We have developed and validated three linked mathematical models: one of the human circadian pacemaker that includes the influence of light and of non-photic processes; one of performance and alertness that includes the effects of circadian rhythms, sleep/wake homeostasis, and sleep inertia; and one of the physiology underlying sleep/wake regulation. Together these models estimate and predict the effects of sleep/wake timing, light exposure, circadian phase, and some pharmaceuticals on performance and alertness. Performance and alertness measures are modeled independently to reflect differences in the underlying physiological processes and effects of sleep/wake on each measure. CPSS, the software that implements this model, has been used by NASA employees and NASA consultants to design light countermeasures for both astronaut pre-launch schedules and in-flight schedules.

Our specific aims were to: (1) Replace the current assumption that an individual sleeps the entire time when scheduled to sleep with probabilities of sleep and wake during scheduled sleep times; (2) Improve daily assessment of sleep and sleep disruption using actigraphy data; (3) Add statistical features including confidence limits to the predictions; (4) Update the software per astronaut and ground crew requests for specific features and reports. These projects address NASA's objectives to improve the design of individual countermeasures to reduce sleep disruption and improve performance and alertness in space and on Earth. As part of these overall goals, we published a novel scheduling algorithm called Shifter that automatically designs optimal light countermeasures for user-defined work and sleep schedules. Both experimental and field studies have shown that light interventions minimize fatigue while improving performance and sleep. This work required several mathematical and computational advances, including the development of a novel schedule representation and scheduling algorithm. The utility of this scheduling software extends beyond NASA-related schedules to include any operational setting that relies on work scheduled outside the typical 9am to 5pm shift, including night and rotating shift-work, transmeridian travel, and the design of work schedules for medical residents to improve performance and meet new national guidelines for restricted work hours.

To individualize model predictions, we are developing a statistical framework based on easily collected trait information (e.g., age, chronotype) that has been shown to correlate with differences in sleep timing, circadian phase, and performance and alertness. Although changes in physiology have been correlated with specialized questionnaire results (e.g., habitual sleep time is highly correlated with circadian phase), our current results suggest that a simple alteration of the model output using demographic information is not accurate. The work of Dr. Phillips (NSBRI post-doctoral fellow) has quantified mechanisms that may underlie individual differences in physiologically-determined sleep timing and self-reported chronotype (e.g., owl or lark ). The use of sleep aids during NASA missions is indicative of the difficulty with initiating and maintaining sleep that astronauts experience during space flight. To assess the ability of individuals to conform to scheduled work hours, Dr. Phillips has integrated the circadian and performance model with a model of the physiological mechanisms that control sleep/wake transitions. This combined model dynamically predicts whether an individual is awake or asleep across a simulated protocol and also allows for predictions of sleep efficiency and the likelihood of falling asleep during scheduled wake periods. This physiologically-based model can be readily extended to incorporate pharmaceutical effects. Using this model, we have now successfully incorporated the effects of melatonin and caffeine at different times and dosages.

Actigraphy is an inexpensive and less intrusive alternative to polysomnography and/or sleep/wake diaries to determine an individual's sleep/wake schedule. Prior iterations of the mathematical model relied on user input to generate sleep/wake schedules. Based on the work from this project, we can now use actigraphy to determine the actual sleep/wake schedule of an individual and use this information as input to our mathematical models. We recently completed a project in collaboration with two NSBRI investigators, Drs. Lockley and Barger. In this project, pattern recognition algorithms were used to identify the level of performance impairment in an individual based on a single session of neurobehavioral testing (rather than multiple hours of testing) under both controlled in-patient laboratory conditions and real world conditions, including during the NASA Phoenix Mars study. We continue to work with NASA and NSBRI personnel to meet their requests regarding use of the models and software.

The previous model assumption that an individual sleeps the entire time that is available to them during a scheduled sleep episode has been improved by the recent incorporation of actigraphy as an input to the mathematical model of the actual sleep/wake times experienced by the individual. The use of actigraphy as a tool to record sleep has improved confidence levels on the daily assessment of sleep when compared to the use of sleep logs or diaries and also has reduced the user requirements for maintaining daily logs. The interface between actigraphy and the software enables faster and possibly more accurate predictions of circadian phase and performance parameters. The Shifter software now includes optimal countermeasure design, so that countermeasures can be planned for times of predicted poor performance and alertness. The schedule and countermeasure design program allows users to interactively design schedules and implement mathematically optimized light countermeasures (including intensity, duration and timing within the wake episode) to minimize worker fatigue. This scheduling software will be valuable to those who work at night, on rotating schedules, on non-24-hr schedules, or on extended duty schedules. The software allows individuals to design countermeasures for their assigned work schedules so that their sleep/ wake rhythms will be adjusted to ensure optimal performance at desired times, with respect to both scheduled work events and their circadian phase. Improving sleep duration and quality can also decrease the risk of accidents and errors, as well as decrease the long-term risks of cardiovascular, metabolic, immune, and psychological pathologies. We continue to work with NASA and NSBRI personnel to meet their requests regarding use of the models and software. We continue to work with Dr. Dorit Donoviel, Associate NSBRI Research Director, and Marti Fleming, NSBRI commercialization consultant, to promote commercialization of the work. The mathematical modeling efforts and software have also been used in educational programs and in the popular press to teach students and teachers about circadian rhythms and sleep and their effects on alertness and performance.

Specific Aim 2 (actigraphy): We have integrated the output from actigraphy software with the input required to run our Circadian Performance Simulation Software (CPSS). CPSS implements our mathematical model of the human circadian pacemaker, performance, and alertness, which includes the key processes of circadian rhythms, sleep/wake homeostasis, and sleep inertia on performance and alertness, as well as the effects of light on circadian rhythms. Pre-processing tools were developed to generate the sleep/wake schedule and light levels from either raw or processed actigraphy data. We have tested the ability to use outpatient actigraphy as input to CPSS to predict circadian phase for individuals under circadian entrained and phase-shift conditions.

Specific Aim 3 (Statistical modeling of individual circadian, sleep, performance, and alertness parameters): We have concentrated on statistical modeling of individual parameters of our circadian, performance, and alertness models. By fitting the model to individual data, rather than group averages, we obtain a set of parameters for the performance and alertness models that are unique to each individual. We can then use other data collected from the individual, such as age, sex, habitual sleep time, morningness/eveningness preference, to determine correlations between model parameters and individual characteristics.

Specific Aim 4 (Work with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake, and countermeasure design to minimize fatigue and performance issues): We have had discussions with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake, and countermeasure design to minimize fatigue and performance issues, as well as incorporating the models into other modeling work performed by NASA. Dr. Barger used the software for NASA supported studies of sleep in ISS and Shuttle crew. We also developed a novel scheduling algorithm that automatically designs optimal light countermeasures for user-defined schedules The scheduling framework is applicable to other work schedules including shift-work and transmeridian travel.

We have developed and validated two linked mathematical models: one of the human circadian pacemaker that includes the influence of light and of non-photic processes, and one of performance and alertness that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia. Together, these models are able to predict the effects of sleep/wake, sleep inertia and circadian phase on performance and alertness. Each performance or alertness measure has a separate equation, reflecting the underlying physiological processes in the effect of sleep/wake on performance and alertness. CPSS, the software implementing this model, has been used by NASA and consultants when designing light countermeasures for astronaut pre-launch schedules as well as for designing in-flight schedules. To further improve this mathematical model and this method for optimal design of countermeasures, our s modeling work will focus on individual, rather than group, predictions and use novel non-linear mathematical and statistical methods. These projects address NASA's objectives to improve the design of individual countermeasures to reduce sleep disruption and improve performance and alertness in space and on Earth. Our current progress includes:

We developed a novel scheduling algorithm called Shifter that automatically designs optimal light countermeasures for user-defined NASA-related schedules. Light interventions have been demonstrated to minimize fatigue, improve performance, and improve sleep in experimental and field studies. Shifter allows individuals who are not circadian experts to design schedules and light interventions within minutes. Previous design of light interventions, such as for the 24.65-hr Mars Day experimental protocol (NASA and NSBRI supported, Dr. Czeisler, PI), took approximately two weeks. The scheduling framework can be applied to non-NASA-related work schedules including shift-work and transmeridian travel. We are currently applying our methods to the design of schedules for medical residents, so as to provide schedules that predict optimal performance at critical times and meet new national guidelines for restricted physician hours. Abstracts have been accepted on this scheduling work and will be presented at national and international meetings.

We are developing new methods to refine the scheduling algorithm for predicting individual differences in circadian phase and performance. We are individualizing predictions based on easily collected trait information (e.g., age, chronotype), and developing a statistical framework for making individual predictions. Experimental evidence demonstrates changes in physiology are well correlated with age and specialized questionnaire results (e.g., habitual sleep time is highly correlated with circadian phase). The work of Dr. Phillips (NSBRI post-doctoral fellow) has quantified mechanisms underlying individual differences in physiologically-determined sleep timing and self-reported (subjective) chronotype (e.g., "owl" or "lark").

Using synergistic support from a NIH "Grand Opportunities" grant to Dr. Klerman, we are developing and populating a database with studies from the BWH Division of Sleep Medicine. The database has enabled a larger data set for our modeling work and will facilitate the building of individual models using demographic data as described above.

To assess the ability of individuals to conform to scheduled work hours, Dr. Phillips is integrating the circadian and performance model with a model of the physiological mechanisms which control sleep-wake transitions. This combined model dynamically predicts wake/sleep state across a simulated protocol, allowing predictions of sleep efficiency, and likelihood of falling asleep during scheduled wake periods. This model has been validated using BWH datasets; several abstracts have been published on this work. Since the model is physiologically based, it is being extended to incorporate pharmaceutical effects, including simulating the effects of melatonin and caffeine at different times and dosages.

We are targeting the use of actigraphy, which is an inexpensive and less intrusive alternative to polysomnography, to determine sleep/wake state and then use the mode to predict circadian phase and performance without other inputs. Several abstracts have been published on this work.

Our NSBRI-funded work is broadly applicable to diverse work environments, ranging from NASA missions to industries such as aviation, transportation, and the military. We are also working with the NSBRI Industry Forum to explore ways to facilitate use of our work in these diverse environments. We continue to work with NASA and NSBRI personnel to meet their requests regarding use of the models and software.

The ease of use of the modeling has been improved by the recent incorporation of actigraphy as input of actual sleep/wake time to the mathematical model. The use of actigraphy as a tool to record sleep has improved the confidence levels on the daily assessment of sleep when compared to the use of sleep logs, as well as reduced the user requirements for maintaining daily logs. The interface between actigraphy and the Circadian Performance Simulation Software (CPSS) enables faster and possibly more accurate predictions of circadian phase and performance parameters. The Shifter software now includes optimal countermeasure design, so that countermeasures can be planned for times of predicted poor performance and alertness. The schedule/countermeasure design program allows users to interactively design schedules and implement mathematically optimal light countermeasures (including intensity, duration and placement) to minimize worker fatigue. This scheduling software will be valuable to those who work at night, on rotating schedules, on non-24-hr schedules or extended duty schedules. The software allows individuals to design countermeasures for their assigned work schedules so that their sleep and wake rhythms will be adjusted for optimal performance at desired times, both with respect to scheduled work events and with their circadian phase. Improving sleep duration and quality can also decrease the risk of accidents and errors, as well as decrease the long-term risks of cardiovascular, metabolic, immune and psychological pathologies.

The mathematical modeling has been used for basic scientific research. Inclusion of mathematical models in the planning process to optimize measures to be studied in experimental protocols enables more efficient use of research resources and directs new research. If the modeling of existing experimental data is found to be unsatisfactory, then model assumptions may need to be revised; this revision may include identification of a new physiological process not previously described.

The mathematical modeling efforts and software have also been used in educational programs and in the popular press to teach students, teachers and health care professionals about circadian rhythms and sleep, work schedules and their effects on alertness and performance.

Specific Aim 2 (actigraphy) We have integrated the output from actigraphy data with the input required to run our Circadian Performance Simulation Software (CPSS), which utilizes our mathematical model of the human circadian pacemaker, performance and alertness, that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia as well as the effects of light on circadian rhythms. Pre-processing tools have been developed to generate the sleep/wake schedule and light levels from raw or processed actigraphy data. The sleep/wake and lighting schedule thus generated may be used as input to CPSS. In the current year, we have been testing the ability of outpatient actigraphy as input to CPSS to predict inpatient circadian phase for individuals with habitual sleep/wake schedules (sleep at night and wake during the day).

Specific Aim 3 (Statistical modeling of individual circadian, sleep, performance and alertness parameters): We have concentrated on statistical modeling of individual circadian parameters of our circadian, performance and alertness models.to obtain a set of parameters for the performance and alertness models unique to each individual. We are using other data collected from the individual, such as age, gender, habitual sleep time, morningness/eveningness preference, etc. to determine correlations between model parameters and individual characteristics. In the current year of this project, we have been working on a new BWH Division of Sleep Medicine Database, supported by both this NSBRI grant and a NIH grant to Dr. Klerman to facilitate the development of individualized models.

Specific Aim 4 (Work with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues.): We have had discussions with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues, as well as incorporating the models into other modeling work performed by NASA.

We have developed and validated a mathematical model of the human circadian pacemaker that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia, and is able to predict the effects of sleep/wake and circadian phase on performance and alertness. The software implementing this model has been used by NASA and consultants when designing light countermeasures for astronaut pre-launch schedules as well as for designing in-flight schedules. To further improve this mathematical model and this method for optimal design of countermeasures, our specific aims are to: (1) Replace the current assumption that an individual sleeps when scheduled to sleep, with probabilities of sleep and wake during those times; (2) Improve daily assessment of sleep and sleep disruption using actigraphy data; (3) Add statistical features including confidence limits to the predictions; (4) Update the software per astronaut and ground crew requests for specific features and reports. The basic mathematical work will focus on individual, rather than group, predictions and use novel non-linear mathematical and statistical measures. These projects address NASA's objectives to improve the design of individual countermeasures to reduce sleep disruption and improve performance and alertness in space and on Earth. Our current progress is as follows:

The major advance in the last year was the publication of a novel scheduling algorithm called Shifter in PLoS Computational Biology. The algorithm automatically designs optimal light countermeasures for user-defined NASA related schedules. These light interventions have been demonstrated to minimize fatigue, improve performance, and improve sleep in experimental and field studies. Developing Shifter required several mathematical and computational advances, including the development of a novel schedule representation and scheduling algorithm. Shifter for the first times allows individuals who are not circadian experts to design schedules and light interventions. Using Shifter, schedules are produced within a minute; previous design of light interventions, such as for the Mars Day protocol, took approximately two weeks with multiple iterations of CPSS and other software. The scheduling framework has also been designed to be applicable to work schedules including shift-work and transmeridian travel. To this end, we have begun applying our methods to the design of schedules for medical residents, so as to provide schedules that predict optimal performance while meeting new guidelines for restricted physician hours.

We are extending our current models and developing new methods to refine the scheduling algorithm for predicting individual differences in circadian phase and performance. To achieve this, we are individualizing predictions based on easily collected trait information (e.g., age, chronotype), and developing a statistical framework for building individual models. Experimental evidence demonstrates changes in physiology are well correlated with age and specialized questionnaire results (e.g., habitual sleep time is highly correlated with circadian phase). This is important because circadian phase correlates with intrinsic circadian period, which we have shown to be the most important input to generating accurate individualized predictions. Further work will involve identifying other demographics which can be used as proxies for physiological parameters.

The use of sleep aids during NASA missions is indicative of the difficulties astronauts face in sleeping during space flight. To assess the ability of individuals to conform to scheduled work hours, we are integrating the circadian and performance model with a model of the physiological mechanisms which control sleep-wake transitions. This combined model dynamically predicts wake/sleep state across a simulated protocol, allowing predictions of sleep efficiency, and likelihood of falling asleep during scheduled wake periods.

Integral to successfully using these methods in space is the ability to assess physiological state, and make individualized predictions. Since sleep is a core issue in space, we are targeting use of actigraphy, which is an inexpensive and less intrusive alternative to polysomnography. Our current work allows us to predict sleep state, circadian phase and performance directly from actigraphy.

Our plan is to use this information as input to our scheduling algorithm, resulting in individualized countermeasure predictions in real-time. The key benefit of these new methods is the ability to investigate different schedules, and to adaptively respond to changes that are unavoidable, such as launch rescheduling due to inclement weather. Our NSBRI-funded work is broadly applicable to diverse work environments, ranging from NASA missions to industries such as aviation, transportation, and the military.

The recent incorporation of actigraphy as input to the mathematical model as a tool to record sleep has improved the confidence levels on the daily assessment of sleep when compared to the use of sleep logs. The interface between actigraphy and the Circadian Performance Simulation Software enables faster and possibly more accurate predictions of circadian phase and performance parameters. The software now also includes optimal countermeasure design, so that countermeasures can be planned for times of predicted poor performance and alertness. The schedule/countermeasure design program allows users to interactively design schedules and implement mathematically optimal light countermeasures (including intensity, duration and placement) to minimize worker fatigue. This scheduling software will be valuable to those who work at night, on rotating schedules, on non-24-hr schedules or extended duty schedules. The software allows individuals to design countermeasures for their assigned work schedules so that their sleep and wake rhythms will be adjusted for optimal performance at desired times, both with respect to scheduled work events and circadian phase. Improving sleep duration and quality can also decrease the risk of accidents and errors, as well as decrease the long-term risks of cardiovascular, metabolic, immune and psychological pathologies.

Mathematical modeling has been used for basic scientific research. Inclusion of mathematical models in the planning process to optimize measures to be studied in experimental protocols enables more efficient use of research resources and directs new research. If the modeling of existing experimental data is found to be unsatisfactory, then model assumptions may need to be revised. This revision may include identification of a new physiological process not previously described. As an example, an additional component (non-linear response to ocular light stimuli) was added to our mathematical model to describe data collected in our clinical research facilities, even before the anatomic and physiologic basis of this component was found. Later experiments validated this mathematical prediction. The proposed mathematical model, based on behavioral experimental findings, had uncovered previously unknown additional physiological processes at the cellular level.

The mathematical modeling efforts and software have also been used in educational programs and in the popular press to teach students and teachers about circadian rhythms and sleep and their effects on alertness and performance.

Specific Aim 2 (actigraphy): We have integrated the output from actigraphy data with the input required to run our Circadian Performance Simulation Software (CPSS), which utilizes our mathematical model of the human circadian pacemaker, performance and alertness, that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia as well as the effects of light on circadian rhythms. Pre-processing tools have been developed to generate the sleep/wake schedule and light levels from raw or processed actigraphy data. The sleep/wake and lighting schedule thus generated may be used as input to CPSS.

Specific Aim 3 (Statistical modeling of individual circadian, sleep, performance and alertness parameters): We have concentrated on statistical modeling of individual circadian parameters of our circadian, performance and alertness models. By fitting the model to individual, rather than grouped data, we obtain a set of parameters for the performance and alertness models unique to each individual. We can then use other data collected from the individual, such as age, gender, habitual sleep time, morningness/eveningness preference, etc. to determine correlations between model parameters and individual characteristics.

Specific Aim 4 (Work with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues.): We have had discussions with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues, as well as incorporating the models into other modeling work performed by NASA. As one example, we developed Shifter, a novel scheduling algorithm that automatically designs optimal light countermeasures for user defined schedules. Appropriate light interventions have been demonstrated to minimize fatigue, improve performance, and improve sleep in experimental and field studies. Developing Shifter required several mathematical and computational advances, including the development of a novel schedule representation and scheduling algorithm. The scheduling framework has also been designed to be applicable to work schedules including shift-work and transmeridian travel.

We have developed and validated a mathematical model of the human circadian pacemaker, performance and alertness that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia. The software implementing this model has been used by NASA and consultants when designing light countermeasures for astronaut pre-launch schedules.

For our current NSBRI project, we have developed a mathematical method for optimal design of countermeasures and schedules after a change in timing of sleep, light or critical tasks. Our specific aims are to: (1) Replace the current assumption that an individual sleeps when scheduled to sleep, with probabilities of sleep and wake during those times; (2) Improve daily assessment of sleep and sleep disruption using actigraphy data; (3) Add statistical features including confidence limits to the predictions; (4) Update the software per astronaut and ground crew requests for specific features and reports. The basic mathematical work will focus on individual, rather than group, predictions and use novel non-linear mathematical and statistical measures. These projects address NASA's objectives to improve the design of individual countermeasures to reduce sleep disruption and improve performance and alertness in space and on Earth.

The software also now includes optimal countermeasure design, so that countermeasures can be planned for times of predicted poor performance and alertness. The schedule/countermeasure design program that allows a user to interactively design a schedule and to automatically design a mathematically optimal countermeasure regime (intensity, duration and placement). This will be valuable to those who schedule people who work at night, on rotating schedules, on non-24-hr schedules or extended duty schedules. Individuals can design countermeasures for their assigned work schedules so that their sleep and wake rhythms will be adjusted for optimal performance at desired times. In addition, if the countermeasure design includes shifting the circadian rhythms to be appropriately aligned with environmental time, then performance, alertness and sleep will all improve. Improving sleep duration and quality can also decrease the risk of accidents and errors, as well as cardiovascular, metabolic, immune and psychological pathologies.

The mathematical modeling has been used for basic scientific research. Inclusion of mathematical models in the planning process to optimize measures to be studied in experimental protocols enables more efficient use of research resources and directs new research. If the modeling of existing data is unsatisfactory, then the model assumptions need to be revised. This revision may include identification of a new physiological process not previously described. As an example, an additional component (non-linear response to ocular light stimuli) was added to the circadian rhythms component of our mathematical model to describe data collected in our clinical research facilities, even before the anatomic and physiologic basis of this component of the mathematical model was found. Later experiments validated this mathematical finding. The mathematical model had demonstrated that previously unknown additional physiological processes were involved.

The modeling work on the differential effects of different wavelength of light on circadian rhythms and alertness can be used for designing artificial (indoor) lighting systems that can maximize circadian or alerting response.

The mathematical modeling efforts and CPSS have also been used in educational programs and in the popular press to teach students and teachers about circadian rhythms and sleep and their effects on alertness and performance.

Specific Aim 2 (Improve the assessment of sleep and sleep disruption through the development of an improved actigraphy-to-sleep/wake classification algorithm): We developed a new sleep-wake classification algorithm that has been tested on actigraphy data collected under various conditions. A new algorithm exploits the assumption that there exists an implicit bimodal distribution in the actigraphy time series and that the extraction of this can clearly segregate sleep and wake cycles. A statistical bimodal distribution is not possible to obtain directly from the time series due to the presence of both inflated zeros and over-dispersion that result in a zero-inflated negative binomial distribution. Therefore we have developed a new method by which bimodal distribution obtained can be indirectly related to sleep-wake epochs of the activity count data.

Specificity, sensitivity, accuracy and predictive value of sleep and wake are determined by comparing epoch by epoch of actigraph data with that of polysomnographic data to test the algorithm Both high specificity and sensitivity are obtained by the application of present algorithm; in contrast, most of the published algorithms have either high sensitivity or specificity, but not both. This algorithm also has high sensitivity and specificity using the data collected from two different instruments (Actiwatch and Motionlogger): therefore this algorithm works well irrespective of the instrument or the way data is collected.

As future work, actigraphy data collected under habitual conditions, under conditions of quiet extended wake (which is difficult to distinguish from sleep), under conditions of forced desynchrony of sleep/wake cycle and circadian rhythms (to explore the effect of circadian phase) and under conditions of extended sleep (which induces insomnia or quiet wake during scheduled sleep, which is also difficult to differentiate from sleep) will be determined to further test the performance of the algorithm and will be compared with other threshold algorithms. Two manuscripts of this work are in preparation.

Specific Aim 3 (Statistical modeling of individual circadian, sleep, performance and alertness parameters): We have begun the first step in this work by concentrating on statistical modeling of individual circadian parameters.

Specific Aim 4 (Work with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues.): We have had discussions with NASA and NSBRI personnel to revise features of our current software to meet their specifications for administratively scheduling sleep, wake and countermeasure design to minimize fatigue and performance issues, as well as incorporating the models into other modeling work performed by NASA.

During space missions, circadian rhythms and sleep are disrupted, both for those working in space and for those on Earth. Astronaut problems with sleep, circadian rhythms and performance have been reported, and NASA data indicate that sleeping pills are among the most commonly used drugs in space. Therefore, it is imperative that schedules and countermeasures are designed to optimize individual performance, alertness and quality sleep relative to operational requirements.

We have developed and validated a mathematical model of the human circadian pacemaker, performance and alertness that includes the key processes of circadian rhythms, sleep/wake homeostasis and sleep inertia. The software implementing this model has been used by NASA to design light countermeasures for astronaut pre-launch schedules. In our previous NSBRI project, we developed a mathematical method for optimal design of countermeasures and schedules after a change in timing of sleep, light or critical tasks.

Specific Aims

1. Replace the current assumption that an individual sleeps when scheduled to sleep, with probabilities of sleep and wake during those times;

2. Improve daily assessment of sleep and sleep disruption using actigraphy data;

3. Add statistical features including confidence limits to the predictions; and

4. Update the software per astronaut and ground crew requests for specific features and reports.

The basic mathematical work will focus on individual, rather than group, predictions and will use novel, non-linear mathematical and statistical measures. This project addresses NASA's objectives to improve the design of individual countermeasures to reduce sleep disruption and improve performance and alertness in space and on Earth.