TW201641075A - EOG-based sleep staging method, computer program product with stored programs, computer readable medium with stored programs, and electronic apparatuses - Google Patents

Abstract

This invention mainly discloses an EOG-based sleep staging method which is used to improve of an accuracy of automatic sleep judging result. The method may be executed by a processor and comprises steps of reading a to-be-test data having a variation depending on time of EOG generated by a user during a sleep process; acquiring a plurality of eye movement characteristics based on a variation of the EOG of a period, with the characteristics including an eye movement ratio, a blink count, a low frequency power ratio, a high frequency power ratio, an alpha rhythm ratio, a spindle rhythm ratio, a delta rhythm ratio and an average amplitude of eletrooculogram (EOG) signals; and determining a sleep state of the period is a "WAKE stage", a "REM stage", a "S1 stage", a "S2 stage", or a "SWS stage" based on the EOG characteristics. Furthermore, a computer program product with stored programs, a computer readable medium with stored programs, and electronic apparatuses.

Description

Eye-based telecommunications sleep staging method, computer program product of internal storage program, internal storage Program computer can read recording media and electronic devices

The present invention relates to a sleep staging method; in particular, to a sleep staging method based on eye-moving telecommunications, a computer program product of a stored program, a computer readable recording medium and an electronic device of a built-in program.

Sleep is related to metabolic and memory consolidation. Poor long-term sleep quality may turn into physical and mental problems such as dyssomnia. In severe cases, medical diagnosis can be used to diagnose sleep problems.

In the sleep diagnosis and treatment process, a multi-channel sleep physiology recorder (PSG) is usually used to record physiological states such as electroencephalogram (EEG), electromyography (EMG) and electrooculogram (EOG) during sleep (eg In the figure 1, the left and right eye corners, the chin, and the top of the electrode Z can respectively obtain a left eye movement telecommunications L, a right eye movement telecommunications R, a muscle movement telecommunications E, and a brain wave telecommunications C) for medical personnel. Comprehensive interpretation of different periods of sleep, such as: WAKE stage, REM stage, S1 stage, S2 stage, and slow wave sleep (SWS) Stage), etc., in order to understand sleep problems and give advice.

Taking an electrooculogram as an example, please refer to the sleep state electro-oculogram shown in Figures 2a~2g, where C and C' can represent brain wave signals, E for EMG signals, and L and R respectively for left. , the right eye electrogram signal, wherein the L and R signals are in opposite phase; the 2a picture can be judged to read the sleep state Fast eye movement period (REM, <6Hz) and alpha wave (alpha, 8~13Hz); from Fig. 2b, the fast eye movement period (REM) and the corresponding brain wave map are not derived from the brain (artifacts) ); the blinks can be read from the 2c map; the first phase of sleep in the slow eye movement (SEM, 0.2-0.6 Hz) can be read from the 2d map; the slow eye can be read from the 2e map. Spindle and K complex (K) appearing in the second phase of sleep; the slow-wave sleep period (SWS, low-frequency high-amplitude slow-wave, 0~4Hz) can be judged from the 2nd map, of which the electro-oculogram Synchronized with the brain wave map; the fast eye movement period (REM) can be read from the 2g map.

However, the above-mentioned manual interpretation method is very time consuming and easy to misjudge; in addition, in the above recording process, it is easy to cause disturbance to the sleep quality of the subject due to factors such as the location and method of the test, and the interpretation result sometimes fails. Accurately reflecting the user's true sleep condition, there is a need to develop a sleep staging method that degrades interference.

Among them, since the heart rate (frequency of heart beat) signal has the characteristics of low interference to the user, observable complete sleep cycle, and convenient recording, it is gradually developed to use the heart rate signal to interpret sleep staging. The examples can be considered as "Chen Junyou, Automatic Sleep Interpretation Method Based on Heart Rate Variability, National Institute of Successful Medical Information, 2012". However, each person's heart rate characteristics are not the same. For example, the heart rate of a good athlete is usually lower than that of the average person, and it will affect the interpretation accuracy of sleep stage (about 70%). Therefore, it is necessary to develop other automatic sleep interpretation techniques. .

In view of this, it is necessary to improve the shortcomings of the prior art described above to meet practical needs and improve its practicability.

The invention provides a sleep staging method based on eye movement telecommunications, which can improve the interpretation accuracy.

The invention further provides a computer program product for storing a program, which can be loaded and executed by a computer to improve the interpretation accuracy of the sleep staging.

The invention further provides a computer readable recording medium for the internal storage program, which can be loaded and executed by the computer to improve the interpretation accuracy of the sleep staging.

The present invention further provides an electronic device that can be loaded into a computer program and executed to improve the interpretation accuracy of sleep staging.

The invention discloses a sleep staging method based on eye-moving telecommunications, which is executed by a processor, and the steps include: a preparation step of reading a data to be tested, the data to be tested includes eye movement telecommunications generated by a user during sleep a change in time; a step of extracting a plurality of eye movement features according to the change of the eye movement telecommunications in a period of time, the plurality of eye movement characteristics including an eye movement ratio, a blinking number, a low frequency power ratio, and a a high frequency power ratio, an alpha wave ratio, a spindle wave ratio, a delta wave ratio, and a telecommunication average amplitude; and a decision step of determining a sleep state of the time period according to the plurality of eye movement characteristics is a waking period, a fast Eye movement, first sleep, second sleep, or slow wave sleep.

The decision step may use a decision tree to determine the sleep state, and the decision tree may include the following stages: a first stage, determining whether the eye movement ratio is greater than a first threshold, and if the determination is yes, performing a second In the stage, if the determination is no, a third stage may be performed; the second stage determines whether the delta wave ratio is greater than a second threshold and the high frequency power ratio is less than a third threshold. If the determination is yes, It can be determined that the sleep state belongs to the slow wave sleep period. If the judgment is no, the sleep state can be determined to belong to the second phase of sleep; in the third stage, it is determined whether the eye movement ratio is greater than a fourth threshold value, and if it is determined to be Yes, a fourth stage can be performed. If the determination is no, a fifth stage can be performed, wherein the fourth threshold is less than the first threshold; and in the fourth stage, the spindle wave ratio RS is determined to be greater than a fifth threshold. And if the low frequency power ratio is less than a sixth threshold value, if the determination is yes, a sixth stage may be performed. If the determination is no, the sleep state may be determined to belong to the fast eye movement period; Determining whether the alpha wave ratio is greater than a seventh threshold and whether the number of blinks is greater than an eighth threshold. If the determination is yes, the sleep state may be determined to belong to the awake period, and if the determination is no, a seventh phase may be performed. The sixth stage, It can be determined whether the uniformity of the telecommunication is greater than a ninth threshold. If the determination is yes, the sleep state can be determined to belong to the first phase of the sleep. If the determination is no, the sleep state can be determined to belong to the second phase of the sleep; In the seventh stage, it can be determined whether the alpha wave ratio is greater than a tenth threshold. If the determination is yes, the sleep state can be determined to belong to the second phase of the sleep. If the determination is no, the sleep state can be determined to belong to the fast eye movement period. Where the tenth threshold value is less than the seventh threshold value.

The first threshold may range from 0.55 to 0.65, the second threshold may range from 0.19 to 0.25, and the third threshold may range from 0.01 to 0.05, the fourth threshold The value may range from 0.3 to 0.5, the fifth threshold may range from 0.2 to 0.4, and the sixth threshold may range from 0.12 to 0.19, and the range of the seventh threshold may be From 0.3 to 0.5, the eighth threshold may range from 4 to 6, the ninth threshold may range from 0.25 to 0.35, and the tenth threshold may range from 0.12 to 0.19.

The eye movement telecommunications may include a left eye movement telecommunications and a right eye movement telecommunications, and the correlation coefficient of the left eye movement telecommunications and the right eye movement telecommunications in the same time period may be calculated as the eye movement ratio.

Before the correlation coefficient is calculated, the left eye movement telecommunication and the right eye movement telecommunication may first pass through a fourth-order 0 to 6 Hz band pass filter.

The left eye movement telecommunications and the right eye movement telecommunications may be selected as a sampling signal, and an opening speed may be calculated according to the sampling signal, and the number of times the opening speed exceeds one threshold value in the same period may be accumulated as the number of the blinks .

The calculation method of the opening speed is as follows: Where V is the opening speed, For the first derivative of the sampled signal versus time t, T is a sampling period; EOG(k.T) is the sampling signal of the kth period; EOG((k+1).T) is the k+1th Periodic sampling signal.

The sampling signal can be converted into a frequency domain signal, and the ratio of 0 to 4 Hz of the frequency domain signal in the same time period to 0 to 30 Hz can be calculated as the low frequency power ratio; the frequency domain signal can also be calculated in the same time period. The ratio of 0 to 30 Hz in the range of 0 to 30 Hz is used as the ratio of the high frequency power; the ratio of the alpha wave characteristic of the sampling signal in the same period can also be calculated as the ratio of the alpha wave; the sampling signal can also be calculated. The ratio of the spindle wave characteristics in the same period is used as the spindle wave ratio; the ratio of the δ wave characteristics of the sampling signal in the same period can also be calculated as the delta wave ratio.

According to the eye movement telecommunication sleep staging method, an average value of the left amplitude of the left eye movement and the right eye movement of the same time period can be used as the average value of the telecommunication.

A normal step may be performed between the capturing step and the determining step, the regular step taking the eye movement ratio, the number of blinks, the low frequency power ratio, the high frequency power ratio, the alpha wave ratio, the spindle wave ratio, The delta wave ratio and the telecommunication average amplitude are sequentially used as a regular target, and the values in the regular target are sorted, and the values can be equally divided into ten levels, and the average value of the highest level is calculated as a high value, and the calculation is performed. The average of the values in the lowest level is used as a low value, and the low value and the high value can be set to 0 and 1, respectively, so that all values in the regular target are between 0 and 1.

The invention further discloses a computer program product for storing a program, which can be completed after the computer loads the computer program and executes it.

The invention further discloses a computer readable recording medium with a built-in program, which can be completed when the computer loads the program and executes it.

The invention further discloses an electronic device capable of loading a computer program in a computer readable recording medium for performing the above method.

The computer can read the recording medium and the electronic device based on the eye-moving sleep staging method, the computer program product of the internal storage program, and the internal storage program, and the decision tree can determine the sleep state of the time period according to the plurality of eye movement characteristics. Belongs to the awake period, rapid eye movement, and sleep In the first phase, the second phase of sleep or the slow wave sleep period, the "easy to obtain accurate automatic sleep staging results" can be achieved, which can be further applied in medical clinical or home health care and other fields.

[study]

C‧‧‧ Brainwave Telecommunications

E‧‧‧Music telecommunications

L‧‧‧Left eye telecommunication

R‧‧‧Right Eye Telegraph

Z‧‧‧electrode

〔this invention〕

A‧‧‧Television average

B‧‧‧Blinks

M‧‧‧ eye ratio

P1‧‧‧ awake period

P2‧‧‧ rapid eye movement

P3‧‧‧Sleep first period

P4‧‧‧Sleep second period

P5‧‧‧Slow wave sleep period

R _α ‧‧‧α wave ratio

R _δ ‧‧‧δ wave ratio

R _S ‧‧‧ spindle ratio

R _{P0-4 ‧‧‧Low} frequency power ratio

R _{P13-22 ‧‧‧High} frequency power ratio

S1‧‧‧Preparation steps

S2‧‧‧ capture steps

S3‧‧‧ decision steps

S4‧‧‧ formal steps

S31‧‧‧ first stage

S32‧‧‧ second stage

S33‧‧‧ third stage

S34‧‧‧ fourth stage

S35‧‧‧ fifth stage

S36‧‧‧Sixth stage

S37‧‧‧ seventh stage

T1‧‧‧ first threshold

T1‧‧‧ second threshold

T3‧‧‧ third threshold

T4‧‧‧ fourth threshold

T5‧‧‧ fifth threshold

T6‧‧‧ sixth threshold

T7‧‧‧ seventh threshold

T8‧‧‧ eighth threshold

T9‧‧‧ ninth threshold

T10‧‧‧ Tenth threshold

Figure 1 is a schematic diagram of electrodes and signals used to record brainwave, electromyography, and electro-oculograms.

Figure 2a shows the user's sleep state electro-oculogram (1).

Figure 2b is the user's sleep state electro-oculogram (2).

Figure 2c shows the user's sleep state electro-oculogram (3).

Figure 2d is the user's sleep state electro-oculogram (4).

Figure 2e shows the user's sleep state electro-oculogram (5).

Figure 2f is the user's sleep state electro-oculogram (6).

The 2g image is the user's sleep state electro-oculogram (7).

Figure 3 is a flow chart showing the operation of the embodiment of the eye movement telecommunication based sleep staging method of the present invention.

Figure 4 is a schematic diagram of the decision tree classification of the decision steps of the present invention.

Figure 5a is an interpretation of the sleep staging of the expert.

Fig. 5b is a sleep staging interpretation result of the eye movement telecommunication-based sleep staging method embodiment of the present invention without smoothing.

Figure 5c is a sleep staging interpretation result of the smoothed processing of the eye movement telecommunication based sleep staging method embodiment of the present invention.

The above and other objects, features and advantages of the present invention will become more <RTIgt; Telematics (EOG) Signals are electrical signals generated by eye movements in daily life, such as the human eye during the WAKE stage, the REM stage, and the first stage of sleep ( In the sleep process such as S1 stage), S2 stage and SWS stage, different electrical signals generated near the eyelids can be measured by electrodes, but not limited thereto. It will be understood by those of ordinary skill in the art to which the invention pertains.

The "Eye Movement Ratio" as used throughout the present invention refers to the ratio of the two-eye activity calculated from the two-eye telecommunication, which is understood by those of ordinary skill in the art to which the present invention pertains.

The "Eye Blink Count" as used throughout the present invention refers to the number of single eye blinks read by one eye telecom, which is understood by those of ordinary skill in the art to which the present invention pertains.

The "Low Frequency Power Ratio" as used throughout the present invention refers to a ratio of 0 to 4 Hz to 0 to 30 Hz signals calculated by one eye movement telecommunications, which is a common knowledge in the technical field to which the present invention pertains. Can understand.

The "High Frequency Power Ratio" in the full text of the present invention refers to a ratio of 13 to 22 Hz to 0 to 30 Hz signals calculated from an eye-moving telecommunications, which is a common knowledge in the technical field to which the present invention pertains. Can understand.

The "Alpha Rhythm Ratio" as used throughout the present invention refers to the ratio of alpha waves calculated by one eye telecom, which can be understood by those of ordinary skill in the art to which the present invention pertains.

The "Spindle Ratio" as used throughout the present invention refers to the ratio of the spindle wave calculated from the two-eye telecommunication telecommunications, which is understood by those of ordinary skill in the art to which the present invention pertains.

The "Delta Rhythm Ratio" as used throughout the present invention refers to the ratio of the delta wave calculated by the one-eye mobile telecommunications, which is in the technical field of the present invention. Usually the knowledge person can understand.

The "Average Amplitude of EOG Signals" as used throughout the present invention refers to the average amplitude of a particular time (eg, window, window) calculated by two-eye telemetry, and is a technique of the present invention. Those with ordinary knowledge in the field can understand.

Please refer to FIG. 3, which is a flowchart of the operation of the embodiment of the eye movement telecommunication sleep staging method of the present invention. The embodiment of the eye-moving telemetry-based sleep staging method can utilize a computer system as an execution architecture. In this embodiment, the computer system can execute a sleep staging program by a processor. The embodiment of the eye-moving telemetry-based sleep staging method is not limited thereto; the embodiment of the eye-moving telemetry-based sleep staging method may include a preparation step S1, a step S2, and a decision step S3, as explained below.

Referring to FIG. 3 again, the preparation step S1 reads a data to be tested, and the data to be tested includes changes in eye movement telecommunication generated by the user during sleep. In this embodiment, the user can use the measuring instrument (eg, multi-channel sleep physiology recorder, etc.) to obtain eye movement telecommunications of the left and right eye channels (LOC, ROC) during sleep (eg, 1st). The figure measures the signals from the left and right corners of the eye. The eye movement signal is a continuous digital signal with a sampling frequency of 1K Hz and a resolution of 16-bits. The eye movement signal can be down sampling first. Up to 128 Hz; in addition, eye-moving telecommunications in the frequency range of 0.5 to 35 Hz can be intercepted, for example, the eye-moving telecommunications can be processed by a band-pass filter; and the eye-motion signal can be taken for 2 seconds. Interval is cut into several non-overlapping segments for 256-point fast Fourier transform (FFT); in addition, 15 2-second segment spectrum averaging can be used as a 30-second period (epoch) for convenience. Follow-up process, but not limited to this.

Referring to FIG. 3 again, the step S2 captures a plurality of eye movement features according to the change of the eye movement telecommunication in a period of time. The eye movement feature includes an eye movement ratio, a blinking number, and a low frequency. Power ratio, a high frequency power ratio, an alpha wave ratio, a spindle ratio, A delta wave ratio and a telecommunication average. The eye movement features are exemplified below.

For example, please refer to Table 1 first, first explain the calculation method of the eye movement ratio M. Since the eye movements of the left and right eye channels can present signals of opposite phases, the correlation coefficient between the signals can be utilized (correlation) Coefficient) as an important feature to separate the main eye movement period (eg, WAKE, REM, S1) from the rest period (eg, S2, SWS), the correlation coefficient can be calculated by assuming the left and right eyes The sample values of the telecommunication are the two sets of samples {x ₁ , x ₂ ,..., x _n } and {y ₁ , y ₂ ,..., y _n }, and the sample averages are respectively , The sample standard deviations are S _x and S _y , respectively, and the sample covariance S _{xy of the} two sets of samples can be defined as shown in the following formula (1a):

Thus, the correlation coefficient r can be defined as shown in the following equation (1b): Wherein, the correlation coefficient r can be used to indicate the degree of correlation between the two sets of samples, and the value is between -1 and +1.

In addition, a 2-second window can be designed to detect the main eye movements in a single epoch. The left-eye telecommunication and the right-eye telecommunication can pass through a fourth-order 0 to 6 Hz bandpass filter, such as a low pass. Butterworth filter to exclude the artifacts in the corresponding electroencephalogram (EGG) that are not derived from the brain, in order to retain the available eye movement features; then, the correlation coefficient of the left and right eye movement telecommunications can be used as a threshold The detection is counted. If the correlation coefficient is lower than the threshold, the time interval (window) can be regarded as an eye movement segment.

In the above, continue to explain the calculation method of the blink number B. Since the blink phenomenon is a general eye dynamic sample that can be found during the waking period, the number of blinks in a single period can be counted as a feature to detect sudden wake-up at night. In the case of the situation, and the time period of the missing wake-up period (WAKE) can be corrected, the steps of the blink detection algorithm are as follows: (I) The right eye movement telecommunication can pass the fourth-order 0 to 6 Hz band pass filter, such as: low A Butterworth filter is used to exclude the artifacts and high frequency noise from the brain that are not derived from the brain (EGG) to preserve the available eye movement features; (II) according to the sampling signal Calculate an EOG velocity, and its first-order approximation is as shown in the following equation (2): Where V is the opening speed, For the first derivative of the sampled signal versus time t, T is a sampling period, EOG(k.T) is the sampling signal of the kth period; EOG((k+1).T) is the k+1th The sampling signal of the period; (III) accumulating the number of times the opening speed exceeds one threshold value in the same period as the number of blinks, for example, finding the closed eye speed immediately after an opening speed, wherein the second speed is greater than the The threshold value (for example, 0.5 second), if its peak value is greater than the threshold value, it can be regarded as a blink of an eye to calculate the number of blinks B. For the definition of open eye and closed eye, the general knowledge in the technical field belongs to relevant literature (such as : "Jammes, B., Sharabty, H., & Esteve, D. (2008). Automatic EOG analysis: A first step toward automatic drowsiness scoring during wake-sleep transitions. Somnologie-Schlafforschung und Schlafmedizin, 12 (3), 227 -232. "Thesis" can be understood and will not be described here.

In the above, continue to explain the power ratio in the above-mentioned eye-moving telecommunications, such as the calculation method of the low-frequency power ratio R _P0-4 and the high-frequency power ratio R _P13-22 , and the frequency bands of the eye-moving telecommunications after FFT The average power will be concentrated, so each frequency band occupies 0 to 30 Hz of all powers as a feature, as shown in the following equation (3): Where i and j are respectively the lower limit frequency and the upper limit frequency of a specific frequency band; accordingly, 0 to 4 Hz can be defined as a low frequency band, and 13 to 22 Hz is defined as a high frequency band for calculating the same time period. A ratio of 0 to 4 Hz to 0 to 30 Hz is taken as the low frequency power ratio R _P0-4 , and a ratio of 13 to 22 Hz in the same period of time to 0 to 30 Hz is taken as the high frequency power ratio R _P13-22 .

Continuing to explain the persistence ratio in the above-mentioned eye movement telecommunications, such as the above-described alpha ratio R _α , the spindle wave ratio R _S , and the delta wave ratio R _δ . The alpha wave ratio R _α is the ratio of the alpha wave window in an epoch to all windows, and two eighth-order bandpass Butterworth filters (eg, 8 to 13 Hz, 22 to 30 Hz) can be used, except for the conventional 8 In addition to the alpha band of 13 Hz, the beta band of 22 to 30 Hz can also be used as a feature due to the high power band of 22 to 30 Hz during the awake period. The signals can be combined and detected by a threshold (eg, 0.5). If the absolute value of the combined signal is greater than the threshold relative to the original signal, the time interval (window) is visible. For the alpha wave.

According to the above, the spindle ratio R _S is the ratio of the spindle window in the epoch to all the windows, and the sigma band FFT and Butterworth bandpass filtering of 12 to 15 Hz can be used. Calculating the spindle wave ratio RS, wherein the FFT is used to find whether the power of the sigma wave band (12 to 15 Hz) is high, and the filtered signal is used to detect any sudden increase in amplitude, if both If the height is high, the time interval (window) can be regarded as the spindle wave, and those having ordinary knowledge in the technical field are concerned with relevant literature (for example: "Duman, F., Erdamar, A., Erogul, O., Telatar, Z., & Yetkin, S. (2009). Efficient sleep spindle detection algorithm with decision tree. Expert Systems with Applications, 36(6), 9980-9985. "Thesis" can be understood, and will not be described here.

Deck, the above-described [alpha] wave ratio R _α, spindles ratio R _S is similar to the [delta] wave ratio R _δ lines a period (Epoch) in SWS (slow wave stage) windows as a percentage of all the windows, can be 0.5 to 2 Hz The third-order bandpass Butterworth filter, if the amplitude of the filtered signal is greater than a threshold (eg, 0.2), the time interval (window) can be regarded as the SWS, so that the SWS phase can be separated between other periods. .

In addition, the calculation method of the telecommunication amplitude A is explained, and the average value of the absolute amplitudes of all the data points in a single time period (epoch) can be calculated by the left and right eye movements as a feature. The EOG energy in the awake period and the first phase of sleep is higher than the EOG energy in the second phase of sleep and slow wave sleep.

Referring to FIG. 3 again, the decision step S3 is based on the eye movement ratio M, the number of blinks B, the low frequency power ratio R _P0-4 , the high frequency power ratio R _P13-22 , and the alpha wave ratio. R _α , the spindle wave ratio R _S , the delta wave ratio R _δ and the telecommunication average A determine that the user's sleep state during the period is the WAKE stage P1 and the REM stage P2. S1 stage P3, S2 stage P4 or SWS stage P5. The decision step S3 can use a decision tree to interpret the sleep state, as explained below.

Please refer to FIG. 4, which is a schematic diagram of the decision tree classification of the decision steps of the present invention. The decision tree of the decision step S3 includes a first phase S31, a second phase S32, a third phase S33, a fourth phase S34, a fifth phase S35, a sixth phase S36, and a seventh phase. S37. In the first stage S31, it is determined whether the eye movement ratio M is greater than a first threshold T1. If the determination is yes, a second phase S32 is performed, and if the determination is no, a third phase S33 is performed; S32, determining whether the delta wave ratio R _{δ is} greater than a second threshold value T2 and the high frequency power ratio RP13-22 is less than a third threshold value T3. If the determination is yes, determining that the slow wave sleep period belongs to the sleep phase P4, if the determination is no, it is determined that the second phase of sleep P3 belongs to the sleep stage; the third stage S33 determines whether the eye movement ratio M is greater than a fourth threshold value T4, and if the determination is yes, the fourth stage is performed. S34, if the determination is no, the fifth stage S35 is performed, wherein the fourth threshold value T4 is less than the first threshold value T1; the fourth stage S34 determines that the spindle wave ratio RS is greater than a fifth threshold value T5 and _Whether the low frequency power ratio R _{P0-4 is} less than a sixth threshold value T6 is satisfied. If the determination is yes, the sixth stage S36 is performed, and if the determination is no, the rapid eye movement period P5 belonging to the sleep stage is determined; the fifth stage S35 , which [alpha] is determined based wave ratio R _α is larger than a seventh threshold T7 and the number of blinks B is greater than an eighth threshold value T8. If the determination is yes, it is determined that the awake period P1 belongs to the sleep stage, and if the determination is no, a seventh stage S37 is performed; the sixth stage S36 is determined Whether the uniform amplitude A of the telecommunications is greater than a ninth threshold T9, and if it is judged as YES, it is determined that the first phase of sleep P2 belongs to the sleep stage, and if the judgment is no, the second phase of sleep P3 belonging to the sleep stage is determined; and the seventh In step S37, it is determined whether the alpha wave ratio R _α is greater than a tenth threshold value T10. If the determination is yes, the sleep second phase P3 is determined to belong to the sleep stage, and if the determination is no, the fast eye movement period P5 belonging to the sleep stage is determined. Wherein the tenth threshold T10 is less than the seventh threshold T7.

In this embodiment, the first threshold value T1 may range from 0.55 to 0.65, the second threshold value T2 may range from 0.19 to 0.25, and the third threshold value T3 may range from 0.01 to 0.05. The fourth threshold value T4 may range from 0.3 to 0.5, the fifth threshold value T5 may range from 0.2 to 0.4, and the sixth threshold value T6 may range from 0.12 to 0.19. The value T7 may range from 0.3 to 0.5, the eighth threshold T8 may range from 4 to 6, and the ninth threshold T9 may range from 0.25 to 0.35. The threshold value of T10 can range from 0.12 to 0.19, but not limited to this.

Please refer to FIG. 3 again, wherein the embodiment of the eye movement telecommunication sleep staging method of the present invention performs a regular step S4 between the capturing step S2 and the decision step S3, with the eye movement ratio M, The number of blinks B, the low frequency power ratio R _P0-4 , the high frequency power ratio R _P13-22 , the alpha wave ratio R _α , the spindle wave ratio R _S , the delta wave ratio R _{δ ,} and the telecommunication amplitude A As a regular target, sort the values in the regular target, divide the value into ten grades, calculate the average of the values in the highest grade (10%) as a high value, and calculate the value in the lowest grade. The average value is taken as a low value, and the low value and the high value are respectively set to 0 and 1, so that all the values in the regular target are between 0 and 1, that is, the eye movement ratio M, The number of blinks B, the low frequency power ratio R _P0-4 , the high frequency power ratio R _P13-22 , the alpha wave ratio R _α , the spindle wave ratio R _S , the delta wave ratio R _{δ ,} and the telecommunication amplitude A are set to 0 to 1. .

Later, some smoothing rules can be used to improve the accuracy, as shown in Table 2.

Please refer to Figures 5a, 5b, and 5c. Figure 5a shows the results of the sleep staging interpretation by the expert. Figure 5b shows the unsmoothed sleep of the embodiment of the eye-moving telemetry-based sleep staging method of the present invention. The staged interpretation result, the 5th figure is the sleep stage interpretation result of the smoothed processing of the embodiment of the eye movement telecommunication based sleep staging method of the present invention. By 5a, 5b It can be seen that the results of the sleep staging interpretation of the embodiment of the eye-moving telemetry-based sleep staging method of the present invention without smoothing are not significantly different from the results of the sleep staging interpretation by the expert; the figures 5a and 5c show that The sleep-scheduled interpretation result of the smoothing-based sleep staging method of the invention is more similar to the sleep-scheduled interpretation result of the artificial interpretation, and the agreement rate of the two can be as high as 83.08%. The calculation manner may be: the percentage of the eye movement telecommunication sleep staging method embodiment of the present invention is the same as the manual interpretation result in all 30 epoch numbers.

Therefore, the embodiment of the eye movement telecommunication sleep staging method of the present invention can improve the accuracy of sleep staging interpretation, and can be further applied to the fields of medical clinical or home health care.

In addition, the embodiment of the eye-moving telemetry-based sleep staging method of the present invention can also be written into a computer program (such as the above-mentioned sleep staging program) by using a programming language (such as C++, Java, etc.), and the program code (Program Code) The writing method of the present invention is well known to those skilled in the art, and a computer program product for generating a built-in program can be completed. After the computer is loaded into the program and executed, the method embodiment of the present invention can be completed.

In addition, the computer program product can also be stored in a computer-readable recording medium such as a variety of memory, memory card, hard disk, optical disc or USB flash drive, etc., when the computer loads the above program and executes After that, the embodiment of the eye movement telecommunication sleep staging method of the present invention can be completed as a basis for the cooperative operation of the software and hardware of the computer system of the present invention.

Moreover, the present invention further discloses an electronic device, which may be a device having a data processing function, such as a computer or a smart phone, for executing an application (eg, an App, etc.) ), and the above embodiment of the eye movement telecommunication based sleep staging method is completed.

The main features of the above embodiments of the present invention are as follows: The technical features of the present invention, the computer-readable readable recording staging method, the computer program product of the internal storage program, and the internal storage program of the present invention are listed below: Reading the data to be tested, the data to be tested includes The change of the eye movement telecommunication generated by the user during the sleep; according to the change of the eye movement telecommunication in a period of time, the plurality of eye movement features are included, and the plurality of eye movement characteristics include the eye movement ratio and the number of eye movements The low frequency power ratio, the high frequency power ratio, the alpha wave ratio, the spindle wave ratio, the delta wave ratio, and the telecommunication average amplitude; the decision tree determines that the sleep state of the time period belongs to the waking period and the fast eye movement period according to the plurality of eye movement characteristics , the first phase of sleep, the second phase of sleep or the slow wave sleep period. Therefore, the above embodiments of the present invention can improve the accuracy of sleep staging interpretation, and can achieve the effect of "accurate automatic sleep staging results", and can be further applied to medical clinical or home health care and other fields.

While the invention has been described in connection with the preferred embodiments described above, it is not intended to limit the scope of the invention. The technical scope of the invention is protected, and therefore the scope of the invention is defined by the scope of the appended claims.

S1‧‧‧Preparation steps

S2‧‧‧ capture steps

S3‧‧‧ decision steps

S4‧‧‧ formal steps

Claims (17)

A method for sleep staging based on eye-moving telecommunications is performed by a processor, the steps comprising: a preparation step of reading a data to be tested, wherein the data to be tested includes eye movement telecommunications generated by the user during sleep a step of extracting a plurality of eye movement features according to the change of the eye movement telecommunications in a period of time, the plurality of eye movement characteristics including an eye movement ratio, a blinking number, a low frequency power ratio, and a high frequency power Ratio, an alpha wave ratio, a spindle wave ratio, a delta wave ratio, and a telecommunication average amplitude; and a decision step of determining, according to the plurality of eye movement characteristics, a sleep state of the time period belongs to an awake period, a rapid eye movement period, A first phase of sleep, a second phase of sleep, or a slow wave sleep period. The eye movement telecommunication sleep staging method according to claim 1, wherein the decision step uses a decision tree to interpret the sleep state, the decision tree comprising the following phase: a first phase, determining whether the eye movement ratio is If the determination is yes, a second phase is performed, and if the determination is no, a third phase is performed; and the second phase determines that the delta wave ratio is greater than a second threshold and the high frequency power Whether the ratio is less than a third threshold value, if the determination is yes, it is determined that the sleep state belongs to the slow wave sleep period, and if the determination is no, the sleep state is determined to belong to the second phase of the sleep; in the third stage, the eye is determined Whether the dynamic ratio is greater than a fourth threshold, if the determination is yes, a fourth phase is performed, and if the determination is no, a fifth phase is performed, wherein the fourth threshold is less than the first threshold; the fourth phase, Analyzing the spindles ratio R _S is greater than a fifth threshold value and the low frequency power ratio is less than a sixth threshold value is satisfied, if it is determined that performing a sixth stage, when the determination is NO, the identification The sleep state belongs to the rapid eye movement period; in the fifth stage, it is determined whether the alpha wave ratio is greater than a seventh threshold value and the number of blinks is greater than an eighth threshold value. If the determination is yes, the sleep state is determined to belong to the waking period. If the judgment is no, a seventh stage is performed; the sixth stage determines whether the telecommunications average is greater than a ninth threshold, and if the determination is yes, the sleep state is determined to belong to the first phase of the sleep, and if the judgment is no Determining that the sleep state belongs to the second phase of sleep; and in the seventh phase, determining whether the alpha wave ratio is greater than a tenth threshold value, and if the determination is yes, determining that the sleep state belongs to the second phase of the sleep, if it is determined to be No, the sleep state is determined to belong to the fast eye movement period, wherein the tenth threshold value is less than the seventh threshold value. The eye movement telecommunication sleep staging method according to claim 2, wherein the first threshold value ranges from 0.55 to 0.65, and the second threshold value ranges from 0.19 to 0.25, the third threshold The value ranges from 0.01 to 0.05, the fourth threshold ranges from 0.3 to 0.5, the fifth threshold ranges from 0.2 to 0.4, and the sixth threshold ranges from 0.12 to 0.19. The seven thresholds range from 0.3 to 0.5, the eighth threshold ranges from 4 to 6, the ninth threshold ranges from 0.25 to 0.35, and the tenth threshold ranges from 0.12 to 0.19. The eye movement telecommunications-based sleep staging method according to claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, and the left eye movement telecommunications and the right eye movement telecommunications are calculated The correlation coefficient in the same period is taken as the eye movement ratio. According to the eye movement telecommunication sleep staging method described in claim 4, wherein the left eye movement telecommunications and the right eye movement telecommunications pass before four before the correlation coefficient is calculated Step 0 to 6 Hz bandpass filtering. According to the eye movement telecommunication sleep staging method of claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement telecommunications and the right eye movement telecommunications As a sampling signal, an opening speed is calculated according to the sampling signal, and the number of times the opening speed exceeds one threshold in the same period is accumulated as the number of blinks. According to the eye movement telecommunication sleep staging method described in claim 6, wherein the opening speed is calculated as follows: Where V is the opening speed, For the first derivative of the sampled signal versus time t, T is a sampling period; EOG(k.T) is the sampling signal of the kth period; EOG((k+1).T) is the k+1th Periodic sampling signal. According to the eye movement telecommunication sleep staging method of claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement telecommunications and the right eye movement telecommunications As a sampling signal, the sampling signal is converted into a frequency domain signal, and a ratio of 0 to 4 Hz of the frequency domain signal in the same period of time is calculated as 0 to 30 Hz as the low frequency power ratio. According to the eye movement telecommunication sleep staging method of claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement telecommunications and the right eye movement telecommunications As a sampling signal, the sampling signal is converted into a frequency domain signal, and a ratio of the frequency domain signal of 13 to 22 Hz in the same period of time to 0 to 30 Hz is calculated as the high frequency power ratio. The eye movement telecommunication sleep staging method according to claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement power The right eye movement and the right eye are selected as a sampling signal, and the ratio of the alpha wave characteristics of the sampling signal in the same period is calculated as the alpha wave ratio. According to the eye movement telecommunication sleep staging method of claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement telecommunications and the right eye movement telecommunications As a sampling signal, the ratio of the spindle wave characteristics of the sampling signal in the same period is calculated as the spindle wave ratio. According to the eye movement telecommunication sleep staging method of claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, the left eye movement telecommunications and the right eye movement telecommunications As a sampling signal, the ratio of the delta wave characteristics of the sampling signal in the same period is calculated as the delta wave ratio. The eye movement telecommunications-based sleep staging method according to claim 1, wherein the eye movement telecommunications comprises a left eye movement telecommunications and a right eye movement telecommunications, and the left eye movement telecommunications and the right eye movement telecommunications according to the same period of time The absolute value of the amplitude is calculated as an average value as the average amplitude of the telecommunications. The eye movement telecommunication sleep staging method according to claim 1, wherein the step of extracting and the step of determining further perform a regular step, the regular step, the eye movement ratio, the number of blinks, the The low frequency power ratio, the high frequency power ratio, the alpha wave ratio, the spindle wave ratio, the delta wave ratio, and the telecommunication average amplitude are sequentially used as a regular target, and the values in the regular target are sorted, and the values are equally divided. Ten grades, the average value of the highest level is calculated as a high value, and the average value of the lowest level is calculated as a low value, and the low value and the high value are respectively set to 0 and 1. , so that all values in the regular target are between 0 and 1. A computer program product of a built-in program, when the computer is loaded into the computer program and executed, the method of any one of claims 1 to 14 can be completed. A computer capable of reading a recording medium, when the computer loads the program and executes The method of any one of claims 1 to 14 can be completed. An electronic device capable of loading a computer program readable by a computer readable recording medium for performing the method of any one of claims 1 to 14.