CN116134534A - Sleep assistance management based on optimized availability - Google Patents

Abstract

The melatonin optimization system detects the user's hormone sensitivity through sleep architecture monitoring and recommends personalized melatonin doses for each user based on the user's behavior, needs, and health conditions. Determining an optimized melatonin dosage requires accurate prediction of the non-interfering sleep onset latency of the upcoming sleep session, such that the dosage may be based on the difference between the user's desired sleep onset latency and the predicted non-interfering sleep onset latency. The system may use a general crowd-based sleep onset latency prediction model or a machine learning model trained to be personalized for each user.

Description

Sleep assistance management based on optimized availability

Cross Reference to Related Applications

The present application claims the benefit of U.S. provisional application No.63/054197 filed 7/20/2020. This application is incorporated herein by reference.

Background

1. Technical field

The disclosed concepts relate to methods and systems for effectively administering sleep assistance, and in particular to methods and systems for optimizing sleep by administering exogenous melatonin.

2. Background art

Of the 4023 U.S. adults investigated in the new consumer report, 27% reported that they were asleep or difficult to sleep most of the night, and 68% -or estimated 1.640 million americans-had sleep trouble at least once a week. Interestingly, about 5.2% of the 2002 NHIS candidate health/supplementation and candidate drug supplementation survey responders reported using melatonin, and 27.5% of these users reported insomnia as the cause of taking supplemental drugs, regardless of efficacy that has been demonstrated. Because people in the united states and other countries want to combat sleep problems, melatonin supplements are powerful in the over-the-counter market, gaining global value in the united states dollar 10 billion in 2020.

Indeed, exogenous melatonin has become one of the most frequently required over-the-counter sleep aids due to its regulatory role in the internal timing of biological rhythms, including promoting/regulating sleep. Melatonin is marketed to promote overall sleep time, to alleviate fatigue from jet lag, or to balance jet lag and circadian rhythms from shift work. There is evidence that melatonin can reduce the time required for a patient with delayed sleep syndrome to fall asleep, as well as help reset the sleep-wake cycle of the body.

Melatonin is secreted during darkness and less during the day. Thus, the use of exogenous melatonin is affected by light in addition to other user factors and behaviors, including age, stress, physical activity, diet, and other hormonal effects. In some studies it has been shown that melatonin in the dosage range (0.5-6 mg) in different formulations (fast and slow release) administered at different times before sleep time (0.5-2 hours) can improve some subjective and objective sleep parameters, as measured by actigraphy or polysomnography, but there is contradictory data. In a literature review of melatonin effectiveness, the general findings of studies conducted on fully healthy volunteers do not support the use of melatonin to initiate sleep or sleep efficiency and daytime sleepiness or somnolence. Inconsistent study results may be due in part to inter-individual variability. Despite the huge global market and the report of melatonin use by millions of us adults, the problem of proper dosage and formulation of melatonin for regulating circadian rhythm and sleep problems has not been solved. Accordingly, there is room for improvement in methods and systems for determining a personally relevant dosage regimen of melatonin.

Disclosure of Invention

Accordingly, in one exemplary embodiment, it is an object of the present invention to provide a melatonin optimization system for optimizing the effectiveness of exogenous melatonin in achieving a desired sleep outcome for a user, the system comprising: a user interface configured to accept information entered into the user interface regarding the user's health, self-reporting behavior, and desired sleep results; a sleep architecture detection module configured to perform monitoring of a sleep architecture of a user and detect hormone sensitivity of the user by monitoring; a behavior detection module configured to detect and collect information about user behavior to define detected user behavior; an initial dose algorithm module configured to define an initial recommended dose of melatonin for a user; a validity assessment module configured to determine a result difference between a desired sleep result of the user and a measured sleep result of the user; and a recommendation engine configured to define user intervention to reduce the result variance. The initial dose algorithm module is configured to define an initial recommended dose of melatonin based on information input to the user interface, and the recommendation engine is configured to define an intervention based on the resulting differences, the monitoring of sleep architecture, the detected user behavior, and the initial recommended dose of melatonin.

These and other objects, features, and characteristics of the present invention, as well as the methods of operation and functions of the related elements of structure and the combination of parts and economies of manufacture, will become more apparent upon consideration of the following description and the appended claims with reference to the accompanying drawings, all of which form a part of this specification, wherein like reference numerals designate corresponding parts in the various figures. It is to be expressly understood, however, that the drawings are for the purpose of illustration and description only and are not intended as a definition of the limits of the invention.

Drawings

Fig. 1 shows a schematic diagram of a melatonin optimization system in accordance with one exemplary embodiment of the disclosed concepts;

fig. 2 shows a schematic diagram of a more detailed variation of the melatonin optimizing system shown in fig. 1 in accordance with one exemplary embodiment of the disclosed concepts;

FIG. 3 shows a diagram of a two-process model illustrating sleep alertness;

FIG. 4 is a schematic diagram of a multimodal input sleep onset latency prediction module that may be included in the validity assessment module of either of the systems shown in FIGS. 1 and 2, according to one exemplary embodiment of the disclosed concepts; and

FIG. 5 is a flowchart including steps of a method for training a machine learning model to predict sleep onset latency of a user of any of the systems depicted in FIGS. 1 and 2, according to one exemplary embodiment of the disclosed concepts.

Detailed Description

As used herein, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise.

As used herein, the expression that two or more components or assemblies are "coupled" shall mean that the portions are connected or operated together, either directly or indirectly (i.e., through one or more intermediate portions or components), so long as a link occurs.

As used herein, the term "number" shall mean one or an integer greater than one (i.e., a plurality).

As used herein, the term "controller" will refer to a plurality of programmable analog and/or digital devices (including associated memory components or portions) capable of storing, retrieving, executing, and processing data (e.g., software routines and/or information used by such routines), including, but not limited to, field Programmable Gate Arrays (FPGAs), complex Programmable Logic Devices (CPLDs), programmable Systems On Chips (PSOCs), application Specific Integrated Circuits (ASICs), microprocessors, microcontrollers, programmable logic controllers, or any suitable processing device or apparatus. The storage portion may be any one or more of various types of internal and/or external storage media, such as, but not limited to, RAM, ROM, EPROM(s), EEPROM(s), FLASH, etc., which provide storage registers, i.e., a non-transitory machine-readable medium, for storing data and program code, such as in an internal memory area of a computer, and may be volatile memory or non-volatile memory.

As used herein, the term "intervention" refers to a recommended set of doses and/or behaviors of exogenous melatonin for a person seeking to alter many features of his or her sleep.

As used herein, the term "machine learning model" will mean a software system that develops and builds a mathematical model based on sample data (referred to as "training data") in order to make predictions or decisions without being explicitly programmed, including but not limited to computer software systems whose development has been trained to identify patterns from training data sets, and subsequently algorithms to identify patterns from training data sets in other data sets.

Directional phrases used herein, such as, for example, but not limited to, top, bottom, left, right, upper, lower, front, rear, and derivatives thereof, relate to the orientation of the elements shown in the drawings and are not limiting upon the claims unless expressly recited therein.

As described in greater detail herein in connection with various specific exemplary embodiments, the disclosed concepts provide methods and systems for effectively administering exogenous melatonin in order to meaningfully accelerate sleep onset. It should be noted that the administration of exogenous melatonin to achieve a timed function of sleep rather than hypnotic effects; that is, the administration of exogenous melatonin affects the time that a person falls asleep, but does not affect the total time that a person sleeps. Research literature indicates that the response to exogenously administered (i.e., ingested) melatonin is greatest when endogenous levels (i.e., natural body production levels) are not normally present, i.e., during the day. Conversely, the effect of administering melatonin during the time that it has been produced by the body (i.e., at night) is minimal. Melatonin causes a phase delay (shift to a later time) of the biological night of the body defined by the circadian rhythm of the body when taken late night/morning, and exogenous melatonin causes a phase advance (shift to an earlier time) of the biological night of the body defined by the circadian rhythm of the body when ingested in afternoon or evening.

The condition during melatonin administration appears to be very important and may determine the effectiveness of any given dose, particularly with respect to acute changes in Core Body Temperature (CBT) and somnolence. Thus, the disclosed concept provides a system for optimizing the effectiveness of melatonin treatment that detects hormone sensitivity through sleep architecture monitoring and personalizes dosing for the user's behavior, needs, and health. As used herein, the terms "dose," "amount," and "administration" include both the amount of melatonin taken and the time of melatonin intake relative to the desired sleep event or time of day.

Fig. 1 is a schematic diagram of a melatonin optimizing system 10 for maximizing melatonin effectiveness based on user needs, characteristics, and behavior, and fig. 2 is a schematic diagram of system 100, system 100 being a more detailed variation of system 10, according to one exemplary embodiment of the disclosed concepts. Fig. 2 also depicts some of the decisions performed by the components of systems 10 and 100. It should be noted that the behavior detection module 15 is depicted as being separate from the recommendation engine 16 in fig. 1, but as being part of the recommendation engine 16 in fig. 2, and that either configuration may be used without departing from the scope of the disclosed concepts. The system 10, 100 gathers information about the user's behavior, characteristics, sleep architecture, and desired sleep outcome as inputs to the system and then analyzes the inputs to make suggestions to help the user achieve the desired sleep outcome. In the exemplary embodiment, the data collection and analysis tasks of systems 10, 100 may be performed by any type of controller or computing system having input/output, processing, and memory capabilities, and any number or combination of such controllers or computing systems may be used without departing from the scope of the disclosed concepts.

Referring to fig. 1 and 2, the user interface 11 allows the user to provide information to the system 10, 100 regarding his or her age, sex, general sleeping time, timing and dosage of caffeine and alcohol intake, health conditions (e.g., hypertension, insomnia, diabetes, beta blocker intake, jet lag), and desired sleep outcome(s) (i.e., sleep onset latency, sleep/wake time, sleep duration, etc.). Such a user interface 11 may for example, but is not limited to, take the form of a mobile phone application or an internet-based portal. The initial dose algorithm module 12 provides initial melatonin doses and timing advice to the user based on age/gender/condition/need/desired sleep time data entered by the user into the user interface 11. The initial dose algorithm module 12 may be, for example and without limitation, pre-programmed to utilize widely available data regarding recommended doses based on problems recognized in a general population. Table 1 below provides a non-limiting example of widely available data that provides recommended multiple melatonin doses based on the user's specific conditions, needs, desired sleep outcome, etc.:

TABLE 1 recommended melatonin dosage for specific situations

The sleep architecture detection module 13 includes a plurality of sensors and/or trackers and a plurality of algorithms for detecting the user's sleep architecture. Any type of sensor, tracker, or other device/method for collecting sleep architecture may be used without departing from the disclosed concepts. Non-limiting examples of devices that may be used to detect sleep architecture include wrist-worn devices, mattresses with sleep trackers, and user-entered sleep logs. The effectiveness evaluation module 14 compares the user's desired sleep results (as indicated by the user's input to the user interface 11) with actual sleep results (as detected by the sleep architecture detection module 13), and the recommendation engine 16 (described in more detail below) then directs the user to change the dosage/timing of melatonin intake or to perform different interventions to change behavioral aspects based on the findings of the effectiveness evaluation module 14.

The activity detection module 15 collects and processes the sensor data and the user's self-reported information (input to the user interface 11) to detect information about the user's activity, including but not limited to the user's exposure to light, physical activity, food intake, and stress level. In one exemplary embodiment of the disclosed concept, the behavior detection module 15 includes at least: wearable sensors for exposure detection, pressure detectors, activity trackers, and inputs from the user interface 11 for self-reporting of food/alcohol/caffeine intake. The wearable sensor for exposure detection may, for example, but not limited to, detect and collect information about the duration, intensity, and timing of both sunlight and artificial light the user is exposed to throughout the day and until the user's sleep time. The pressure detector may be, for example, but not limited to, a wearable device that detects user physiological data such as heart rate variability and/or skin conductance. The evaluation of the pressure level may be performed by an algorithm on the device or by a third party through an Application Programming Interface (API) call from the pressure detector. The activity tracker may be, for example, but not limited to, a wearable device that detects the number of minutes a user engages in aerobic activity, wherein the aerobic activity is characterized by a heart rate level of the user that is between 55% and 85% of a maximum heart rate of the user, the maximum heart rate (maxHR) being calculated using the following equation (1):

maxhr=207-0.7 (age of user) (1)

The user interface for self-reporting of food/alcohol/caffeine intake enables the user to report the amount and time of food, alcohol and/or caffeine intake throughout his or her day.

The recommendation engine 16 comprises a melatonin dosage adjustment module 17 and a behavior modification module 18, the behavior modification module 18 evaluating melatonin dosage effectiveness and analyzing detected user behavior and user-stated preferences to recommend melatonin dosage adjustments and/or behavior modification to the user. More specifically, the dose adjustment module 17 determines the difference between the user's desired sleep onset latency and the measured sleep onset latency (or any other selected sleep metric) and recommends interventions based on the effectiveness of any previous melatonin interventions, current daytime activity, and previous night sleep. If the recommended intervention is melatonin administration and the previous melatonin dosage has reached a predetermined maximum allowable level, then only behavioral changes are recommended. The behavior modification module 18 calculates the difference between the recommended behavior for the user and the measured behavior of the user. If the difference in behavior exceeds a predetermined threshold, an intervention is recommended.

Behavior modification advice is provided only to the actual behavior being monitored. For example, and without limitation, if a particular implementation of the system 10 being used does not include a light sensor, the behavior modification module 18 will not provide an exposure recommendation. A non-limiting list of behavior change recommendations that may be provided by behavior change module 18 includes: instructing the user to perform an outdoor activity such as walking for at least 30 minutes per day, instructing the user to turn off artificial light at least two hours before the desired sleep onset time, instructing the user to perform a respiratory exercise to help reduce the user's stress level, and instructing the user to avoid ingestion of caffeine after a particular time of day.

Referring again to the effectiveness module 14 of the system 10, 100, in addition to comparing the user's desired sleep outcome with the actual sleep outcome, the effectiveness module 14 may also use a variety of statistical or empirical techniques to predict non-interfering sleep onset (i.e., naturally occurring sleep onset) and use the predicted non-interfering sleep onset as part of the assessment of the effectiveness of any current intervention. The recommendation engine 16 may then use the predicted non-interfering sleep onset to determine recommended interventions for the upcoming night (or other sleep cycle). Referring to FIG. 3, one non-limiting example of a technique that validity module 14 may use to predict sleep onset is a two-process model of sleep alertness. The two-process model of sleep alertness uses the time of self-wake, the time of self-fall and the time of day to generate a sleep-dependent curve (process S) and a sleep-independent circadian curve (process C). The process S curve shows that alertness is highest when waking up and steadily decreases during the day, reaching its lowest level at the beginning of sleep (represented by the lowest value on the process S curve). The process S reverses at the beginning of sleep and is called S' until it wakes up. The S' portion of the process S curve indicates that the homeostasis sleep liabilities are alleviated during sleep.

Still referring to fig. 3, the predicted alertness level at time of day is represented as the arithmetic sum of the S and C curves (depicted as dashed line s+c in fig. 3). Sleep Onset Latency (SOL) at any given time of day can then be predicted using equation (2) below:

y ＝ 0.56*10 ^0.12x (2)

where x is the predicted alertness at a particular time of day (i.e., found by locating the s+c curve point corresponding to the time of day), and y is the sleep onset latency. It should be appreciated that the constants 0.56 and 0.12 in equation (2) are derived using known parameter estimation methods, where these constants are derived by defining a best fit relationship between the observed alertness level and sleep onset latency of several study subjects. More specifically, for the lowest level of predicted alertness where x=1, sleep onset latency y≡0.5 minutes is found.

Validity module 14 may include a SOL prediction module 140 that incorporates a machine learning model 145 (both SOL prediction module 140 and machine learning model 145 are described in more detail with respect to fig. 4 and 5) to predict sleep onset latency n hours prior to a user's desired sleep time. The preferred sleep time of the user may be represented by the user through the user interface 11 or may be automatically detected from historical sleep architecture data collected by the sleep architecture detection module 13 by averaging the detected sleep times during a week of use and, if different sleep routines exist in the two parts of the week, averaging the sleep times on weekdays and weekends, respectively. In one non-limiting exemplary embodiment of the disclosed concept, the machine learning model 145 may be configured to implement personalized SOL predictions tailored to a particular user of the system 10, 100, rather than predicting sleep onset latencies using equation (2) above, equation (2) is considered suitable for a general population based on data compiled from several study subjects.

Referring to fig. 4 and 5, fig. 4 shows a multimodal input SOL prediction module 140 including a machine learning model 145, and fig. 5 shows steps of a method 200 for training the machine learning model 145 to predict sleep onset latency, the machine learning model 145 may be trained to make user-personalized SOL predictions by analyzing sleep architecture data collected by the sleep architecture detection module 13. The effect of training the machine learning model 145 for personalized SOL prediction may be to find constants that are more personalized to the user than the constants 0.56 and 0.12 used in equation (2), i.e., y=0.56×10 ^0.12x And even define relationships different from those defined in equation (2) to predict the sleep onset latency of the user. The accuracy of the personalized SOL prediction function that allows for multiple inputs (as done by the multi-modal input SOL prediction module 140) may be maximized by minimizing the value of the error function defined in equation (3) below:

where X represents all required parameters (all of which are parameters selected as inputs to the multi-modal input SOL prediction module 140) and K is the number of days in the training phase. It should be appreciated that there are many known techniques for solving the minimization of the error function, such as, but not limited to, gradient descent techniques, and that any known optimization algorithm or other technique for minimizing the error function defined in equation (3) may be used without departing from the scope of the disclosed concepts.

Still referring to fig. 4 and 5, to train the machine learning model 145 to predict the sleep onset latency of the user, a new user is first instructed at step 201 of the method 200 to use the system 10 or 100 during a baseline period (e.g., one week) without taking any melatonin. Next, the user is instructed at step 202 to take different doses of melatonin throughout a subsequent test period (e.g., one week), wherein the dose is defined as a minimum, maximum, or intermediate dose (e.g., the intermediate dose is defined as the intermediate between the maximum and minimum doses). The baseline period and the test period together constitute a data collection period. It should be appreciated that the durations of the baseline period and the test period may be longer or shorter than one week, and that the durations of the baseline period and the test period may be different from each other without departing from the scope of the disclosed concepts. The test period or baseline period of the user may also be omitted entirely without departing from the disclosed concepts, such that the machine learning model 145 may be trained using comparable available data from dedicated study trials, wherein sleep architecture data is collected for a plurality of study subjects experiencing the baseline period and the test period; however, it should be appreciated that SOL predictions made by machine learning model 145 in this case are not personalized to the user of system 10 or 100.

In FIG. 4, SOL prediction module 140 is depicted as including a two-process (2-P) SOL prediction module. The 2-P SOL prediction module uses an algorithm that considers two processes (i.e., the process S curve and the process C curve shown in fig. 3, respectively) of steady-state sleep drive and circadian timing in predicting sleep onset latency. However, it should be appreciated that a unified performance model, a SAFTE model, or any other biological model that allows for steady state sleep drive and circadian timing may be used in place of the 2-P model in the SOL prediction module 140 without departing from the scope of the disclosed concepts. Referring to fig. 5, it is determined at step 203 whether a multi-modal SOL prediction is to be found in step 205, i.e. a SOL prediction that considers behavior in addition to steady state sleep drive and circadian timing. If a multi-modal SOL is to be found, at step 204, data about the user's behavior collected by the behavior detection module 15 is provided to the machine learning model 145. If a multi-modal SOL prediction is not sought, process 200 proceeds from step 203 to step 205.

At step 205, the machine learning model 145 analyzes the baseline period data, the test period data, and the behavior detection module 15 data (if applicable) for each day of the data collection period and compares the data to the user detected sleep architecture collected by the sleep architecture detection module 13 for each day of the data collection period to determine a pattern indicative of SOL. As previously described with respect to steps 201 and 202 of process 200, if personalized SOL predictions are not sought, then instead of providing user baseline data and test data to machine learning model 145, subject baseline data and test data may be provided at step 205 such that machine learning model 145 may compare the subject baseline data and test data to the subject's sleep architecture data in order to determine a pattern indicative of SOL.

Once the machine learning model 145 is trained using the method 200 to identify patterns and associations between sleep architecture data and baseline period data, test period data, and behavior data (if applicable), the machine learning model 145 is able to make reliable SOL predictions based on newly provided data about the user's sleep architecture, the recent melatonin doses taken by the user, and the recent behavior of the user (if applicable) from the previous night (or other sleep period). It should be appreciated that there are several techniques available for constructing a machine learning model such as machine learning model 145 used in SOL prediction module 140, including but not limited to decision tree regressors, generalized linear models, and other regression modeling task solutions, and that any technique for constructing a machine learning model may be used to construct machine learning model 145 without departing from the scope of the disclosed concepts.

In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" or "comprises" does not exclude the presence of elements or steps other than those listed in a claim. In the device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. In any device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The mere fact that certain elements are recited in mutually different dependent claims does not indicate that a combination of these elements cannot be used to advantage.

Although the invention has been described in detail for the purpose of illustration based on what is currently considered to be the most practical and preferred embodiments, it is to be understood that such detail is solely for that purpose and that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover modifications and equivalent arrangements that are within the spirit and scope of the appended claims. For example, it is to be understood that the present invention contemplates that, to the extent possible, one or more features of any embodiment can be combined with one or more features of any other embodiment.

Claims (15)

1. A melatonin optimization system for optimizing the effectiveness of exogenous melatonin in achieving a desired sleep outcome for a user, the system comprising:

a user interface configured to accept information entered into the user interface regarding the user's health, self-reporting behavior, and desired sleep outcome;

a sleep architecture detection module configured to perform monitoring of the user's sleep architecture and detect hormone sensitivity of the user through the monitoring;

a behavior detection module configured to detect and collect information about the user's behavior in order to define the detected behavior of the user;

an initial dose algorithm module configured to define an initial recommended dose of melatonin for the user;

a validity assessment module configured to determine a result difference between the desired sleep result of the user and a measured sleep result of the user; and

a recommendation engine configured to define interventions for the user to reduce the outcome differences,

wherein the initial dose algorithm module is configured to define the initial recommended dose of melatonin based on the information input to the user interface, and

wherein the recommendation engine is configured to define the intervention based on the outcome differences, the monitoring of the sleep architecture, the detected behavior of the user, and the initial recommended dose of melatonin.

2. The melatonin optimization system of claim 1, wherein the desired sleep outcome is sleep onset latency.

3. The melatonin optimizing system of claim 1, wherein the self-reporting behavior includes information about food intake, alcohol intake, and caffeine intake.

4. The melatonin optimization system of claim 1, wherein the behavior detection module comprises a pressure detector configured to detect user physiological data including at least one of heart rate variability and skin conductance.

5. The melatonin optimizing system of claim 1,

wherein the behavior detection module comprises a pressure detector comprising a device configured to communicate with remote pressure detection software via an Application Programming Interface (API),

wherein the remote pressure detection software is configured to determine a pressure level of the user.

6. The melatonin optimization system of claim 1, wherein the behavior detection module comprises a light sensor configured to determine a duration, intensity, and timing of exposure of the user to both sunlight and artificial light throughout the day and until a sleeping time of the user.

7. The melatonin optimizing system of claim 1,

wherein the intervention comprises changing a current dose of melatonin recommended to the user,

wherein the change in the current dose of melatonin recommended to the user is based on: the effectiveness of any prior melatonin intervention, the user's current day's behavior detected by the behavior detection module, and the user's sleep architecture of the previous night detected by the sleep architecture detection module.

8. The melatonin optimization system of claim 1, wherein the recommendation engine limits the intervention to include only changes in the user's behavior if a currently recommended melatonin dosage has reached a predetermined maximum level.

9. The melatonin optimizing system of claim 1,

wherein the validity assessment module comprises a machine learning model,

wherein the machine learning model is trained to provide non-interfering sleep onset latency predictions for upcoming sleep sessions of the user based on data collected by the sleep architecture detection module regarding the user's most recent sleep session.

10. The melatonin optimizing system of claim 9,

wherein the machine learning model has been provided with training to personalize the non-interfering sleep onset latency prediction for the user,

wherein the training of the machine learning model includes providing sleep architecture data to the machine learning model from a baseline period when the user is not using exogenous melatonin, and providing sleep architecture data to the machine learning model from a test period when the user is using a different dose of exogenous melatonin.

11. The melatonin optimization system of claim 9, wherein the training of the machine learning model further comprises providing behavioral data of the user associated with the baseline period from the behavioral detection module to the machine learning model, and providing behavioral data of the user associated with the test period from the behavioral detection module to the machine learning model.

12. The melatonin optimizing system of claim 9,

wherein the machine learning model has been provided with training to predict non-interfering sleep onset latencies for the user,

wherein the training of the machine learning model comprises: sleep architecture data from a study subject at a baseline period when the study subject is not using exogenous melatonin is provided to the machine learning model, and sleep architecture data from a study subject at a test period when the study subject is using a different dose of exogenous melatonin is provided to the machine learning model.

13. The melatonin optimization system of claim 9, wherein the recommendation engine is configured to define the intervention based on the non-intervention sleep onset latency prediction provided by the machine learning model.

14. The melatonin optimization system of claim 10, wherein the recommendation engine is configured to define the intervention based on the non-intervention sleep onset latency prediction provided by the machine learning model.

15. The melatonin optimization system of claim 12, wherein the recommendation engine is configured to define the intervention based on the non-intervention sleep onset latency prediction provided by the machine learning model.

Images