera and the Arduino board signal and performs offline analysis using piezoelectric signals and infrared camera video to determine the mouse's state.

[0088] Storage unit: Stores the non-contact mouse sleep monitoring and analysis method described above.

[0089] In one embodiment, the monitoring box includes a fully transparent acrylic box body 6, a box cover 7, and a monitoring body 1. The monitoring body 1 includes an outer sleeve and a piezoelectric sensor disposed inside the outer sleeve. The thickness of the outer sleeve is preferably 60 μm, and the thickness of the piezoelectric sensor is preferably 28 μm. The piezoelectric sensor is a PVDF piezoelectric sensor, and the thin-film working mode of the PVDF piezoelectric sensor is d31. The effective sensitivity of this working mode is 1000 times higher than that of the traditional working mode d33, thereby amplifying the piezoelectric signal of mouse activity and improving the accuracy and sensitivity of signal detection.

[0090] like Figure 1 , Figure 2 As shown, the fully transparent acrylic box 6 of the monitoring box has multiple cutouts 12 on its left and right sides, 5cm above the ground, for suspending the monitoring body 1. In one embodiment, there are eight cutouts 12. The left side is connected to a triangular food trough 8 and a water trough 9, and cutouts are made at the connection points so that mice can eat and drink. Bedding can be laid on the monitoring body 1 to provide a suitable living environment for the mice and to a certain extent isolate the mice's excrement from the monitoring body 1. In addition, the bottom surface is designed to be retractable to facilitate the cleaning of fallen wood planing bedding and other garbage. Most of the garbage is still poured out from the top. The front of the box 6 is provided with a hinge 11 5cm above the ground, which can be opened inward to facilitate the removal and removal of the bottom panel and prevent the mice from escaping. In addition, to facilitate use with in vivo electrophysiology and fiber optic photometry equipment, a rectangular cutout 10 of 5cm x 5cm is provided in the cover 7. In a preferred embodiment, the monitoring body 1 is suspended along the long axis, which can convert the stress generated by the mouse activity that is consistent with the polarization direction of the PVDF piezoelectric sensor film into stress along the long axis of the PVDF piezoelectric sensor film.

[0091] like Figure 4 As shown, in one embodiment, the preprocessing mechanism includes a low-pass filter module 2, a signal amplification module 3, an amplitude modulation module 4, and a voltage follower module 5. The low-pass filter module 2, the signal amplification module 3, and the voltage follower module 5 operate based on a CA3140E operational amplifier, and the amplitude modulation module 4 operates based on an NE5532P operational amplifier. The effectiveness of these modules has been verified through simulation and debugging. The modules within the preprocessing mechanism are connected by wires. Furthermore, by measuring the pin voltage, the 0V-5V voltage is mapped to ports 1-1024, facilitating subsequent data acquisition using an Arduino board. Finally, the voltage follower module 5 amplifies the current and inputs it to the Arduino board to communicate with the computer, achieving digital communication. The Arduino board is a Mega 2560, which has good signal transmission and reception capabilities and can support 100Hz data communication.

[0092] In one embodiment, the infrared camera has a frame rate of 720P, a focal length of 2.9mm, a field of view of 90 degrees, and captures images without distortion. It can support 30Hz shooting and data transmission. The infrared camera is equipped with a photosensitive element. When there is sufficient light, the infrared camera captures visible light to form an RGB image without color distortion. When the light is weak, it will turn on the infrared supplementary light and capture infrared light for imaging, which is convenient for recording and shooting in dark environments. Since the mouse eye cannot observe infrared light, infrared light will not affect its physiological activities.

[0093] In one embodiment, the signal processing unit performs real-time data reading and processing based on a Matlab program and provides a judgment on the sleep-wake state of the mouse. The reading frequency of the piezoelectric signal is 30Hz, the effective data frequency range is 0.5Hz-10Hz, the infrared imaging acquisition frequency is 3Hz, and the data acquired during the program operation will be stored in a user-specified location for easy experimental recording and offline re-analysis.

[0094] This invention has many other embodiments, and all technical solutions formed by equivalent transformation or equivalent transformation fall within the protection scope of this invention.

Claims

1. A non-invasive method for monitoring and analyzing sleep in mice, characterized in that: The method comprises the following steps: S1: Real-time monitoring of mice, and recording piezoelectric signals and infrared camera video; S2: Online analysis of piezoelectric signals to determine the state of the mice; S21: Setting a minimum resolution time t1 for each interval to determine, and setting a sampling period of t2 from each determination time to pre-process the piezoelectric signal data in the sampling period; S22: Threshold analysis of the amplitude of the pre-processed data to determine whether the mice are in a resting state, if yes, step S23 is entered, if not, it is determined that the mice are in a non-resting state; S23: Further analysis of the power spectrum of the autocorrelation curve of the piezoelectric signal data to determine whether the mice are in a REM sleep state or a NREM state; S3: Offline analysis of piezoelectric signals and infrared camera video to determine the state of the mice; S31: Setting a minimum resolution time t1 for each interval to determine, and setting a sampling period of t2 from each determination time to align the time axis of the piezoelectric signal and the infrared camera video in the sampling period; S32: Traversing the infrared camera frames in the sampling period, calculating the frame difference, setting a third threshold, and determining whether the mice are in a non-resting state by judging whether the frame difference is greater than the third threshold, if yes, it is determined that the mice are in a non-resting state, if not, it is determined that the mice are in a resting state, and step S33 is entered; S33: Pre-processing the piezoelectric signal data in the sampling period; S34: Threshold analysis of the amplitude of the pre-processed piezoelectric signal data to determine whether the mice are in a resting state, if yes, step S35 is entered, if not, it is determined that the mice are in a non-resting state; S35: Further analysis of the power spectrum of the autocorrelation curve of the piezoelectric signal data to determine whether the mice are in a REM sleep state or a NREM state; S36: After completing the entire timing determination, traversing the analysis results, taking the state with the highest frequency in the current 1 second and the time period of t4 seconds before and after it as the current state of the mice, and taking the data starting point or ending point if the time before or after it is less than t4 seconds.

2. The non-contact mouse sleep monitoring, analyzing method according to claim 1, characterized in that: In steps S22 and S34, a first threshold is set, and threshold analysis of the pre-processed data is performed by the first threshold, if the amplitude of the piezoelectric signal data is greater than the first threshold, it is determined that the mice are in a non-resting state, otherwise it is determined that the mice are in a resting state.

3. The non-contact mouse sleep monitoring, analyzing method according to claim 1, characterized in that: In steps S23 and S35, a second threshold is set, if the power of the frequency component with the maximum power in the power spectrum of the autocorrelation curve of the piezoelectric signal data is less than the second threshold, it is determined that the mice are in a NREM sleep state, otherwise it is determined that the mice are in a REM sleep state.

4. The non-contact mouse sleep monitoring, analyzing method according to claim 1, characterized in that: The step S31 comprises: Obtaining the initial recording time of the piezoelectric signal and the time consumption of recording each piezoelectric signal through the piezoelectric signal; Obtaining the initial shooting time of the infrared camera and the time consumption of shooting each frame through the infrared camera video; Comparing the initial recording time of the piezoelectric signal and the initial shooting time of the infrared camera, and cutting off the part of any signal that is ahead of the other signal to align the start time of the two signals; The number of corresponding piezoelectric signals recorded in each image shooting period is obtained, and the corresponding piezoelectric signals are down-sampled so that the sampling numbers of the two signals are uniform; The delayed part of any signal relative to the other signal is cut off, so that the ending time of the two signals is aligned, thereby aligning the time axis of the piezoelectric signal and the infrared camera video.

5. The non-contact mouse sleep monitoring, analyzing method according to claim 1, characterized in that: The step S2 further comprises: S24: Set every time t3 as a check period, judge whether the transition frequency of the resting state and the non-resting state of the mouse in each check period is greater than the fourth threshold value; if yes, the minimum resolution time t1 needs to be increased and the step S2 is recycled again, otherwise the step S2 does not need to be recycled.

6. The non-contact mouse sleep monitoring, analyzing method according to claim 5, characterized in that: The step S3 further comprises: S37: Set every time t3 as a check period, judge whether the frame difference of the image frame in each check period exceeds the fifth threshold value, if yes, the minimum resolution time t1 needs to be increased and the steps S2, S3 are recycled again, otherwise the steps S2, S3 do not need to be recycled.

7. A non-invasive mouse sleep monitoring, analyzing system, characterized in that: Comprise: The data acquisition unit comprises an infrared camera device, a monitoring box, a piezoelectric signal acquisition device and an arduino board; the infrared camera device comprises an infrared camera for shooting infrared camera video; the monitoring box comprises a monitoring body for providing a mouse state monitoring space; the piezoelectric signal acquisition device comprises a piezoelectric sensor for acquiring piezoelectric signals and a preprocessing mechanism for preprocessing the piezoelectric signals; the arduino board is signal connected with the preprocessing mechanism, for receiving the preprocessed piezoelectric signals and sending to the signal processing module at an adjustable frequency; The signal processing unit comprises an online signal processing computer and an offline signal processing computer; the online signal processing computer is signal connected with the arduino board, and analyzes online through the piezoelectric signals to judge the mouse state; the offline signal processing computer is signal connected with the infrared camera device and the arduino board at the same time, and analyzes offline through the piezoelectric signals and the infrared camera video to judge the mouse state; The storage unit stores the non-contact mouse sleep monitoring and analysis method as claimed in any one of claims 1-6.

8. The non-contact mouse sleep monitoring, analyzing system of claim 7, wherein: The monitoring body comprises an outer envelope and a piezoelectric sensor arranged inside the outer envelope, the piezoelectric sensor is a PVDF piezoelectric sensor, and the film working mode of the PVDF piezoelectric sensor is d31.

9. The non-contact mouse sleep monitoring, analyzing system of claim 7, wherein: The preprocessing mechanism comprises a low-pass filter module, a signal amplification module, an amplitude modulation module and a voltage follower module, the low-pass filter module, the signal amplification module and the voltage follower module work based on the CA3140E operational amplifier, the amplitude modulation module works based on the NE5532P operational amplifier, and the effectiveness of the simulation debugging is verified, and the modules in the preprocessing mechanism are connected through wires.

10. The non-contact mouse sleep monitoring, analyzing system of claim 7, wherein: The signal processing unit performs real-time data reading and processing based on the Matlab program, and gives the sleep-wake state judgment of the mouse, the reading frequency of the piezoelectric signal is 30Hz, the effective data frequency interval is 0.5Hz-10Hz, and the infrared photography image acquisition frequency is 3Hz.

Images