CA2680882A1 - Predicting human cognitive performance - Google Patents

Abstract

An apparatus and method for predicting cognitive performance of an individual based on factors including preferably sleep history, the time of day, and the individual's activities. The method facilitates the creation of predicted cognitive performance curves that allow an individual to set his/her sleep times to produce higher levels of cognitive performance. The method also facilitates the reconstruction of past cognitive performance levels based on sleep history.

Description

Agent Ref: 67803/00007 1 Predicting Human Cognitive Performance

2 I. TECHNICAL FIELD

3 This invention relates to a method for predicting cognitive performance of an individual

4 preferably based on that individual's prior sleep/wake history, the time of day, and tasks (or activities) being performed by the individual.

6 II. BACKGROUND OF THE ART

7 Maintenance of productivity in any workplace setting depends upon effective cognitive 8 performance at all levels from command/control or management down to the individual soldier 9 or worker. Effective cognitive performance in turn depends upon complex mental operations.
Many factors have been shown to affect cognitive performance (e.g., drugs or age). However, 11 of the numerous factors causing day to day variations in cognitive performance, two have been 12 shown to have the greatest impact. These two factors are an individual's prior sleep/wake 13 history and the time of day.

14 Adequate sleep sustains cognitive performance. With less than adequate sleep, cognitive performance degrades over time. An article by Thorne et al. entitled "Plumbing Human 16 Performance Limits During 72 hours of High Task Load" in Proceedings of the 24th DRG
17 Seminar on the Human as a Limiting Element in Military Systems, Defense and Civil Institute of 18 Environmental Medicine, pp. 17-40 (1983), an article by Newhouse et al.
entitled "The Effects of 19 d-Amphetamine on Arousal, Cognition, and Mood After Prolonged Total Sleep Deprivation"
published in Neuropsychopharmacology, vol. 2, pp. 153-164 (1989), and another article by 21 Newhouse et al. entitled "Stimulant Drug Effects on Performance and Behavior After Prolonged 22 Sleep Deprivation: A Comparison of Arrtphetamine, Nicotine, and Deprenyl"
published in 23 Military Psychology, vol. 4, pp. 207-233 (1992) all describe studies of normal volunteers in 24 which it is revealed that robust, cumulative decrements in cognitive performance occur during continuous total sleep deprivation as measured by computer-based testing and complex 26 operational simulation. In the Dinges et al. article entitled "Cumulative Sleepiness, Mood 27 Disturbance, and Psychomotor Vigilance Performance Decrements During a Week of Sleep 28 Restricted to 4-5 Hours Per Night" published in Sleep, vol. 20, pp. 267-277 (1997), it is revealed 29 that on fixed, restricted daily sleep amounts, cumulative reduced sleep also leads to a cognitive performance decline. Thus, in operational settings, both civilian and military, sleep deprivation 31 reduces productivity (output of useful work per unit of time) on cognitive tasks.

21922696.1 1 Agent Ref: 67803/00007 1 Thus, using computer-based cognitive performance tests, it has been shown that total 2 sleep deprivation degrades human cognitive performance by approximately 25%
for each 3 successive period of 24 hours awake. However, it also has been shown that even small 4 amounts of sleep reduce the rate of sleep loss-induced cognitive performance degradation.
Belenky et al. in their article entitled "Sustaining Performance During Continuous Operations:
6 The U.S. Army's Sleep Management System," published in 20"' Army Science Conference 7 Proceedings, vol. 2, pp. 657-661 (1996) disclose that a single 30-minute nap every 24 hours 8 reduces the rate of cognitive performance degradation to 17% per day over 85 hours of sleep 9 deprivation. This suggests that recuperation of cognitive performance during sleep accrues most rapidly early in the sleep period. No other factor besides the amount of sleep contributes 11 so substantially and consistently to the normal, daily variations in cognitive performance.

12 In addition to sleep/wake history, an individual's cognitive performance at a given point in 13 time is determined by the time of day. In the early 1950s, Franz Halberg and associates 14 observed a 24-hour periodicity in a host of human physiologic (including body temperature and activity), hematologic, and hormonal functions, and coined the term 'circadian' (Latin for'about a 16 day') to describe this cyclic rhythm. Halberg showed that most noise in experimental data came 17 from comparisons of data sampled at different times of day.

18 When humans follow a noctumal sleep/diurnal wake schedule (for example, an 8-hour 19 sleep/16-hour wake cycle, with nightly sleep commencing at approximately midnight), body temperature reaches a minimum (trough) usually between 2:00 AM and 6:00 AM.
Body 21 temperature then begins rising to a maximum (peak) usually between 8:00 PM
and 10:00 PM.
22 Likewise, systematic studies of daily human cognitive performance rhythms show that speed of 23 responding slowly improves across the day to reach a maximum in the evening (usually 24 between 8:00 PM and 10:00 PM) then dropping more rapidly to a minimum occurring in the early morning hours (usually between 2:00 AM and 6:00 AM). Similar but somewhat less 26 consistent rhythms have been shown from testing based on various cognitive performance 27 tasks. Thus, superimposed on the effect of total sleep deprivation on cognitive performance 28 noted above was an approximately t10% variation in cognitive performance over each 24-hour 29 period.

Various measures have been shown to correlate, to some extent, with cognitive 31 performance. These include objective and subjective measures of sleepiness (or its converse, 32 alertness). Some individuals familiar with the art use "sleepiness" to indicate the opposite of 33 "alertness" (as is the case in the present document). "Drowsiness" often is used 21922696.1 2 Agent Ref: 67803/00007 1 interchangeably with "sleepiness" although some familiar with the art would argue that 2 "sleepiness" pertains specifically to the physiological need for sleep whereas "drowsiness"
3 refers more to the propensity or ability to fall asleep (independent of physiological sleep need) 4 or the subjective feeling of lack of alertness. The term "fatigue" has been used as a synonym for "sleepiness" by the lay population, but those familiar with the art do not consider "fatigue" to 6 be interchangeable with "sleepiness - rather, "fatigue" is a broad term that encompasses more 7 than just the effects of sleep loss per se on performance. Likewise, "cognitive performance" has 8 been defined as performance on a wide variety of tasks, the most commonly used being 9 vigilance tasks (tasks requiring sustained attention). From vigilance and other tasks, some researchers use accuracy as their measure of cognitive performance, while others use reaction 11 time (or its inverse, speed). Still others use a measure that is calculated as speed multiplied by 12 accuracy, that is the amount of useful work performed per unit of time (also known as 13 throughput). Those familiar with the art generally agree that vigilance tasks are appropriate 14 measures of cognitive performance under conditions of sleep deprivation, and that either reaction time (speed) or some measure that takes reaction time into account (e.g., throughput) 16 is a valid and reliable way of ineasuring cognitive performance.

17 The Multiple Sleep Latency Test (MSLT) is a widely accepted objective measure of 18 sleepiness/alertness. In the MSLT, individuais try to fall asleep whiie lying in a darkened, quiet 19 bedroom. Various physiological measures used to determine sleep or wakefulness are recorded (eye movements, brain activity, muscle tone), and time taken to reach the first 30 21 seconds of stage 1(light) sleep is determined. Shorter latencies to stage I
are considered to 22 indicate greater sleepiness (lower alertness). Sleep latencies under 5 minutes are considered 23 to be pathological (i.e., indicative of a sleep disorder or sleep deprivation). During both total and 24 partial sleep deprivation, latency to sleep on the MSLT (alertness) and performance decline (i.e., sleepiness as measured by MSLT increases). However, although there is a correlation 26 between MSLT-determined sleepiness/alertness and cognitive performance (greater sleepiness 27 as indexed by MSLT corresponding to poorer cognitive performance), this correlation has never 28 been shown to be perfect and for the most part is not strong. As a result, the MSLT is a poor 29 (i.e., unreliable) predictor of cognitive performance.

Subjective measures of sleepiness/alertness also have been shown to correlate (albeit 31 weakly) with cognitive performance. Hoddes et al. in their article entitled "Quantification of 32 Sleepiness: A New Approach" published in Psychophysiology, vol. 10, pp. 431-436 (1973) 33 describe the Stanford Sleepiness Scale (SSS), a subjective questionnaire used widely to 21922696.1 3 Agent Ret. 67803/00007 1 measure sleepiness/alertness. In the SSS, individuals rate their current level of 2 sleepiness/alertness on a scale from 1 to 7, with 1 corresponding to the statement, "feeling 3 active and vital; alert; wide awake" and 7 corresponding to the statement "almost in reverie; sleep 4 onset soon; losing struggle to remain awake." Higher SSS scores indicate greater sleepiness. As with the MSLT, during both total and partial sleep deprivation, scores on the SSS increase.
6 However, as with MSLT, the correspondence between SSS-determined sleepiness/alertness 7 and cognitive performance decrements is weak and inconsistent. As a result, the SSS also is a 8 poor predictor of cognitive performance. Some other examples of subjective measures of 9 sleepiness/alertness include the Epworth Sleepiness Scale described by Johns in his article entitled "Daytime Sleepiness, Snoring, and Obstructive Sleep Apnea" published in Chest, vol.
11 103, pp. 30-36 (1993) and the Karolinska Sleepiness scale described by Akerstedt and Gillberg 12 in their article entitled "Subjective and Objective Sleepiness in the Active Individual" published in 13 International Journal of Neuroscience, vol. 52, pp. 29-37 (1990). The correspondence between 14 these subjective measures and cognitive performance also is weak and inconsistent.

In addition, factors modifying cognitive performance may not correspondingly affect 16 objective or subjective measures of sleepiness/alertness, and vice versa.
For example, the 17 Penetar et al. article entitled "Amphetamine Effects on Recovery Sleep Following Total Sleep 18 Deprivation" published in Human Psychophannacology, vol. 6, pp. 319-323 (1991) discloses 19 that during sleep deprivation, the stimulant drug d-amphetamine improved cognitive performance but not sleepiness/alertness (as measured by the MSLT). In a similar study, 21 caffeine given as a sleep deprivation countermeasure maintained elevated cognitive 22 performance for over 12 hours while the effects on subjective sleepiness, vigor and fatigue 23 transiently improved but then decayed. Thorne et al. in their article entitled "Plumbing Human 24 Performance Limits During 72 hours of High Task Load" in Proceedings of the 24th DRG
Seminar on the Human as a Limiting Element in Military Systems, Defense and Civil Institute of 26 Environmental Medicine, pp. 17-40 (1983) describe how cognitive performance continues to 27 decline over 72 hours of sleep deprivation whereas subjective sleepiness/alertness declined 28 over the first 24 hours but subsequently leveled off. The findings that cognitive performance 29 and measures of sleepiness/alertness are not always affected in the same way indicate that they are not interchangeable. That is, measures of sleepiness/alertness cannot be used to 31 predict cognitive performance, and vice versa.

32 Methods and apparatuses related to alertness detection fall into five basic categories: a 33 method/apparatus for unobtrusively monitoring current alertness level; a method/apparatus for 21922696.1 4 Agent Ref. 67803/00007 1 unobtrusively monitoring current alertness level and providing a warning/alarm to the user of 2 decreased alertness and/or to increase user's alertness level; a method/apparatus for 3 monitoring current alertness level based on the user's responses to some secondary task 4 possibly with an alarm device to warn the user of decreased alertness and/or to increase user's alertness level; methods to increase alertness; and a method/apparatus for predicting past, 6 current, or future alertness.

7 These methods and apparatuses that unobtrusively monitor the current alertness level are 8 based on an "embedded measures" approach. That is, such methods infer 9 alertness/drowsiness from the current level of some factor (e.g., eye position or closure) assumed to correlate with alertness/drowsiness. Issued patents of this type include U.S. Patent 11 No. 5,689,241 to J. Clarke, Sr., et al. disclosing an apparatus to detect eye closure and ambient 12 temperature around the nose and mouth; U.S. Patent No. 5,682,144 to K.
Mannik disclosing an 13 apparatus to detect eye closure; and U.S. Patent No. 5,570,698 to C. Liang et al. disclosing an 14 apparatus to monitor eye localization and motion to detect sleepiness. An obvious disadvantage of these types of methods and apparatuses is that the measures are likely 16 detecting sleep onset itself rather than small decreases in alertness.

17 In some patents, methods for embedded monitoring of alertness/drowsiness are combined 18 with additional methods for signaling the user of decreased alertness and/or increasing 19 alertness. Issued patents of this type include U.S. Patent No. 5,691,693 to P. Kithil describing a device that senses a vehicle operator's head position and motion to compare current data to 21 profiles of "normal" head motion and "impaired" head motion. Waming devices are activated 22 when head motion deviates from the "normal" in some predetermined way. U.S.
Patent No.
23 5,585,785 to R. Gwin et al. describes an apparatus and a method for measuring total handgrip 24 pressure on a steering wheel such that an alarm is sounded when the grip pressure falls below a predetermined "lower limit" indicating drowsiness. U.S. Patent No. 5,568,127 to H. Bang 26 describes a device for detecting drowsiness as indicated by the user's chin contacting an alarm 27 device, which then produces a tactile and auditory warning. U.S. Patent No.

5,566,067 to J.
28 Hobson et al. describes a method and an apparatus to detect eyelid movements. A change in 29 detected eyelid movements from a predetermined threshold causes an output signal/alarm (preferably auditory). As with the first category of methods and apparatuses, a disadvantage 31 here is that the measures are likely detecting sleep onset itself rather than small decreases in 32 alertness.

21922696.1 5 Agent Ref: 67803/00007 1 Other alertness/drowsiness monitoring devices have been developed based on a 2 "primary/secondary task" approach. For example, U.S. Patent No. 5,595,488 to E. Gozlan et al.
3 describes an apparatus and a method for presenting auditory, visual, or tactile stimuli to an 4 individual to which the individual must respond (secondary task) while performing the primary task of interest (e.g., driving). Responses on the secondary task are compared to baseline

6 "alert" levels for responding. U.S. Patent No. 5,259,390 to A. MacLean describes a device in

7 which the user responds to a relatively innocuous vibrating stimulus. The speed to respond to

8 the stimulus is used as a measure of the alertness level. A disadvantage here is that the

9 apparatus requires responses to a secondary task to infer alertness, thereby altering and possibly interfering with the primary task.

11 Other methods exist solely for increasing alertness, depending upon the user to self-12 evaluate alertness level and activate the device when the user feels drowsy. An example of the 13 latter is U.S. Patent No. 5,647,633 and related patents to M. Fukuoka in which a 14 method/apparatus is described for causing the user's seat to vibrate when the user detects drowsiness. Obvious disadvantages of such devices are that the user must be able to 16 accurately self-assess his/her current level of alertness, and that the user must be able to 17 correctly act upon this assessment.

18 Methods also exist to predict alertness level based on user inputs known empirically to 19 modify alertness. U.S. Patent No. 5,433,223 to M. Moore-Ede et al.
describes a method for predicting the likely alertness level of an individual at a specific point in time (past, current or 21 future) based upon a mathematical computation of a variety of factors (referred to as "real-22 world" factors) that bear some relationship to alterations in alertness.
The individual's Baseline 23 Alertness Curve (BAC) is first determined based on five inputs and represents the optimal 24 alertness curve displayed in a stable environment. Next, the BAC is modified by alertness modifying stimuli to arrive at a Modified Baseline Alertness Curve. Thus, the method is a 26 means for predicting an individual's alertness level, not cognitive performance.

27 Another method has been designed to predict "work-related fatigue" as a function of 28 number of hours on duty. Fletcher and Dawson describe their method in an article entitled "A
29 Predictive Model of Work-Related Fatigue Based on Hours of Work" published in Journal of Occupational Health and Safety, vol. 13, 471-485 (1997). In this model a simplifying 31 assumption is made - it is assumed that length of on-duty time correlates positively with time 32 awake. To implement the method, the user inputs a real or hypothetical on-duty/off-duty 33 (work/rest) schedule. Output from the model is a score that indicates "work-related fatigue."
21922696.1 6 Agent Ref: 67803/00007 1 Although this "work-related fatigue" score has been shown to correlate with some performance 2 measures, it is not a direct measure of cognitive performance per se. It can be appreciated that 3 the fatigue score will be less accurate under circumstances when the presumed relationship 4 between on-duty time and time awake breaks down - for example when a person works a short shift but then spends time working on projects at home rather than sleeping or when a person 6 works long shifts but conscientiously sleeps all the available time at home.
Also, this method is 7 obtrusive in that the user must input on-duty/off-duty information rather than such information 8 being automatically extracted from an unobtrusive recording device. In addition, the model is 9 limited to predictions of "fatigue" based on work hours. Overall, this model is limited to work-related situations in which shift length consistently correlates (inversely) with sleep length.

11 Given the importance of the amount of sleep and the time of day for determining cognitive 12 performance (and hence estimating productivity or effectiveness), and given the ever-increasing 13 requirements of most occupations on cognitive performance, it is desirable to design a reliable 14 and accurate method of predicting cognitive performance. It can be appreciated that increasing the number of relevant inputs increases cognitive performance prediction accuracy. However, 16 the relative benefits gained from such inputs must be weighed against the additional 17 burdens/costs associated with their collection and input. For example, although certain 18 fragrances have been shown to have alertness-enhancing properties, these effects are 19 inconsistent and negligible compared to the robust effects of the individual's sleep/wake history and the time of day. More important, the effect of fragrances on cognitive performance is 21 unknown. Requiring an individual to keep a log of exposure to fragrances would be time 22 consuming to the individual and only result in negligible gains in cognitive performance 23 prediction accuracy. In addition, while the effects of the sleep/wake history and the time of day 24 on cognitive performance are well known, the effects of other putative alertness-altering factors (e.g., job stress), how to measure them (their operational definition), and their direction of action 26 (cognitive performance enhancing or degrading) are virtually unknown.

27 An important and critical distinction between the present invention and the prior art is that 28 the present invention is a model to predict performance on tasks with a cognitive component. In 29 contrast, previous models involving sleep and/or circadian rhythms (approximately 24-hour) focused on the prediction of "alertness" or "sleepiness." The latter are concepts that specifically 31 relate to the propensity to initiate sleep, not the ability to perform a cognitive task.

32 Although sleepiness (or its converse, alertness) could be viewed as an intervening 33 variable that can mediate cognitive performance, the scientific literature clearly shows that 21922696.1 7 Agent Ref: 67803/00007 1 cognitive performance and alertness are conceptually distinct, as reviewed by Johns in the 2 article entitled "Rethinking the Assessment of Sleepiness" published in Sleep Medicine 3 Reviews, vol. 2, pp. 3-15 (1998) and as reviewed by Mitler et al. in the article entitled "Methods 4 of Testing for Sleepiness" published in Behavioral Medicine, vol. 21, pp.
171-183 (1996).
Thomas et al. in the article entitled "Regional Cerebral Metabolic Effects of Prolonged Sleep 6 Deprivation" published in Neurolmage, vol. 7, p. S130 (1998) reveal that 1-3 days of sleep loss 7 result in reductions in global brain activation of approximately 6%, as measured by regional 8 cerebral glucose uptake. However, those regions (heteromodal association cortices) that 9 mediate the highest order cognitive functions (including but not limited to attention, vigilance, situational awareness, planning, judgment, and decision making) are selectively deactivated by 11 sleep loss to a much greater extent -- up to 50% - after three days of sleep loss. Thus, 12 decreases in neurobiological functioning during sleep restriction/deprivation are directly 13 reflected in cognitive performance degradation. These findings are consistent with studies 14 demonstrating that tasks requiring higher-order cognitive functions, especially those tasks requiring attention, planning, etc. (abilities mediated by heteromodal association areas) are 16 especially sensitive to sleep loss. On the other hand, brain regions such as primary sensory 17 regions, are deactivated to a lesser degree. Concomitantly, performance (e.g., vision, hearing, 18 strength and endurance tasks) that is dependent on these regions is virtually unaffected by 19 sleep loss.

Consequently, devices or inventions that predict "alertness" perse (e.g., Moore-Ede et al.) 21 putatively quantify the brain's underlying propensity to initiate sleep at any given point in time.
22 That is, devices or inventions that predict "alertness" (or its converse "sleepiness") predict the 23 extent to which sleep onset is likely. The present invention differs from such approaches in that 24 the nature of the task is accounted for - i.e., it is not the propensity to initiate sleep that is predicted. Rather, the present invention predicts the extent to which performance of a particular 26 task will be impaired by virtue of its reliance upon brain areas most affected by sleep deprivation 27 (heteromodal association areas of the brain). The most desirable method will produce a highly 28 reliable and accurate cognitive performance estimate based on the sleep/wake history of an 29 individual, the time of day, and the amount of time on a particular task.
III. DISCLOUSRE OF THE INVENTION

31 A method for determining cognitive performance levels based on a data series having at 32 least one wake state and at least one sleep state in accordance with the invention includes 33 analyzing the data series to select the function; determining a cognitive performance capacity 21922696.1 8 Agent Ref: 67803/00007 1 using the selected function; modulating the cognitive performance capacity with a time of day 2 value; providing the modulated value; when a piece of data within the data series is 3 inconclusive, then selecting at least two functions, determining a cognitive performance capacity 4 using each function, modulating each cognitive performance capacity with a time of day value, and providing each modulated value; and repeating the above steps at least once.

6 A method for obtaining a cognitive performance level in accordance with invention 7 includes transmitting a data series representing wake states and sleep states of an individual to 8 an extemal device, and receiving at least one predicted cognitive performance level from the 9 external device.

A method for providing a cognitive performance level in accordance with the invention 11 includes receiving a data series representing at least one wake state and at least one sleep 12 state, selecting a function based on the data series, determining a cognitive performance 13 capacity using the selected function, modulating the cognitive performance capacity with a time 14 of day value, and providing the modulated value.

An apparatus for predicting cognitive performance in accordance with the invention 16 includes an input that receives data, a data analyzer connected to the input to select a 17 calculation function responsive to the received data, a calculator connected to the data analyzer 18 to calculate a cognitive performance capacity using the calculation function, an evaluator 19 connected to the input to determine a task value based on the received data, a memory that stores modulation data, and a modulator connected to the memory, the evaluator, and the 21 calculator.

22 A system for monitoring the cognitive performance level of each of a plurality of individuals 23 in accordance with the invention includes a plurality of data collectors, each of the data 24 collectors is attached to a respective individual, a receiver in communication with each of the plurality of data collectors, a data analyzer connected to the input to select a calculation function 26 responsive to the received data for at least one of the plurality of data collectors, a calculator 27 connected to the data analyzer to calculate a cognitive performance capacity using the 28 calculation function for at least one of the plurality of data collectors, an evaluator connected to 29 the input to determine a task value based on the received data for at least one of the plurality of data collectors, a memory that stores modulation data, and a modulator connected to the 31 memory, the evaluator, and the calculator.
21922696.1 9 Agent Ref: 67803/00007 1 An apparatus for providing a cognitive performance level in accordance with the invention 2 includes means for receiving a data series having at least one wake state and at least one sleep 3 state, means for selecting a function based on the data series, means for determining a 4 cognitive performance capacity using the selected function, means for storing a series of time of day values, means for modulating the cognitive performance capacity with a corresponding time 6 of day value, and means for providing the modulated value.

7 An apparatus for obtaining a cognitive performance level from an external device in 8 accordance with the invention includes means for transmitting a data series having at least one 9 wake state and at least one sleep state to the external device, and means for receiving at least one predicted cognitive performance level from the external device.

11 A feature of the present invention is that it provides a numerical representation of 12 predicted cognitive performance with an immediate ergonomic and economic advantage, i.e., an 13 indication of productivity or effectiveness of an individual. Another feature of the present 14 invention is that it does not require or use measurements/computations that are indirect, intermediate, inferential or hypothetical concomitants of cognitive performance. Examples of 16 the latter are alertness, sleepiness, time to sleep onset, body temperature and/or other 17 physiological measures that vary with time. A further feature of the invention is that it accounts 18 for transient or adventitious variations in cognitive performance from any source as a result of 19 how that source affects the sleep/wake history (e.g., age) and/or physiological time of day (e.g., shift work). In effect, such sources are not treated as having effects on cognitive performance 21 independent of the sleep/wake history and/or the time of day, and as such do not require 22 separate measurement, tabulation, and input into the method.

23 One objective of this invention is to provide an accurate method for predicting cognitive 24 performance of an individual.

A further objective is to provide a method that facilitates prediction of the effects of 26 possible future sleep/wake histories on cognitive performance (forward prediction).

27 Another objective is to provide a method that facilitates retrospective analysis of likely 28 prior cognitive performance based on the individual's sleep/wake history, the time of day, and 29 the activities done by the individual.

Another objective is to provide a method for coordination and optimization of available 31 sleep/wake time in order to obtain net optimal predicted cognitive performance for an individual 32 and/or a group of individuals.

21922696.1 10 Agent Ref. 67803100007 1 It can be appreciated that an implicit advantage and novelty of the method is its 2 parsimony. The method uses those factors possessing maximal predictive value (as 3 demonstrated empirically) as continuously updated inputs. Thus, the model will be simple to 4 implement. Other models predicting "alertness" require the user to track multiple input variables (e.g., caffeine, alcohol ingestion, light/dark exposure, diurnal type), rather than presenting these 6 inputs as optional "attachments" to a standard, simplified model based on those factors 7 accounting for maximum cognitive performance change. For example, in accordance with a 8 segment of the present method, the effects of age on cognitive performance are accounted for 9 implicitly via the empirically derived effects of age on sleep. That is, sleep quality degrades with age. The inherent degradation in sleep quality with aging would implicitly result in a prediction 11 of degraded cognitive performance (since in the present method degraded sleep results in a 12 prediction of degraded cognitive performance), even if an individual's age were unknown.
13 Therefore, age need not constitute a separate (independent) input variable to a cognitive 14 performance prediction model.

The invention also provides other significant advantages. For example, an advantage of 16 this invention is the elimination of a need for empirical evaluation.

17 Another advantage of this invention is obtaining an accurate prediction of cognitive 18 performance of an individual. The advantage may be achieved by a method incorporating three 19 factors that have been empirically demonstrated to exert a significant effect on cognitive performance, namely, (1) the individual's sleep/wake history, (2) the time of day ("day" herein 21 referring to a 24-hour period including both nighttime and daylight hours), and (3) the 22 individual's time on a particular task.

23 Another advantage achieved by this invention is an accurate prediction of current 24 cognitive performance.

Another advantage achieved by this invention is that it is capable of providing a real time 26 prediction of cognitive performance.

27 Yet another advantage achieved by this invention is a prediction of future expected 28 cognitive performance throughout the day based on hypothetical future sleep/wake periods.

29 An additional advantage achieved by this invention is a retrospective analysis of cognitive performance at given times.

31 A further advantage of the invention is that a particular cognitive performance prediction is 32 not based on normative data (i.e., does not require a "look-up table" for output), but rather is 21922696.1 11 Agent Ref. 67803/00007 1 calculated directly based on each individual's sleep/wake information, the time of day, and the 2 time on a task.

3 A further advantage of the invention is that it can be used to optimize the individual's 4 future sleep/wake schedule based on a fixed mission/work schedule. Previous methods and apparatuses are directed toward modifying the work schedule and/or mission to "fit the 6 individual." In most situations, however, work schedules and/or missions are fixed. Thus, 7 modifying the work schedule or mission to suit the individual is impractical or impossible. A
8 more reasonable approach incorporated in the present method is to allow the individual to 9 adjust his/her sleep/wake periods to meet work/mission demands. Thus, the current method presents a more practical alternative by providing a means to regulate work hours to a directly 11 applicable metric (cognitive performance) instead of regulating work hours by time off duty or by 12 using indirect measures of cognitive performance such as alertness.

13 A feature of this invention is the provision of a graphical representation that translates an 14 individual's sleep/wake history and the time of day into an immediately useful, self-evident index. A prediction of cognitive performance, unlike a prediction of "alertness" or "sleepiness,"
16 requires no further interpretation.

17 The method for predicting human cognitive performance in accordance with the invention 18 accomplishes the above objectives and achieves the above advantages. The method and 19 resulting apparatus are easily adapted to a wide variety of situations and types of inputs.

In accordance with an aspect of the invention, an individual sleep/wake history is inputted 21 into a processing device. The processing device classifies the individual pieces of sleep/wake 22 history data as either sleep or wake. Based on the classification of data, the processing device 23 selects and calculates a cognitive performance capacity corresponding to the present state of 24 the individual, the cognitive performance capacity may be modified by a time of day value to adjust the cognitive performance capacity to a predicted cognitive performance. The predicted 26 cognitive performance represents the ability of the individual to perform cognitive tasks. The 27 predicted cognitive performance may be displayed for a real-time indication or as part of a 28 curve, printed out with the information that could have been displayed, and/or stored for later 29 retrieval and/or use. The calculation of the cognitive performance capacity is made based on functions that model the effect of the interrelationship of sleep and being awake on cognitive 31 performance. The time of day function models the effect of an individual's circadian rhythms on 32 cognitive performance.

21922696.1 12 Agent Ref: 67803/00007 1 In accordance with the underlying method of the invention, the method can be 2 accomplished with a wide variety of apparatus. Examples of the possible apparatus 3 embodiments include electronic hardware as either dedicated equipment or equipment intemal 4 to a computer, software embodied in computer readable material for use by computers, software resident in memory or a programmed chip for use in computers or dedicated 6 equipment, or some combination of both hardware and software. The dedicated equipment 7 may be part of a larger device that would complement the dedicated equipment's purpose.
8 Given the following enabling description of the drawings, the invention should become 9 evident to a person of ordinary skill in the art.

IV. BRIEF DESCRIPTION OF THE DRAWINGS

11 Figure 1(a) is a conceptual diagram representation of the invention including the fine-12 tuning alternative embodiment. Figure 1(b) graphically shows the combination of output from 13 the functions represented by Figure 3(a) with time of day modulation to derive predicted 14 cognitive performance.

Figure 2 is a block diagram representation of the wake, sleep, delay, and sleep inertia 16 functions for calculating predicted cognitive performance capacity.

17 Figure 3(a) graphically illustrates the effect of being awake and asleep on cognitive 18 performance capacity over a 24-hour period. Figure 3(b) is an enlarged view of circled portion 19 3(b) of Figure 3(a), and graphically shows the delay function with respect to cognitive performance capacity. Figure 3(c) is an enlarged view of circled portion 3(c) of Figure 3(a), and 21 graphically shows the sleep inertia function with respect to cognitive performance capacity.

22 Figures 4(a)-(b) depict a detailed flowchart showing the steps of the method of the 23 invention.

24 Figure 5 illustrates time on task effects across a 10-minute Psychomotor Vigilance Task (PVT) sessions at two hour increments during 40 hours of total sleep deprivation.

26 Figure 6 depicts a functional representation of an alternative embodiment.

27 Figure 7(a) illustrates a block diagram of structural components for the preferred 28 embodiment. Figure 7(b) illustrates a block diagram of an alternative set of structural 29 components.

Figures 8(a)-(b) depict a detailed flowchart showing the steps of an alternative 31 embodiment.

21922696.1 13 Agent Ref: 67803l00007 1 V. MODES FOR CARRYING THE DESCRIBED EMBODIMENTS

2 The present invention now is described more fully hereinafter with reference to the 3 accompanying drawings, in which preferred embodiments of the invention are shown. This 4 invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this 6 disclosure will be thorough and complete, and will fully convey the scope of the invention to 7 those skilled in the art. The present invention will now be described more fully hereinafter with 8 reference to the accompanying drawings, in which preferred embodiments of the invention are 9 shown. Like numbers refer to like elements throughout.

As will be appreciated by one of skill in the art, the present invention may be embodied as 11 a method, data processing system, or computer program product. Accordingly, the present 12 invention may take the form of an entirely hardware embodiment, an entirely software 13 embodiment or an embodiment combining software and hardware aspects.
Furthermore, the 14 present invention may take the form of a computer program product on a computer-usable storage medium having computer-usable program code means embodied in the medium. Any 16 suitable computer readable medium may be utilized including hard disks, CD-ROMs, optical 17 storage devices, or magnetic storage devices.

18 Computer program code for carrying out operations of the present invention may be 19 written in an object oriented programming language such as Java , Smalltalk or C++. However, the computer program code for carrying out operations of the present invention may also be 21 written in conventional procedural programming languages, such as the "C"
programming 22 language.

23 The program code may execute entirely on the user's computer, as a stand-alone 24 software package; on a remote computer; or it may execute partly on the user's computer and partly on a remote computer. In the latter scenario, the remote computer may be connected 26 directly to the user's computer through a LAN or a WAN (Intranet), or the connection may be 27 made indirectly through an external computer (for example, through the Internet using an 28 Internet Service Provider). The invention may be implemented as software that may be resident 29 on a stand-alone device such as a personal computer, a PAL device, a personal digital assistant (PDA), an e-book or other handheld or wearable computing devices (incorporating Palm OS, 31 Windows CE, EPOC, or future generations like code-named products Razor from 3Com or 32 Bluetooth from a consortium including IBM and Intel), or a specific purpose device having an 33 application specific integrated circuit (ASIC).

21922696.1 14 Agent Ref: 67803/00007 1 The present invention is described below with reference to flowchart illustrations of 2 methods, apparatus (systems) and computer program products according to an embodiment of 3 the invention. It will be understood that each block of the flowchart illustrations, and 4 combinations of blocks in the flowchart illustrations, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general 6 purpose computer, special purpose computer, or other programmable data processing 7 apparatus to produce a machine, such that the instructions, which execute via the processor of 8 the computer or other programmable data processing apparatus, create means for 9 implementing the functions specified in the flowchart block or blocks.

These computer program instructions may also be stored in a computer-readable memory 11 that can direct a computer or other programmable data processing apparatus to function in a 12 particular manner, such that the instructions stored in the computer-readable memory produce 13 an article of manufacture including instruction means which implement the function specified in 14 the flowchart block or blocks. Examples of how the software can be stored for use are the following: in random access memory (RAM); in read only memory (ROM); on a storage device 16 like a hard drive, disk, compact disc, punch card, tape or other computer readable material; in 17 virtual memory on a network, a computer, an intranet, the Intemet, the Abilene Project, or 18 otherwise; on an optical storage device; on a magnetic storage device;
and/or on an EPROM.
19 The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be 21 performed on the computer or other programmable apparatus to produce a computer 22 implemented process such that the instructions which execute on the computer or other 23 programmable apparatus provide steps for implementing the functions specified in the flowchart 24 block or blocks.

The present invention involves a method for predicting cognitive performance at a given 26 time in the past, present, or future as a consequence of the amount of sleep and wakefulness 27 up to that time as a function of the time of day and the workload for a particular individual. The 28 method calculates a numerical estimate of cognitive performance for an individual as a 29 continuous function of time. The calculations (described below) are based on empirically derived direct mathematical relationships among (1) the continuous decrement of cognitive 31 performance during wakefulness; (2) restoration of cognitive performance during sleep; (3) 32 cyclic variation in cognitive performance during the course of the day; and (4) variations in 33 cognitive performance due to whether and what activities are occurring.

21922696.1 15 Agent Ref: 67803/00007 1 In accordance with the invention, a numeric value indicating predicted cognitive 2 performance at a given moment in time is provided as shown in Figures 1 (a)-4(b). As shown in 3 Figure 1(a), predicted cognitive performance equals the output of a series of calculations and/or 4 determinations obtained in three general steps, using functions empirically derived from direct measurements of cognitive performance under scientifically controlled conditions. The first 6 step, as shown in Figure 2, preferably uses a set of functions to calculate an initial value 7 referred to as the level of cognitive performance capacity as graphically depicted in Figures 8 3(a)-(c). Once the level of cognitive performance capacity is calculated, the second step 9 preferably calculates or uses a previously calculated time of day modulator M represented as G8 in Figure 1(b) and S8 in Figure 4(b). The third step preferably calculates a task modulator T
11 represented as S9-S10(b) in Figure 4(b). Alternatively, the second and third steps may be 12 switched and/or combined. The fourth step preferably involves the mathematical combination of 13 the results from the first through third steps yielding a predicted cognitive performance, shown 14 as a block diagram in Figure 1(a) and graphically represented in Figure 1(b), which illustrates the combination of the cognitive performance capacity and the time of day modulator.

16 There are four functions relating to the sleep/wake history used to calculate the level of 17 cognitive performance capacity as shown in Figures 2-4(b). The wake function w(t) quantifies 18 empirically derived relationships between the time awake and degradation of cognitive 19 performance. The sleep function s(t) quantifies empirically derived relationships between the time asleep and maintenance and/or recuperation of cognitive performance. In addition to these 21 two primary functions that operate during the bulk of the time awake or asleep there are two 22 other functions that operate briefly during the transition from one state to the other. They 23 include the delay of recuperation function d(t) and the sleep inertia function i(t). The delay of 24 recuperation function d(t) represents the relationship between the wake to sleep transition and the recuperation of cognitive performance. This function operates during the initial period of 26 sleep following being awake, known as stage 1 sleep, as shown in Figure 3(b). The sleep 27 inertia function i(t) represents the relationship between the sleep to wake transition and 28 cognitive performance. This function operates during the initial period of time being awake after 29 being asleep as shown in Figure 3(c).

The function representing the time of day effects on cognitive performance is used to 31 calculate a modulating factor M. The time of day function describes empirically derived 32 relationships between the time of day (point in time within a 24-hour period) and the variation in 33 cognitive performance over the course of the day as exemplified by G8 in Figure 1(b).

21922696.1 16 Agent Ref. 67803/00007 1 The function representing the task/activity impact on cognitive performance is used to 2 calculate a modulating factor T. The task function describes the impact of the performance of a 3 task and/or an activity upon cognitive performance preferably based upon, for example, the 4 intensity, the length, the complexity, and the difficulty associated with the particular task and/or activity. Figure 5 illustrates the impact of performing a task across a 10-minute Psychomotor 6 Vigilance Task (PVT) session at two hour increments during 40 hours of total sleep deprivation.
7 For each PVT session, except the last one, there was an improvement from trial 10 of one PVT
8 session to trial 1 of the next PVT session.

9 A mathematical operation, shown in Figure 1(b) as multiplication, is used to combine the results from the first, second, and third steps into a single predicted cognitive performance curve 11 E in the fourth step.

12 Using the preferred embodiments, predicted cognitive performance E can theoretically 13 reach an index level of 120, but only when cognitive performance capacity C
is an index level of 14 100 (i.e., 20 minutes after awakening from a sleep period in which cognitive performance capacity C was fully restored) and simultaneously the time of day function M
is at its acrophase.
16 Although possible, in practice this situation is unlikely.

17 The inputted data S2 into the method includes a representation of an individual's 18 sleep/wake history and task information. The sleep/wake history is a time series or temporal 19 record based on local clock time. Each successive period, interval or epoch identifies one of two mutually exclusive states, namely wake or sleep. The task information is a series of 21 information regarding what the individual is or is not performing in the way of activities/tasks and 22 alternatively the intensity, the difficulty, the length and/or the complexity of the activity/task may 23 be included in the task information. Both the sleep/wake history and the task information are 24 not necessarily "historical" in the sense of occurring in the past, but may for example be hypothetical, projected, idealized, or contemplated. The latter in particular are appropriate for 26 the predictive uses of this method.

27 The gold standard for measuring sleep and wakefulness is polysomnography (PSG).
28 PSG sleep scoring is based on the concurrent recording, or at least recording in such a way as 29 allows the latter synchronization (typically with time-stamping or time-linking) of the data, of electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG).
These 31 signals are typically then visually inspected on an epoch-by-epoch basis (each epoch 32 traditionally is 30 seconds in length for PSG) to determine an individual's stage of sleep or 33 wakefulness. Polysomnographic sleep scoring distinguishes between wake, non-rapid eye 21922696.1 17 Agent Ref: 67803/00007 1 movement sleep (NREM) and rapid eye movement sleep (REM), with NREM sleep being further 2 distinguished into four stages (stages 1, 2, 3, and 4) on the basis of characteristic EEG markers.
3 PSG is not a practical method for determining sleep and wakefulness in applied settings (e.g., 4 while driving, working, or on the battlefield), because PSG requires that individuals be attached to sensors or electrodes that connect with a recording device, and currently the only accepted 6 method for scoring PSG is by visual inspection of the recorded EEG, EOG, and EMG results.

7 Presently, if a computer is used for scoring PSG, then typically a human reviews the 8 results for accuracy in the scoring, because computer scoring has not been approved by the 9 American Sleep Disorders Association. Also, recently, researchers have been exploring whether spectrally analyzed PSG or similar data using Fast Fourier Transforms might provide a 11 better measurement of sleep in humans than PSG scoring.

12 A preferred method of determining sleep from wakefulness would be a device that is 13 portable, unobtrusive, reliable, and whose recordings can be scored automatically. One such 14 method is monitoring of movement activity, or actigraphy. The movement activity device is typically worn on the non-preferred wrist, but may be placed elsewhere on an individual (e.g., 16 the ankle). When worn on the non-preferred wrist, these devices have been shown to 17 accurately quantify sleep and wakefulness as compared to the standard provided by PSG
18 (reliabilities as high as 90%).

19 The most widely used method of scoring actigraphy data is an algorithm developed by Cole and associates and described in their article entitled "Automatic Sleep/Wake Identification 21 from Wrist Actigraphy" published in Sleep, vol. 15, pp. 461-469 (1992).
Successful actigraphy 22 sleep-scoring algorithms such as the Cole et al. algorithm (also known as the Cole-Kripke 23 algorithm) are for use with conventional (number-of-zero-crossings) actigraphs, and some 24 algorithms account for the number of counts above a certain threshold.
These algorithms are limited to making simple sleep vs. wake distinctions, and cannot distinguish sleep stage changes 26 (e.g., Stage 1 to Stage 2, or Stage 2 to REM) within sleep itself.
Consequently, such algorithms 27 cannot discriminate recuperative sleep (stages 2, 3, 4, and REM) from non-recuperative sleep 28 (stage 1).

29 More recently, digital signal processing (DSP) actigraphs have begun to be developed.
Because the DSP actigraph will provide much more information than just the conventional number 31 of zero crossings or counts above threshold (this and other information provided by a conventional 32 actigraph will, however, be retained), it shows promise for distinguishing between different sleep 33 stages. Thus, s{eep scoring systems for DSP wili not only replace, but will also make irrelevant, 21922696.1 18 Agent Ref: 67803/00007 1 the Coie-Kripke algorithm. A sleep scoring system for the DSP will be developed as the DSP
2 database of empirical data from use of DSP actigraphs increases.

3 Other algorithms and methodologies for automated actigraphy scoring have been 4 developed by, for example, Jean-Louis et al., 1996; Sadeh et al., 1989; and Zisapel et al., 1995.
Each of these scoring systems shows considerable promise, especially for scoring the 6 actigraphically recorded sleep/wake states of individuals with sleep disorders or other medical 7 disorders. Available scoring systems mainly differ along technical aspects, for example, the 8 extent to which activity counts in previous and subsequent epochs influence the scoring of the 9 current epoch; and variation among mathematical principles underlying each scoring system.
As one of ordinary skill in the art will realize, any actigraph scoring system is capable of 11 providing the sleep/wake data input for the method of this invention.

12 The sleep/wake history will preferably take the form of a data series. The sleep/wake 13 history may include past, present, and/or future (predicted) sleep/wake patterns of an individual.
14 The sleep/wake history is a representation of a state of an individual as either being asleep or awake and is divided into epochs. The epochs are the same length, but that length could be of 16 any time period as dictated by restraints of the method and apparatus used to collect data 17 and/or the desired precision of the sleep/wake pattern. The PSG or similar scoring can be 18 converted into a sleep/wake history for an individual.

19 It can be appreciated that the accuracy of the cognitive performance prediction is directly related to the accuracy of the sleep/wake history input and the sleep scoring system used to 21 interpret the sleep/wake states of an individual. One possible source of inaccuracy may arise 22 from the temporal resolution of the input epoch or interval. That is, the shorter the input epoch, 23 the higher the temporal resolution and consequent moment-to-moment accuracy of the 24 sleep/wake input. For example with actigraphy, past experience indicates that the most effective length of an epoch is one minute. Another source of inaccuracy may arise from 26 ambiguity in the sleep/wake discrimination itself. In the event that the history input is ambiguous 27 (i.e., the sleep or wake state is uncertain), the calculation of predicted cognitive performance 28 can be performed twice concurrently, once for each possible state (sleep or wake), resulting in a 29 dual output representing the possible range of expected cognitive performance. One of ordinary skill in the art will appreciate that the dual output can be further divided if there is more than one 31 ambiguity in the sleep/wake history. Such treatment in executing the functions expressed below 32 is included as a component of this method and any implementing apparatus.

21922696.1 19 Agent Ref: 67803/00007 1 The method of this invention is not limited with regard to time or technique: on-line/real-2 time vs. off-line/post-hoc; or incremental, iterative vs. discrete solutions to continuous forms of 3 those equations.

4 A preferred embodiment of the method encompasses a mathematical model that expresses predicted cognitive performance capacity E at time t as a modulation of the current 6 cognitive performance capacity C by a time of day function M by a task function T. It can be 7 written as a general description in its simplest form as:

8 E= C o M o T Equation 1 9 where V represents a mathematical operator. Any mathematical operator may be used to combine cognitive performance capacity C, day of time function M, and task function T. The 11 form and nature of time of day function M and/or task function T dictate the exact operator that 12 is most desirable. There may be two different operators used to express the predicted cognitive 13 performance capacity E such that the first V may be one mathematical operator and the second 14 V may be a second mathematical operator. Altematively, the modulations could be performed in two steps where two of the items are modulated with the resulting modulated value being 16 modulated with the third of the items. Most preferably, Equation 1 a below would be used to 17 combine cognitive performance capacity C, day of time function M, and task function T.

18 E =C* M' T Equation 1 a 19 In the altemative, Equation 1 b below could also be used to combine cognitive performance capacity C, time of day function M, and task function T.

21 E= C+ M+ T Equation lb 22 Cognitive performance capacity C represents a function of sleep/wake history, that is 23 C = w(t) + s(t) + d(t) + i(t) Equation 2 24 where w(t), s(t), d(t), and i(t) are the instantaneous values of the wake, sleep, delay, and sleep inertia functions at time t. Time of day function M represents a function of the time of day, such 26 that 27 M = m(t) Equation 3 28 where m(t) is the instantaneous value of the time of day function at time t. Task function T
29 represents a function of the impact of performing or not performing a task when the individual is awake, such that 21922696.1 20 Agent Ref 67803/00007 1 T = t(t) Equation 4 2 In keeping with the invention, a four-step process may be performed after either an initial 3 setting of the starting time t, the starting cognitive performance capacity C, and the time of the 4 last transition tLs when appropriate in S1 of Figure 4(a) where these data can be entered in any order. In the first step, the level of cognitive performance capacity C at time t may be calculated 6 based on an individual's sleep/wake history using functions w(t), s(t), d(t), and i(t) as 7 represented by S3-S7e in Figures 4(a)-(b). In the second step, time of day modulator M may be 8 calculated using the time of day function as represented by S8 in Figure 4(b). According to an 9 aspect of the invention, the second step can be performed once to provide a series of data points in time sequential order for multiple executions of the first step. In the third step, task 11 modulator T may be calculated using the task function as represented by S9a through SlOc in 12 Figure 4(b). In the fourth step, predicted cognitive performance E may be derived from the 13 combination of cognitive performance capacity C and time of day modulator M
resulting in 14 cognitive performance capacity C being modulated by time of day modulator M
being modulated by task modulator T as illustrated by S11 in Figure 4(b).

16 First Step: Calculation (or Determination) of Cognitive Performance Capacity C

17 Figure 2 is a schematic flow diagram representing the use of the functions described 18 below. Examples of the calculations discussed are graphically illustrated in Figures 3(a)-(c).
19 Figures 4(a)-(b) are a detailed flowchart of the steps in the method. As a preferred embodiment of the model, cognitive performance capacity C is herein assigned index values preferably 21 having a total range of zero to 120. The ranges in this application are intended to encompass 22 the end points of the stated numerical range. However, cognitive performance capacity C may 23 be scaled to other values or units for specific applications, for example, zero to 100.

24 In the preferred embodiment, only one of the four functions w(t), s(t), d(t), and i(t) operates at any given interval of time, and the others are equivalent to zero in Equation 2 as represented 26 by S7a through S7d. Functions w(t) and s(t) describe the non-transition states, while functions 27 d(t) and i(t) describe the transition states. For instance in a non-transition state, when the 28 individual is awake, function s(t) is set to zero, and when the individual is asleep, function w(t) is 29 set to zero. Likewise, during specific intervals of transition from wake to sleep and vice versa, only one of the transition functions d(t) or i(t) operates, the other being set equal to zero. When 31 there is a change between sleep and wake, or vice versa, a time counter tLS
is reset to keep 32 track of the time in the present state for determining decision rules for the transition functions 33 d(t) and i(t) as shown in Figure 4(b).

21922696.1 21 Agent Ret 67803/00007 1 (1) Wake function (w(t)) 2 The wake function S7a represents the depletion of cognitive performance capacity with 3 the passage of time awake. It is based on evidence that (1) near-100%
cognitive performance 4 is maintained from day to day when individuals obtain eight hours of sleep each night; and (2) cognitive performance appears to decline by approximately 25% for every 24 hours of 6 wakefulness.

7 In S7a, the wake function w(t) calculates the current value of cognitive performance 8 capacity C resulting from the decay in cognitive performance capacity that occurs over an 9 interval of time from t-1 to t, which in the preferred embodiment is the length of one epoch. As noted above, this calculation is performed independent of and prior to modulation of cognitive 11 performance capacity C by the time of day function M in S9. A generalized form of the wake 12 function is given by the equation:

13 CW = w(t) Equation 5 14 where wake function w(t) may be any negative-valued function decreasing with t. More preferably, the wake function w(t) is a linear function depleting performance at a constant rate, 16 and, most preferably, the wake function w(t) is expressed at time t as follows:

17 w(t) = Ct_, - kW Equation 5a 18 where the interval of wakefulness is from t-1 to t (in epochs) and the decay in performance per 19 minute is k,H. Thus, if t-1 to t is not one minute, then kW is appropriately adjusted. The total range of kw is any positive real number, and preferably kW is a range of .003 to .03 of a point per 21 minute, and most preferably kW is equal to approximately 1 point per hour or 0.017 of a point per 22 minute. The value kW is based on empirical data showing that cognitive performance declines 23 by approximately 25 points for every 24 hours of continuous wakefulness.
Equation 4a is 24 represented in Figures 2 and 4(b) at S7a. An example is illustrated as the wake function in Figure 3(a), for an initial cognitive performance capacity of 100 index points, a decay rate of 26 0.017 of a point per minute, over an interval of 16 hours (960 minutes).

27 (2) Sleep function (s(t)) 28 The sleep function S7c restores cognitive performance capacity with the passage of time 29 asleep. The sleep function s(t) is based on empirical evidence that the recuperative value of sleep on cognitive performance accumulates in a nonlinear manner. That is, the rate of 31 cognitive performance capacity recuperation is higher initially during sleep and slows as the 21922696.1 22 Agent Ref. 67803/00007 1 time asleep accumulates. Other data indicates that sleep beyond a certain point confers little or 2 no additional benefit for cognitive performance and the rate of recuperation approaches zero.
3 Thus, for example, two hours of sleep are not twice as recuperative as one hour of sleep. The 4 sleep function increases cognitive performance capacity at a rate that depends on the current level of cognitive performance capacity -- the lower the initial cognitive performance capacity, 6 the more rapidly recuperation accumulates. In other words, preferably the slope of a tangential 7 line for a particular cognitive performance capacity index level is the same each time that index 8 level is reached during different sleep periods.

9 For example, following a full day (16 hours) of wakefulness, during ensuing nighttime sleep recuperation accumulates rapidly early in the night. As cognitive performance capacity is 11 restored across the sleep period, the rate of recuperation declines.
Following sleep deprivation, 12 initial cognitive performance capacity is even lower than it would be following a normal 16-hour 13 day, and the rate of recuperation is even higher than at the beginning of recovery sleep. During 14 chronic partial sleep deprivation, cognitive performance capacity may not be completely restored each night despite this more rapid initial recuperation rate.

16 The sleep function calculates the current value of cognitive performance capacity C
17 resulting from the recovery of capacity that occurs while an individual is asleep over an interval 18 of time T (from t-1 to t). As noted above, this calculation is performed independent of, and prior 19 to, modulation of C by the time of day function M and modulation by the task function T. A
generalized form of the sleep function is given by the equation:

21 CS = s(t) Equation 6 22 where sleep function s(t) may be any positive-valued function increasing with t, and more 23 preferably the sleep function s(t) is an exponential function. This is based on empirical data 24 showing that cognitive performance restoration during sleep is nonlinear, with the rate of recuperation highest initially and gradually slowing as sleep continues. Thus, the most 26 preferred sleep function is an exponential function, which in its discrete form is stated as:

27 Ct = C t_, + (100 - Ct_1) I kS Equation 6a 28 where the interval of sleep is from t-1 to t (in minutes), the maximum cognitive performance 29 capacity value is 100 index points, Ct_, is cognitive performance capacity in the period preceding time t, and ks is the recuperation "time constant". In other words, ks is the time required to fully 31 restore cognitive performance capacity C if it was restored at a constant rate equal to the initial 32 slope of the curve. The recuperation time constant ks is derived empirically from partial sleep 21922696.1 23 Agent Ref. 67803/00007 1 deprivation data and is selected based on the length of the epoch. In accordance with the 2 preferred embodiment, ks is equal to any positive real number. For example, ks may be in the 3 range of 100 to 1000 with an epoch length of one minute, and, more particularly may be 4 approximately 300 with an epoch length of one minute. However, the optimum values for ks will depend at least in part on the length of the epoch. Equation 6a is represented in Figures 2 and 6 4(b) as S7c. A graphical example is illustrated as the sleep function in Figure 3(a), using an 7 initial cognitive performance capacity level of 100 index points, and using a time period of one 8 minute and ks = 300, the effect of eight hours of sleep following 16 hours of wakefulness.

9 (3) Delay function d(t) for wake to sleep transitions The delay of recuperation function d(t) defines the duration of an interval after sleep onset 11 during which recuperation of cognitive performance capacity from the sleep function is delayed.
12 During this interval, the wake function degradation of cognitive performance capacity continues 13 as represented by S7d in Figure 4(b). By preventing immediate accumulation of cognitive 14 performance capacity at the beginning of a sleep period or following awakenings from sleep, this delay adjusts the cognitive performance capacity calculation S6b.

16 The delay of recuperation function is based upon empirical studies showing that the first 17 few minutes of sleep are generally comprised of stage 1 sleep, which is not recuperative for 18 sustaining cognitive performance capacity. Frequent arousais to wake or stage 1 sleep (sleep 19 fragmentation) drastically reduce the recuperative value of sleep on cognitive performance capacity. Available data suggest that five minutes is the approximate length of time required to 21 return to recuperative sleep (stage 2 or deeper sleep) following an arousal to wake or stage 1 22 sleep. If many hours of sleep are obtained without interruption, then the delays make only a 23 small difference in overall restoration of cognitive performance capacity.
If sleep is interrupted 24 with frequent awakenings, the delays in recuperation after each awakening will accumulate, and thus substantially reduce total cognitive performance capacity restored during the total sleep 26 period.

27 The delay function specifies the duration of a sleep interval during which application of the 28 sleep function is postponed and a transitional formula is applied. A
generalized form of the 29 delay function for wake to sleep transitions is expressed as a decision rule:
d(t): IF (t - tLs) kd 31 THEN Ct = d(t) 32 ELSE Ct = s(t) Equation 7 Agent Ref: 67803/00007 1 where LS stands for last state change, and thus the wake to sleep transition time tLs denotes the 2 time of the last wake interval preceding a contiguous series of sleep intervals. This decision 3 rule is shown in Figures 2 and 4(b) as S6b, S7c and S7d taken together. For calculating 4 cognitive performance capacity during the interval kd, cognitive performance capacity Ct is evaluated by a transitional formula Ct = d(t). After kd has elapsed, Ct =
s(t). Note that if 6 wakefulness ensues before the end of kd, then Ct never reverts to s(t). That is the sleep 7 function is not applied during the brief sleep interval.

8 It is believed that the preferred range for kd is from 0 to 30 minutes, more preferably kd 9 equals about five minutes measured from the time of sleep onset before recuperation is derived from sleep. Preferably d(t) equals w(t). One of ordinary skill in the art wifl realize there are a 11 variety of factors that influence the length of kd. Thus a more preferred delay function may be 12 expressed as:

13 d(t): IF (t - tLs) <_ 5 14 THEN Ct = w(t) ELSE Ct = s(t) Equation 7a 16 The effects of delayed recovery on cognitive performance capacity, as embodied by Equation 17 7a, are graphically illustrated in detail in Figure 3(b).

18 As one of ordinary skill in the art will appreciate, PSG or similar scoring is able to classify 19 when stage 1 sleep occurs. The conversion of PSG or similar scoring data would then convert the occurrences of stage 1 sleep into wake data for the sleep/wake history.
Consequently, 21 when the sleep/wake history is based on converted PSG or similar scoring data, the delay 22 function d(t) is not necessary for the determination of an individual's cognitive performance 23 capacity. Alternatively, the delay function could be determined based upon when the individual 24 entered stage 2 or deeper sleep instead of using the kd value, and that once stage 2 or deeper sleep is reached then the sleep function s(t) would be used.

26 Altematively, the delay function d(t) may simply maintain the cognitive level of Ct at the 27 beginning of the delay period, i.e., CtLs.

28 (4) Sleep inertia function i(t) for sleep to wake transitions 29 The sleep inertia function i(t) defines the duration of an interval after awakening from sleep during which manifest cognitive performance capacity is suppressed below the actual current 31 level. The sleep inertia function i(t) is based upon empirical data showing that cognitive 21922696.1 25 Agent Ref: 67803/00007 1 performance is impaired immediately upon awakening, but improves primarily as a function of 2 time awake. It is also based on positron emission tomography studies showing deactivated 3 heteromodal association cortices (those areas that mediate this cognitive performance) 4 immediately upon awakening from sleep, followed by reactivation of these areas over the ensuing minutes of wakefulness. That is, actual cognitive performance recuperation realized 6 during sleep is not apparent immediately after awakening. The data indicate that 20 minutes is 7 the approximate length of time required for cognitive performance capacity to return to levels 8 that reflect actual recuperation accrued during sleep.

9 A sleep inertia delay value K. specifies the duration of the interval after awakening during which manifest cognitive performance capacity may be transitionally suppressed below the 11 sleep-restored cognitive performance capacity level. During this interval, a transitional function 12 bridges from an initial level to that determined by the wake function alone. A generalized form 13 of the sleep inertia function for sleep to wake transitions is expressed as a decision rule:

14 i(t): IF (t - tLs) < k;
THEN C, = i(t) 16 ELSE Ct = w(t) Equation 8 17 where the sleep to wake transition time tLs denotes the time of the last sleep interval preceding 18 a contiguous series of wake intervals. For calculating cognitive performance capacity during the 19 interval k;, Ct is evaluated by a transitional formula Ct = i(t). After ki has elapsed, Ct= w(t).
Equation 8 is represented in Figures 2 and 4(b) as S6a, S7a and S7b taken together.

21 The preferred range for K. is from 0 to about 60 minutes, and preferably in the range of 22 about 10 to about 25 minutes, and most preferably between 18 and 22 minutes.

23 The sleep inertia function i(t) may be any function over the interval 0 to k;, preferably any 24 negatively accelerated function. A preferred sleep inertia function i(t) is a simple quadratic equation. This function preferably suppresses cognitive performance capacity by 10% to 25%
26 immediately upon awakening, and most preferably by 25%. The function recovers 75% of the 27 suppressed cognitive performance capacity in the first 10 (or about half of k;) minutes after 28 awakening and 100% of the suppressed cognitive performance capacity usually by 20 minutes 29 after awakening, after which the wake function resumes. These values are based on empirical data concerning the transition from sleep to wake. These studies show that cognitive 31 performance is impaired immediately upon awakening from sleep, that the bulk of this 32 impairment dissipates within the first few minutes of awakening, and that approximately 20 21922696.1 26 Agent Ref: 67803/00007 1 minutes is required for performance to be fully restored. Using the preferred 25% suppression 2 of cognitive performance capacity and 20 minute recovery time, the preferred form of the sleep 3 inertia function is expressed as a decision rule:

4 i(t): IF (t - tLS) < 20 THEN Ct = CSW * [0 .75 + 0.025 (t - tLS) - (0.025 (t - tLS) )2]

6 ELSE Ct = w(t) Equation 8a 7 where CSw is cognitive performance capacity at the end of the sleep period calculated by the 8 sleep function at the sleep to wake transition time tLS. This decision rule is shown in Figures 2 9 and 4(b) as S6a, S7a, and S7b taken together. Equation 8a illustrates an initial suppression of 25% and ki equal to 20 minutes, and a negatively accelerated ramp bridging the interval until the 11 wake function w(t) resumes its effects. The effect of the sleep inertia function i(t) on cognitive 12 performance capacity, as embodied by Equation 8a, is graphically illustrated in Figure 3(c).

13 An alternative variant of the sleep inertia function i(t) is a linear equation based on k; equal 14 to 10 minutes and an initial 10% decrease in cognitive performance capacity. The resulting decision rule is then:

16 i(t): IF (t - tLs) < 10 17 THEN Ct = CsW *[0.9 + (t - tLS) / 100]

18 ELSE Ct = w(t) Equation 8b 19 As one of ordinary skill in the art will realize, both Equations 8a and 8b can be adjusted for a change in the value of k; and amount of the initial suppression of cognitive performance 21 capacity.

22 Second Step: Calculation of the time of day modifier M
23 (1) Time of day function m(t) 24 The time of day function m(t) shown at S8 in Figure 4(b) describes the cyclical 24-hour variation in cognitive performance. The time of day function m(t) is based on empirical data 26 showing that under constant routine and/or total sleep deprivation conditions (i.e., with 27 sleep/wake history controlled), cognitive performance capability oscillates between 28 approximately 5% to approximately 20% peak to peak over a 24-hour period.
This effect is 29 commonly attributed to circadian rhythms of individuals. Output from this function modulates the current cognitive performance capacity prediction C (calculated in the first step) according to 29 922696.1 27 Agent Ret 67803/00007 1 the time of day. The result of this modulation is the predicted cognitive performance capacity E.
2 A generalized form of the time of day function is given by 3 M = m(t) Equation 9 4 where m(t) can be any rhythmic function with a base period of 24 hours, and, preferably, m(t) is the sum of two sinusoids, one with a period of 24 hours and the second with a period of 12 6 hours, which provides a biphasic circadian component. This function may be based on 7 empirical data showing that a considerable proportion of variability seen in cognitive 8 performance measurements can be accounted for by two such sinusoidal waveforms. As 9 previously noted, the peak in empirically observed cognitive performance capacity occurs usually between 8:00 PM and 10:00 PM, and the trough occurs usually between 2:00 AM and 11 6:00 AM, providing a variation of approximately 5% to approximately 20%
each day. A
12 secondary trough occurs usually around 3:00 PM. Using these values for the preferred form of 13 function m(t), the resulting function accounts for the empirically demonstrated asymmetry of 14 daily cognitive performance rhythms, with a mid-aftemoon decrease.

The descriptive form of the function m(t), including its offset and amplitude values varies 16 with the operator selected for the third step. The computed value of the function can be 17 expressed either as an additive percentage of cognitive capacity (dependent or independent of 18 the current value of cognitive performance capacity C) or as a multiplicative dimensionless 19 scalar. The preferred form of the function, using the multiplicative operator, is expressed as m(t) = F+(A, * cos(2II(t - V,) / P,) + A2 * cos(211(t - V2) / P2)) Equation 9a 21 where F is an offset, t is the time of day, P, and P2 are periods of two sinusoids, V, and V2 are 22 the peak times of day in time units or epochs past midnight, and A, and A2 are amplitudes of 23 their respective cosine curves. This function may be used to modulate the previously calculated 24 cognitive performance capacity C. Equation 9a is shown as S8 in Figures 1(a) and 4(b) and graphically illustrated as G8 in Figure 1(b). As shown in Figure 4(b), t is an input in the time of 26 day function m(t) for each epoch of data.

27 For example in a preferred embodiment the variables are set as follows: t is the number of 28 minutes past midnight, P, is equal to 1440 minutes, P2 is equal to 720 minutes, V, is equal to 29 1225, and V2 is equal to 560. Further, when A, and A2 are represented as scalars, their amplitudes are in a range from 0 to 1, and more preferably are in a range from 0.01 to 0.2, and 31 most preferably A1 is equal to 0.082 and A2 is equal to 0.036. Further in this example F is equal 32 to either 0 or 1, and more preferably F is equal to 1. The resulting value of the time of day 21922696.1 28 Agent Ref: 67803/00007 1 function m(t), in this example, is in the range of 0 to 2, and preferably in the range of 0.8 to 1.2, 2 and most preferably in the range of 0.92 to 1.12.

3 As mentioned above, the second step may, for example, be preformed on the fly, for 4 example, in real time or be previously calculated prior to the first step.

Third Step: Calculation of the Time on Task Modulator T

6 In the preferred embodiment, only one of the two functions g(t) and h(t) operates during 7 any period in which the individual is awake with the other function being equivalent to zero.
8 However, when the individual is asleep then both functions g(t) and h(t) are equal to one 9 (Equation 12) as represented by S9a through S10c in Figure 4(b) or zero (Equation 12a). The selection of the function preferably is based upon whether the individual is performing a task or 11 not is illustrated by S9b through S10b and the individual is awake is illustrated by S9a and 12 S10c. As such, the time on task modulator may be calculated prior to (as illustrated in Figure 13 8(a)) or after steps S7a and S7b instead as a separate branch as illustrated in Figure 4(b).

14 (1) Rest function g(t) The rest function g(t) is illustrated as S10a in Figure 4(b). The rest function g(t) preferably 16 represents the restoration of cognitive performance capacity due to an individual resting and 17 relaxing between tasks and/or activities. The rest function g(t) preferably does not provide the 18 same amount of restoration that occurs during sleep as discussed above with respect to the 19 sleep function s(t). A generalized form of the rest function is given by the equation:

t(t) = g(t) Equation 10 21 where g(t) may be any positive-valued function. Alternatively, the rest function g(t) may be 22 expressed as follows:

23 g(t) = z * s(t) Equation 10a 24 where z is a scalar, which preferably is in a range of 0 to 1, and tLS will preferably represent the length of the resting and/or inactivity period.

26 (2) Work function h(t) 27 The work function h(t) is illustrated as S10b in Figure 4(b). The work function h(t) 28 preferably represents declination of cognitive performance capacity due to an individual 29 performing a task(s) and/or an activity(ies). In SlOb, the work function h(t) calculates the task modulator T resulting from the decay in cognitive performance capacity that occurs over an 21922696.1 29 Agent Ref: 67803/00007 1 interval of time from t-I to t, which in the preferred embodiment is the length of one epoch. A
2 generalized form of the work function is given by the equation:

3 t(t) = h(t) Equation 11 4 where h(t) may be any negative-valued function decreasing with t. More preferably, the work function h(t) is a linear function depleting performance at a constant rate.
Alternatively, the work 6 function h(t) may be an exponential function that, for example, may increase the depletion rate 7 the longer the task is performed and/or activity is done. Another or further altemative is that the 8 type of task, i.e., the difficulty, the complexity, and/or the intensity will impact the depletion rate 9 per epoch. The greater the difficulty, the complexity, and/or the intensity of the task, then the greater the depletion rate per epoch will be.

11 Altematively, both or just one of the rest function and the work function may be impacted 12 by the time of day modulator M such that prior to being modulated with the cognitive 13 performance capacity C and the time of day modulator M, the task modulator T is modulated 14 based upon the time of day as represented by the time of day modulator M. A
further alternative is for the time of day modulator M to be used twice in Equations 1, 1a, and 1b above.
16 (3) Asleep function 17 A generalized form of the task function when the individual is asleep is 18 t(t) =1 Equation 12 19 where the modulation is performed using multiplication, because the task function T will not impact the individual's cognitive performance index. Alternatively, if the task modulator is added 21 to the other functions, then the task function will take the following form 22 t(t) = 0 Equation 12a 23 Fourth Step: Calculation of Predicted Cognitive Performance 24 The overall process of calculating predicted cognitive performance capacity E is illustrated schematically in Figures 1(a) and 4(a)-(b). The time of day function M and the task function T
26 modulate the cognitive performance capacity C derived from the individual's sleep/wake history 27 to generate the final predicted cognitive performance E as shown in, for example, Figure 1(a).
28 In the third step, predicted cognitive performance E is derived from the combination of cognitive 29 performance capacity C, time of day function M, and task function T. In its most general form:
E=COMOT Equation1 21922696.1 30 Agent Ref. 67803/00007 1 where V is any mathematical operation for combining cognitive performance capacity C, time of 2 day function M, and task function T. The conventional choice of operations for providing this 3 combination is addition or multip{ication. Depending on the form of time of day function m(t) and 4 task function T(t) selected above, the same numerical value of predicted cognitive performance E can be generated by either operation. Most preferably the combination is performed with 6 multiplication S11, represented as:

7 E= C'' M * T Equation 1 a 8 In Equation 1a, the predicted cognitive performance E is the modulation of the current cognitive 9 performance capacity C with a value centered around the number one representing the current value of the time of day modulator M and the task modulator T.

11 As noted above, the preferred numerical representation of cognitive performance capacity 12 C is a value ranging from zero to 100 to represent an index (or a percentage) of cognitive 13 performance capacity available for a particular individual. However, predicted cognitive capacity 14 E can meaningfully exceed 100 under certain circumstances due to time of day modulation about the current value of cognitive performance capacity C. A possible example of such a 16 circumstance would be a sleep period resulting in an index level of 100 cognitive performance 17 capacity C and terminated at the evening peak (and after sleep inertia has dissipated).

18 Alternatively, if a percentage representation is used while retaining a 100% scale, either 19 the predicted cognitive capacity E may be truncated/clipped at 100% or 0 to 120% may be scaled to 0 to 100%. Either choice will maintain a maximum of 100%. This most likely will be 21 implemented as scaling 120% to 100% and then truncating/clipping any predicted cognitive 22 capacity E to 100% if the prescaled value exceeds 120%.

23 As shown in Figure 1, the method repeats for each new epoch of data. For each iteration 24 of the method, one time unit equal to the length of an epoch may be added to time t preferably in the form of a counter S13 as exemplified in Figure 4(b). The counter step S13 may occur, for 26 example, as illustrated in Figure 4(b), at the same time as S11 or S12, or after S12.

27 In the preferred embodiment described above, the sleep inertia function i(t) is applied to 28 cognitive performance capacity C prior to modulation of cognitive performance capacity C by the 29 time of day modulator M and/or task modulator T. An alternative embodiment applies the sleep inertia function i(t) not to cognitive performance capacity C, but to predicted cognitive capacity 31 E, that is, subsequent to the modulation of cognitive performance capacity C by time of day 32 modulator M and/or task modulator T.

21922696.1 31 Agent Ref: 67803/00007 1 Also in the preferred embodiment described above, the wake function w(t) is set to zero 2 when the sleep inertia function i(t) is applied. Another alternative embodiment applies the sleep 3 inertia function i(t) and the wake function w(t) simultaneously. When the sleep inertia function 4 i(t) and the wake function w(t) become equal to each other or the sleep inertia function i(t) becomes greater than the wake function w(t), then cognitive performance capacity C is 6 calculated (or determined) using the wake function w(t).

7 The preferred embodiment may be further modified to account for the effects of narcotics 8 or other influences that will impact the cognitive capacity as shown in Figure 6. Further 9 modification to the preferred embodiment will allow for the inclusion of jet lag and similar time shifting events by, for example, compressing or expanding the 24 hour period of the time of day 11 function M(t) over a period of days to realign the time of day function M(t) to the adjusted 12 schedule.

13 The preferred embodiment may be modified to include the testing of the individual at 14 regular intervals to collect additional data and adjust the current cognitive performance index to reflect the results of the test. A test that may be used is the PVT or similar reaction time 16 measurement test(s). The current cognitive performance index at the time of the test then is 17 adjusted preferably along with the underlying weights of variables discussed above in 18 connection to the Equations such that the method and/or apparatus is fine-tuned to reflect a 19 particular individual's recuperation and/or depletion of cognitive performance capacity.

Another altemative embodiment is the removal of the third step from the preferred 21 embodiment. Like the other alternative embodiments, this alternative embodiment may be 22 combined in a variety of ways with the other altemative embodiments.

24 The preferred embodiment may be realized as software to provide a real-time current state of an individual's cognitive performance and the capability upon demand to extrapolate 26 future levels of cognitive performance. A flowchart representing the steps to be performed by 27 the software in the preferred embodiment is shown in Figures 4(a)-(b) and for an altemative 28 embodiment, to be described later, in Figures 8(a)-(b).

29 The software may be implemented as a computer program or other electronic device control program or an operating system. The software is preferably resident in a device, e.g. an 31 actigraph, attached to the individual or in a stand-alone device such as a personal computer, a 32 PAL device, a personal digital assistant (PDA), an e-book or other handheld or wearable 21922696.1 32 Agent Ref: 67803/00007 1 computing devices (incorporating Palm OS, Windows CE, EPOC, or future generations like 2 code-named products Razor from 3Com or Bluetooth from a consortium including IBM and 3 Intel), a specific purpose device receiving signals from a device, e.g. an actigraph, attached to 4 an individual or human input from human analysis or observation. The software could be stored, for example, in random access memory (RAM); in read only memory (ROM);
on a 6 storage device like a hard drive, disk, compact disc, punch card, tape or other computer 7 readable material; in virtual memory on a network, a computer, an intranet, the Internet, the 8 Abilene Project, or otherwise; on an optical storage device; on a magnetic storage device;
9 and/or on an EPROM. The software may allow for the variables in the equations discussed above to be adjusted and/or changed. This capability will allow users to adjust the variables 11 based on empirical knowledge and also leam the interrelationship between the variables.

12 The software implementation onto the measuring device such as an actigraph will convert 13 any decimal numbers used in calculations into integers that are appropriately scaled as is well 14 known to those skilled in the art. Further the integers would then be approximated such that minimal error would be created, for example, approximation for the Cole-Kripke algorithm 16 weighting factors become 256, 128, 128, 128, 512, 128, and 128, respectively. Using linear 17 approximation will simplify the binary arithmetic and the corresponding assembly code for 18 software implementation.

19 In software, the time of day modulator would be embodied as a table with one hour steps resulting in 24 rows using 8-bit unsigned integers. The intervening steps would be interpolated 21 from the one hour steps to provide 15-minute steps. This simplification provides sufficient 22 resolution for available displays. A pointer system would be utilized to retrieve the appropriate 23 data to calculate the time of day modulator. Depending on a myriad of factors, one of ordinary 24 skill in the art will most likely choose a multiplicative modulation to achieve appropriate scaling or an additive modulation for less complex but more rapid evaluation, i.e., if speed is a concem.
26 The main disadvantage with the additive modulation is that there will be an approximately 3%
27 error compared to the 1% error using the multiplicative modulation in this invention. This 28 system will allow the time of day function to be uploaded when the measuring/recording device, 29 such as an actigraph, is initialized and reduce the repetitive computational burden that would exist if a cosine table was used and the time of day function was calculated from the cosine 31 table for each epoch.

32 The preferred embodiment, as shown in Figure 7(a), may also be realized by a stand-33 alone device or a component add-on to a recording device. The stand-alone device is separate 21922696.1 33 Agent Ref: 67803/00007 1 from the device or other means of recording an individual's sleep history.
In contrast, the 2 component add-on to a recording device includes modifying the recording device to include the 3 component add-on to provide one device that both records and analyzes an individual's sleep 4 history.

A suitable stand-alone device preferably includes a physical input connection, e.g., an 6 input port (input means 20) to be physically connected to an input device, e.g., a keyboard, data 7 entry device, or a data gathering device such as an actigraph.
Alternatively, the physical 8 connection may occur over an information network. Alternatively, the input port may be an 9 interface to interact with a user. Alternatively, the physical input connection may be realized by a wireless communication system including telemetry, radio wave, infrared, PCS, digital, 11 cellular, light based system, or other similar systems. The wireless communication system has 12 an advantage in that it eliminates the need for a physical connection like cables/wires, plug-ins, 13 etc. which is particularly convenient when monitoring a mobile subject. The data gathering or 14 data entry device provides a sleep history that may include past, present and/or predicted/anticipated sleep patterns of an individual. Input means 20 embodies S1 for initial 16 inputting of information and S2 for the continual or one-time loading of data depending upon the 17 implementation selected.

18 The stand-alone device further preferably includes a data analyzer (interpretation means 19 30). The data analyzer performs S3-S6b. Interpretation means 30 analyzes the input data by performing different analysis functions. Interpretation means 30 compares the present input 21 data to the last input data to determine if there has been a change from sleep to wake or wake 22 to sleep; and if so, then set a time counter to the time for the last state, S3 and S4a in Figure 23 4(a). Interpretation means 30 also classifies the inputted data, as represented by S5 in Figure 24 4, to then be able to select or generate at least one of the following calculation functions responsive to the composition of the input data: 1) wake function, 2) sleep function, 3) delay 26 function, and 4) sleep inertia function as depicted by S6a-S7d in Figure 4(b). Interpretation 27 means 30 may be realized by an appropriately programmed integrated circuit (IC). One of 28 ordinary skill in the art will realize that a variety of devices may operate in concert with or be 29 substituted for an IC like a discrete analog circuit, a hybrid analog/IC or other similar processing elements.

31 The stand-alone device further preferably includes a calculator (determination means 40).
32 Determination means 40 may be implemented by appropriately programming the IC of the 33 interpretation means or it may be implemented through a separate programmed IC.
21922696.1 34 Agent Ref: 67803/00007 1 Determination means 40 calculates the cognitive performance capacity factoring in the 2 sleep/wake history and the current state using the function selected by interpretation means 30, 3 S7a-S7d in Figure 4(b).

4 The interpretation means 30 and determination means 40 may be combined into one combined means or apparatus.

6 The stand-alone device further preferably includes a first memory 60 that stores 7 modulation data including a modulating data series or curve preferably representing a time of 8 day curve. The stand-alone device further preferably includes a second memory 50 that holds 9 data for the creation of a data series or a curve representing cognitive performance capacity C
over time t. The first memory 60 and the second memory 50 may be any memory or storage 11 method known to those of ordinary skill in the art. The second memory 50 is preferably a first-12 in-first-out memory providing means for adding the value from the determination means 40 to 13 the end of the data series or the curve. The first memory and the second memory may be 14 combined as one memory unit. As one of ordinary skill in the art will realize there may be a memory to store the various intermediary values necessary for calculating cognitive 16 performance capacity C and predicted cognitive performance E as required to implement this 17 invention as either hardware or software.

18 The stand-alone device also preferably includes, as a separate IC or in combination with 19 one of the previously mentioned ICs, a modulator (modulation means 70) embodying S8-S9 shown in Figure 4(b). Modulation means 70 receives the present cognitive performance 21 capacity calculated by determination means 40 and calculates the time of day value from data 22 stored in the first memory 60. Modulation means 70 modulates the first data series or curve 23 (cognitive performance capacity) with the time of day value. The modulation preferably is 24 performed by matching the timing sequence information relating to the data series or the curves based on the latter of midnight and the length of time from the initial input of data as preferably 26 determined by the number of epochs and the initial starting time related to the first entered 27 sleep/wake state. Modulation means 70 may modulate a series of cognitive performance 28 capacity values with the time of day function if the second memory 50 exists to store the 29 cognitive performance capacity values.

As is well known by one of ordinary skill in the art, a counter or other similar functioning 31 device and/or software coding may be used in the stand-alone device to implement S11 shown 32 in Figure 4(b).

21922696.1 35 Agent Ref: 67803/00007 1 The stand-alone device may also include a display to show a plotted modulated curve 2 representing the modulation result over time, as stored in a memory, e.g. a first-in-first-out 3 memory, or a numerical representation of a point on the modulated curve at a selected time 4 from the modulation means 70 representing the predicted cognitive performance E. The numerical representation may take the form of a gauge similar to a fuel gauge in a motor vehicle 6 or the number itself. The stand-alone device, as an alternative or in addition to the display, may 7 include a printer or communication port to communicate with an extemal device for printing 8 and/or storage of a representation of predicted cognitive performance E.

9 The stand-alone device instead of having dedicated hardware may provide the storage space and processing power to execute a software program and accompanying data files. In 11 this case, the stand-alone device may be a desktop computer, a notebook computer, or similar 12 computing device. The software program handles the receiving of the data representing sleep 13 history from an outside source through a communication port or via a computer network such as 14 intranets and the Internet, and then performs the necessary analysis and processing of the method described herein. The storage space may be a memory in the form of computer 16 readable material storing at least the time of day curve and possibly the input data, which may 17 also be resident in the random-access-memory (RAM) of the computer given its temporary use.
18 The input data and the resulting produced data indicating various cognitive performance levels 19 of an individual may also be saved to a more permanent memory or storage than is available in RAM.

21 An alternative embodiment modifies the input port 20 to receive some form of raw data, 22 i.e., prior to being sleep scored, representing sleep activity of an individual. In this embodiment, 23 the interpretation means 30 would then sleep score the raw data as part of the data analysis 24 performed by it. A third memory to store the weighting factors required for sleep scoring, if a table is used for them, eise the sleep scoring function will implicitly include the weighting factors 26 and the third memory will be unnecessary.

27 Another alternative embodiment provides for the interpretation means 30 to filter the 28 sleep/wake data such that for the first kd number of sleep epochs after a wake epoch are 29 changed to wake epochs. In keeping with the invention, the filtering may be accomplished a variety of ways. The preferred way is to add a decision step prior to S3 in Figure 4(a) such that 31 if DSw is a sleep epoch and t - tLs <_ kd, then S3-S6a will be skipped and S7a will occur. The 32 result is that the decision rule represented as d(t) in Equation 6 above would be eliminated, and 33 S6b and S7d would be unnecessary in Figures 4(b) and 8(b).

21922696.1 36 Agent Ref: 67803/00007 1 One of ordinary skill in the art will appreciate that the stand-alone device is broad enough 2 to cover a computer/workstation connected to the Intemet or other computer network. A user 3 would transmit their sleep/wake history over such network to the stand-alone device for 4 obtaining a predicted cognitive performance based on the transmitted data.
The interface of the stand-alone device may allow the user to adjust the variables discussed above in connection 6 with the method to learn the interrelationship between the variables and the predicted cognitive 7 performance. Preferably, the range of allowable adjustment of the variables would be that of 8 the respective ranges discussed in connection with each of the variables above.

9 The component add-on to the measuring device may have similar components to the stand-alone device described above and shown in Figure 7(a). Preferably the component add-11 on is contained in one integrated chip to minimize the space needed to house it and/or is 12 implemented as software as part of a designed measuring device. The input means 20 13 becoming, for example, a wire or other type of connector. However, the add-on component may 14 include more than one electrical component, e.g., a memory chip and an IC.
The component add-on may transmit the predicted cognitive performance to a remote device for further 16 analysis.

17 The apparatus for accomplishing the third step is illustrated as part of Figure 7(b). The 18 additional components preferably include a task input means 20' for receiving information 19 regarding the task that may either be manually provided through some sort of data entry mechanism such as a keyboard, touch pad, a button or set of buttons, a touch screen or other 21 similar mechanisms, or through analysis of the data collected by the attached device.
22 Alternatively, the task input means 20' may be a part of or similar to the input means 20.
23 Preferably, a determination means 40' for calculating the task modulator based on what is 24 received from the task input means 20'. The determination means 40' preferably is in communication with the modulation means 70', which is the modulation means 70 with the 26 added modulation of the task modulator. As with the device described in connection with Figure 27 7(a), the various components of Figure 7(b) may be consolidated into one or a series of 28 combination components. Additionally, the components in Figures 7(a) and 7(b) may also not 29 be directly connected but separated into different devices.

The alternative embodiment described above involving real-time adaptation of the above-31 described method may be implemented by the addition of a routine such that the individual is 32 notified to press a button for recalibration on the recording device. Based upon the individual's 33 response time, the individual's current level of cognitive performance is determined and 21922696.1 37 Agent Ref. 67803/00007 1 adjustments are accordingly made to the above-described Equations to allow for the recent 2 activity of the individual to lead to the determined cognitive performance.

3 Both the software and/or hardware are envisioned as being able to operate and function in 4 real-time. For the purposes of this invention, real-time is understood to represent a continuous stream and analysis of cognitive performance level for each epoch of sleep/wake data entered.
6 Thus, the software and/or hardware will both provide to an individual or some other entity the 7 present cognitive performance level based on the data from the last entered epoch of 8 sleep/wake data entered into either the software or hardware. Most sleep scoring systems 9 make the sleep/wake determination based on data from epochs on either side of the epoch being analyzed. Consequently, there would be a delay in providing information to the user.
11 As one of ordinary skill in the art will appreciate from this description, the described 12 method is able to accept a continuous stream of data from either individual epochs or groups of 13 epochs. If blocks of time are entered, then after initial transitions the first few epochs are 14 govemed by the appropriate transition function with the appropriate time of solid sleep or wakefulness being used in the non-transition functions.

16 As a feature of the invention, the sleep/wake data may comprise the time at which a state 17 change occurs from sleep to wake or wake to sleep. The sleep/wake data may also comprise 18 the duration of the individual's wake state and the duration of the individual's sleep state. In 19 order to generate the predicted cognitive performance curve, the sleep/wake data may be extrapolated and/or expanded into a series of individual epochs. As discussed above an epoch 21 represents a predetermined length of time. Thus the sleep/wake data may be presented in 22 conventional units of time or may be presented in epochs. For example, if the sleep/wake data 23 was sleep for 10 epochs and wake for 3 epochs, in generating the cognitive performance 24 capacities, epochs 1 through 10 may represent the sleep state and epochs 11 through 13 may represent the wake state.

26 In accordance with an aspect of the invention, the predicted cognitive performance E at a 27 particular time q may be determined using either the predicted cognitive performance E or the 28 cognitive performance capacity C at time r as a base point where r can be before or after time q.
29 From the base point determining the cognitive performance capacities for the time points between times q and r where there is a change in state.

31 As shown in Figures 8(a)-(b), the steps are substantially the same as the preferred 32 embodiment with changes made to the wake and sleep functions, consequently the definition of 21922696.1 38 Agent Ref. 67803/00007 1 the variables is the same as the preferred embodiment except as noted. The equations 2 described below and the steps shown in Figures 8(a)-(b) are for the situation when the initial 3 cognitive value is prior in time to the desired predicted cognitive value.
Each element of 4 sleep/wake data is classified as either sleep or wake.

If the sleep/wake data represents the wake state, then the impact of the task function t(t) 6 is determined. Alternatively, the task function t(t) may be modulated by the time of day function 7 M prior to modulating the wake function wm(t) or the sleep inertia function i(t). Next, a selection 8 is made between two functions as to which is applicable based on the following decision rule:

9 tFOt<k;
THEN Ct = i(t) 11 ELSE Ct = wm(t) Equation 13 12 where Ot represents the amount of time in the current state, i.e., t - tLS.
The sleep inertia function 13 i(t) is used only if the last data entry is the wake state for a period of time is less than or equal to 14 k;. Thus the same sleep inertia function i(t) as used in the preferred embodiment is also used in this alternative embodiment after being modulated by the task function t(t).
The modified wake 16 function wm(t) takes into account that the sleep inertia function i(t) provides a delay of k; when a 17 curve is formulated, such that after an individual recovers from the initial suppression of 18 cognitive performance capacity the individual retums to the level of cognitive performance 19 capacity of the last epoch the individual was asleep prior to waking.
Accounting for this delay provides the following:

21 wm(t) = Ct_, - kW(At - k;) Equation 14 22 Alternatively, the modified wake function wm(t) may begin at a point where an undelayed wm(t) 23 intersects the sleep inertia function i(t). The wake function wm(t) is modulated by the task 24 function t(t) under either altemative.

If the sleep/wake data represents the sleep state, then a selection is made between two 26 functions as to which is applicable based on the following decision rule:

27 IFOt<kd 28 THEN Ct = d(t) 29 ELSE Ct = sm(t) Equation 15 21922696.1 39 Agent Ref: 67803/00007 1 The delay function d(t) is used only if the last data entry is the sleep state for a period of time is 2 less than or equal to kd. Thus the same delay function d(t) as used in the preferred embodiment 3 is also used in this altemative embodiment. The modified sleep function s,(t) takes into account 4 the delay function for a period of time equal to kd. Accounting for the delay function d(t) provides the following:

6 sm(t) = ((Ct., - (kw * kd)) + (100 - (100 - Ct_1) (1 -1/kg)et_ka)) Equation 7 where the first part of the equation represents the delay function d(t) and the second part 8 represents the recovery of cognitive performance capacity C (f(t) portion of S7c').

9 A summation of the time components of the sleep/wake data is performed as each piece of sleep/wake data is handled with respect to the calculation of the cognitive performance 11 capacity or prior to modulation of the final cognitive performance capacity with the time of day 12 function m(t). The latter is shown in Figures 8(a)-(b). After the new cognitive performance 13 capacity Ct is calculated, the method repeats to handle the next piece of sleep/wake data if the 14 present piece is not the last piece. After the last piece the predicted cognitive performance E is calculated based on Equation 1 above and as detailed in the preferred embodiment.

16 Alternatively, the task function t(t) may be included at the same time of the time of day 17 function m(t) instead of for each set of wake states by moving S9b through S10b to a position 18 similar to that illustrated in Figure 4(b).

19 It should be noted again that this method includes the processes and calculations based on Equations 1 through 12 expressed in their general form, with an alternative being the 21 removal of the task function elements. Embodiments shall apply functions relating the variables 22 involved according to empirical knowledge, resulting in specific expressions of those equations, 23 as illustrated in the text and Figures 1-4(b) above (but not confined to these), which may be 24 changed or refined according to the state of empirical knowledge.
VI. INDUSTRIAL APPLICABILITY

26 There are a variety of potential applications of this invention. In its simplest application, 27 the method according to the invention may be used to predict the impact of various idealized 28 (i.e., unfragmented) amounts of nightly sleep on predicted cognitive performance E. Another 29 practical application uses the method to predict the cognitive performance in an individual with fragmented sleep, either due to a sleep disorder such as sleep apnea or due to environmental 31 disturbances such as airplane or train noises. Another practical application uses the method to 21922696.1 40 Agent Ref 67803/00007 1 predict the cognitive performance E of an individual changing his/her schedule for night shift 2 work.

3 In another application, the method is used to retrospectively predict cognitive 4 performance E in a commercial motor vehicle operator involved in a driving collision/traffic accident. In this application, the method is used first to predict an individual's level of cognitive 6 performance E across some interval based on that individual's current work and sleep/wake 7 schedule.

8 Another similar application is using the method to re-schedule sleep and wakefulness in 9 order to optimize predicted cognitive performance E over an interval for a commercial motor vehicle operator. In this example, first we model a driver's predicted cognitive performance E
11 based on his current sleep/wake schedule. The driver's current sleep/wake schedule is 12 generated around the maximum duty hours allowed under the Federal Highway Administration's 13 (FHWA) hours-of-service regulations. These regulations allow the driver to obtain a maximum 14 15 hours on-duty (maximum 10 hours driving plus five hours on-duty but not driving) followed by a minimum eight hours off-duty. The driver may continue this on/off-duty cycling until 60 hours 16 on-duty has been accumulated - at which point the driver must take time off until seven days 17 has elapsed since he commenced duty. An altemative work schedule also allowed under 18 current FHWA regulations is based on a schedule of 12 hours on-duty and 12 hours off-duty 19 with the underlying assumption that the driver sleeps eight of his 12 hours off-duty. The use of this invention will allow the selection of the schedule that allows for maximizing the driver's 21 cognitive performance levels throughout a period of time.

22 Although described above in connection with a variety of specific activities, this invention 23 has many other applications. The method for predicting cognitive performance will provide 24 critical information for managing both individual and group productivity.
For example, in military operational planning, this method will enable commanders to determine precisely, based on 26 prior sleep history and duties performed, each soldier's current and predicted level of cognitive 27 performance. Commanders can also input a likely sleep/wake and work schedules and thereby 28 predict a soldier's cognitive performance throughout an impending mission.
Throughout 29 conduct of the mission itself, the latter cognitive performance predictions (originally based on likely sleep/wake and work schedules) can be updated based on actual sleep acquired and 31 work performed by an individual soldier. The ability to project future cognitive performance will 32 allow commanders to optimize troop performance during continuous operations by, for example, 33 planning sleep/wake and duty schedules around the mission to optimize cognitive performance, 21922696.1 41 Agent Ref: 67803/00007 1 selecting those troops or combinations of troops whose predicted cognitive performance will be 2 maximal at a critical time, etc. This method will assist in maximizing productivity at both the 3 individual level and unit level.

4 This invention may be employed in a variety of commercial applications covering many occupational areas for purposes of optimizing output (productivity). The invention provides 6 managers with the capability to plan operations and regulate work hours to a standard based on 7 objective cognitive performance predictions. This is in contrast to the frequently used method of 8 regulating work hours by time off-duty (a relatively poor predictor of sleep/wake patterns and 9 performance of tasks, and consequently a poor predictor of cognitive performance) or by generating alertness/sleepiness predictions (which, as noted above, do not always correspond 11 to cognitive performance). The invention can be "exercised" in hypothetical sleep/wake and 12 duty scenarios to provide an estimate of cognitive performance under such scenarios. To the 13 extent that optimizing cognitive performance is of interest to the general public, there is a 14 possibility for use in a variety of applications.

This invention also may be used in conjunction with drugs to alter the sleep/wake cycle of 16 an individual and/or optimize or minimize the cognitive performance level of an individual as 17 needed and/or desired.

18 This invention also can work conjunctionally with the concepts of particle swarm 19 theory/algorithms. Particle swarm algorithms are routinely used to optimize the throughput of containers through a ship port or to optimize the use of workers within a work group to perform 21 tasks over a given period. An example of an application is the planning of a mission for an army 22 unit by its commander.

23 The method may also be used to gauge and evaluate the cognitive performance effects of 24 any biomedical, psychological, or other (e.g., sleep hygiene, light therapy, etc.) treatments or interventions shown to improve sleep. Examples of these include but are not limited to patients 26 with overt sleep disorders, circadian rhythm disorders, other medical conditions impacting sleep 27 quality and/or duration, poor sleep hygiene, jet lag, or any other sleep/wake problem. Currently, 28 the efficacy of treatments for improving sleep is determined by comparing baseline 29 polysomnographic measures of nighttime sleep and some measure of daytime alertness (e.g., the MSLT, the Maintenance of Wakefulness Test (MWT), the Stanford Sleepiness Scale or the 31 Karolinska Sleepiness Scale) with the same measures obtained after treatment. Both treatment 32 efficacy and the likely impact on performance during waking periods are inferred from the 33 results on the daytime alertness tests. For example, the Federal Aviation Administration 21922696.1 42 Agent Ref: 67803100007 1 currently requires any commercial pilots diagnosed with sleep apnea to undergo treatment.
2 Such treatment is followed by daytime alertness testing on a modified version of the MWT.
3 During the MWT, pilots are put in a comfortable chair in a darkened room and instructed to try to 4 remain awake for extended periods. If the pilots are able to avoid overt sleep under these sleep-conducive conditions then they are deemed fit for duty. The inference is that the minimal 6 ability to maintain wakefulness at a discrete point in time translates into the ability to operate an 7 aircraft safely (i.e., it is inferred that alertness is equivalent to cognitive performance). However, 8 sleep deprivation can affect cognitive performance even when it does not result in overt sleep, 9 particularly during an alertness test when for various reasons the individual may be highly motivated to stay awake.

11 In contrast, the current method allows cognitive performance to be estimated directly from 12 measured sleep parameters considered in conjunction with the time of day and performance of 13 tasks. The advantages of this method over current methods for evaluating treatment efficacy 14 are: (1) the motivations and motivation levels of the patients being tested cannot affect results (cognitive performance determinations); and (2) the method allows numerical specification and 16 prediction of cognitive performance across all projected waking hours rather than indicating 17 alertness at a discrete, specified point in time. Thus, the method provides a continuous scale 18 for gauging cognitive performance across time rather than providing only a minimal "fitness for 19 duty" determination based on the patient's ability to maintain EEG-defined wakefulness at a specific time.

21 The method may also be used clinically as an adjunct for diagnosing sleep disorders such 22 as narcolepsy and idiopathic CNS hypersomnolence. Equally important, it may also be used to 23 differentiate among sleep disorders. The latter is critical to the course of treatment, and 24 consequent treatment efficacy depends on a valid and reliable diagnosis.
For example, sleep apnea and periodic limb movements during sleep are characterized by nighttime sleep 26 disruption (i.e., partial sleep deprivation) accompanied by daytime cognitive performance 27 deficits. In contrast, narcolepsy and idiopathic hypersomnolence tend to be characterized by 28 apparently normal nighttime sleep, but accompanied by daytime cognitive performance deficits.
29 Based on the apparently normal nighttime sleep in the latter two groups, the invention would predict relatively normal cognitive performance. Thus, a discrepancy between predicted 31 cognitive performance (based on the current invention) and observed or measured cognitive 32 performance could be used to distinguish one sleep disorder from another.
For example, 33 narcolepsy, idiopathic hypersomnolence, or other CNS-related causes of daytime cognitive 21922696.1 43 Agent Ref: 67803/00007 1 performance deficits (where no sleep deficit is apparent) could be distinguished from sleep 2 apnea, periodic limb movements, or other causes of daytime cognitive deficits (where impaired 3 sleep is evident).

4 Although the present invention has been described in terms of particular preferred embodiments, it is not limited to those embodiments. Alternative embodiments, examples, and 6 modifications which would still be encompassed by the invention may be made by those skilled 7 in the art, particularly in light of the foregoing teachings.

8 Furthermore, those skilled in the art will appreciate that various adaptations and 9 modifications of the above-described preferred embodiments can be configured without departing from the scope and spirit of the invention. Therefore, it is to be understood that, within 11 the scope of the appended claims, the invention may be practiced other than as specffically 12 described herein.

21922696.1 44

Claims (22)

1. An apparatus for predicting cognitive performance comprising:
an input connection that receives data, a data analyzer connected to said input connection to select a calculation function responsive to the received data, a calculator connected to said data analyzer to calculate a cognitive performance capacity using the calculation function, an evaluator connected to said input connection to determine a task value based on the received data, a memory that stores modulation data, and a modulator connected to said memory, said evaluator, and said calculator.

2. The apparatus according to claim 1, wherein said modulator modulates the cognitive performance capacity with the modulation data with the task value to generate a predicted cognitive performance.

3. The apparatus of any of claims 1 or 2, further comprising a display connected to said modulator.

4. The apparatus of any of claims 1-3, further comprising a transmitter connected to said modulator.

5. The apparatus of any of claims 1-4, wherein the modulation data represents time of day variations over a 24-hour period.

6. The apparatus of any of claims 1-5, wherein said data analyzer includes a sleep scoring system.

7. The apparatus of any of claims 1-6, wherein said input connection includes a wireless receiver.

8. The apparatus of any of claims 1-7, wherein said input connection includes a keyboard.

9. The apparatus of any of claims 1-8, further comprising a first-in-first-out memory connected to said modulator, said first-in-first-out memory stores an output of said modulator.

10. The apparatus of any of claims 1-9, further comprising a cognitive performance capacity adjustor connected to said calculator.

11. The apparatus of any of claims 1-10, further comprising a data collector connected to said input connection.

12. The apparatus of claim 11, wherein said data collector includes an actigraph.

13. The apparatus of any of claims 11 or 12, wherein said data collector includes a polysomnogram.

14. An apparatus for predicting cognitive performance comprising:
a polysomnogram, a task value input means, a data analyzer connected to said polysomnogram to select a calculation function responsive to the data received from said polysomnogram, a calculator connected to said data analyzer to calculate a cognitive performance capacity using the calculation function, an evaluator connected to said task value input means to determine a task value based on data entered into said task value input means, a memory that stores modulation data, and a modulator connected to said memory, said evaluator, and said calculator.

15. The apparatus according to claim 14, wherein said modulator modulates the cognitive performance capacity with the modulation data with the task value to generate a predicted cognitive performance.

16. The apparatus of any of claims 14 or 15, further comprising a display connected to said modulator.

17. The apparatus of any of claims 14-16, further comprising a transmitter connected to said modulator.

18. The apparatus of any of claims 14-17, wherein the modulation data represents time of day variations over a 24-hour period.

19. The apparatus of any of claims 14-18, wherein said data analyzer includes a sleep scoring system.

20. The apparatus of any of claims 14-19, wherein said task value input means includes a keyboard.

21. The apparatus of any of claims 14-20, further comprising a first-in-first-out memory connected to said modulator, said first-in-first-out memory stores an output of said modulator.

22. The apparatus of any of claims 14-21, further comprising a cognitive performance capacity adjustor connected to said calculator.