A system equipped with a sensing unit for detecting the sleep state of a subject.

Abstract

To provide a device, system, and method for detecting a sleep disorder that may occur in a subject.SOLUTION: A sensing unit includes: an accelerometer which is adapted to measure motions of the head and / or mandible of a subject; and a gyroscope which is adapted to measure the motions of the mandible of the subject, and transmits data of the rotational motion measured from the gyroscope and data of the acceleration measured from an accelerometer to a data analysis unit. The data analysis unit derives multiple measured rotation values and acceleration values from the sampled data of the rotational motion and acceleration, collates the measured rotation values and acceleration values with one or more of N mandibular motion classes indicative of a sleep event, generates diagnostic information for diagnosing a sleep disorder related to the sleep event on the basis of one or more of the collated N mandibular motion classes indicative of the sleep event, and creates a report containing the diagnostic information for diagnosing a sleep disorder related to the detected sleep event.SELECTED DRAWING: None

Description

Technical Field

[0001] The present invention relates to an apparatus, a system, and a method for detecting sleep disorders that may occur in a subject.

Background Art

[0002] The most common method for evaluating sleep disorders, more specifically sleep-disordered breathing, is the polysomnogram (PSG) test in a laboratory. To undergo this test, an overnight stay in a dedicated facility under the supervision of a trained technician is required. However, this method is costly and time-consuming and cannot keep up with the pace of demand. In a PSG test, various types of sensors are used to record multiple physiological signals (EEG, EMG, ECG, thermistor, pressure, video, etc.). Subsequently, medical experts evaluate the data obtained from these sensors.

[0003] In the art, other systems have been considered. US 2017 / 0265801 relates to a teeth-grinding detection system for detecting teeth-grinding and chewing. The system includes an accelerometer attached to the jaw. This senses and records changes in acceleration at the start and end of clenching the jaw. By processing the data from the accelerometer and comparing it with an accelerometer threshold, movements related to teeth-grinding are distinguished from other movements of the head.

[0004] US 2017 / 0035350 also relates to a teeth-grinding detection system. The system includes two accelerometers attached to the masseter muscles. The first accelerometer is attached to the skin of the left masseter muscle, and the second accelerometer is attached to the skin of the right masseter muscle. Teeth-grinding is detected when the recorded data of the two accelerometers are substantially equal.

[0005] US 2007 / 273366 relates to a sleep disorder detector system. The system includes a distance measuring device by detecting the emitted magnetic field. The device can be attached to a support placed on the head to measure mouth movements. The data from the device is processed to detect respiratory diseases during sleep such as snoring.

[0006] The problem with these known systems is that the head movement and mandibular movement of the subject wearing the sensing unit are considered separately from each other. A similar problem applies to the head and mandibular positions calculated from movements measured by accelerometers. However, data from accelerometers is limited and can be affected by movements of other body parts, such as the chest and trachea during breathing. Therefore, the relationships between these various movements and positions are not adequately considered for analyzing sleep disorders and can negatively impact diagnoses based on the measured data stream.

[0007] In fact, mandibular movement can be induced by either respiratory or non-respiratory movements. Therefore, head movement during sleep can cause mandibular movement. Mandibular movement can be considered both mechanically related to tracheal traction and under the control of the brain. Thus, mandibular movement is either passively induced by the respiratory movement of tracheal traction or directly controlled by the brain. This traction is the tensile force exerted by the chest on the human head. Since the respiratory muscles are controlled by the brain and the chest moves with breathing, this tensile force is at the human respiratory frequency. Therefore, if the head moves at this respiratory frequency, the mandible attached to the head will follow the movement imposed on the head at that respiratory frequency. This is a passive movement following the movement of the head. Even if the head does not move, or does not move most of the time, mandibular movement is similarly directly and actively controlled by the brain. The brain controls mandibular movement by stimulating the mandibular muscles. Therefore, it is useful to be able to distinguish between mandibular movements controlled by the brain and mandibular movements controlled by tracheal traction attachment. A system is needed that can more accurately interpret signals from the brain and more accurately identify sleep disorders. [Overview of the project]

[0008] The object of the present invention is to provide a data processing system for temporally relating measured values ​​of the movement and position of a subject's head and mandible in a sensing unit and analysis of measurement data.

[0009] In particular, the present invention relates to a system comprising (equally, in combination) a sensing unit related to sleep disorders that may occur in a subject and a processing unit for processing the data thereof. Herein, the sensing unit includes a gyroscope adapted to measure the movement of the subject's mandible. Surprisingly, the inventors have discovered that the rotation of the mandible can be captured using the gyroscope, thereby allowing for the evaluation of brainstem activity that controls mandibular movement during sleep.

[0010] In some embodiments, the present invention relates to a system comprising a sensing unit and a data processing device, e.g., a processing unit, relating to a sleep disorder that may occur in a subject. The sensing unit includes an accelerometer adapted to measure the movement of the subject's head and / or mandible, and a gyroscope adapted to measure the movement of the subject's mandible. The sensing unit is adapted to produce a measurement signal based on the measurements taken. The processing unit includes first and second inputs for receiving first and second time streams of measurement signals from the accelerometer and gyroscope, respectively.

[0011] Accordingly, provided herein is a system for characterizing sleep disorders in subjects having a head and mandible, which includes a gyroscope and a data analysis unit connected by a data link. In certain embodiments, the system is characterized by including the following: - A gyroscope configured to measure the rotational movement of the subject's mandible; - Data analysis unit and data link (the data link is configured to transmit rotational motion data measured from the gyroscope to the data analysis unit; Here, the data analysis unit includes a memory unit configured to store N mandibular movement classes, where N is an integer greater than 1, and at least one of the N mandibular movement classes indicates the onset of a sleep disorder; - Here, each j(1< j < N) A mandibular motion class comprises the j-th set of rotational values, where each j-th set of rotational values ​​represents at least one velocity, velocity change, frequency, and / or amplitude of mandibular rotation associated with the j-th class; - Here, the data analysis unit includes a sampling element configured to sample rotational motion data measured during the sampling period, thereby obtaining the sampled rotational motion data; - Here, the data analysis unit is configured to derive a plurality of measured rotational values ​​from the sampled rotational motion data; and, - Herein, the data analysis unit is further configured to match the measured rotation values ​​with the N mandibular movement classes.

[0012] In some embodiments, the system includes an accelerometer adapted to measure acceleration, which indicates the movement and / or position of the subject's head and / or mandible. - The data link is further configured to transmit acceleration data measured from the accelerometer to the data analysis unit; - Here, each j(1 < j < N) A mandibular movement class includes the j-th set of acceleration values, where each j-th set of acceleration values ​​represents at least one movement of the mandible or head associated with that j-th class; - Here, the sampling element is configured to sample acceleration data measured during the sampling period, thereby obtaining the sampled acceleration data; - Here, the data analysis unit is configured to derive a plurality of measured acceleration values ​​from the sampled acceleration data; and, - Herein, the data analysis unit is further configured to match the measured acceleration values ​​with the N mandibular movement classes.

[0013] In some embodiments, the system further includes a magnetometer, which is adapted to measure magnetic field data. Variations in the magnetic field data indicate the movement of the subject's head and / or mandible. - The data link is further configured to transmit magnetic field data measured from the accelerometer to the data analysis unit; - Here, each j(1 < j < N) A mandibular motion class includes the j-th set of magnetic field data values, where each j-th set of magnetic field data values ​​represents at least one velocity or velocity change of mandibular or head movement associated with that j-th class; - Here, the data analysis unit includes a sampling element configured to sample magnetic field data measured during the sampling period and thereby obtain the sampled magnetic field data; - Here, the data analysis unit is configured to derive a plurality of measured magnetic field values ​​from the sampled magnetic field data; and, - Herein, the data analysis unit is further configured to match the measured magnetic field values ​​with the N mandibular movement classes.

[0014] In some embodiments, the gyroscope and, optionally, an accelerometer and / or magnetometer or a part thereof are included in a sensing unit which can be attached to the mandible of the subject.

[0015] In some embodiments, one or more of the N mandibular movement classes are characterized by a predetermined frequency range.

[0016] In some embodiments, the analysis unit is configured to identify the movement of the subject's head based on the data from the gyroscope, accelerometer, and / or magnetometer.

[0017] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is awake, and a plurality of the N mandibular movement classes indicate that the subject is asleep.

[0018] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is in the N1 sleep state, and at least one of the N mandibular movement classes indicates that the subject is in the REM sleep state. Optionally, at least one of the N mandibular movement classes indicates that the subject is in the N2 sleep state, and / or at least one of the N mandibular movement classes indicates that the subject is in the N3 sleep state.

[0019] In some embodiments, one or more of the N mandibular movement classes indicate obstructive apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and / or central hypopnea.

[0020] In some embodiments, one of the N mandibular movement classes indicates teeth grinding. Here, the measured rotational movement data indicates a mandibular movement amplitude of at least 1 mm, a frequency established in the range of 0.5 to 5 Hz during at least three respiratory cycles if the movement is phasic, or a mandibular movement amplitude that is continuously and forcibly greater than 1 mm for at least 2 seconds.

[0021] There is further provided a method for assisting in characterizing sleep disorders in a subject having a mandible, including the following steps. - A step of receiving, by a data analysis unit, data of rotational movement from a gyroscope disposed on the mandible of the subject via a data link; - A step of storing, in a memory unit included in the data analysis unit, N mandibular movement classes (where N is an integer greater than 1, and at least one of the N mandibular movement classes indicates the onset of a sleep disorder); - Here, each j (1 < j <N) A mandibular motion class includes the j-th set of rotational values, where each j-th set of rotational values ​​represents at least one velocity, velocity change, frequency, or amplitude of mandibular rotation associated with that j-th class; - A step of sampling rotational motion data during the sampling period using the sampling elements included in the data analysis unit, thereby obtaining the sampled rotational motion data; - Using the data analysis unit, derive multiple measured rotational values ​​from the sampled rotational motion data; and, - A step of using the data analysis unit to compare the measured rotation values ​​with the N mandibular movement classes.

[0022] In some embodiments, the method further includes the following steps. - A step of measuring acceleration using an accelerometer (the acceleration indicates the movement and / or position of the subject's head and / or mandible); - A step of transmitting acceleration data measured from the accelerometer to the data analysis unit using the data link; - Here, each j(1 < j < N) A mandibular movement class includes the j-th set of acceleration values, where each j-th set of acceleration values ​​represents at least one movement of the mandible or head associated with that j-th class; - A step of sampling acceleration data measured during the sampling period using a sampling element, thereby obtaining the sampled acceleration data; - A step of deriving multiple measured acceleration values ​​from sampled acceleration data using the data analysis unit; and - A step of using the data analysis unit to compare the measured acceleration values ​​with the N mandibular movement classes.

[0023] In some embodiments, the method further includes the following steps. - A step of measuring magnetic field data using a magnetometer (the fluctuations in the magnetic field data indicate the movement of the subject's head and / or mandible); - A step of transmitting the measured magnetic field data from the magnetometer to the data analysis unit using the data link; - Here, the j(1 < j < N) A mandibular motion class includes the j-th set of magnetic field data values; each j-th set of magnetic field data values ​​represents at least one velocity or velocity change of mandibular or head movement associated with that j-th class; - A step of sampling magnetic field data measured during the sampling period using a sampling element included in the data analysis unit, thereby obtaining the sampled magnetic field data; - A step of deriving multiple measured magnetic field values ​​from the sampled magnetic field data using the data analysis unit; and - A step of using the data analysis unit to compare the measured magnetic field values ​​with the N mandibular movement classes.

[0024] In some embodiments, the method further includes the step of using the analysis unit to identify the movement of the subject's head based on the data from the gyroscope, accelerometer, and / or magnetometer.

[0025] In some embodiments, at least one of the N mandibular movement classes represents bruxism. Here, the measured rotational movement data represents a mandibular movement amplitude of at least 1 mm, and if the movement is phase-dependent, at a frequency established in the range of 0.5–5 Hz during at least three respiratory cycles, or a mandibular movement amplitude greater than 1 mm sustained and tonic for at least two seconds. [Brief explanation of the drawing]

[0026] Next, the present invention will be described in more detail using drawings illustrating the system and its operation. The system can be described as a sensing unit system and a device or unit for processing the sensed data. The drawings are as follows.

[0027] [Figure 1] The system according to the present invention.

[0028] [Figure 2A-2B] Two streams showing the changing position of a person's head while they are lying on their bed.

[0029] [Figure 3A-3B] Stream was captured by the sensing unit while grinding his teeth.

[0030] [Figure 4] Loop gain.

[0031] [Figure 5] Identification of micro-awakening after pretreatment.

[0032] [Figure 6] The measured signal after applying bandpass filtering.

[0033] [Figure 7] A signal indicating mild arousal.

[0034] [Figure 8] An example of the first and second measurement signal streams in obstructive sleep apnea.

[0035] [Figure 9] An example of the first and second measurement signal streams in obstructive hypopnea.

[0036] [Figure 10] An example of the first and second measurement signal streams in mixed sleep apnea.

[0037] [Figure 11] An example of the first and second measurement signal streams in central apnea.

[0038] [Figure 12] An example of the first and second measurement signal streams in central hypopnea.

[0039] [Figure 13] An example of the first and third measurement signal streams in respiratory effort-related arousal (RERA).

[0040] [Figure 14] Spectrogram of the frequency distribution of mandibular bone movement.

[0041] [Figure 15] An example of the procedure for feature extraction, data processing, and data description.

[0042] [Figure 16-17] Analysis of mandibular movement data acquired using a magnetic sensor.

[0043] [Figure 18] An example of a method for automatically detecting sleep stages from mandibular movement data acquired using a gyroscope and accelerometer. This method is explained in detail in Example 18.

[0044] In Figure 1, the following numbering is used: 1: Sensing unit; 2: Accelerometer; 3: Gyroscope; 4: Magnetometer; 5: Oxygen meter; 6: Thermometer; 7: Audio sensor; 8: Electromyography unit; 9: Photoelectric pulse wave recording; 10: Data processing unit; 11-1: First input; 11-2: Second input; 11-3: Third input; 11-4: Fourth input; 12: Identification unit; 13: Analysis unit. Detailed description of the invention

[0045] Before describing the systems and processes of the present invention, it should be understood that, naturally, such systems and methods or combinations thereof may vary, and therefore this is not limited to the specific systems and methods or combinations described. It should also be understood that the scope of the terms is limited only by the appended claims, and therefore the terms used herein are not intended to be limiting.

[0046] As used herein, the singular forms "one" and "this" include both singular and plural references unless the context clearly indicates otherwise.

[0047] As used herein, the terms “contains,” “includes,” and “consist of” are synonymous with “encompassing,” “contains,” “includes,” or “contains,” and are inclusive or free form, and do not exclude additional, unquoted members, elements, or steps of the method. As used herein, the terms “constitutes,” “consist of,” and “composed of” will be understood to include the terms “consist of” and “consist of.”

[0048] The enumeration of numerical ranges by endpoints includes all numbers and fractions within each range, as well as the enumerated endpoints.

[0049] When referring to measurable values ​​such as parameters, quantities, durations, etc., the terms “about” or “approximately” as used herein mean that such variation includes variation from a specified value appropriate for performing in the disclosed aspects and embodiments, and within a variation of + / -10%, preferably + / -5%, more preferably + / -1%, and even more preferably + / -0.1% or less of that specified value. It should be understood that the values ​​themselves to which the modifiers “about” or “approximately” refer are also specifically and preferably disclosed.

[0050] The terms “one or more” or “at least one” are self-explanatory and, by further example, include references to any one of the members, or any two or more of the members, for example, three or more, four or more, five or more, six or seven or more of any of the members, up to all members.

[0051] All references cited herein are incorporated herein in their entirety by reference. In particular, the teachings of all references specifically mentioned herein are incorporated herein by reference. Unless otherwise defined, all terms used herein, including technical and scientific terms, have the meanings generally understood by those skilled in the art. Definitions of terms, as provided herein by further guidance, are included for a better understanding of these teachings.

[0052] The following sections define various aspects in more detail. Each aspect as defined may be combined with one or more other aspects unless otherwise specified. In particular, any feature that is shown to be favorable, specific, or advantageous may be combined with any other feature that is shown to be favorable, specific, or advantageous.

[0053] This invention relates to the measurement and evaluation of mandibular movement in subjects during sleep. The mandible, or mandible, is located below the maxilla and forms the mandible. It is the only movable bone in the human skull (excluding the ossicles of the middle ear). During movement, the mandible rotates around the temporomandibular joint, where it connects to the skull in front of the ear (temporal bone). During mandibular movement, the relationship between the length and tension of the muscle fibers fixed to the mandible changes. This can lead to stiffening of the upper airway in subjects at risk of instability during sleep. This movement is activated under the agonist / antagonist muscles that move the mandible up and down, thereby opening and closing the mouth, respectively. These agonist / antagonist muscles are innervated by motor nerves originating from the trigeminal nerve nucleus in the brainstem (pontine center) and supported by the motor branches of these nerves.

[0054] Provided herein is a system for characterizing sleep disorders in a subject having a head and mandible. The system includes a gyroscope, which is configured to measure the rotational movement of the mandible of the subject, which is an activity in which the gyroscope is particularly well suited, as observed by the inventors. The gyroscope can be used to assess brainstem activity that stimulates mandibular movement during sleep in a manner that keeps the upper airway (pharynx) open and prevents sleep apnea. The movable bone of the mandible rotates like a lever, stretching pharyngeal muscle fibers attached to the arch of the mandible, directly or indirectly (via the hyoid bone, which is the second movable bone), including the tongue.

[0055] The movement of the gyroscope, to some extent, represents central drive. This means that the trigeminal nerve nucleus in the pons acts to finely displace the mandible under the influence of higher centers involved in the sleep mechanism (sleep stages), in relation to the respiratory center, which is also located in the brainstem. As a result, the provision of a gyroscope within the sensing unit can be used to assess various sleep-related activities by examining rotational mandibular displacement, which may include respiration, sleep stages, or other events (e.g., movement or motor onset). In addition to metrics derived directly or indirectly from measurements, values ​​measured by a gyroscope positioned to measure the rotational movement of the subject's mandible (such as the velocity and amplitude of the mandibular gyroscope signal) can be used to obtain an assessment of central drive originating from the trigeminal nerve nucleus.

[0056] The inventors have found that other sensing units are unsuitable for measuring and evaluating mandibular movement as provided herein. For example, inertial sensors such as accelerometers are unsuitable for measuring rotational displacement of the mandible because they are limited in their ability to measure linear acceleration. Measurements by accelerometers can be affected by body and head movements, such as those of the chest and trachea during respiration, making it difficult to distinguish the source of the data and adding unnecessary noise and complexity to the system. As a result, existing systems for analyzing sleep disorders do not adequately consider the relationship between possible body and head movements. This negatively impacts diagnoses based on the measured data stream. The inventors have found that mandibular rotation carries the necessary information for accurate evaluation, and furthermore, such movements can be accurately recorded by a gyroscope.

[0057] The system further includes a data analysis unit and a data link. The data link provides a communication path between the gyroscope and the data analysis unit. While a data link using wired communication is certainly possible, the data link is preferably a wireless data link, for example, to improve the comfort of the subject.

[0058] Rotational motion data is transmitted from the gyroscope to the data analysis unit via the data link. The data link is of a general nature and includes a configuration for transferring data wirelessly or via a wired connection.

[0059] The data analysis unit comprises a memory unit, a data storage device such as a hard drive, solid-state drive, or memory card. The memory unit is configured to store several (N) mandibular movement specific patterns (classes), where N is an integer greater than 1. At least one of the N mandibular movement classes indicates the onset of a sleep disorder. Preferably, the N mandibular movement classes include multiple movement classes that show various mandibular movements. Each j(1 < j <N) A mandibular motion class consists of the j-th set of rotational values, where each j-th set of rotational values ​​represents at least one velocity, velocity change, frequency, and / or amplitude of mandibular rotation associated with the j-th class.

[0060] The rotational motion data measured or recorded by the gyroscope is linked to the mandibular motion class as follows:

[0061] The data analysis unit includes a sampling element configured to sample rotational motion data measured during a sampling period. Thus, sampled rotational motion data is obtained. The information contained in the signal recorded by the gyroscope can then be extracted for further analysis. It should be understood that while specialized hardware can certainly be provided, in some embodiments the data analysis unit may be included in a general-purpose computing device such as a personal computer or smartphone.

[0062] The data analysis unit is configured to derive multiple measured rotational values ​​from sampled rotational motion data and to match the measured rotational values ​​with N mandibular motion classes. Preferably, deriving measured rotational values ​​from sampled rotational motion data includes one or more of the following steps: discretization, time averaging, fast Fourier transform, etc. The matching can also be fully or partially automated by providing a machine learning model. Thereafter, the data analysis unit is configured to learn several statistical and / or physical metrics to capture the characteristics of the signal in the frequency and time domains and to identify patterns in the rotational signal for specific events such as sleep stages or respiratory effort. Thus, providing a machine learning model can provide automatic interpretation of the relevant information and / or matching the characteristic data with the onset of sleep disorders. Thus, the examination of mandibular motion during sleep provides information on respiratory control states in response to changes in the permeability or resistance of airflow in the upper airway, regardless of whether it relates to a series of changes in head position. Furthermore, analysis of the nature of mandibular movement using the system according to the present invention can detect the onset of non-respiratory movements that occur repeatedly during sleep, such as teeth grinding or chewing, or other characteristics such as oral and facial movement disorders. Swallowing and sucking movements in infants can also be clearly identified. Swallowing movements can also be detected in adults. This makes it possible to distinguish between wakefulness and micro-wakefulness.

[0063] In some embodiments, one or more mandibular movement classes exhibit a separate large mandibular movement (IMM). Since IMMs are associated with microarousal or respiratory distress-induced arousal, microarousal can be effectively inferred from the aforementioned measurements and analyses.

[0064] In some embodiments, the process of matching the measured rotational values ​​with N mandibular movement classes utilizes artificial intelligence methods, such as random forests.

[0065] In some embodiments, the system further includes an accelerometer, which is adapted to measure acceleration (including acceleration fluctuations) indicating the movement and / or position of the subject's head and / or mandible. The inventors have found that the accelerometer is particularly well suited for measuring head movement and position. Adding an accelerometer to the system makes it possible to further evaluate the behavior of the mandible during sleep. In particular, the inventors have found that measuring acceleration can explain unexpected changes in the motion, amplitude, and / or velocity of the gyroscope. Thus, measurements by the accelerometer can be used to supplement the measurements by the gyroscope.

[0066] The measured or recorded acceleration data is transmitted by the accelerometer to the data analysis unit via the data link. In these embodiments, each j(1 < j < N) A mandibular movement class consists of the j-th set of acceleration values. Each j-th set of acceleration values ​​or metrics represents at least one mandibular or head movement associated with the j-th class. The sampling element is configured to sample acceleration data measured during the sampling period. After sampling, the measured acceleration data is called sampled acceleration data. The data analysis unit is configured to derive multiple measured acceleration values ​​from the sampled acceleration data, for example, by discretization and, at choice, time averaging. Information contained in the signals recorded by the accelerometer may be extracted for further analysis. The data analysis unit is further configured to match the measured acceleration values ​​with N mandibular movement classes. The matching process is understood to include automatically determining the mandibular movement class closest to the measured acceleration value. The matching can be fully or partially automated with a machine learning model. This allows for the automatic interpretation of the matching of relevant information and / or characteristic data with the onset of sleep disorders.

[0067] The inventors have discovered that the accelerometer is particularly sensitive to head movements. Furthermore, the gyroscope and accelerometer enable efficient identification of head movements from mandibular movements, thereby improving the detection of sleep disorder onset. As a result, providing a gyroscope and accelerometer in a single system improves the sensitivity and accuracy of the system and can be used to evaluate new information that could not be interpreted from measurements taken by the gyroscope or accelerometer alone. For example, changes in head position stimulated by central nervous system activation may affect the rotational movement of the mandibular bone, which may be misinterpreted as a change in the degree of mouth opening and closing. Therefore, the combination of a gyroscope and accelerometer allows for the distinction of head movements from jaw movements. Given the superior unexpected capabilities provided by this combination, the presence of a gyroscope cannot be considered a substitute for other sensing devices, such as a second accelerometer.

[0068] In some embodiments, the system further includes a magnetometer adapted to measure magnetic field data. Variations in the magnetic field data indicate the direction and / or position of movement of the subject's head and / or mandible. Adding a magnetometer to the system makes it possible to further evaluate the behavior of the mandible during sleep. It can be understood that providing a magnetometer to the system helps to evaluate the orientation of the sensing unit, similar to a compass. Thus, its primary function is understood to be not limited to the scope of the system, but the magnetometer is not intended to function as a unit for measuring distance as intended in the art.

[0069] The data link is further configured to transmit measured or recorded magnetic field data from the magnetometer to the data analysis unit. < j <N) A mandibular motion class includes the j-th set of magnetic field data values. Each j-th set of magnetic field data values ​​represents at least one velocity or velocity change of mandibular motion or head motion associated with the j-th class. The data analysis unit includes a sampling element configured to sample magnetic field data measured during a sampling period. Thus, sampled magnetic field data is obtained. The data analysis unit is configured to derive a plurality of measured magnetic field values ​​from the sampled magnetic field data. The data analysis unit is further configured to match the measured magnetic field values ​​with N mandibular motion classes.

[0070] In a particular form, the magnetometer may include two parts: one part is attached to the patient's forehead, and the other part is attached to the patient's mandible. The inventors have found this to be a particularly effective configuration for detecting the movement of the mandible.

[0071] In some embodiments, signals generated from the magnetometer, gyroscope, accelerometer, and / or further sensors are transmitted, for example, via time-division multiplexing and / or carrier waves of different frequencies through a single physical medium.

[0072] In some embodiments, the gyroscope, and / or accelerometer, and / or magnetometer, or parts thereof, are included in a sensing unit. The sensing unit can be attached to the mandible of the subject. This is an embodiment with very compact form elements, is easy to apply, and improves patient comfort. The provision of the accelerometer and / or magnetometer is understood not to replace the function of the gyroscope, but to realize a new interpretation that would only be possible by the combination of the gyroscope and one or more additional sensing devices such as the accelerometer and / or magnetometer. Preferably, first, the interpretation of the acquired signal is associated with the data from the gyroscope, and in the second step, it is supplemented with the data from the accelerometer and / or magnetometer. For example, first, the angular velocity of the mandible can be analyzed using the data from the gyroscope, and a comprehensive cycle can be reached in the cycle analysis. Furthermore, the data from the accelerometer can be used to provide information about the conditions under which cycles are produced (e.g., the origin of activation (cortical and subcortical), the endotype (dynamics of respiratory disturbance), and the type of muscle masticatory activity (more or less tonic or phase)). The combination of data also allows for new assessments that would be impossible with data from a single sensing unit alone. For example, a precise description of the onset type opens up the possibility of predicting the onset or recurrence of sleep disturbances, or changes in respiration (e.g., peripheral capillary oxygen saturation SpO2).

[0073] Preferably, the size of the sensing unit is a maximum of 5 cm in length, 2 cm in thickness, and 1 cm in height. This reduces interference with the subject's normal sleep.

[0074] In some embodiments, one or more of the N mandibular movement classes are associated with a predetermined frequency range. In other words, in these embodiments, one or more of the N mandibular movement classes include mandibular movements occurring within a predetermined frequency range. Preferably, at least two of the N mandibular movement classes are associated with a predetermined frequency range that includes a predetermined A frequency range and a predetermined B frequency range, and the predetermined A frequency range and the predetermined B frequency range do not overlap.

[0075] In some embodiments, at least one predetermined frequency range consists of frequencies of 0.15 Hz to 0.60 Hz, 0.25 Hz to 0.50 Hz, or 0.30 Hz to 0.40 Hz. This is the frequency range of the signal indicating the subject's respiration.

[0076] In some embodiments, the system further includes one or more auxiliary components selected from a list including an oxygen meter and / or a thermometer and / or an audio sensor and / or an electromyography unit and / or a pulse photoplethysmograph. Preferably, these auxiliary components are connected to the analysis unit via a data link for operation.

[0077] In some embodiments, the analysis unit is configured to identify the subject's head movement based on data from a gyroscope and / or accelerometer and / or magnetometer. Preferably, the head movement includes rotation, for example, rotation around an axis passing through the center of the subject's head. In this case, preferably, at least one of the N mandibular movement classes indicates a change in head position. This allows for efficient differentiation between general head movement and movement of the mandible itself. In these embodiments, the system preferably includes both an accelerometer and a gyroscope.

[0078] In some embodiments, the analysis unit is adapted to apply one or more preprocessing steps to gyroscope data and / or accelerometer data and / or magnetometer data. One or more preprocessing steps are selected from a list including: applying a bandpass filter, applying a lowpass filter, exponential moving average, and / or calculating the frequency entropy of the gyroscope data and / or accelerometer data and / or magnetometer data. Applying lowpass filtering improves the detection of minute arousal.

[0079] In some embodiments, the analysis unit may include an interpretation module configured to interpret specific parameters for measuring sleep quality and the extent of sleep-disordered breathing. Sleep quality parameters may include, for example, total sleep time (TST), sleep onset latency (SOL), first awakening from sleep onset (WASO), arousal index, sleep efficiency (SE), REM ratio, non-REM sleep, REM sleep latency, and other sleep quality indicators. Measurements related to sleep-disordered breathing may include hourly incidence during sleep and cumulative duration of respiratory effort. The analysis unit may be configured to report the interpreted subject-specific parameters. Such reporting may include providing output to a device such as a computer or smartphone. Such reporting may also include providing a visual or textual report of the subject-specific parameters, for example, in the form of hypnotism.

[0080] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is awake, and multiple of the N mandibular movement classes indicate that the subject is asleep. By incorporating the “sleeping” and “awake” classifications into the method, it is ensured that measurements taken while the subject is awake or asleep are interpreted accordingly. This interpretation can be performed using an interpretation module.

[0081] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is in sleep state N1, and at least one of the N mandibular movement classes indicates that the subject is in sleep state REM. Optionally, at least one of the N mandibular movement classes indicates that the subject is in sleep state N2, and / or at least one of the N mandibular movement classes indicates that the subject is in sleep state N3.

[0082] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is in sleep state N2.

[0083] In some embodiments, at least one of the N mandibular movement classes indicates that the subject is in sleep state N3.

[0084] In some embodiments, one or more of the N mandibular movement classes are associated with the detection of sleep stages. The detection of sleep stages may be further carried out to establish the subject's unique sleep pattern. The detection of sleep stages is preferably automated at different levels of resolution.

[0085] In a preferred embodiment, sleep patterns may include the following (classified by increasing complexity): (1) Two-class scoring (i.e., binary): when detecting the subject's state of wakefulness or sleep; (2) Three-class scoring: when classifying the subject's state of wakefulness, non-REM sleep stages, or sleep stages including REM sleep stages; (3) Four-class scoring: When classifying sleep stages, including the subject's state of wakefulness, light sleep (N1 and N2) stages, deep sleep (N3) stage, or REM sleep stage; (4) 5-class scoring: When classifying all sleep stages, including the subject's arousal state, N1 sleep stage, N2 sleep stage, N3 sleep stage, and REM sleep stage.

[0086] Examples 18 and 19 illustrate exemplary methods for automatic detection of sleep stages using three-class scoring.

[0087] In some embodiments, at least one of the N mandibular movement classes exhibits cortical activation.

[0088] In some embodiments, at least one of the N mandibular movement classes exhibits subcortical activation.

[0089] In some embodiments, one or more of the N mandibular movement classes represent obstructive apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and / or central hypopnea.

[0090] In some embodiments, one of the N mandibular movement classes represents bruxism. The measured rotational movement data also represents a mandibular movement amplitude of at least 1 mm, and if the movement is phase-dependent, a frequency established in the range of 0.5–5 Hz during at least three respiratory cycles, or a mandibular movement amplitude of more than 1 mm sustained and tonic for at least two seconds.

[0091] Bruxism during sleep is frequently reported by 5-10% of adults. It is often intermittent, fluctuates over time, sometimes disappearing for several weeks before recurring, and can occur during the night or for several consecutive nights. Bruxism is often recognized by the sleeper's partner as an unpleasant, loud grinding sound. This can lead to facial or temporal pain and signs of tooth enamel wear in the subject. Its origin is not well understood, but obstructive sleep apnea is considered one possible cause.

[0092] In some embodiments, one or more of the N mandibular movement classes represent loop gain, muscle gain that mobilizes the mandible during apnea or hypopnea or effort periods, a post-activation passive collapse point, and / or a pre-activation arousal point.

[0093] Further provided herein is a method for assisting in the characterization of sleep disorders, such as sleep-disordered breathing (SDB), in subjects with a mandible, the method comprising the following steps: - A data analysis unit receives rotational motion data from a gyroscope placed on the mandible of the subject via a data link; - A step in which the memory unit included in the data analysis unit stores N mandibular movement classes (Note: N is an integer greater than 1, and at least one of the N mandibular movement classes indicates the onset of a sleep disorder (e.g., the onset of sleep-disordered breathing (SDB)). < j < N) A mandibular motion class consists of the j-th set of rotational values, where each j-th set of rotational values ​​represents at least one velocity, velocity change, frequency, or amplitude of mandibular rotation associated with that j-th class); - A step of sampling rotational motion data during the sampling period using the sampling element included in the data analysis unit (thus obtaining sampled rotational motion data); - Using the data analysis unit, a step of deriving multiple measured rotational values ​​from the sampled rotational motion data; and, - A step of matching the measured rotational values ​​with N mandibular movement classes using the data analysis unit. Therefore, sleep disorders in patients can be detected comfortably and efficiently.

[0094] In some embodiments, the method further includes the following steps. - A step of measuring acceleration using an accelerometer (the acceleration indicates the movement and / or position of the subject's head and / or mandible); - A step of transmitting acceleration data measured from the accelerometer to the data analysis unit using the data link; - A step of sampling acceleration data measured during the sampling period using a sampling element, thereby obtaining the sampled acceleration data; - A step of deriving multiple measured acceleration values ​​from sampled acceleration data using the data analysis unit; and - A step of matching the measured acceleration values ​​with N mandibular movement classes using the data analysis unit. Note: Each j(1 < j < N) A mandibular movement class includes the j-th set of acceleration values, where each j-th set of acceleration values ​​represents the movement of at least one mandible or head associated with the j-th class. Using both an accelerometer and a gyroscope allows for effective identification of mandibular movement from the movement of the entire head.

[0095] In some embodiments, the method further includes the following steps. - A step of measuring magnetic field data using a magnetometer (the fluctuations in the magnetic field data indicate the movement of the subject's head and / or mandible); - A step of transmitting the magnetic field data measured from the magnetometer to the data analysis unit using the data link; - A step of sampling magnetic field data measured during the sampling period using a sampling element included in the data analysis unit, thereby obtaining the sampled magnetic field data; - A step of deriving multiple measured magnetic field values ​​from sampled magnetic field data using the data analysis unit; and - A step of matching the measured magnetic field values ​​with N mandibular movement classes using the data analysis unit. Note: j(1 < j < N) A mandibular motion class consists of the j-th set of magnetic field data values; each j-th set of magnetic field data values ​​represents at least one velocity or velocity change of mandibular motion or head movement associated with that j-th class.

[0096] In some embodiments, the method further includes the step of using the analysis unit to identify the movement of the subject's head based on data from a gyroscope and / or accelerometer and / or magnetometer.

[0097] In some embodiments, at least one of the N mandibular movement classes indicates bruxism. Furthermore, the measured rotational movement data exhibits a mandibular movement amplitude of at least 1 mm, a frequency established in the range of 0.5–5 Hz during at least three respiratory cycles if the movement is phase-dependent, or a mandibular movement amplitude greater than 1 mm sustained and tonically for at least two seconds. The combination of these parameters indicates bruxism, thereby enabling effective detection of bruxism. In some embodiments, the frequency range is 1.0–4.5 Hz, or 1.5–4.0 Hz, or 2.0–3.5 Hz, or 2.5–3.0 Hz.

[0098] The following describes specific embodiments for matching data (preferably sampled rotation, acceleration, and / or magnetic field data) to N mandibular movement classes. These embodiments include extracting features from the aforementioned data. These features include measured rotation values ​​and optionally measured acceleration values ​​and / or measured magnetic field values. After extracting the features, they are matched to one or more mandibular movement classes. Preferably, the mandibular movement classes to be matched to the features include central hypopnea, normal sleep, and obstructive hypopnea. Preferably, these features are matched to the mandibular movement classes using SHAP scores, and the match is interpreted and explained.

[0099] In some embodiments, the features are selected from a non-exhaustive list including: the central trend (mean, median, and mode) of the MM (i.e., mandibular motion meaning rotation, acceleration, and / or position as measured using a gyroscope, accelerometer, and / or magnetometer) amplitude; the MM distribution (raw or envelope signal): skewness, kurtosis, IQR, percentiles 25, 75, and 90; extreme values: minimum, maximum, percentiles 5 and 95 of the MM amplitude; variation trend: linear trend and coefficients of tensor product-based spline coefficients (S1, 2, 3, 4) from a generalized additive model for evaluating MM as a function of time; and the duration of each onset. It should be understood that such features refer to measured rotational, acceleration, and / or magnetic values, whether sampled and / or discretized. Preferably, the values ​​are sampled and discretized. It should be understood that the list illustrates exemplary embodiments and is not intended to limit the system.

[0100] In some embodiments, feature extraction includes isolating an event. An event is a series of mandibular motion data (preferably sampled rotational, acceleration, and / or magnetic data) that may be attributable to a single movement of the head and / or mandible. One specific type of event is normal breathing, e.g., normal breathing for a predetermined duration. The predetermined duration may be, for example, 2 to 20 seconds, or 5 to 15 seconds, 30 seconds, or 10 seconds. The range of the duration can be adapted to the intended use; for example, 30 seconds is suitable for identifying sleep stages, 10 seconds for identifying sleep bruxism or microawakenings, and 20 seconds for identifying respiratory events.

[0101] In some embodiments, the features are extracted according to the following steps 1 to 4. 1. Obtain sampled mandibular motion data. The mandibular motion data includes sampled rotational values ​​and optionally sampled acceleration values ​​and / or sampled magnetic field values. Preferably, the sampling rate is 1.0 to 100.0 Hz, or 2.0 to 50.0 Hz, or 5.0 to 25.0 Hz, preferably 10.0 Hz. Preferably, the obtained sampled mandibular motion data is obtained between 10.0 minutes and 12.0 hours, or between 20.0 minutes and 4.0 hours, or between 30.0 minutes and 2.0 hours. 2. Characterize the timestamps of mandibular movement events. 3. For each timestamp ti, perform the following steps: 3.a.ti confirm whether it is the beginning of the mandibular movement event. 3.b.ti is the beginning of mandibular movement events, - Assign ti to t_begin, then search for the end of the mandibular movement event (t_end); and, - Add indices to t_begin and t_end; 4. For each mandibular movement event E, perform the following steps: 4.a. Calculate the duration of the event. dt = (t_end - t_begin) 4.b. Measure the statistical distribution of mandibular movement data sampled during the event. Preferably, this involves calculating one or more features selected from a list including minimum, maximum, mean, median, mode, percentiles 5, 25, 75, 90, 95, skewness, kurtosis, and IQR. Additionally or alternatively, a General Additive Model (GAM) nonlinear model is used to estimate the MM amplitude and / or position using a spline function with respect to time t, and the coefficients of the spline function are extracted. Additionally or alternatively, a simple linear model is fitted to extract intersections and gradients from the mandibular motion, including amplitude and / or position. Optionally, connect all features. Next, mandibular movement events are compared with mandibular movement classes.

[0102] In some embodiments, exploratory data visualization, one-way analysis of variance, and paired Student's t-tests with Bonferroni correction are used to match mandibular movement events with mandibular movement classes. Preferably, in this procedure, the significance level is set to p=0.0001 to 0.01, more preferably p=0.001.

[0103] In some embodiments, machine learning methods (e.g., extreme gradient boosting, deep neural networks, convolutional neural networks, random forests) are used to classify the measured mandibular motion data into mandibular motion classes.

[0104] In some embodiments, the employed random forest algorithm uses 20 to 5000, or 100 to 2000, or 200 to 1000, or 500 decision trees. In some embodiments, each decision tree is constructed on a random subset of the features.

[0105] In some embodiments, during model development (i.e., training of the artificial intelligence model), the measured mandibular movement data is randomly divided into two subsets: the larger subset for model development and the smaller subset for model validation. In some embodiments, the larger set contains 60–80% or 70% of the measured mandibular movement data. In some embodiments, the smaller set contains 20–40% or 30% of the mandibular movement data. Preferably, a comprehensive minority oversampling technique (SMOTE) is used on the training set before developing the model.

[0106] In some embodiments, Lundberg's Shapley Additive Explanation (SHAP) method is used in model development to evaluate the contribution of multiple features to classification. Thus, the SHAP method can be used to interpret and explain the predictions made by the adopted machine learning model.

[0107] Certain aspects of this disclosure may be expressed alternatively or additionally as follows: In some embodiments, the system includes a sensing unit and a device for processing data related to disturbances that may occur during a subject's sleep. The processing device includes an identification unit adapted to identify first and second measurement signal streams. The first signal has a frequency located within a first predetermined frequency range. The second signal has a value of at least one intrinsic characteristic that characterizes head and / or mandibular movement located within a second predetermined range. The first and second predetermined frequency ranges consist of frequency values ​​of head and mandibular movement, respectively, that characterize the subject's sleep state. The identification unit is adapted to produce a trigger signal after observing that the first and second signals identified in the first and second streams are present for a first predetermined period. The identification unit is also adapted to identify a third signal in the first and second measurement signal streams, where the frequency and / or value of the at least one intrinsic characteristic represents the mandibular movement and changes in head position of the subject. The identification unit is connected to an analysis unit adapted to be activated under the control of the trigger signal. The analysis unit is also adapted to compare a third signal with a profile characterizing frequencies and / or values ​​associated with sleep disorders and to produce the results of the comparison. The present invention is based on the concept that, during a subject's sleep, their respiratory movements are controlled by a neural center in the brain, which in turn controls the muscles of the head and mandible attached to it, and these muscles then position the subject's head and mandible. The accelerometer and gyroscope each provide a time stream of measurement signals characterizing the movement of the head and mandible. By using the identification unit, it is possible to identify from these streams of measurement signals those that characterize the subject's sleep state, activate the analysis unit, and analyze sleep disorders affecting the subject during actual sleep.

[0108] Therefore, as can be seen, the movement of the mandible is not determined solely by the movement of the chest, but also directly by the central nervous system in the brain, which controls the muscles attached to it and positions the mandible. These also control the position of the head.

[0109] In fact, since tracheal traction, which is inevitably at the respiratory frequency, can cause head movement, measurements using both an accelerometer and a gyroscope are preferable. In practice, the gyroscope is more sensitive to the rotational movement of the mandible, which is acted upon by its own muscles under the direct control of the brain, than to the head movement that tracheal traction can produce, as indicated by an accelerometer. In addition to respiratory movement, when the central nervous system is activated, there is another signal with a large amplitude that is measured. However, the movement due to tracheal traction is damped by the elasticity of the tissues connecting the mandible to the rest of the head, and thus the movement can be transmitted passively. Thus, this is a spinal drive, i.e., a relatively imperceptible reflex of the diaphragm that produces tracheal traction, while the antagonist / acting muscles of the mandible provide direct movement, particularly through the action of the driving branches of the trigeminal nerve directly from the brain, i.e., trigeminal nerve drive. The gyroscope can effectively measure the rotational movement of the mandible, which is caused by the muscles of the mandible and is a result of the brain's direct action on the mandible. Therefore, by combining the signals from the accelerometer and gyroscope, the detection of the origin and nature of mandibular bone movement can be improved, and the determination of whether a person is sleeping can be improved.

[0110] Preferably, the sensing unit includes a magnetometer adapted to measure the movement of the subject's head and / or mandible. The device or unit includes a third input for receiving a third time stream of measurement signals from the magnetometer. The analysis unit is adapted to integrate the measurement signals from the magnetometer with the third signal. By using the magnetometer, the absolute positions of the head and mandible can be measured.

[0111] Preferably, the sensing unit includes an oxygen meter and / or a thermometer and / or an audio sensor and / or an electromyography unit and / or a pulse photoplethysmograph. The identification device or unit includes fourth and / or fifth and / or sixth and / or seventh and / or eighth inputs for receiving fourth and / or fifth and / or sixth and / or seventh and / or eighth time streams of measurement signals from the oxygen meter, thermometer, audio sensor, electromyography unit and pulse photoplethysmograph, respectively. The analysis unit is adapted to integrate the measurement signals from the oxygen meter, thermometer, audio sensor, electromyography unit and pulse photoplethysmograph with the third signal. The identification device or unit is then adapted to associate the measurement signals from the oxygen meter and / or thermometer and / or audio sensor and / or electromyography unit and / or pulse photoplethysmograph with the third signal. These measurement signals from oxygen meters, and / or thermometers, and / or audio sensors, and / or electromyography units allow for the inclusion of more measurement signals, thus improving the reliability of the sleep disorder analysis.

[0112] Preferably, the first predetermined range of frequencies is 0.15 Hz to 0.60 Hz. The identification unit is adapted to identify the first signal over the duration of at least two respiratory cycles of the subject. The second predetermined range consists of the amplitude value of the rotational movement of the mandible, which is, for example, on the order of 1 / 10th of a millimeter of normal breathing. The frequency range of 0.15 Hz to 0.60 Hz characterizes a situation in which the subject's head is, so to speak, semi-immobile, and therefore reflects whether the subject is asleep or in the process of falling asleep.

[0113] Preferably, the analysis unit is adapted to identify from among the third signals signals that characterize head rotation around at least one axis extending through the subject's head in the first and second streams. Head rotation is often closely associated with sleep-wakefulness, micro-awakening, cortical activation and / or subcortical activation, indicating sleep disturbances.

[0114] Further provided herein is a method for automatically detecting sleep stages from mandibular rotational movement data recorded by a gyroscope. The method may be a machine learning-based method according to one or more embodiments described herein. The method preferably includes the following steps: - A step of providing rotational motion data sampled from at least one subject; the sampled data may be provided by one or more sampling and processing methods as described herein. - A step of inputting the provided data into a machine learning classifier and calculating a prediction score; and - A step to determine the sleep stage based on the calculated score.

[0115] As can be seen, preferred embodiments of other methods described herein are also preferred embodiments of methods for the automatic detection of sleep or sleep stages. The data of the Method may be used in other methods that may be essentially therapeutic and input into such devices.

[0116] In some embodiments, sleep stages may include the following classes (classified by increasing complexity): (1) Two-class scoring (i.e., binary) for detecting the wakefulness or sleep state of the subject; (2) A three-class scoring system for classifying the subject's state of wakefulness, non-REM sleep phase, or sleep phase including REM sleep; (3) A four-class scoring system for classifying the sleep stages of subjects, including their state of wakefulness, light sleep (N1 and N2) stages, deep sleep (N3) stage, or REM sleep stage; and (4) A five-class scoring system to classify all sleep stages, including the subject's state of wakefulness, N1 sleep stage, N2 sleep stage, N3 sleep stage, and REM sleep stage.

[0117] Examples 18 and 19 illustrate and discuss methods for automatically detecting sleep stages using a three-class scoring system.

[0118] Apart from the detection of sleep-related disorders, the systems and methods described herein can also be used for the following exemplary applications: detection of sleep stages and / or monitoring of sleep quality in healthy subjects, older adults, or subjects suffering from abnormal sleep patterns. The detection of sleep disorders, whether inherently clinical or psychological, can be used to adjust treatment or tailor it to the needs of the subject. Furthermore, studying the impact of sleep behavior on the clinical outcomes of chronic diseases may provide new insights into the diseases and the effectiveness of treatments. The systems described herein can also be used in combination with other systems or methods. These systems may be used, at their discretion, inherently therapeutic (e.g., respiratory devices (CPAP, BiPAP, adaptive support ventilation), mandibular advancement orthoses, and oral devices, devices for percutaneous or implanted nerve and / or muscle stimulation, devices for correcting the posture and / or position of the body and / or head during sleep, etc.). In some embodiments, an alarm may be connected to the system. Alternatively, the system may be connected to or comprise a device having an alarm function.

[0119] The present invention may also be described by the following numbered embodiments, which, in addition or alternatively, are equivalent to the term “system” unless the context clearly indicates otherwise.

[0120] Embodiment 1 A combination comprising a sensing unit and a device for processing data (e.g., a processing unit related to disturbances that may occur during a subject's sleep). The sensing unit includes an accelerometer adapted to measure the movement of the subject's head and / or mandible, and a gyroscope adapted to measure the movement of the subject's mandible. The sensing unit is adapted to produce a measurement signal based on the measurements taken. The device includes first and second inputs for receiving first and second time streams of measurement signals from the accelerometer and gyroscope, respectively. The processing device includes an identification unit adapted to identify the first and second measurement signal streams. The first signal has a frequency located within a first predetermined frequency range. The second signal has a value of at least one intrinsic characteristic that characterizes the movement of the head and / or mandible located within a second predetermined range consisting of such value. The first predetermined frequency range and the second predetermined range consist of frequency values ​​of the movement of the subject's head and mandible, respectively, that characterize the subject's sleep state. The identification unit is adapted to produce a trigger signal after observing that the first and second signals identified in the first and second streams are present for a first predetermined period. The identification unit is also adapted, after producing the trigger signal, to identify a third signal in the first and second measurement signal streams, where the frequency and / or value of at least one unique characteristic represents the mandibular movement and head position changes of the subject. The identification unit is connected to an analysis unit adapted to be activated under the control of the trigger signal. The analysis unit is also adapted to compare the third signal with a profile characterizing frequencies and / or values ​​associated with sleep disorders and to produce the results of the comparison.

[0121] Embodiment 2 The combination according to Embodiment 1, characterized in that the sensing unit includes a magnetometer adapted to measure the movement of the subject's head and / or mandible, the device or unit includes a third input for receiving a third time stream of measurement signals from the magnetometer, and the analysis unit is adapted to integrate the measurement signals from the magnetometer with the third signal.

[0122] Embodiment 3 A combination according to Embodiment 1 or 2, characterized in that the sensing unit includes an oxygen meter and / or a thermometer and / or an audio sensor and / or an electromyography unit and / or a pulse photoplethysmograph, and the identification device or unit each includes a fourth and / or fifth and / or sixth and / or seventh and / or eighth input for receiving fourth and / or fifth and / or sixth and / or seventh and / or eighth time streams of measurement signals from the oxygen meter, thermometer, audio sensor, electromyography unit and pulse photoplethysmograph, and the analysis unit each is adapted to integrate the measurement signals from the oxygen meter, thermometer, audio sensor, electromyography unit and pulse photoplethysmograph into a third signal.

[0123] Embodiment 4 A combination of any one of embodiments 1 to 3, characterized in that the first predetermined range consisting of the frequencies is 0.15 Hz to 0.60 Hz, and the identification unit is adapted to identify the first signal over the duration of at least two respiratory cycles of the subject.

[0124] Embodiment 5 The combination according to any one of embodiments 1 to 4, characterized in that the second predetermined range consisting of the aforementioned values ​​includes at least one amplitude value of head movement indicating a change in head position.

[0125] Embodiment 6 A combination of any one of embodiments 1 to 5, characterized in that the analysis unit is adapted to identify from among the third signals a signal characterizing head rotation around at least one axis extending through the subject's head in the first and / or second stream.

[0126] Embodiment 7 A combination of any one of embodiments 1 to 6, characterized in that the identification unit is adapted to identify the movements of first and second signal streams characterizing the movement of the mandible and changes in head position of a subject, and the analysis unit is adapted to remove at least one feature used to identify the information characterizing the movement from the signal stream of the movement.

[0127] Embodiment 8 A combination of any one of embodiments 1 to 7, characterized in that the processing apparatus is adapted to apply preprocessing to the first and / or second streams by applying a bandpass filter and / or a lowpass filter and / or an exponential moving average and / or a calculation of the signal frequency entropy thereto.

[0128] Embodiment 9 A combination of Embodiment 7 or 8, depending on Embodiment 7, characterized in that the analysis unit is adapted to verify, during a second period, particularly 30 seconds, whether the at least one characteristic used to identify information characterizing the motion has a value that characterizes either a sleep state or a wake state, and the analysis unit is adapted to produce first data items indicating a sleep state and a wake state, respectively, if the at least one characteristic used to identify the information characterizing the motion and removed from the analyzed signals of the first and second received streams has a value that describes either a sleep state or a wake state.

[0129] Embodiment 10 A combination of any one of Embodiments 7, 9, or 8, depending on Embodiment 7, characterized in that the analysis unit is used to identify information characterizing the motion during a second period, particularly 30 seconds, and is adapted to verify whether the frequency and / or at least one characteristic removed from the analyzed signals of the first and second received streams has values ​​characterizing the sleep state of N1 and the REM sleep state, respectively, and the analysis unit is adapted to produce second and third data items indicating the sleep state of N1 and the REM sleep state, respectively, if the frequency and / or at least one characteristic characterizing the motion and removed from the analyzed signals of the first and second received streams has values ​​representing the sleep state of N1 and the REM sleep state, respectively.

[0130] Embodiment 11 A combination of Embodiments 7, 9, or 10, wherein the analysis unit is used to identify information characterizing the motion during a second period, particularly 30 seconds, and is adapted to verify whether at least one characteristic removed from the analyzed signals of the first and second received streams has values ​​characterizing sleep states N2 and N3, respectively, and the analysis unit is adapted to produce fourth and fifth data items indicating sleep states N2 and N3, respectively, if at least one characteristic characterizing the motion and removed from the analyzed signals of the first and second received streams, and used to identify the information, has values ​​representing sleep states N2 and N3, respectively.

[0131] Embodiment 12 A combination of any one of embodiments 1 to 11, characterized in that the analysis unit is adapted to verify, during a third period, particularly 3 to 15 seconds, whether at least one intrinsic characteristic of the analyzed signals of the first and second received streams has levels that characterize cortical activation and subcortical activation, respectively, and if the analysis unit is adapted to produce a sixth data item indicating cortical activation and subcortical activation, respectively.

[0132] Embodiment 13 A combination of any one of Embodiments 1 to 12, characterized in that the analysis unit is adapted to verify whether at least one intrinsic characteristic of the analyzed signal has levels that characterize obstructive apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and central hypopnea, and the analysis unit is also adapted to produce seventh, eighth, and ninth data items indicating obstructive apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and central hypopnea, respectively, if the at least one intrinsic characteristic of the analyzed signals of the first and second streams has levels that represent obstructive apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and central hypopnea, respectively.

[0133] Embodiment 14 A combination of any one of embodiments 1 to 13, characterized in that the identification unit identifies in first and second streams the value of at least one intrinsic characteristic indicating a frequency value and / or a variation not observed during sleep, produces a neutralization signal when such variation is observed, and supplies the neutralization signal to the analysis unit for neutralization.

[0134] Embodiment 15 A combination of any one of embodiments 1 to 14, characterized in that the analysis unit is adapted to verify whether at least one intrinsic characteristic of the analyzed signals of the first and second streams has increased by more than 1 mm at a frequency established in the range of 0.5 to 5 Hz during at least three respiratory cycles, if the movement is phase-dependent, or if it has increased by more than 1 mm sustained and rigidly for at least two seconds, and to produce a 10th data item indicating teeth grinding during the verification.

[0135] Embodiment 16 A combination of any one of embodiments 1 to 15, characterized in that the analysis unit is adapted to capture one or more values ​​of first and second streams that provide access to the calculation of loop gains of muscle gains mobilizing the mandible during periods of apnea, hypopnea, or effort, from the post-initiation passive collapse point and / or pre-initiation awakening point. Examples [Example 1]

[0136] In Example 1, Figure 1 is referenced. Figure 1 shows a system according to the present invention. The system includes a sensing unit 1 and a device 10 for processing data related to disturbances that may occur in a subject during sleep, preferably a processing unit. The sensing unit preferably includes an accelerometer 2 adapted to measure the movement of the subject's head and / or mandible in three dimensions. The sensing unit also preferably includes a gyroscope 3 adapted to measure the rotational movement of the subject's mandible in three dimensions. According to one preferred embodiment, the sensing unit 1 also includes a magnetometer 4, particularly in the form of a compass, and / or an oxygen meter 5, and / or a thermometer 6, and / or an audio sensor 7, and / or an electromyography unit 8, and / or a pulse photoplethysmograph 9. Other sensors, such as a sweat sensor or a nasal pressure sensor, may also form part of the sensing unit. The pulse photoplethysmograph operates by transmission or reflection and provides access to the calculation of pulse frequency and changes in arterial tone.

[0137] The sensing unit is preferably small in size, for example, a maximum of 5 cm in length, 2 cm in thickness, and 1 cm in height, so as not to disturb the subject's normal sleep. The sensing unit is preferably very small in overall size, lightweight, flexible, and allows for good ergonomics. The signals produced by the sensing unit are very suitable for decoding using artificial intelligence. The diagnostic power of the measurements obtained by the sensing unit is comparable to that of a complete polysomnography recording. Mandibular movement may preferentially occur along an axis, for example, the anterior-posterior axis, while the subject's head is turned to the right. Movement along other axes can be measured similarly. The sensing unit is preferably intended for single use for hygienic reasons, but can of course be readjusted and reused.

[0138] Preferably, the head position is measured based on values ​​measured along three axes by accelerometer 2. Since the accelerometer measures the value of acceleration relative to Earth's gravity, if there was no initialization phase during the application of the sensing unit to the human head, it is preferable to integrate these measurements over time to obtain the head position, which is the relative position. The position can be expressed, for example, as the values ​​of the pitch angle, roll angle, and yaw angle in Euler angles, or as a gain by a 15° tranche. The head position can also be expressed in the following terms: standing, lying down, left, right, supine.

[0139] The following table shows various angle values ​​and their estimated head positions: JPEG0007851639000001.jpg39115

[0140] The magnetometer 4 is added to sense the orientation of the head, especially when the movement occurs perpendicular to the gravity vector. By combining the values ​​measured by the accelerometer and the magnetometer, the distance traveled can be calculated, thereby obtaining the absolute value of the head position.

[0141] Regarding head movement, mandibular movement is preferably measured on three axes with the help of measurements from accelerometer 2. Mandibular movement is also measured with the help of gyroscope 3.

[0142] There are various types of movements of the head and mandible, and the resulting positional changes. In the case of the mandible, the movement is, for example, rotational movement at the respiratory frequency. However, in cases of teeth grinding, chewing, or oral dyskinesia, lateral movement is possible during sleep, and the condyle of the mandible rotates within the glenoid fossa of the temporomandibular joint, but these movements are not on the same axis as those in respiratory movement.

[0143] In the case of the head, the outcome of its movement is probabilistic; that is, the position the head occupies at the end of the movement cannot be predicted after activation. The amplitude of its movement and positional changes have different values. Therefore, if the amplitude of the head movement is large, the positional change of the mandible measured by the gyroscope cannot be examined, and in such cases the subject will be awake, and no information about the subject's sleep disorder can be obtained. Small amplitudes of mandibular movement captured by the gyroscope are observed when they are due to respiratory movement. Changes in the yaw angle are related to the head and indicate that the head is rotating from left to right. Changes in the pitch angle are related to flexion or extension of the head, in addition to the fact that they provide information about mandibular movement, despite the use of other parameters. These values ​​of the captured signals are analyzed with the help of the analysis unit described below.

[0144] Mandibular movement can be imposed by respiratory movement as much as non-respiratory movement. Therefore, head movement during sleep can cause mandibular movement. Mandibular movement can be caused by tracheal traction or the human brain. Tracheal traction is the traction force exerted by the chest on the head. This traction is at the human respiratory frequency. Therefore, when the head moves at the respiratory frequency, the mandible, which is attached to the head, moves at the same respiratory frequency, following its movement imposed by the head. This is a passive movement that follows the movement of the head. Mandibular movement can be directly and actively controlled by the brain, in which case the head does not move. When the brain controls mandibular movement, it is the muscles of the mandible that are directly stimulated. Therefore, it is useful to be able to clearly distinguish between mandibular movement controlled by the brain and mandibular movement controlled by tracheal traction.

[0145] Independent mandibular movements (IMM) during brain activation (e.g., at the end of a period of respiratory effort, during coughing or spitting, or again when speaking during sleep) are distinguished from respiratory mandibular movements (RMM) caused by the subject's breathing. There are also mandibular movements caused by teeth grinding and chewing. RMM-type mandibular movements are directly controlled by the subject's brain and do not lead to head movements. RMM-type movements are also caused by tracheal traction and are combined with head movements at the respiratory frequency. When RMM-type movements stop, normalize, or begin, it is useful to observe whether the head moved at that time with the help of accelerometer measurements. Teeth grinding-type movements often move the head and are followed by activation as shown by the accelerometer. This is because it actually captures this large amplitude movement in contrast to the relatively fine rotational movements of the mandible that are clearly shown by the gyroscope.

[0146] The apparatus 10 according to the present invention for processing data related to sleep disorders includes a first time stream F1 of the measurement signal from an accelerometer 2, i.e., a first input 11-1 for receiving measured acceleration data. This includes a second time stream F2 of the measurement signal from a gyroscope 3, i.e., a second input 11-2 for receiving measured rotational motion data. It may also include a third time stream F3 of the measurement signal from a magnetometer 4, i.e., a third input 11-3 for receiving magnetic field data. If the sensing unit also includes an oxygen meter, the identification device includes a fourth time stream F4 of the measurement signal from the oxygen meter, i.e., a fourth input adapted to receive oxygen meter data. If the sensing unit also includes a thermometer, the identification device includes a fifth time stream F5 of the measurement signal from the thermometer, i.e., a fifth input adapted to receive thermometer data. If the sensing unit also includes an audio sensor, the identification device includes a sixth time stream F6 of the measurement signal from the audio data, i.e., a sixth input adapted to receive audio data. If the sensing unit also includes an electromyography (EMG) unit, the identification device also includes a seventh input adapted to receive the seventh time stream F7 of the measurement signals from the EMG unit, i.e., EMG data. If the sensing unit also includes a pulse photoplethysmograph, the identification device also includes an eighth input adapted to receive the eighth time stream F8 of the measurement signals from the pulse photoplethysmograph, i.e., photoplethysmography data. In other words, measurement data from various sensors is transmitted from these sensors to the analysis unit via a data link.

[0147] Since the various streams mentioned above may be time-division multiplexed and / or each carried by a carrier wave of a different frequency, the various inputs must not be physically different. Therefore, the various input streams may be transmitted over a single data link.

[0148] The apparatus includes a data analysis unit, which includes an identification unit 12 adapted to identify a first signal whose frequency falls within a first predetermined range of frequencies and a second signal whose value falls within a second predetermined range of values ​​in first and second measurement signal streams F1 and F2. The first and second predetermined ranges each consist of vibrational values ​​of the movement of the subject's head and mandible that characterize the subject's sleep state. If the sensing unit includes a magnetometer 4, the identification unit 12 is also adapted to identify a third signal in a third measurement signal stream F3 whose value falls within a third predetermined range of the orientation of the subject's head that may be observed during sleep. The identification unit is adapted to produce a trigger signal after observing that the first and second signals identified in the first and second streams are present for a first predetermined period. After producing the trigger signal, the identification unit is adapted to identify a third signal in the first and second measurement signal streams whose frequency and / or value characterize the movement of the subject's mandible and / or changes in head position. The identification unit is connected to an analysis unit 13 which is adapted to be activated under the control of the trigger signal. The analysis unit is also adapted to compare the third signal with a profile that characterizes frequencies and / or values ​​associated with sleep disorders and to produce the results of the comparison.

[0149] In particular, the identification unit may be included in a data analysis unit which also includes a memory unit. The memory unit is configured to store N mandibular movement classes, where N is an integer greater than 1. At least one of the N mandibular movement classes indicates the onset of sleep-disordered breathing. < j <N) A mandibular motion class consists of the j-th set of rotational values, where each j-th set of rotational values ​​represents at least one velocity, velocity change, frequency, and / or amplitude of mandibular rotation associated with the j-th class. Each j-th mandibular motion class also optionally includes the j-th set of acceleration values ​​and / or the j-th set of magnetic field data values. The data analysis unit includes a sampling element configured to sample rotational motion data measured during a sampling period, and optionally measured acceleration data and / or measured magnetic field data, thereby obtaining sampled rotational motion data, and optionally sampled acceleration data and / or sampled magnetic field data. The data analysis unit is configured to derive a plurality of measured rotational values ​​from the sampled rotational motion data, and optionally derive a plurality of measured acceleration values ​​and / or measured magnetic field values ​​from the sampled acceleration data and / or sampled magnetic field data. The data analysis unit is further configured to match the measured rotational values ​​with the N mandibular motion classes. Optionally, the data analysis unit is further configured to match the measured acceleration and / or magnetic field values ​​with the N mandibular movement classes. Thus, the onset of sleep-disordered breathing is effectively detected.

[0150] Regarding the data link, the device and the sensing unit preferably communicate wirelessly, although cable connection is equally possible. The device is preferably part of a computer located in a data processing center. Wireless communication is performed, for example, with the help of a telephone network, and the sensing unit is equipped with, for example, a Bluetooth system capable of communicating with a telephone. Thus, a stream of measurement signals produced by the sensing unit is transmitted to the device.

[0151] This invention is based on the facts described below. That is, the movement of the mandible is determined not only by the movement of the chest, but also, as shown in the literature, by the control of muscles attached to the mandible, the role of which is determined by direct control from the brain's neural centers that position the mandible. As observed, the position of the head, and especially changes in the head during sleep, can stop or initiate all mandibular movement, completely independently of the aforementioned chest movement. That is, the movement of the mandible can follow the movement of the chest only if the position of the head allows it and does not fix it. Thus, the movement of the head can act on or paralyze the movement of the mandible. In this sense, it is nothing more than an incidental sign of brain activation that characterizes microarousal or arousal and can otherwise influence mandibular movement.

[0152] In fact, head movements affect upper airway patency by applying compressive force when the head folds during sleep or by activating / deactivating muscle motor units of the upper airway. These head movements during sleep alter upper airway patency, so it is necessary to understand and superimpose mandibular movements over time. Therefore, correctly analyzing these mandibular movements can be interpreted in terms of fluctuations in respiratory control, starting with the airflow produced by the sleeping subject. In other words, sensing and analyzing mandibular movements considering the head position and its changes during sleep, whether in a micro-awakening or waking state, involves taking into account brain control for positioning or repositioning the mandible by activating / deactivating the muscles attached to the mandible. Aside from brain activation, head position movements at respiratory frequencies are produced by tracheal traction, while mandibular movements at the same frequencies are directly determined by the nervous system.

[0153] By activating the movable bone formed by the mandible like a lever, brain control attempts to stiffen the upper airway by activating the muscles of the tongue and the pharyngeal area attached to it in order to avoid apnea. For this purpose, brain control relies on the muscles that raise and lower the mouth, and open and close the mouth, during sleep, at the respiratory frequency. Brain control can also act on the muscles that push the mandible forward at the respiratory frequency, or even act on a combination of these muscle groups involved in movement in different directions.

[0154] Changes in head position during sleep are often accompanied by awakenings or micro-awakenings, which are recorded, for example, by electrodes placed on the scalp. This records activity in the brain's cortex. However, if there is a change in the movement of the mandible, scalp electrodes may not record activation. This is because the activation remains subcortical and is sometimes purely autonomic in the brainstem. These head movements occur completely independently of chest movements.

[0155] A distortion of the neck occurs because the head position no longer coincides with the body position, or because the change in head position is an incidental sign of spontaneous or involuntary turning over under the control of the nervous system. Analyzing the movement of the mandible in the vertical and horizontal planes as a function of the resulting head position provides information about the level of respiratory effort, particularly its amplitude, employed by the brain's central control in response to changes in resistance to airflow through the airways. Respiratory events are considered to be increased effort when control from the nervous system increases, and central nervous system-dependent when control from the nervous system decreases. The brain control that allows an organism to break apnea requires the mandibular lever to act upward in the vertical plane and forward in the horizontal plane, ideally with the head aligned axially with the body, to prevent compression of the upper airway. (Micro) arousal itself is identified by independent large mandibular movements (IMM). Their duration is measured and clearly distinguished from following mandibular movements, whether respiration or non-respiration. [Example 2]

[0156] In Example 2, Figures 2A and 2B are referenced.

[0157] During sleep, the results of analysis that influence the measurement data stream from the sensing unit provide information, and changes in head position indicated by signals from the accelerometer often serve as markers of these state changes, particularly in the brain's control state. Figures 2A and 2B show the stream during which the head position of a person lying on their side in bed changes. This movement cannot be superimposed on the movement of the mandible during wakefulness, and therefore on the state of consciousness during chewing, vocalization, or swallowing, as has been studied by specialists outside of sleep medicine. The latter is related to chewing, vocalization, and swallowing problems studied in dentistry, oral medicine, maxillofacial surgery, orthodontics, logopedic, etc., in conscious subjects who are not in a sleep state.

[0158] Figure 2A shows, from left to right, the change in head position, first from a first position where the head is turned to the left, to a second position where the head is turned to the right. Subsequently, a change to a third position where the head is turned to the left again is observed. The first stream F1 produced by the accelerometer is related to the three axes (Fx, Fy, Fz) of the three-dimensional space in which the measurement is taken. The second stream F2 produced by the gyroscope is also related to these three axes. It is clearly visible that these two streams have high-amplitude peaks at the moment of head rotation. As can also be seen, when the head is in the first position, streams F1 and F2 have larger amplitude fluctuations and fluctuations in control intensity than indicated by Reference 1, particularly in the vertical y direction of stream F1, indicating an increase in brain control state. Furthermore, this is also seen in the stream Ft, which shows chest movement. Thus, the analysis unit can infer from these streams that the person is exhibiting increased and fluctuating respiratory effort.

[0159] When the head rotates and is in the second position, we can see that the amplitude of stream F1 is significantly reduced, to the same extent as stream F2. As shown in reference 2, the level of stream F1 decreases, which indicates that the mouth is open. We can also see that airflow F5 decreases, and oxygen flow may be lost (reference 3). As shown in reference 4, we can see that the amplitude of stream F2 also decreases. This indicates that the brain's control amplitude is lost. All of this indicates that the amplitude of effort decreases and breathing is affected (see airflow F5). This further activates the brain and produces commands. The head changes position again, turning to the left. After this, we can see that the amplitude of stream F2 increases and airflow F5 increases. Thus, as can be seen, brain control tends to normalize breathing.

[0160] Figure 2B shows that even slight changes in head position can trigger brain control. Figure 2B shows the changes when the head is rotated slightly to the right. Stream F1 initially shows an increase in brain control and respiratory effort, as indicated by arrow 1. As shown in Reference 2, as head position changes, the accelerometer (F1) can be seen to show an increase in amplitude and frequency indicating brain activation. Brain activation is clearly visible for 30 seconds in streams F8 (EEG) and F7 (EMG), and is magnified here (Reference 2). Next, the level of stream F1 (Reference 3) shows a brain control state with decreased amplitude and can be seen to be mandible elevated (mouth closed).

[0161] Surprisingly, the technology employed by the system according to the present invention provides information on the nature of mandibular movement during sleep, its central origin, and its control by the nervous system. Specifically, while it is necessary to maintain ventilation by stiffening the pharynx, thereby maintaining the subject's oxygenation, the head's extremities must ideally remain aligned with the body, particularly the trunk, during sleep. Therefore, mandibular movement must be interpreted as a function of the head's position and its changes. Otherwise, it will be impossible to understand why this movement stops, starts, or changes amplitude during sleep. [Example 3]

[0162] In Example 3, Figures 3A and 3B are referenced.

[0163] The techniques provided herein can be applied to the detection of bruxism. Known bruxism diagnoses involve administering electromyography (EMG) of the masseter and anterior temporalis muscles, and possibly the anterior temporalis muscle, during a laboratory polysomnography study. This study further includes audio-video recording. Because the demand for sleep recording is not proportional to the recording capacity of sleep laboratories, this study is costly, cumbersome, and difficult to access. The recordings are difficult to reproduce due to their overnight duration and the difficulty involved. Furthermore, tracking bruxism requires recording over several nights, as it is not systematically reproduced every night and may occur intermittently. Therefore, it is necessary to conduct the study at the subject's home under real-life conditions, without disrupting the natural progression of sleep. Results need to be provided quickly to optimally control bruxism and verify the effectiveness of treatment.

[0164] Currently, bruxism is not detected at home because there are no technical solutions to do so. Proposed solutions, such as surface electromyography (EMG) of the masseter and temporalis muscles, do not provide a definitive diagnosis. In fact, the only recordings of EMG activity of the masseter or temporalis muscles can be affected by nocturnal feeding activity or by the fatty media of the muscles interfering with the capture of their EMG activity. In the laboratory, video recording can confirm that mandibular movement and the resulting EMG activity are associated with bruxism.

[0165] The technical solution proposed by the present invention comprises recording the movement of the mandible, preferably in space, on three main axes with the help of the sensing unit, and then performing algorithmic analysis of the signals with the help of the analysis unit. This analysis makes it possible to identify mandibular movements that occur specifically and exclusively during the onset of bruxism, established by the detection of RMMA (rhythmic masseter activity), i.e., phase-dependent but sometimes tonic activity, during surface electromyography of the masseter muscle. The signal stream produced by the sensing unit is analyzed on the three axes, thereby also capturing lateral movements that are imposed during bruxism and can contribute to enamel wear. The mandibular movements known as bruxism are the result of the simultaneous action of agonist and antagonist muscles, including not only the mandibular ascending and descending group such as the anterior temporalis muscle, but also the medial and lateral pterygoid muscles, both of which are vastus medialis and vastus lateralis muscles.

[0166] Figures 3A and 3B show streams captured by the capture unit during bruxism. The recorded muscle EMG activity, observed as streams F7D and F7G, has been confirmed to contribute to mandibular movement. Typical features of masseter and / or anterior temporalis electromyographic activity are reflected in mandibular movement, which is also the etiology of bruxism. The latter are superimposed in the form of modulated signals on the tonic (persistent) or phase (rhythmic) electromyographic episodes of bruxism that produce them. The duration of this cycle or episode can be calculated.

[0167] Prior to the onset of bruxism, the effort period indicated by arrow 1 can be easily identified by analysis of mandibular movement and transient arousal indicated by arrow 2, accompanied by cortical or simply subcortical autonomic activation. Whether cortical or not, the activation is reflected only in changes in EEG cortical wave frequencies, as shown in stream F8, for example, or is not shown in EEG subcortically, and is well characterized by prior mandibular movement, often preceding the onset of bruxism, as described in the literature. As can be seen, the peaks in phase and / or tonic activity of the masseter muscle occur simultaneously with extreme positions of mandibular movement. This clearly confirms the relationship between muscle recruitment and the movement of the mandibular ossicles. In Figure 3A, in stream F1, the effort period indicated by arrow 1 is seen, followed by activation indicated by arrow 2, and then mandibular movement due to bruxism, indicated by arrow 3. Figure 3B is a magnified view of the 10-second period indicated by arrow K in the upper right of Figure 3A. Figure 3B shows the synchronization between EMG activity in the right masseter muscle (F7D) and the left masseter muscle (F7G) and mandibular movement during bruxism.

[0168] Here, we can see that the reactivated activity of the right masseter muscle (F7D) in stream F7 (EMG) is synchronized with the activity of the left masseter muscle (F7G) and the mandibular movement activity due to bruxism. As clearly shown in this figure, after a period of effort clearly shown in F1Z and F2X, the position of the mandible with abnormal amplitude changed with respiratory frequency. In F1Z, large movements accompanied by head movement follow, and in the gyroscope F2X, after movements indicating cortical activation, four rotational movements at high frequency (1 Hz) corresponding to the onset of bruxism follow. Subsequently, the period of effort is reproduced.

[0169] The movements of the head and mandible, analyzed through their inherent characteristics, namely, the frequency and morphological characteristics of the signal stream in particular, can be distinguished as a function of their production mechanisms and sequentially ordered over time. These characteristics can be observed, for example, by analyzing the amplitude, area, or gradient of the measured signal. They are, for example: • Movements related to respiratory effort, followed by • Movements associated with temporary cortical or subcortical activation, followed by Movements associated with teeth grinding or chewing that can be clearly distinguished, such as the number of attacks during a teeth grinding cycle, the length of the cycle between two attacks, and the duration of the attacks.

[0170] The movement of the mandible is produced by the action / antagonism of the muscles that move the mandible up and down. The latter is directly controlled by the core of the cranial nerve center of the trigeminal nerve drive branch. Here, the movement of the mandible can be sensed by the change in the angle the mandible exhibits while moving relative to a plane, for example, when moving perpendicular to a horizontal plane.

[0171] Mandibular movement may begin or cease only when the head position changes, even if chest movement continues. Changes in head position always occur simultaneously with cortical or subcortical micro-awakenings and thus disrupt control by the cranial nervous system. Mandibular movement may continue at the sleep respiratory rate even when abdominal and chest movement ceases, i.e., when the diaphragmatic muscles, which perform chest and abdominal expansion during spinal nerve-controlled inspiration, become non-functional or cease to function. Mandibular movement may also be exerted in another plane, such as the horizontal plane, in the form of front-to-back or back-to-front movement, i.e., in a plane other than the rostral-caudal traction plane in which tracheal traction is affected.

[0172] Similar to the second stream supplied by the gyroscope, the first stream supplied by the accelerometer shows a tension phase motion at the respiratory frequency of the mandibular position in an upward direction, i.e., the direction opposite to the direction in which the traction force produced by tracheal traction is exerted. These upward and forward movements are produced by the anterior temporalis and masseter muscles, and by the contraction of the pterygoid muscles, respectively, and particularly by the upper muscle group.

[0173] As respiratory effort begins and central respiratory control increases, the amplitude of mandibular movement increases, and the direction of mandibular movement may also be in a plane other than the vertical plane, which was the plane of tension. This is due to the action of certain muscle groups that are recruited more than others, such as the pterygoids. Movement at respiratory frequencies may occur in a more horizontal direction that is captured by an inertial unit, which consists of an accelerometer and a gyroscope. In fact, if the effort is observed only in the vertical plane, the effort period may escape signal analysis. The movement may also occur primarily in one direction (vertical or horizontal) rather than in another direction.

[0174] The shape of respiratory movements, particularly their acceleration gradient, changes as a function of the recruited muscle groups. During vertical movement, when the masseter muscle is active, the direction of inhalation is upward, opposite to the direction observed when the antagonist muscle is dominant, causing a decrease in movement, and this situation can lead to a change in the waveform of the movement.

[0175] The analysis of the stream supplied by the sensing unit allows for verification that the movement of the mandible during inhalation is downward when the activity of the descending muscles is dominant, and upward when the activity of the ascending muscles is dominant. This information is obtained by analyzing the captured changes in velocity and acceleration. This provides access to the level and nature of the subject's response to avoid respiratory onset, and the degree of mandibular ascending muscle recruitment to stabilize the upper airway. [Example 4]

[0176] In Example 4, Figure 4 is referenced.

[0177] The observed streams are the identification of four features that explain the behavior of the mandible during onset. From these features, the subject can understand the mechanism of respiratory onset at a specific head position during a specific stage of sleep, and the brain's response to self-release. In addition to explaining the progression of the onset, information on the risk of short-term and long-term recurrence can be identified. These features have predictive values, for example, when the amplitude of the response to the disorder, called loop gain, is high, i.e., when the response to the disorder is high. Figure 4 shows the loop gain. In this figure, arrow 1 indicates the collapse point of stream F1, that is, the solution where the exercise of brain control is lost, so that the mandible passively lowers under the influence of local anatomical constraints, such as weight determined by the subject's obesity. Arrow 2 shows the movement of the mandible that is also seen in stream 2. The amplitude between peaks of mandibular movement at the beginning of arrow 2 is low. Subsequently, the mandible lowers while the mouth is about to open. This can be seen at the level of the lowering stream 1, and the amplitude between peaks will increase. Next, the level of stream 1 reaches the level indicated by 3, which corresponds to the awakening point, followed by a much larger amplitude peak indicated by arrow 4. This large amplitude movement allows for the measurement of the loop gain, accompanied by the closing of the mouth, as indicated by the peaks of streams F1 and F2, and the maximum value that stream 1 reaches in between, even though the mouth has closed again. The loop gain indicates the response to the fault. This is calculated as the ratio of the difference between the notable points indicated by arrows 4 and 3 for the numerator and arrows 3 and 1 for the denominator.

[0178] Spontaneous recurrence of short episodes of apnea, particularly in a central form, is seen at high risk. The duration of episodes can be predicted by evaluating upper airway muscle gain, especially phase gain. Low muscle gain indicates a higher risk of longer episodes than high gain. The awakening point, the lowest point of mandibular position immediately before activation that terminates the episode, also allows for prediction of the duration of episodes. If this position does not descend sufficiently, there is a risk of recurrence, and in some cases, it occurs periodically. Furthermore, the influence of anatomical constraints, such as those related to weight and local accumulation of adipose tissue in the upper airway, can be measured, particularly by calculating the mandibular position based on the accelerometer measurements, during minute awakening or immediately after awakening when the latter is still dominant over the central nervous system, or during mandibular descent (collapse point). [Example 5]

[0179] In Example 5, Figures 5 and 6 are referenced.

[0180] The stream of measurement signals produced by the sensing unit may contain noise that affects the measurement signals. This can be used to preprocess the stream when it is received by the device. The principle of this preprocessing is simply to produce an enhanced signal. Those skilled in the art have been able to find from analysis that during a certain period of enhanced brain control, the position of the mandible, and therefore its velocity and acceleration, change periodically at approximately the same value between 0.15 Hz and 0.60 Hz, at the same frequency as the respiratory rate. For example, by low-pass filtering of the measurement signals from the accelerometer and gyroscope, it is possible to separate the signals related to micro-arousal by retaining only the lower frequencies in the respiratory frequency band. Figure 5 shows that by applying this preprocessing, micro-arousal representing activation is suppressed compared to the period of enhanced brain control. A clear peak is seen in the signal during each micro-arousal. The application of this preprocessing can be made, for example, by applying a 6th-order Butterworth filter, which is well known in the field of digital signal processing.

[0181] Conversely, by filtering one of the captured signals using a bandpass filter corresponding to the respiratory frequency band, it is possible to ensure a period of enhanced brain control. Figure 6 shows the result of applying this type of filter to the signal from the gyroscope. From this figure, it can be seen that the signal value is high during the effort period.

[0182] The characteristics used to identify information about the measurement signal stream include, for example, the following: - The position of the head and mandible (e.g., roll angle, pitch angle, and yaw angle) - Acceleration along the axes of the mandible and head - Rotational speed along the axes of the mandible and head - Standards of rotational velocity around one or more axes of the mandible and head (in space, if vector u has coordinates (x, y, z), its standard is written as follows: (x 2 + y 2 + Z 2 ) 0.5 ) - Standard acceleration along one or more axes of the mandible and head - The value measured over 10 or 30 seconds, or the median of the values ​​defined by two activations. - The average of values ​​measured over 10 or 30 seconds, or values ​​defined by two activations. - The maximum value measured over 10 or 30 seconds, or the value defined by two activations. - The minimum value measured over 10 or 30 seconds, or the minimum value defined by two activations. - Standard deviation of the value measured over 10 or 30 seconds, or the value defined by two activations. - Exponential moving average of measured values ​​(half-lives of 5, 60, 120, and 180 seconds) - Fourier transform and integral over all frequencies of the measured values, the respiratory frequency band (0.15~0.60 Hz), and the low frequency band (0~0.10 Hz). - Identification of the Fourier transform and the maximum energy frequency or second maximum energy frequency of the measured value. - Shannon entropy over a 90-second window of measurement - Time offset of rotational velocity and acceleration signals of the mandible, head, and other features to take past and future into account.

[0183] It is equally possible to combine the above methods with each other.

[0184] Once characteristics are identified in the stream of measured signals, the analysis unit can proceed with their analysis. For this purpose, artificial intelligence is used, for example, by calling a random forest type algorithm. The features thus extracted from the entire set of signal fragments from which the polysomnography results are known are injected into the algorithm in parallel with the expected results to produce a model that enables the classification of new fragments by pattern recognition type.

[0185] A signal pattern is a specific state of a signal sequence that can be represented physically or mathematically through parameters. Pattern recognition is the process of identifying (classifying) a specific pattern in a signal using an automated learning algorithm based on previously acquired information or statistical parameters extracted from the signal.

[0186] Deep learning is an automated machine learning technique that includes models inspired by the structure of the human brain, called artificial neural networks. These networks consist of multiple layers of neurons that enable the extraction of information from data and the production of results. This technique is highly effective for unstructured types of data such as images, sequences, and biological signals.

[0187] Automated learning (or statistical learning) is a field of artificial intelligence that aims to improve performance in solving tasks without explicitly programming each task by applying statistical modeling techniques that give machines (computers) the ability to learn information from data.

[0188] Artificial intelligence (AI) is a set of technologies aimed at enabling machines to simulate intelligent activity.

[0189] The development of these models can proceed, for example, as follows: 1) 200 subjects will be equipped with sensing units and simultaneously undergo polysomnography, a reference clinical test in the field of sleep. 2) Next, each random forest model is trained using signals captured from 40 of these subjects. The signals from the sensing units and a subset of features obtained after the preprocessing step are injected into the random forest algorithm in combination with reference results from the sleep examination to produce a classification model based on this input data. 3) Next, the remaining subjects are used for model validation. Signals from the sensing units corresponding to these subjects are injected into the model produced in the previous step to produce results, and these results are compared with the results obtained from polysomnography. If the results obtained from the model and the results obtained from polysomnography are considered to be in good agreement, the model is considered valid. Otherwise, development is restarted from step 2 of that section.

[0190] To ensure reliable identification of disturbances occurring during a subject's sleep, it is preferable to observe when the subject has actually entered a sleep stage. Once it is detected that the subject is actually in a sleep stage, it becomes possible to establish the subject's sleep stage so that the signals present in the stream of measurement signals can be correctly interpreted. Upon falling asleep, a respiratory oscillation frequency of, for example, 0.15 Hz to 0.60 Hz of the mandible is expected. To confirm a stable sleep state, this respiratory frequency needs to be present continuously for several tens of seconds. [Example 6]

[0191] Example 6 describes the various sleep stages of the subject. In particular, Table 1 (described below after the example) shows the various sleep stages of the subject and their relationship to the movement of the mandible and head position of the subject. Essentially characterizing the awake state is that the mandible makes unpredictable movements, whereas in the subject, sleep states without sleep disturbances are characterized by rotational movements of the mandible at respiratory frequencies. To detect the awake state and each sleep state, the analysis unit preferably operates using a 30-second analysis window and preprocesses the first and second streams using a bandpass filter and / or exponential moving average. To extract profiles characterizing the awake state and each sleep state, for example, normalized mean levels are considered. These levels are actually higher for the awake state than for the sleep state.

[0192] Furthermore, during sleep, stages N1, N2, N3, and REM (rapid eye movement) are distinguished. In the N1 sleep stage, changes in mandibular movement with respect to respiratory rate are observed, and in adults, fluctuations in peak amplitude are often seen over a limited period of several minutes. Head position is generally stable, but mandibular position remains unpredictable or may change periodically. To detect the N1 sleep stage using a processing device, it is preferable to use a 30-second analysis window to ensure continuity of movement. Preprocessing of the first and second streams by calculating the entropy of the signal frequency can be used. The normalized mean level is considered in the first approach as a profile characterizing the N1 sleep stage, but other approaches can also be used to improve the accuracy of the analysis. In the N1 stage, the normalized mean level will be higher than in the N2 or N3 stages.

[0193] The analysis unit is adapted to verify whether the normalized mean and variation of the amplitude and frequency of the received first and second streams during a second period, particularly 30 seconds, are at levels characteristic of the sleep state of N1. If the normalized mean and variation of the amplitude and frequency of the received first and second streams are at levels characteristic of the sleep state of N1, the analysis unit is adapted to produce a second data item indicating the sleep state of N1.

[0194] During sleep stages N2 and N3, fluctuations in the amplitude and / or frequency of brain control decrease progressively from N2 to N3. Therefore, in typical subjects, there is virtually no mandibular or head movement during these stages. To detect sleep stages N2 or N3 using a processing device, a 30-second analysis window is preferably used to ensure continuity of movement. Preprocessing with a low-pass or band-pass filter is also preferably used. The normalized mean level is considered in the initial approach as a profile characterizing the N2 sleep stage. The normalized mean level decreases progressively during stages N2 and N3, respectively. The normalized median level can also be used to identify the N2 or N3 stage or other statistical measurement techniques.

[0195] The analysis unit is adapted to verify whether the normalized mean and / or normalized median of the received first and second streams have levels that characterize sleep state N2 and sleep state N3, respectively, during the second period, particularly 30 seconds. If the normalized mean and / or normalized median of the received first and second streams have levels that characterize sleep state N2 and sleep state N3, respectively, the analysis unit is adapted to produce fourth and fifth data items indicating sleep state N2 and sleep state N3, respectively.

[0196] In humans, the REM phase is characterized by unpredictable mandibular movement. To detect this type of phase using a processing device, it is preferable to use a 30-second analysis window to ensure continuity of movement. In adults, this type of movement with unpredictable frequency and / or amplitude often lasts longer in the REM phase than in the N1 phase. The duration of such movement during the N1 phase is often limited to a few minutes. Due to the opening of the mouth, the direction of mandibular movement during brain activation is often negative. During the REM phase, fluctuations in mandibular movement at respiratory frequencies are observed, with non-periodic fluctuations in amplitude between peaks. Head position usually does not change during the REM phase. Detection is performed in a similar manner to the N1 phase, with the aim of observing respiratory instability during mandibular movement. The REM phase often begins without cortical activation that can be captured by EEG and without head movement. Therefore, the accelerometer measures nothing, but the gyroscope observes changes in mandibular rotation. This highlights the importance of having both gyroscope and accelerometer signals to correctly observe the transition to the REM phase. The end of the REM phase is often closely related to brain activation observed by the accelerometer and gyroscope. The accelerometer and gyroscope observe independent mandibular movement (IMM) and, if applicable, head movement. Normalized mean levels can be considered as a first approach. For example, amplitude and frequency variations can also be examined. Detection of REM during the first 15 minutes of sleep allows for the diagnosis of hypersomnia. [Example 7]

[0197] In Example 7, Figure 14 is referenced.

[0198] For example, comparative analysis during sleep of fluctuations in signal amplitude and / or motion frequency values, and / or other statistical characteristics of the signal, either individually or grouped into a classifier, such as a random forest type classifier, can be applied to practice statistical inference and distinguish between different stages. For this purpose, Figure 14 shows a spectrogram of the distribution of mandibular motion frequencies for distinguishing stages. In Figure 14, the vertical axis represents amplitude density and the horizontal axis represents frequency. These specific characteristics of each sleep stage can also be identified by machine deep learning. This algorithmic and / or statistical approach can also be used to characterize respiratory onset and non-respiratory motion onset.

[0199] The table below shows examples of fluctuations in the amplitude levels of mandibular rotation signals at various stages of sleep. In this table, "interval" refers to the interval between the upper level (percentile 2.5) and the lower level (percentile 97.5), "amplitude" refers to the difference between the maximum and minimum values, and "variation" is a measure of the spread of the values ​​considered. The measurements are based on 1000 samples acquired over 30 seconds at each stage. JPEG0007851639000002.jpg46131

[0200] The analysis unit is adapted to verify whether the normalized mean and variation of the amplitude and frequency of the received first and second streams during a second period, particularly 30 seconds, are at levels characteristic of a REM sleep state. If the normalized mean and variation of the amplitude and frequency of the received first and second streams are at levels characteristic of a REM sleep state, the analysis unit is adapted to produce a third data item indicating a REM sleep state.

[0201] The identification unit is adapted to identify motion signals characterizing the rotation of the subject's mandible and / or head movement in the first and second streams. The analysis unit is adapted to analyze these motion signals, for example, by applying a bandpass filter and an exponential moving average or a measurement of the signal frequency entropy to these motion signals. For example, by applying this bandpass filter to the first and second streams of the supplied signals for a first observation period of 30 seconds, with a half-life equal to, for example, 5, 60, 120, or 180 seconds, to the respiratory frequency and this exponential moving average, the analysis unit can observe whether the signal is unstable. In this case, an awakened state is observed. On the other hand, if the signal is stable, a sleep state is observed.

[0202] The analysis unit is adapted to apply an exponential moving average over a second period of 30 seconds to 15 minutes, particularly 3 minutes, of the first and second streams as a profile characterizing sleep states. In some analyses, the second period may be 30 minutes. The analysis unit is adapted to verify during the second period whether the exponential moving average has a substantially constant value, and if the value is substantially constant or not, to produce a first data item indicating a sleep state or a wakefulness state, respectively.

[0203] The identification unit is adapted to identify motion signals characterizing the rotation of the subject's mandible and head in the first and second streams. The analysis unit is adapted to calculate the entropy of the frequencies of these motion signals. By applying this entropy function to the first and second streams of the supplied signals, for example, with an analysis window of 90 seconds and an observation period of 30 seconds, the analysis unit can observe a normalized average level. If the level is high, an N1 or REM sleep state is observed as a function of the level value.

[0204] The identification unit is adapted to identify motion signals characterizing the rotation of the subject's mandible and / or head movement in the first and second streams. The analysis unit is adapted to apply a bandpass or lowpass filter to these motion signals. By applying the bandpass filter, for example, at the respiratory frequency, or the lowpass filter (e.g., less than 0.10 Hz), to the first and second streams of the supplied signals, and setting the observation period to 30 seconds, the analysis unit can observe normalized mean and / or median levels. As a function of these levels, sleep status N2 or N3 is observed.

[0205] Brain activation in the form of micro-arousal can last for 3–15 seconds and can be cortical or subcortical. Brain activation leading to arousal lasts for 15 seconds or more. Cortical cerebral activation during REM sleep may be characterized by repeated mandibular descents. In the case of cortical activation, the corticomedulla reflex is activated, and multiple sudden movements of the mandible with large amplitude or long duration are observed. The reflex amplifies the movement. In the case of subcortical activation, this reflex is not activated, and only one sudden movement of smaller amplitude may be observed, accompanied by a discontinuity in frequency relative to the respiratory frequency on which the mandible acted. This movement may have a much lower amplitude and shorter duration than when the corticobulbar reflex was activated. Therefore, this movement is often less conspicuous, and its identification may be aided by the detection of accompanying head movements that may occur only over very short distances. [Example 8]

[0206] In the further embodiment, Table 2 is referenced. Table 2 shows the characteristics of cortical and subcortical brain activation. As a result of cortical activation, the mandible suddenly opens and closes with a large amplitude for 3 to 15 seconds. When this cortical activation occurs during sleep, it is generally accompanied by a change in the subject's head position. The analysis unit analyzes the amplitude and duration of this movement in a 10-second window using first and second data streams.

[0207] A characteristic of subcortical activation is the discontinuity in the frequency and shape of the mandibular movement changes. The mandibular bone remains stable in most cases. The analysis unit analyzes the amplitude and duration of this movement in a 10-second window using the first and second data streams. This analysis can be performed similarly for continuous variables.

[0208] Accordingly, the analysis unit verifies whether the amplitudes of the received first and second streams of signals during the third period, particularly for 3 to 15 seconds, have levels that characterize cortical and subcortical activation, respectively. If the amplitudes of the received first and second streams have levels that characterize cortical and subcortical activation, respectively, the analysis unit is adapted to produce a sixth data item indicating cortical and subcortical activation, respectively.

[0209] To detect the presence or absence of respiratory events or non-respiratory movements, the analysis unit analyzes changes in mandibular position, amplitude between peaks of mandibular movement, fluctuations in amplitude between peaks of mandibular movement indicating variations in brain-controlled amplitude, and the frequency of mandibular movement. If low amplitude is observed, i.e., if the amplitude corresponding to the amplitude observed during normal respiratory movement is observed in the presence of stable central (mandibular movement occurring with a continuous and stable degree of mouth opening), there is no occurrence that warrants considering sleep disturbances.

[0210] If a high respiratory control amplitude is observed, for example, an amplitude corresponding to a movement greater than 0.3 mm, i.e., an amplitude change greater than the amplitude change observed during normal movement, an increase in motor or respiratory effort that may indicate sleep disturbance is presumed.

[0211] If a significant decrease in the amplitude of respiratory control is observed, i.e., if the centrality of control is stable or unstable, and for example for at least 10 seconds or two respiratory cycles, the onset of central respiratory onset is suspected to be at least on the order of 0.1 mm or zero.

[0212] By measuring the gain of upper airway muscle response during the onset of symptoms, obstructive, or marked, respiratory effort can be diagnosed, in contrast to central characteristics such as the absence of respiratory effort or respiratory effort falling below normal levels. This analysis allows for the characterization of apnea and hypopnea as obstructive or central. Normal levels of respiratory effort are measured in advance for each stage of sleep, during periods of normal breathing during sleep.

[0213] Changing head position can alter the configuration of sleep-wake transition events, with or without altering sleep stages. The gain of muscle response during onset is calculated by measuring the amplitude change between peaks during phased mandibular movement at the respiratory frequency during onset. This is a measure of the amplitude difference between peaks at the beginning and end of the onset period during phased movement, and can already be calculated from a single respiratory cycle that provides the gain value. This change can be minimal, on the order of 0.1 mm or less, but can reach 3 cm. This change may be accompanied by a change in the absolute position of the mandible, meaning the mouth is slightly open when phased displacement is applied. This change can occur in any direction between horizontal and vertical, considering head position during sleep. [Example 9]

[0214] In Example 9, Table 3 is referenced. Table 3 shows typical brain control behavior for detecting respiratory events and non-respiratory movement onset. As can be seen, to detect obstructive apnea or hypopnea, the analysis unit uses, for example, the median and / or mean values ​​of the first and second streams of measurement signals. To increase the reliability of the analysis, at least two respiratory cycles or a 10-second observation time would be desirable. Obstructive apnea or hypopnea is characterized by large brain control amplitudes at respiratory rates that can be repeated periodically or aperiodically, ultimately resulting in large mandibular movements during brain activation. In particular, the distribution of mandibular movement amplitude values ​​in the target stream is analyzed.

[0215] To detect respiratory effort-related arousal (RERA), the analysis unit proceeds as described above. To detect central apnea or hypopnea, the observation period will also be at least two respiratory cycles or 10 seconds. [Example 10]

[0216] In Example 10, Figures 4 and 7 are referenced.

[0217] The condition of teeth grinding is detected, for example, using the median, mean, maximum, or other statistical values ​​of mandibular rotation and its acceleration over a 30-second observation period.

[0218] Brain control after brain activation, with or without changes in head position, • Stable at low amplitude; • If the central aggravation is high or “enhancing”; if the mouth closes and the latter is obstructive, the onset is modified; • Low, or "reducing," centrally enhanced; open-mouthed, onset imposed, and obstructive; • Low, or "reducing," centrally decreasing; onset is imposed and central; • When the lateral pterygoid muscle is recruited, it is strengthened centrally, to a high or "enhancing" degree.

[0219] When the head position remains unchanged but the level of respiratory control changes, the position of the mandible and its changes continue to provide information about the level of respiratory control. Figure 7 shows signals indicating cortical cerebral activation and plots the mandibular movement measured by the accelerometer and gyroscope. Moving from left to right in Figure 7, the first thing seen are several oscillations indicating mandibular movement at a constant frequency. This movement of the mandible is caused by breathing with some effort. The subject needs to make an effort to allow air to pass through the upper airway. This can be confirmed by the amplitude of the signal from the gyroscope. Reference 1, in particular, shows micro-arousal caused by cortical activation. Next is a strong oscillation showing a larger amplitude movement following brain activation. What is seen next is an increase in the level of the signal. This indicates that the mouth is closed and the mandible has risen by a few tenths of a millimeter. If the mandible rises and the respiratory oscillation of that movement is still detected with a larger amplitude than normal, it can be inferred that there is a persistent obstructive onset as observed here. When the amplitude of movement decreases, respiratory control does not increase beyond normal levels, and respiratory effort can be said to be normalized. Reference 2 shows that micro-arousal triggered by subcortical activation produces signals with lower amplitude than cortical activation.

[0220] The series of results obtained after signal processing by the analysis unit can be presented, for example, in the following way. • Hypnosis: The stages of sleep and the evolution of wake / sleep transitions during recording; • The start and end times of the recording, the time spent in bed, and / or lying down; • Total sleep time; various effectiveness indicators; • Sleep fragmentation; for example, the number and indicators of micro-awakenings and wakefulness (activation), and the number and indicators of wake / sleep transition changes; • Number and indicators of respiratory and non-respiratory movement onset; For example, if central respiratory episodes accompanied by periodic, regular, and gradually increasing or decreasing fluctuations in cerebral control amplitude are repeated, and the duration of these episodes exceeds 40 seconds, it may be possible to suspect that the type of periodic respiration is progressing, possibly in a state of heart failure; • Periodic events may be obstructive episodes, or they may be periodic recurring obstructive episodes (for example, when loop gain is high or arousal is high); • Repeated subcortical activation unrelated to respiratory effort suggests a link to limb movement.

[0221] Head position affects the frequency and nature of respiratory and non-respiratory movement onset during sleep. Changes in head position during sleep always occur simultaneously with activation from the brain. During the latter, the head finds a new position, and the mandible, which moves with a large amplitude through multiple repetitions during this change, then finds a new position to receive respiratory drive again, the amplitude of which becomes a measure of the level of central control. Thus, as shown in Figure 7, there is an onset association. Therefore, active respiratory control, involving central activation, possible changes in head position, and consequently possible changes in mandibular position, as well as changes in the amplitude of mandibular respiratory movement, is integrated into the brain. The relationship between activity and brain activation was investigated. If the level of control activity changes, for example through changes in the amplitude between peaks of mandibular respiratory movement, this is first a result of central activation and changes in the control state in the brainstem. Furthermore, respiratory activity is captured by the gyroscope during mandibular rotational movement, while central activation is captured by the accelerometer during head linear movement.

[0222] Changes in head position and the resulting cerebral activation determine the risk of an outbreak, whether respiratory or non-respiratory movement-induced, by altering the control level and therefore the type of outbreak. Changes in head position inevitably involve brain activation, which can alter the state of airflow within the airways, and in particular, by altering the retention of the muscles in the upper airway ducts, the new orientation of the head may expose the airways to mechanical crushing forces.

[0223] The movement and repositioning of the mandible during cortical or subcortical activation in the onset of the disease can be explained as follows: (1) The mandible is passive, or while central motor control is suppressed for the duration of several respiratory cycles, the mandible decreases in deactivation without the tonic and / or phase-dependent support of muscular tissue. After the mouth closes, the relaxation of its position by passively opening the mouth, i.e., the mandible, is no longer supported because the tonic portion of muscular tissue that is considered to support it is lost by a variable distance but with a significant gradient (>1 / 10 mm / s). Measuring the distance between the lowest point recorded before closing the mouth and the subsequent change in gradient is a marker of passive disintegration of the pharynx when there is a loss of central control following activation. This situation can last for a time corresponding to several (up to 5) respiratory cycles. Because the central control (persistence of the toning portion) of the muscle tissue that controls the position of the mandible is not lost, this relaxation does not occur, and the mouth may remain closed or effectively closed. (2) Next, the mandible exhibits a phased and / or tonic morphology of muscle response gains, which re-determine the position of the mandible with the respiratory control frequency during the onset, before any new activation is induced. This is followed by the resumption of muscle activity controlling the position and movement of the mandible. This resumption of muscle activity may manifest as a change in gradient representing the new position, with or without respiratory movement, i.e., with or without a phase component, i.e., with an amplitude between at least one peak measurable beyond the background noise of the measurement (>0.05 mm). This latter movement indicates the resumption of respiratory movement, i.e., a change in respiratory frequency, and thus indicates respiratory effort, which allows us to identify the obstructive onset as central or mixed, according to the rules of assessment. Central movement allows us to identify whether the degree of mouth opening is stable, increasing, or decreasing, while the amplitude between peaks of respiratory movement reflects the degree of current effort. (3) The central or amplitude point is at its lowest, and from there, the action of closing the mouth is executed. The first action determined by activation is similar to the arousal threshold. This movement may be downward, for example, in the case of REM or when the mouth cannot open because the respiratory effort is exerted by the activity of the lateral pterygoid muscle and the masseter muscle that hold the mouth in a forward and high position. This movement is proof of activation. The latter may be cortical or subcortical, or subcortical and then cortical. If the mandible does not open much during onset, it may be due to the activity of the lateral pterygoid muscle and may suddenly open during activation. However, in most cases, since the mouth opens during onset, it is brutally closed by activation. (4) As shown in FIG. 4, the mandible position points continue at the maximum distance during activation from the arousal point. The distance separating them is a measure of the amplitude of the mandibular movement during activation. Its value is measured and compared with the level of respiratory effort developed during onset before being activated through the change in the amplitude of the respiratory movement from the start of the resumption of effort to the arousal point. The ratio of these values is a measure of the degree of loop gain of the mandible.

Example 11

[0224] In Example 11, FIG. 8 is referred to. FIG. 8 shows an example of the first measurement signal stream F1 (measured by an accelerometer) and the second measurement signal stream F2 (measured by a gyroscope) in a situation where the subject suffers from obstructive apnea. In this figure, F5n indicates nasal flow, and F5th indicates the thermal flow of the mouth and nose. As observed there, during the period from T1 to T2 and following the apnea indicated by Control 1, the signal is not stable. At the start of this period, it can be seen that the signal supplied by the gyroscope has a smaller amplitude than at the end of the onset. Since it is necessary to fight the obstruction that causes apnea or hypopnea, the central control is strengthened during onset. During the same period from T1 to T2, it can be seen that the accelerometer (Control 2) and the gyroscope (Control 3) show respiratory effort and subsequent brain activation (Control 4).

[0225] As shown in the signal analysis, when there is obstructive apnea between T1 and T2, the movement of the mandible at the respiratory vibration frequency is observed with the accelerometer (F1), and as the result of the position of the mandible increasingly drops from one respiratory cycle to another respiratory cycle, the amplitude increases from the amplitude between peaks simultaneously with the opening of the mouth (A). At the same time (C), it can be seen that the angular velocity of the rotational respiratory movement with the increasing amplitude indicates that the effort itself is increasing. Note the influence of the activation on the movement of the mandible measured with the accelerometer at the height of the character B. The activation causes an upward movement of the mandible, resulting in the closing of the mouth. On this occasion, the mandible reaches a new position. At the level of the character D of the gyroscope, it can be seen that this movement of closing the mouth is not purely rotational. The change in the signal state at the heights of the character B and the character D occurs simultaneously with the resumption of ventilation during the activation of the brain (microarousal).

Example 12

[0226] In Example 12, Figure 9 is referenced. Figure 9 shows an example of the first measurement signal stream F1 and the second measurement signal stream F2 in a situation where the subject is suffering from obstructive hypopnea, indicated by arrow 1. Arrow 0 indicates an aroused state. Figure 9 also shows the sixth stream F6 captured by an audio sensor indicating the presence of snoring, the seventh stream F7 detected by electromyography of the jaw, and the eighth stream F8 detected by electroencephalography. A series of mandibular movements (R) with larger amplitudes are observed for several seconds, each indicating cortical or subcortical activation. These movements are accompanied by changes in the peaks of streams F6, F7, and F8. Indeed, the electromyographic and electroencephalographic signals clearly indicate the presence of brain activation in this case. Obstructive hypopnea is indicated by arrows 2 and 3, where arrow 2 indicates effort and mouth opening, and arrow 3 indicates effort and mandibular rotation. This hypopnea is followed by activation in the form of micro-arousal, indicated by arrow 4. High values ​​of respiratory mandibular movement during micro-awakening reflect high respiratory effort, further emphasized by snoring. Thus, in stream F1 from the accelerometer and stream F2 from the gyroscope, it can be seen that snoring causes the mouth to open and the mandible to rotate. Therefore, Figure 9 shows that brain activity may be recorded by the accelerometer and gyroscope measuring mandibular movement as much as brain activation during periods of effort at respiratory frequencies, but in this case, it shows frequencies that are no longer normally respiratory frequencies. Also in Figure 9, the number 0 indicates the subject's state of wakefulness. [Example 13]

[0227] In Example 13, Figure 10 is referenced. Figure 10 shows examples of the first measurement signal stream F1 and the second measurement signal stream F2 when the subject suffers from mixed-type sleep apnea. As in Figure 8, Figure 10 shows an increase in the angular velocity of the mandible at frequencies corresponding to the respiratory frequency. The number 1 indicates a lack of respiratory flow closely related to the lack of control and effort shown by the number 2, and the subsequent recovery of brain control and effort shown by the number 3. [Example 14]

[0228] In Example 14, Figure 11 is referenced. Figure 11 shows examples of the first measurement signal stream F1 and the second measurement signal stream F2 in a subject suffering from central apnea. Peak F indicates head and mandibular movement during resumption of breathing. It can also be seen that there is virtually no mandibular movement between peaks F. The number 1 indicates the lack of effort indicated by the number 2, and the lack of respiratory flow closely related to the activation and resumption of effort indicated by the number 3. [Example 15]

[0229] In Example 15, Figures 12 and 13 are referenced. Figure 12 shows an example of the first and second measurement signal streams F1 and F2 when the subject suffers from a transient loss of all brain-derived control, characteristic of central hypopnea. This loss is characterized by the mouth opening passively because it is no longer supported by muscles. Thus, as can be seen in streams F1 and F2, the signals show no activity between peaks. On the other hand, at the moment of the peak, a large amplitude of mandibular movement is observed. Towards the end of the peak, movement corresponding to non-respiratory frequencies is seen, which is a result of brain activation and subsequently causes microarousal. Digit 1 indicates a period of hypopnea where a decrease in flow from the thermistor is clearly visible in stream F5. Digits 2 and 3 indicate the loss of mandibular movement in streams F1 and F2 during the period of central hypopnea. Figure 13 shows an example of the first and second measurement signal streams F1 and F2 when the subject experiences prolonged respiratory effort ending in brain activation. The signal from accelerometer F1 shows a large movement of the head and mandible at the position indicated by H. Subsequently, stream F2 remains substantially constant, but the level of F1 from the accelerometer decreases. This indicates that there is mandibular movement in any case, and it is slowly descending. This is followed by a high peak I, which is the result of the change in head position during activation that marks the end of the activity period. The number 1 indicates the effort during this prolonged period, marked by snoring. As indicated by the number 2, this effort can be seen to increase over time. This effort ends with brain activation that results in the head and mandibular movement indicated by the letter I, as indicated by the number 3.

[0230] The analysis unit has a memory model of these various signals, which are the result of processing using artificial intelligence as described above. The analysis unit processes the stream using these results and generates a report on the analysis of the results.

[0231] It was found that the accelerometer is particularly suitable for measuring head movement, while the gyroscope, which measures rotational movement, is particularly suitable for measuring rotational movement of the mandible. Therefore, brain activation that leads to mandibular rotation without changing the position of the head can be detected by the gyroscope. On the other hand, IMM type movement is detected by the accelerometer, especially when the head moves in this case. RMM type movement is detected by the gyroscope, which is very sensitive to it. [Example 16]

[0232] Further examples refer to exemplary procedures for feature extraction, data processing, and data description used in the methods and apparatus provided herein. These procedures are schematically shown in Figure 15.

[0233] Specifically, feature extraction, data processing, and description were performed using the R statistical programming language (8), while machine learning experiments were conducted using the sci-kit learn and SHAP packages in the Python language.

[0234] Twenty-three distinct features were extracted from the raw signal of mandibular movement for each onset, or every 10 seconds of normal respiration. These features include the central trend (mean, median, mode) of MM amplitude; MM distribution (raw or enveloped signal): skewness, kurtosis, IQR, percentiles 25, 75, and 90; extreme values: minimum, maximum, percentiles 5 and 95 of MM amplitude; variability trend: linear trend and coefficients of tensor product-based spline coefficients (S1, 2, 3, 4) from a generalized additive model for evaluating MM as a function of time; and the duration of each onset.

[0235] The influence of various features on the model classification of central hypopnea, normal sleep, and obstructive hypopnea can be explained by the SHAP score. The SHAP score measures the mean marginal contribution of all possible associations with other features to classify the three target labels. A higher SHAP score indicates a greater significant contribution provided by that feature. Lundberg's Shapley Additive Explanation (SHAP) method integrates Shapley's score in cooperative game theory (1953) (Lloyd S Shapley. “A value for n-person games”. In: Contributions to the Theory of Games 2.28 (1953), 307-317.) and local interpretation approaches (Marco Tulio Ribeiro, Sameer Singh, Carlos Guestrin. “Why should i trust you? Explaining the predictions of any classifier”. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM. 2016, 1135-1144.) to provide the best solution to date for explaining black-box models. SHAP theory considers input features as “players” in cooperative games where the “payout” correctly predicts the target label (i.e., central or obstructive hypopnea). The SHAP algorithm allows you to combine each feature with other features in a random order to form a system and assign a payout (SHAP score) to each feature according to its contribution to the overall prediction. The SHAP score is the average change in prediction that the association gains when a new feature is added. Essentially, the SHAP score of a feature is the average marginal contribution of that feature across all possible associations of a particular prediction.

[0236] Specifically, the aforementioned features can be extracted as follows: 1. Load the raw MM data sequence at a sampling rate of, for example, 10 or 25 Hz. The sequence has a significant duration, for example, 30 minutes to 8 hours; 2. Characterizing the timestamps of obstructive and central hypopnea onset; 3. Perform the following steps for each timestamp ti: 3.a. Determine whether ti is the onset of obstructive or central hypopnea. 3.b. If ti is the onset of obstructive or central hypopnea, - Assign ti to (t_begin), then find the end (t_end), and then, - Metrics are assigned to t_begin and t_end, and the raw data series is extracted into a temporary folder named "Onset E"; 4. Perform the following steps for each episode E: 4.a. Calculate the onset period dt = (t_end - t_begin), 4.b. Measure the distribution of parameters measured during the onset of the disease; - Minimum value, maximum value, mean, median, mode, percentiles 5, 25, 75, 90, and 95, skewness, kurtosis, and IQR; - Fit a GAM nonlinear model to estimate the amplitude and / or position of the MM using a spline function at time t, and extract the coefficients of the spline function; - Approximate a simple linear model and extract the intercept and linear gradient; - By matching the measurement data with the mandibular movement class, all features and labels are linked.

[0237] After feature extraction, the extracted features and their corresponding target labels were integrated into a tabular dataset.

[0238] Subsequently, exploratory data visualization, one-way analysis of variance, and pairwise Student's t-tests with Bonferroni correction were performed to classify the characteristics of mandibular movement into three groups (normal breathing, obstructive, and central hypopnea). The significance level was set at a very stringent criterion for null hypothesis testing (p = 0.001)(10).

[0239] For model development, the data were randomly split into two subsets (a large set (70%) for model development and a small set (30%) for model validation). Since the original training set was not balanced between central (minority class) and obstructive hypopnea (majority class), a synthetic minority over-sampling technique (well-known SMOTE, Synthetic Minority Over-sampling Technique) was applied to the training set before model development.

[0240] To classify the three groups using 23 input features, a multi-class classification rule was created. This was composed of a random forest algorithm that combined 500 different decision trees (each constructed with a random subset of 5 features).

[0241] Next, the content of the random forest model was analyzed to evaluate the importance of each feature and the associations that might contribute to classification (the potential combinations between them to distinguish obstructive hypopnea and central hypopnea). To evaluate the contribution of each feature to the prediction, the Lundberg's Shapley additive explanations (SHAP) method, well-known in the art, was adopted.

[0242] These methods enable, among other things, the detection of obstructive hypopnea and central hypopnea.

Example 17

[0243] In further examples, see Figures 16 and 17. These figures illustrate the analysis of mandibular motion data captured by a magnetic sensor. The data analysis itself is similar to the data analysis of mandibular motion data acquired by an accelerometer and / or gyroscope in addition to the magnetic sensor.

[0244] Figure 16 shows 18 of the most important MM signal features derived from magnetometer measurements, ranked by their global impact on the model's predictions. The bars represent the mean SHAP score for each feature, stratified by three target labels: central hypopnea (dark gray), normal (light gray), and obstructive hypopnea (gray)). The SHAP score measures the mean marginal contribution across all possible associations with other features to classify the three target labels. A higher SHAP score indicates a more significant potential contribution from that feature.

[0245] Figure 17 shows the interpretation of disease onset based on extracted features and SHAP scores. Specifically, Figure 17 shows the SHAP score scale and the probability of the target label. Figure 17 consists of two general regions, region a) and region b). Region a) contains extracted features that support the prediction of the target label, and region b) contains extracted features that point away from the target label. [Example 18]

[0246] In a further example, see Figure 18, which illustrates an exemplary method for measuring sleep stages from mandibular movement data captured by a gyroscope and accelerometer. The steps described below correspond to the reference numbers in Figure 18.

[0247] Specifically, these steps are as follows: (1) Using the system of the present invention, which includes a gyroscope and an accelerometer, the movement of the mandible during sleep of the subject is recorded. The resulting data pack includes six channels of raw signals obtained from the three-axis accelerometer and gyroscope sensors. The raw data may further include recordings from other devices suitable for measuring sleep stages, such as EEG, EOG and EMG signals for sleep stages, and six channels of MM signals obtained from the three-axis accelerometer and gyroscope sensors.

[0248] (2) After passing through the preprocessing and feature generation module, the raw data is split into segments of 30 seconds in length. The preprocessing consists of generating time series sampled from the sleep score series at 0.1 Hz and 0.034 Hz ​​(a 30-second sliding window) and time series obtained from the sensor and PSG. The preprocessing is performed in two steps: the series or sequence is split, and then the feature extraction function is applied to each window. Handcrafted feature extraction can be used as input data for machine learning experiments. For example, the feature generation module extracted a set of 1728 features from six channels of MM activity signals using a centrally located sliding window every 30 seconds. The extracted features include signal energy in the low frequency band (0-0.1 Hz), high frequency band (>0.3 Hz), or respiratory frequency band (0.2-0.3 Hz), exponential moving averages with several half-lives, energy entropy in several frequency bands, statistical features applied to the above features: centrality trend (mean, median), extreme values ​​(minimum, maximum), quartiles, standard deviation, and normal normal values ​​for all of the above features.

[0249] (3) The extracted feature set is sent to a machine learning classifier, which produces a soft predictive score (i.e., probability) and binary output for each target label according to a specific classification task. The automated sleep stage task was approached with three levels of complexity. The task targets are the three basic sleep stages: wakeful, non-REM (including N1, N2, and N3), and REM. Feature selection and hyperparameter tuning were performed using cross-validation, where the input data was randomly split into folds at the participant level. The final model was trained on all training sets using only the most relevant features and optimized hyperparameter values. Due to the imbalance in the ratios between target labels, the training data was balanced before each training session using a comprehensive minority oversampling method (SMOTE). The machine learning algorithm, the Extreme GradieNt BoostiNg (XGB) classifier, is employed as the core algorithm for all three classification tasks. The XGB classifier is optimized during the training process by minimizing a regularized objective function that combines a convex loss function (based on the difference between the predicted output and the target output) with a penalty term for the model's complexity. Model Training: The training objective is set to multi-class classification, aiming to classify three target labels according to a specific task. The training involved a dropout-multiple additive regression tree (DART) booster and a histogram-optimized approximate greedy tree construction algorithm. Log loss was selected as the evaluation metric (thus optimizing balanced accuracy across the three target classes). To prevent overfitting, the learning rate (eta, or step size reduction) parameter was set to 0.01. This reduces the feature weights and makes the boosting process more conservative. The model's output represents a softmax function that produces probability scores for each target label. The final decision (assigning only one label every 30 seconds) is then achieved by applying the function of argument that maximizes these three probability scores.

[0250] (4) The most satisfactory solution is adopted for implementation based on the period-by-period agreement between the model's predictions and the reference PSG scoring on an unseen validation dataset. Further quantitative evaluations were performed to verify whether the selected algorithm could provide reliable estimates of sleep quality scores such as TST, sleep efficiency, and REM ratio. The selection of the model was based on the following criteria: Class-by-Class Concordance Evaluation: A normalized confusion matrix allows us to evaluate the class-by-class performance of a model for a specific multi-class classification task. The rows represent the truth derived from manual PSG scoring, while the columns represent the results of automated algorithmic scoring. The diagonal cells of the confusion matrix show the true positive rate for each class. Precision (or positive predictive value) is defined as true positives / (true positives + false positives) and measures the mode's ability to correctly identify positive cases. Recall (also called sensitivity, hit rate, or true positive rate) indicates the usefulness of the model (defined as the proportion of correct classifications among all subject instances), where recall = true positive predictions / all positive instances; The F1 score is a composite metric defined as the harmonic mean of recall and precision for each class. 2 × (precision rate × recall rate) / (precision rate + recall rate) The F1 score has an intuitive meaning, indicating the accuracy of the model (the number of periods correctly classified) and the robustness of the model (a low misclassification rate). The actual data shows an unbalanced ratio between sleep stages, and since all labels are equally important, a classifier that achieves a consistently high F1 score across all classes is employed. Global period-based agreement metrics: Balanced Accuracy (BAC) measures the mean of true positive and true negative rates between target classes. Cohen's Kappa coefficient measures the strength of agreement between model classifications and actual observations (manual PSG scoring). This strength of agreement can be interpreted as six levels: <0: weak, 0-2: slight, 0.2-0.4: normal, 0.41-0.6: moderate, 0.61-0.8: fairly good, 0.81-1: near perfect.

[0251] (5) Predictive data from the selected (3-class task) model passes through the interpretation module. The first submodule (calculation of sleep score) converts the predicted sequence of sleep stages into a quantitative score. The definitions of these quantitative scores are shown in the table below: JPEG0007851639000003.jpg171140

[0252] (6) Creating Hypnotic Stages: A customized function converts a discrete sequence of encoded labels (e.g., 2 = awake, 1 = REM sleep, 0 = non-REM sleep) into hypnotic stages. The graph shows step lines representing the values ​​of individual sleep stages as a function of time, which simulates conventional hypnotic stages obtained from manual PSG scoring. [Example 19]

[0253] In Example 19, the experiment from Example 18 was continued. Specifically, the method shown in Example 18 was performed on a group of 96 participants randomly assigned to a training subset (n=68, 70%) and a validation subset (n=28, 30%). Both subsets represented healthy individuals aged 18–53 years.

[0254] The movement of the mandible was recorded during the subject's sleep using the system of the present invention, which includes a gyroscope and an accelerometer. The resulting data pack includes 6 channels of raw signals obtained from the 3-axis accelerometer and gyroscope sensors. These raw signals were used to develop an automated sleep stage model. Reference data was also recorded using devices suitable for determining sleep stages, such as EEG, EOG, and EMG. The latter data was used to determine the accuracy of the applied model.

[0255] Instead of using deep learning models, we followed a traditional framework. This means hand-drawn feature extraction and structured data-driven algorithms. Hand-drawn feature extraction provides better control and understanding of the input data compared to black-box models like convolutional neural networks. For classification tasks, we employed XGBoost. This algorithm has several advantages over traditional methods (LDA, SVM, RF), including high computational and resource efficiency, enabling very fast training and execution speeds.

[0256] Subject Subsets: Polysomnography (PSG) profiles from the group of 96 participants showed normal sleep activity in both subset groups, with median sleep efficiency of 89.4% and 87.3%. Within each set, the data structure also shows an imbalance in the ratios between the three sleep stages: non-REM sleep was predominant in both groups in most cases, with the exception of the normal wakefulness label (92.3 vs. 7.7 in the training set, and 79.9 vs. 20.1 in the validation set). This suggests that data balancing techniques are necessary in model development and that performance metrics should be carefully interpreted during model validation.

[0257] Three-Class Coring: The current model aims to classify wakefulness (no sleep), non-REM sleep, and REM sleep. The model achieves balanced accuracy across the three classes (82.9%, 74.9%, and 82.5% for wakefulness, non-REM, and REM sleep, respectively). The model also has a considerable agreement strength (Kappa = 0.71). The F1 score is 0.86, which is optimal for detecting wakefulness periods. Based on the distribution of wakefulness, non-REM, and REM instances within the model, it was found that wakefulness instances are well clustered and clearly separated from other instances, while most REM labels are more variable and mixed with other non-REM or wakefulness points, making wakefulness identification easier than distinguishing between non-REM and REM. As suggested by this pattern, nonlinear algorithms such as random forests, XGboost, and deep neural networks may be considered to better separate the three classes.

[0258] Consistent Analysis of Sleep Quality Indicators: A 3-class task sleep stage algorithm can automatically classify data into awake, non-REM, or REM every 30 seconds. The output is then transformed by a second algorithm to provide estimates of sleep quality indicators. These indicators can be categorized into three main categories: a) Time-based indicators, which measure the cumulative time (in minutes) during sleep (TST) or during specific sleep stages such as awake, REM, or non-REM; b) Ratio-based indicators, which are estimated as the percentage of a specific sleep stage (REM, non-REM) over the entire duration of sleep; c) Latency-based indicators, which measure the elapsed time from the start of recording to falling asleep (sleep-onset latency), or the elapsed time from falling asleep to the first REM period (REM latency).

[0259] The quantitative score of the automated sleep stage algorithm was measured according to the table shown in Example 18. The difference between the standard score of the PSG profile and the quantitative score of the automated sleep stage algorithm is shown in the table below: JPEG0007851639000004.jpg20781

[0260] These data demonstrate that a three-class-based scoring algorithm can measure total sleep time with acceptable accuracy (median difference was only -7.15 minutes; 97.5% CI was -20.34 to +4.38) compared to a reference method (manual PSG scoring). This agreement was also suitable for measuring sleep efficiency (median difference was -1.29%; -3.03 to +0.01).

[0261] Conclusion: The usability of mandibular movement recorded during sleep in subjects was investigated using the system of the present invention, which includes a gyroscope and an accelerometer. As shown in the results, automated sleep stage detection based on data measured by a gyroscope configured to measure rotational motion provides superior performance at all three resolution levels (in the case of three-class scoring) compared to the system of the art which includes an accelerometer alone.

[0262] JPEG0007851639000005.jpg215133JPEG0007851639000006.jpg21578

[0263] JPEG0007851639000007.jpg21584

[0264] JPEG0007851639000008.jpg214135JPEG0007851639000009.jpg21423

Claims

1. A system for detecting sleep events of a subject having a head and a mandible, the system including a sensing unit, a data analysis unit, and a data link configured to be attached to the mandible of the subject; wherein the sensing unit includes a gyroscope configured to measure the rotational movement of the mandible of the subject over a period of time, and an accelerometer configured to measure an acceleration indicating the movement and / or position of the head and / or mandible of the subject over a period of time; the data link is configured to transmit data of the rotational movement measured by the gyroscope and data of the acceleration measured by the accelerometer to the data analysis unit; wherein the data analysis unit includes a memory unit configured to store N mandibular movement classes, where N is an integer greater than 1, and one or more of the N mandibular movement classes indicate sleep events related to sleep disorders; wherein each j-th (1 < j < N) mandibular movement class includes a j-th set of rotational values, each j-th set of rotational values indicating at least one speed, speed change, frequency, and / or amplitude of the mandibular rotation associated with the j-th class, and / or each j-th (1 < j < N) mandibular movement class includes a j-th set of acceleration values, each j-th set of acceleration values indicating at least one of the movement of the mandible and / or the movement of the head associated with the j-th class; wherein the data analysis unit includes a sampling element configured to sample data of the rotational movement and data of the acceleration measured during the same period, thereby obtaining sampled data of the rotational movement and the acceleration; wherein the data analysis unit derives a plurality of measured rotational values and acceleration values from the sampled data of the rotational movement and the acceleration; compares the measured rotational values and acceleration values with one or more of the N mandibular movement classes indicating the sleep events; generates diagnostic information for diagnosing sleep events based on one or more of the N mandibular movement classes indicating the sleep events that have been compared, the diagnostic information including at least one of an indicator of a sleep event related to a sleep disorder and a quantitative parameter indicating the range and / or degree of the sleep event; and A system for creating a report containing diagnostic information configured to diagnose a sleep disorder associated with the detected sleep event. **Claim 2** The system according to claim 1, further comprising a magnetometer, the magnetometer being adapted to measure magnetic field data, variations in the magnetic field data indicating movement of the subject's head and / or mandible, where the data link is configured to transmit the magnetic field data measured from the magnetometer to the data analysis unit; where each of the jth (1 < j < N) mandibular movement classes includes a jth set of values of magnetic field data, and each jth set of values of magnetic field data indicates at least one velocity or change in velocity of the movement of the mandible or head associated with the jth class; where the sampling element is further configured to sample the magnetic field data measured over a period of time, thereby obtaining the sampled magnetic field data; where the data analysis unit is further configured to derive a plurality of measured magnetic field values from the sampled magnetic field data; and where the data analysis unit is further configured to match the measured magnetic field values with one or more of the N mandibular movement classes indicative of the sleep event, A system. **Claim 3** The system according to claim 1, wherein one or more of the N mandibular movement classes indicate a change in head position; and the data analysis unit is further configured to identify the movement of the subject's head based on the data of the rotational movement and acceleration, and to identify the movement of the head from the movement of the subject's mandible. **Claim 4** The system according to claim 1, wherein one or more of the N mandibular movement classes are characterized by a predetermined frequency range consisting of the frequency indicating the respiration of the subject, and at least one of the predetermined frequency ranges consists of a frequency of 0.15 Hz to 0.60 Hz. **Claim 5** The system according to claim 1, wherein one or more of the N mandibular movement classes indicate a respiratory disease event associated with a respiratory disease; and the respiratory disease includes obstructive apnea, mixed apnea, obstructive hypopnea, respiratory effort related arousal, central apnea, and / or central hypopnea. **Claim 6** The data analysis unit Analyze the sampled rotational movement and acceleration data over a period of time to observe one or more of the changes in the mandibular position, the position of the head, the peak-to-peak amplitude of the mandibular movement, the variance of the peak-to-peak amplitude of the mandibular movement, the frequency of the mandibular movement, and / or the variance of the frequency of the mandibular movement; and when one or more of the changes in the peak-to-peak amplitude of the mandibular movement and / or the frequency of the mandibular movement are observed, detect the presence of a sleep event The system according to claim 1, configured for this purpose.

7. One or more of the N mandibular movement classes indicate a bruxism event related to bruxism disorder, where the measured rotational movement data shows a mandibular movement amplitude of at least 1 mm, and when the movement is phasic, at a frequency established in the range of 0.5 to 5 Hz during at least 3 respiratory cycles, or shows a mandibular movement amplitude exceeding 1 mm continuously and forcibly for at least 2 seconds, the system according to claim 1.

8. The system according to claim 1, wherein the report including the diagnostic information includes the incidence rate and / or the cumulative period of the sleep event per hour.

9. A method for detecting a sleep event of a subject having a head and a mandible, comprising Measuring the rotational movement of the mandible of the subject using a gyroscope and measuring the acceleration indicating the movement and / or position of the head and / or mandible of the subject during a period of time using an accelerometer, wherein the gyroscope and the accelerometer are arranged on the mandible of the subject; Receiving, by a data analysis unit, the data of the rotational movement measured from the gyroscope and the data of the acceleration measured from the accelerometer via a data link; Storing, in a memory unit included in the data analysis unit, N mandibular movement classes, where N is an integer greater than 1, and one or more of the N mandibular movement classes indicate a sleep event related to a sleep disorder; where each jth (1 < j < N) mandibular movement class includes a jth set of rotational values, and each jth set of rotational values indicates at least one speed, speed change, frequency, or amplitude of the mandibular rotation associated with the jth class, and / or each jth (1 < j < N) mandibular movement class includes a jth set of acceleration values, and each jth set of acceleration values indicates at least one of the mandibular movement and / or the head movement associated with the jth class; Using the sampling element included in the data analysis unit, sampling the data of the rotational movement and the measured acceleration data during the same period, thereby obtaining the sampled rotational movement and acceleration data; Using the data analysis unit, deriving a plurality of measured rotation and acceleration values from the sampled rotational movement and acceleration data; Using the data analysis unit, matching the measured rotation and acceleration values with one or more of the N mandibular movement classes indicating the sleep event; Using the data analysis unit, generating diagnostic information for diagnosing a sleep event based on one or more of the N mandibular movement classes indicating the matched sleep event, the diagnostic information including at least one of an indicator of a sleep event related to a sleep disorder and a quantitative parameter indicating the range and / or degree of the sleep event; And Using the data analysis unit, creating a report including the diagnostic information configured to diagnose a sleep disorder related to the detected sleep event A method comprising.

10. The method according to claim 9, comprising Measuring magnetic field data using a magnetometer (fluctuations in the magnetic field data indicate movements of the subject's head and / or mandible); Using the data link to transmit the measured magnetic field data from the magnetometer to the data analysis unit; Here, each j-th (1 < j < N) mandibular movement class includes a j-th set of values of magnetic field data; each j-th set of values of magnetic field data indicates at least one velocity or velocity change of the mandible or head movement associated with the j-th class; Using the sampling element included in the data analysis unit, sampling the magnetic field data measured over a certain period, thereby obtaining the sampled magnetic field data; Using the data analysis unit, deriving a plurality of measured magnetic field values from the sampled magnetic field data; and Using the data analysis unit, matching the measured magnetic field values with one or more of the N mandibular movement classes indicating the sleep event A method further comprising.

11. The method according to claim 9, wherein one or more of the N mandibular movement classes indicate a change in head position; the method further comprises the step of using the data analysis unit to identify the head movement of the subject based on rotational motion and acceleration data, and identifying the head movement from the mandibular movement of the subject.

12. The method according to claim 9, wherein one or more of the N mandibular movement classes are characterized by a predetermined frequency range consisting of frequencies representing the subject's respiration, and at least one of the predetermined frequency ranges consists of frequencies from 0.15 Hz to 0.60 Hz.

13. The method according to claim 9, wherein one or more of the N mandibular movement classes indicate a respiratory disease event associated with a respiratory disease; the respiratory disease includes obstructive apnea, mixed apnea, obstructive hypopnea, respiratory effort-related arousal, central apnea, and / or central hypopnea.

14. The method according to claim 9, wherein one or more of the N mandibular movement classes exhibit a bruxism event associated with a bruxism disorder, wherein the measured rotational movement data exhibits a mandibular movement amplitude of at least 1 mm, and if the movement is phase-dependent, at a frequency established in the range of 0.5 to 5 Hz during at least three respiratory cycles, or a mandibular movement amplitude of more than 1 mm sustained and tonically for at least two seconds.

15. The method according to claim 9, Using the data analysis unit, the steps include: analyzing sampled rotational motion and acceleration data over a certain period and observing one or more changes in mandibular position, head position, peak amplitude of mandibular motion, variance of peak amplitude of mandibular motion, frequency of mandibular motion, and / or variance of frequency of mandibular motion; and If changes in the peak amplitude and / or frequency of mandibular movement are observed using the aforementioned data analysis unit, the presence of a sleep event is detected. Methods that further include this.

Images