Surgery – Diagnostic testing – Detecting brain electric signal
Reexamination Certificate
2002-06-14
2003-09-23
Hindenburg, Max F. (Department: 3736)
Surgery
Diagnostic testing
Detecting brain electric signal
C600S546000, C600S301000, C128S920000, C340S575000
Reexamination Certificate
active
06625485
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to analysis of electroencephalograms, and more particularly concerns a method for quantifying and monitoring human alertness.
2. Description of Related Art
An electroencephalogram (EEG) measures electrical activity of the brain. Since the electrical activity of the brain first started to be recorded, efforts have been made to identify how the EEG data correlates with behavior, mood, mental performance, attention, and vigilance. However, classifying changes in the state of alertness, drowsiness, mental performance, or attention has proven difficult due to the significant variability of the EEG for an individual and between individuals, even under controlled conditions.
The frequencies of the waking EEG have generally been organized into three bands (or bandwidths): the theta band (4 to 7 Hertz (“Hz”)), the alpha band (8 to 13 Hz) and the beta band (14 to 24 Hz). When a person is fully rested and undertakes a cognitive task, the EEG amplitude is relatively small overall, power in the alpha band is suppressed, and the power in the frequencies above 13 Hz (i.e., beta band) tend to increase. As a person becomes tired, the overall amplitude of the EEG and the power in the frequencies below 13 Hz increase (i.e, alpha and theta bands). As sleep onset approaches, first dominant alpha activity followed by distinct theta become apparent. Momentary fluctuations between states are observed in the EEG when a person is tired or sleep deprived and is trying to remain awake.
The real-time quantification of an individual's state of alertness has a number of commercial applications. Shift workers, truck drivers, train operators and other individuals who work during normal hours of maximum sleepiness could be notified when they become too drowsy. In addition, these workers could be notified when a short-nap would be most beneficial in managing their fatigue. Pilots, air traffic controllers and other workers who perform monotonous tasks could be notified when they begin to daydream or are not maintaining an acceptable level of alertness. The EEG of adults and children with Attention Deficit Disorders or Attention Deficit/Hyperactivity Disorders could be monitored to assist them in maintaining a normal state of alertness. The recording and subsequent off-line analysis of the EEG to assess alertness over wider time intervals can be used to assess treatment outcomes for patients with sleep disorders, to titrate dosages of prescription drugs, and to measure the effects of new pharmacological substances on alertness.
The quantification of states of alertness, drowsiness, mental performance, and so on, are complicated by significant differences in the EEG observed between individuals during similar tasks, in addition to changes within individuals as a result of circadian rhythms and other fluctuations. One approach to overcome the issue of variability between individuals in the amplitude and distribution of the EEG power has been to develop paradigms whereby the individual can be used as his or her own control. For example, Gevins et al. (U.S. Pat. No. 5,447,166) monitors neurocognitive workload based on the individual's normative neurocognitive calibration function acquired during a specific battery of tasks. The present inventors have realized that this normalization technique is limited, however, when conditions are not highly controllable. For the monitoring of alertness across the multitude of activities and conditions common to daily living, for example, users would need to be measured during multiple sessions to acquire baseline values representing the continuum from highly alert to sleep onset.
An alternative approach has been to develop a classification model using EEG variables which have been transformed to minimize individual differences. The most common transformation technique standardizes the EEG variables for each observation or time period of data analyzed (i.e., epoch) relative to the mean and standard deviations for the particular session or a baseline session (i.e., Z-scores). This normalization technique is limited, however, when monitoring the continuum from fully alert to sleep onset, because there is no single session or activity that best categorizes the diversity in patterns of distribution of power across the EEG frequency bins, individuals, and tasks. Standardizing the EEG based on the amplitude and distribution of power during sessions where the individual is awake with eyes closed, for example, reduces variability in the alpha range across subjects. However, the present inventors have recognized that this approach does not normalize population differences with respect to increased beta power during cognitive tasks. Misidentification of the state of alertness will inevitably occur along the alertness-drowsiness continuum unless techniques are developed to first identify individual patterns in the distribution of the EEG, and then adapt the model to overcome these differences. The methods of the present invention solve these problems.
Another problem complicating the monitoring of the EEG and classifying the state of alertness along the alertness-drowsiness continuum is the contamination of the EEG recording by artifacts resulting from eye blinks and ocular movements. These artifacts are problematic because the power resulting from eye artifacts cannot be readily differentiated from power contributed by theta activity in the frequency domain. As mentioned previously, the onset of sleep is categorized by increased or dominant activity in the theta range of power between 4-7 Hz. Fast blinks that rewet the eyes typically range from 0.20 to 0.30 seconds in duration and cause contamination by increasing the power in the 3 to 5 Hz bins. Slower eye blinks or closures that occur when a subject becomes tired, range from 0.30 to 1 second in duration and cause contamination by increasing the power in the 1 to 4 Hz bins. Frequency bins above 5 Hz can also be affected, depending on the amplitude and duration of the blink.
Epochs with eye artifact in the EEG can be identified and automatically rejected from the analysis by monitoring significant variations in EEG amplitude. This approach, however, as the present inventors have recognized, results in an excessive amount of data being rejected when eye movements are detected. Alternatively, the power below 5 Hz can be high-pass filtered to eliminate both the eye artifact and the theta activity. The present inventors have recognized that this approach results in the exclusion of variables that are highly correlated with alertness. A number of methods have been developed to remove the eye artifact from the EEG in the time-domain based on comparison with signals acquired simultaneously from electro-oculographic (EOG) recording or eye motion sensors. For applications that acquire EEG during activities of daily living, however, placement of EOG electrodes near the eye(s) would be uncomfortable and cosmetically unacceptable. The use of an eye motion sensor would only be appropriate when the user's movement is limited, such as when seated in the cockpit of an airplane or vehicle. In the acquisition of EEGs during activities of daily living, additional artifacts are typically encountered which can result in a failure to correctly identify the individual's state of alertness. Electromyographic (EMG) activity, spikes and amplifier saturation can cause a substantial increase in power that distorts the entire EEG band. Gross head, eye or body movements result in artifacts that increase the power in the slower frequencies (i.e, less than 4 Hz). Procedures should be developed which automatically detect, as well as monitor, the occurrence of these artifacts in order to maximize the amount and quality of data acquired real-time or off-line analysis.
Another problem encountered when analyzing the EEG in the frequency domain is the non-stationary effects (i.e., momentary fluctuations) of the EEG itself. One approach to reduce the variability of the EEG
Berka Christine
Konstantinovic Zoran R.
Levendowski Daniel J.
Olmstead Richard
Advanced Brain Monitoring, Inc.
Astorino Michael
Hindenburg Max F.
LandOfFree
Method for the quantification of human alertness does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for the quantification of human alertness, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for the quantification of human alertness will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3056084