JP6040874B2 - Sleep stage estimation device - Google Patents

Description

  The present invention relates to a sleep stage estimation device for estimating a sleep stage.

  The stage of sleep is divided into REM sleep, where the brain is awake, and non-REM sleep, where the brain is at rest, depending on the depth of sleep, breathing, electrocardiogram (heart rate), brain wave, blood pressure, skin It is known that biological information such as electrical activity changes. Therefore, various techniques for estimating the sleep stage based on such biological information are being developed.

  Conventionally, as an apparatus for estimating a sleep stage, for example, a sleep depth estimation apparatus described in JP-A-2004-089267 can be cited. This apparatus performs a sleep stage discriminant analysis using basic data obtained from a biological sensor, and estimates a sleep stage using a state transition model.

JP 2004-089267 A

  However, since there are individual differences in the sleep state, the estimation accuracy of the sleep stage may be lowered depending on the person. For this reason, it is required to perform estimation more suitable for the individual of the sleep stage.

  Then, an object of this invention is to provide the sleep stage estimation apparatus which can improve the estimation precision of a sleep stage by personal adaptation.

In order to solve the above-described problems, the present invention provides a sleep stage estimation device that estimates a user's sleep stage, a biological information acquisition unit that acquires biological information at the time of sleep of the user, and a feature amount from the biological information. A feature amount calculating means for calculating, a sleep stage estimating means for estimating the sleep stage of the user from the feature quantity, and an evaluation means for evaluating the sleep state after the user wakes up. When the estimated sleep stage ratio deviates from a predetermined allowable range, the sleep stage discrimination criterion for the user is corrected, and the feature amount calculating means determines that the sleep state evaluation result is In a bad case, when the ratio of the sleep stage estimated by the sleep stage estimation means deviates from a predetermined allowable range, the feature quantity is corrected when calculating the feature quantity for the user.

According to the sleep stage estimation apparatus according to the present invention, when the evaluation result of the sleep state is bad, when the sleep stage ratio estimated by the sleep stage estimation means deviates from the predetermined allowable range, it is suitable for the individual user. Since the sleep stage discrimination standard for the user is corrected as not being the discrimination standard, the discrimination standard can be learned to gradually adapt to the individual, and the sleep stage estimation accuracy can be increased. Further, according to this sleep stage estimation device, when the sleep state evaluation result is bad, when the ratio of the sleep stage estimated by the sleep stage estimation means deviates from the predetermined allowable range, it is suitable for the individual user. Since correction is performed at the time of calculation of the feature amount for the user assuming that the feature amount has not been calculated, the feature amount can be learned to gradually adapt to the individual, and the sleep stage estimation accuracy can be increased.

Moreover, the evaluation means of the sleep stage estimation apparatus according to the present invention may evaluate the sleep state using an evaluation input by the user.
According to this sleep stage estimation apparatus, since the user's subjective evaluation useful for sleep state evaluation can be used, the accuracy of sleep state evaluation can be increased.

Moreover, the evaluation means of the sleep stage estimation apparatus according to the present invention may evaluate the sleep state using a sleep history that is information related to the user's past sleep.
According to this sleep stage estimation device, the sleep state can be evaluated from the comparison with the sleep history in consideration of homeostatics (constancy) of the sleep state, so that the accuracy of the sleep state evaluation can be further increased. .

  According to the sleep stage estimation apparatus according to the present invention, the sleep stage estimation accuracy can be increased by personal adaptation.

It is a block diagram which shows one Embodiment of the sleep stage estimation apparatus which concerns on this invention. It is a graph which shows the example of transition of a user's sleep stage. It is a figure for demonstrating correction | amendment of the discrimination criterion of a sleep stage. It is a table | surface for demonstrating the example of the correction amount of a discrimination | determination reference | standard. It is a figure for demonstrating correction | amendment of a feature-value. It is a table | surface for demonstrating the example which needs correction | amendment of a feature-value. It is a flowchart which shows the flow of control of the sleep stage estimation apparatus which concerns on this embodiment. It is a flowchart which shows the flow of control of a personal adaptation process.

  DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

  The sleep stage estimation device 1 shown in FIG. 1 estimates a user's sleep stage by acquiring biological information. The main sleep stages include a deep sleep stage corresponding to non-REM sleep and a shallow sleep stage corresponding to REM sleep. There is also a wakefulness stage that awakens from sleep. The sleep stage estimation device 1 guides the sleep stage so that the user can be awakened at a preset wake-up time. For example, the induction method includes a method of applying a stimulus to the user.

  FIG. 2 is a graph showing an example of transition of the sleep stage of the user. As shown in FIG. 2, after entering the shallow sleep stage due to falling asleep, the user periodically repeats the deep sleep stage and the shallow sleep stage. Temporary awakening may occur during the light sleep phase. In general, it is known that awakening is poor when awakened in a deep sleep stage, and awakened is good when awakened in a shallow sleep stage. Therefore, the sleep stage estimation device 1 applies a stimulus such as sound or vibration to the user in order to guide the user's sleep stage to the awakening stage during a preset wake-up time.

  As shown in FIG. 1, the sleep stage estimation device 1 includes a calculation unit 2 that controls the entire device in an integrated manner. The calculation unit 2 is an electronic control unit including a CPU [Central Processing Unit], a ROM [Read Only Memory], a RAM [Random Access Memory], and the like. The calculation unit 2 is connected to a biological information acquisition unit (biological information acquisition unit) 3, an evaluation input unit 4, a display unit 5, and a stimulus output unit 6.

  The biological information acquisition unit 3 includes various devices that can acquire biological information when connected to a user. The biometric information of the user includes information on breathing, electrocardiogram (heartbeat), electroencephalogram, blood pressure, and electrodermal activity.

  The evaluation input unit 4 is a device for the user to input a sleep state evaluation after getting up. The evaluation input unit 4 may be a button type or a touch panel type. The display unit 5 is a display for showing the estimation result of the sleep stage estimation device 1 to the user. The display unit 5 and the evaluation input unit 4 may be integrally configured as a user interface.

  The stimulus output unit 6 is a device that gives a stimulus to the user in order to induce the sleep stage. Examples of the stimulus output unit 6 include a speaker that outputs sound, a vibrator that outputs vibration, and an illuminator that emits light.

  The calculation unit 2 includes a feature amount calculation unit (feature amount calculation unit) 10, a sleep stage estimation unit (sleep stage estimation unit) 11, a sleep stage guidance unit 12, a quality estimation unit 13, and an evaluation calculation unit (evaluation unit) 14. , Has a database 15. The database 15 stores sleep histories such as past sleep stage ratios and past sleep times. The database 15 stores the sleep history for each individual.

  The feature amount calculation unit 10 calculates a feature amount from the user's biological information during sleep acquired by the biological information acquisition unit 3. The feature amount is, for example, a heart rate per unit time extracted from heartbeat detection data as biological information, a fluctuation of a heartbeat period, a heartbeat intensity, and the like. In order to estimate the sleep stage in a multi-dimensional manner, feature quantities from various viewpoints are calculated from the biological information.

  The sleep stage estimation unit 11 estimates the sleep stage of the current user based on the feature amount calculated by the feature amount calculation unit 10. The sleep stage estimation unit 11 estimates the user's sleep stage by applying the feature amount to the sleep stage discrimination criterion stored in the database 15. Various known methods can be employed for estimating the sleep stage using such feature quantities. The sleep stage discrimination criterion is initially set based on a predetermined standard model, and is gradually corrected by personal adaptation described later. The estimation result of the sleep stage estimation unit 11 is stored in the database 15 in chronological order.

  The sleep stage guidance unit 12 guides the sleep stage based on the estimation result of the sleep stage estimation unit 11 so that the user can wake up naturally during a preset wake-up time. The sleep stage guidance unit 12 guides the sleep stage by giving the stimulus output unit 6 an appropriate stimulus to the user at an appropriate timing.

  The quality estimation unit 13 estimates the quality of sleep after the user wakes up. The quality of sleep in this embodiment is calculated | required as a ratio of the deep sleep in all the sleep time based on the estimation result of the sleep stage estimation part 11, for example. Note that the quality of sleep may be obtained by measuring the brain wave of the user during sleep and integrating the detection time of the δ wave that appears particularly during deep sleep. Alternatively, an appearance probability model using the deep sleep time of the δ wave as a parameter may be created, and the quality may be obtained from the deep sleep residence time.

  The evaluation calculation unit 14 evaluates the sleep state after the user is awakened. The evaluation calculation unit 14 evaluates the sleep state using an evaluation result (subjective evaluation result) input to the evaluation input unit 4 by the user after awakening. In addition, the evaluation calculation unit 14 may evaluate the sleep state using a sleep history (including past evaluation results) that is information related to the user's past sleep recorded in the database 15. In this case, the evaluation calculation unit 14 performs evaluation using homeostasis (constancy) of the daily sleep state.

  Next, personal adaptation by the sleep stage estimation apparatus 1 according to the present embodiment will be described. The sleep stage estimation device 1 is adapted to the individual user by correcting the sleep stage discrimination criterion and the feature amount based on the sleep result of the sleep stage estimation unit 11 and the evaluation result sleep by the evaluation calculation unit 14 (hereinafter referred to as the sleep stage estimation unit 11). Simply described as personal).

  FIG. 3 is a diagram for explaining correction of the sleep stage discrimination criterion in the sleep stage estimation unit 11. In FIG. 3, the vertical axis represents the speed of breathing, and the horizontal axis represents the speed of heartbeat. Further, the boundary between the deep sleep stage and the shallow sleep stage is shown as B1, the boundary between the shallow sleep stage and the awakening stage is shown as B2, and the boundary between the awakening stage is shown as B3.

  As shown in FIG. 3, the sleep stage estimation unit 11 estimates the sleep stage by plotting intersection points obtained from the respiration rate and the heart rate calculated by the feature amount calculation unit 10 on a discrimination criterion. I do. Note that the discrimination criterion shown in FIG. 3 is a simplified example, and a multidimensional discrimination criterion corresponding to a large number of feature amounts is actually used.

  When the evaluation result of the evaluation calculation unit 14 is bad, the sleep stage estimation device 1 has a ratio of sleep stages estimated by the sleep stage estimation unit 11 (for example, a ratio of the number of times of estimation of a deep sleep stage and a shallow sleep stage) is a predetermined tolerance. When deviating from the range, the sleep stage discrimination criterion for the user is corrected as not being a discrimination criterion suitable for the individual user.

  FIG. 4 is a table for explaining an example of the correction amount of the discrimination criterion for the sleep stage. FIG. 4 shows an example of the correction amount when the sleep stage estimated by the sleep stage estimation unit 11 is mistaken for the actual sleep stage. In FIG. 4, “real” items indicate actual sleep stages, and “false” items indicate sleep stages estimated by the sleep stage estimation unit 11. Further, correction 1, correction 2, and correction 3 shown in FIG. 5 are the number of times that the estimation results are different. As shown in FIG. 4, the sleep level estimation unit 11 increases the correction amount as the number of erroneous sleep stage estimations increases.

  B′1 shown in FIG. 3 is a boundary between the deep sleep stage and the shallow sleep stage in the corrected criterion. When the sleep stage estimation unit 11 estimates that the sleep stage is a shallow sleep stage, if the actual sleep stage is a deep sleep stage, the boundary B1 is moved to the shallow sleep stage side by + σ and used as the boundary B′1 Personal adaptation to the person. The sleep stage estimation device 1 performs learning to adapt the discrimination criterion to the individual user by repeating such correction. It should be noted that the estimation of the sleep stage can be determined based on the user's own evaluation or the result of the sleep stage estimation device with higher reliability.

  Moreover, in this sleep stage estimation apparatus 1, when the evaluation result of the evaluation calculation part 14 is bad, the ratio of the sleep stage estimated by the sleep stage estimation part 11 (for example, the ratio of the estimated number of times of the deep sleep stage and the shallow sleep stage) is calculated. When it deviates from the predetermined allowable range, it is assumed that the feature amount suitable for the individual user cannot be calculated, and correction is performed when calculating the feature amount for the user.

  The sleep stage estimation device 1 adds (variates) the data used for learning to the feature quantity, and operates (corrects) the feature value so that the variance value becomes large. The sleep stage estimation device 1 looks at the state by sequentially manipulating the number of factors that cover 90% of the feature values obtained by principal component analysis among the feature values used for estimation. Note that all feature quantities may be manipulated.

  FIG. 5 is a diagram for explaining the correction of the feature amount. FIG. 5 shows the feature amount before correction as T1 and the feature amount after correction as T′1. As shown in FIG. 5, the feature amount T1 before correction is estimated to be a deep sleep stage, but the corrected feature amount T′1 adapted to the individual is estimated to be a shallow sleep stage. The sleep stage estimation device 1 performs learning to adapt the feature amount to the individual user by repeating such correction.

  FIG. 6 is a table for explaining an example of correction of feature amounts. FIG. 6 shows the estimated number of sleep stages at 6 hours, which is an average sleep time, assuming that the sleep stage is estimated once every 30 seconds. The total estimated number of times in sleep time 6 hours is 720 times. Here, the ratio between the estimated number of deep sleep stages and the estimated number of shallow sleep stages will be described as the ratio of sleep stages (ratio of deep sleep stages to shallow sleep stages).

  In general, since the deep sleep stage is 20% of the total sleep time, when the total number of estimations is 720, the average number of shallow sleep stages is 576, and the number of deep sleep stages is 144. Times. When the sleep stage ratio deviates from a predetermined allowable range (a range including the average), the sleep stage estimation device 1 corrects the feature amount assuming that there is an error in the estimation result. For example, the allowable range can be a range of about ± 3% of the average. This allowable range can be set in consideration of the average and its standard deviation, and the allowable range may be changed (corrected) according to the individual user.

  When the estimated number of deep sleep stages is −5% from the average, the estimated number of shallow sleep stages is 612, and the estimated number of deep sleep stages is 108. In this case, since there is little data in the deep sleep stage, the variance is gradually added to each feature amount as −σ, −1.1σ,... −2.0σ so that the variation becomes small, and the ratio approaches the average. The feature amount is corrected as follows.

  When the estimated number of deep sleep stages is + 5% from the average, the estimated number of shallow sleep stages is 540, and the estimated number of deep sleep stages is 180. In this case, since there is a lot of deep sleep stage data, the variance is gradually added to each feature amount so as to increase the variation, and the feature amount is corrected so that the ratio approaches the average. The corrected feature quantity balanced in this way is calculated as the feature quantity for the user.

  Note that the above-described correction of the discrimination criterion and the feature amount is basically performed every night, but the necessity of correction may be determined using, for example, an average ratio in a 90-minute cycle. In addition, the estimated number of deep sleep stages may be calculated from the sleep state cycle of the individual week, and may be adjusted monthly. Furthermore, the amount of activity during the day (for example, the number of steps) may be input and reflected in the correction content. For example, if a person with an average number of steps during the day is 10,000 steps, and the number of steps counted on a certain day is 15,000 steps, even if the sleep stage ratio deviates from the average, Not applicable. Also, if alcohol is ingested, it is not subject to correction.

  Next, the flow of control of the sleep stage estimation apparatus 1 according to the present embodiment will be described with reference to the drawings. FIG. 7 is a flowchart showing a control flow of the sleep stage estimation apparatus 1.

  As shown in FIG. 7, the sleep stage estimation device 1 acquires the user's biological information by the biological information acquisition unit 3 in step S1. The biometric information acquisition unit 3 acquires various types of biometric information of the user at a predetermined cycle.

  Next, in step S2, the feature amount calculation unit 10 calculates a feature amount from the acquired biological information. The feature amount calculation unit 10 calculates various feature amounts used for estimating the sleep stage from the biological information of the user. At this time, the feature amount calculation unit 10 corrects the feature amount when the feature amount correction for the user is set by the past personal adaptation processing.

  Thereafter, in step S <b> 3, the sleep stage estimation unit 11 estimates the current user's sleep stage using the criterion stored in the database 15 based on the calculated feature amount. The sleep stage estimation unit 11 performs sleep stage estimation in a multiple manner based on a large number of feature quantities. The criterion stored in the database 15 is a criterion based on a preset standard model or a criterion corrected by past personal adaptation processing. The estimation result of the sleep stage estimation unit 11 is stored in the database 15.

  In step S <b> 4, the sleep stage guidance unit 12 determines whether or not it is near a preset wake-up time. The sleep stage estimation device 1 repeats steps S1 to S3 until it is determined that the sleep stage guide unit 12 is near the wake-up time. If the sleep stage estimation device 1 determines that the sleep stage guide unit 12 has reached the vicinity of the wake-up time, the process proceeds to step S5.

  In step S <b> 5, the sleep stage guidance unit 12 performs guidance (control) of the sleep stage for the user. The sleep stage induction unit 12 induces the sleep stage by applying an appropriate stimulus to the user at an appropriate timing by the stimulus output unit 6. The sleep stage guidance unit 12 guides the sleep stage so that the user is naturally awakened at a preset wake-up time. The sleep stage estimation device 1 displays the sleep stage estimation result on the display unit 5.

  In step S6, the sleep stage estimation apparatus 1 performs an evaluation input to the evaluation input unit 3 by the user after awakening. The user inputs a subjective evaluation about the sleep state such as the awake state and the sleep quality. The evaluation calculation unit 14 evaluates the sleep state using the evaluation input by the user and the comparison result with the past sleep history. The evaluation calculation unit 14 evaluates the sleep state with emphasis on the result of the evaluation input by the user.

  In step S7, the quality estimation unit 13 estimates the quality of sleep overnight. The quality estimation unit 13 estimates the quality of sleep using the input evaluation, the ratio of the estimated sleep stage, the brain wave of the user during sleep, and the like. Further, the quality estimation unit 13 estimates the quality of the current sleep using the sleep history such as the ratio of the past sleep stages and the past sleep time stored in the database 15. The sleep stage estimation device 1 displays a sleep quality estimation result on the display unit 5.

  In step S8, the sleep stage estimation device 1 determines whether the evaluation result of the evaluation calculation unit 14 is “good” or “bad”. If the evaluation result of the evaluation calculation unit 14 is “good”, the sleep stage estimation device 1 stores the sleep stage and the sleep quality estimation result in the database 15 and controls the sleep stage appropriately. finish. Note that the sleep quality may be estimated only when the evaluation result of the evaluation calculation unit 14 is “good”, and the estimation result may be stored.

  On the other hand, when determining that the evaluation result of the evaluation calculation unit 14 is “bad”, the sleep stage estimation device 1 determines whether or not the ratio of the sleep stages estimated in step S3 is within a predetermined allowable range. . When the estimated sleep stage ratio is within the allowable range, the sleep stage estimation device 1 stores the sleep stage and sleep quality estimation results in the database 15 and ends the control.

  If the estimated sleep stage ratio is not within the predetermined allowable range, the sleep stage estimation apparatus 1 proceeds to step S10. In step S10, a personal adaptation process for adapting the sleep stage discrimination criteria and feature quantities to the individual is performed.

  FIG. 8 is a flowchart showing the flow of control of personal adaptation processing. As shown in FIG. 8, the sleep stage estimation device 1 determines whether or not the sleep pattern is within an allowable condition in step S <b> 11 of the personal adaptation process.

  The sleep pattern is a bedtime, a sleep time (total time), and the like, and is an index for excluding a case where the sleep state is different from the normal time from a correction target for personal adaptation. Acceptable conditions include that the sleeping time is within 30 minutes of the normal average of the user and that the sleeping time is an error within 30 minutes of the normal average of the user. If the sleep pattern is outside the allowable conditions, the sleep stage estimation device 1 ends the process on the assumption that the sleep pattern is different from normal and should not be personally adapted.

  When the sleep stage estimation device 1 determines that the sleep pattern is within the allowable conditions, the sleep stage estimation device 1 proceeds to step S12. In step S12, the sleep stage estimation device 1 determines whether or not the sleep quality error estimated by the quality estimation unit 13 is within an allowable range. The sleep stage estimation device 1 is, for example, when the sleep quality is low with respect to the sleep time even though the sleep liability is high, or when the sleep average is not changed and the sleep liability is not high, but sleep compared with the daily life If the quality is too high, it is determined that the tolerance is exceeded. Sleep debt refers to a state in which sleep deficiency continues and accumulates.

  When it is determined that the sleep quality error estimated by the quality estimation unit 13 is within the allowable range, the sleep stage estimation device 1 ends the process. On the other hand, if the sleep stage estimation device 1 determines that the estimated error in sleep quality is outside the allowable range, the process proceeds to step S13.

  In step S13, the sleep stage estimation device 1 corrects the sleep stage discrimination criterion. The sleep stage estimation apparatus 1 corrects the discrimination criterion using the evaluation result of the evaluation calculation unit 14 and the error of quality estimation. The sleep stage estimation device 1 corrects the determination criterion so that the ratio of sleep stages (estimated number of deep sleep stages and estimated number of shallow sleep stages) approaches an average. This average may be an average based on a large number of human statistical results, or may be an average modified to fit the individual user.

  In step S <b> 14, the sleep stage estimation apparatus 1 sets feature value correction. As a result, correction is performed at the time of calculating feature values for the user. In the present embodiment, correction is performed by adding an individual correction value (variance value) to the feature amount. The sleep stage estimation device 1 sets the correction of the feature amount so that the sleep stage ratio approaches the average. Thereafter, the sleep stage estimation device 1 ends the process and waits until the next sleep.

  According to the sleep stage estimation apparatus 1 according to the present embodiment described above, when the sleep state evaluation result is bad, when the estimated sleep stage ratio deviates from the predetermined allowable range, it is suitable for the individual user. Since the sleep stage discrimination standard for the user is corrected as not being the discrimination standard, the discrimination standard can be learned to gradually adapt to the individual, and the sleep stage estimation accuracy can be increased.

  Further, according to the sleep stage estimation device 1, when the evaluation result of the sleep state is bad, when the estimated sleep stage ratio deviates from the predetermined allowable range, the feature amount suitable for the individual user is calculated. Since correction is performed at the time of calculating the feature amount for the user, the feature amount can be gradually learned to suit the individual, and the sleep stage estimation accuracy can be improved. As described above, the sleep stage estimation device 1 performs a large correction for correcting the sleep stage discrimination reference and also a fine correction for correcting the feature amount when there is a need to perform personal adaptation correction regarding the sleep stage estimation. By doing so, it is possible to make fine adjustments to the properties of the individual user, and the sleep stage can be estimated with high accuracy. Personal adaptation such as sleep stage threshold correction and feature amount gain adjustment can be adjusted off-line.

  Moreover, since this sleep stage estimation apparatus 1 utilizes the user's subjective evaluation useful as sleep state evaluation, it can improve the accuracy of sleep state evaluation for individual users. In addition to the user's subjective evaluation, the sleep stage estimation device 1 evaluates the sleep state from the comparison with the sleep history in consideration of the homeostatic (homeostasis) of the sleep state. The accuracy can be further increased.

  As mentioned above, although preferred embodiment of this invention was described, this invention is not limited to embodiment mentioned above. For example, the evaluation of the sleep state may be either a user's subjective evaluation or an automatic evaluation using a past sleep history. Moreover, you may employ | adopt another evaluation method. In addition, the sleep stage may be divided not only into a deep sleep stage and a shallow sleep stage, but also into finer stages. Furthermore, the aspect which performs only estimation of a sleep stage and sleep quality, and does not induce the sleep stage by a stimulus may be sufficient. Further, it is not always necessary to provide a display unit. Personal adaptation can be reset to the standard model (initial setting).

  DESCRIPTION OF SYMBOLS 1 ... Sleep stage estimation apparatus 2 ... Calculation part 3 ... Biometric information acquisition part (biological information acquisition means) 4 ... Evaluation input part 5 ... Display part 6 ... Stimulus output part 10 ... Feature-value calculation part (feature-value calculation means) 11 ... Sleep stage estimation unit (sleep stage estimation means) 12 ... Sleep stage induction section 13 ... Quality estimation section 14 ... Evaluation calculation section (evaluation means) 15 ... Database

Claims (3)

A sleep stage estimation device for estimating a user's sleep stage,
Biological information acquisition means for acquiring biological information during sleep of the user;
A feature amount calculating means for calculating a feature amount from the biological information;
Sleep stage estimation means for estimating the user's sleep stage from the feature quantity;
An evaluation means for evaluating a sleep state after the user awakens;
With
The sleep stage estimation means corrects the sleep stage discrimination criterion for the user when the estimated sleep stage ratio deviates from a predetermined tolerance when the sleep state evaluation result is poor ,
The feature quantity calculation means, when the evaluation result of the sleep state is bad, when the ratio of the sleep stage estimated by the sleep stage estimation means deviates from a predetermined allowable range, the feature quantity calculation means for the user A sleep stage estimation apparatus characterized by correcting the feature amount at the time of calculation . The sleep stage estimation apparatus according to claim 1 , wherein the evaluation unit evaluates a sleep state using an evaluation input by the user. The sleep stage estimation apparatus according to claim 1 , wherein the evaluation unit evaluates a sleep state using a sleep history that is information related to the user's past sleep.

Images