Data processing: artificial intelligence – Neural network – Learning task
Reexamination Certificate
2001-06-28
2004-10-12
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Neural network
Learning task
C706S012000, C706S016000
Reexamination Certificate
active
06804661
ABSTRACT:
BACKGROUND
1. The Field of the Invention
This invention relates to signal processing and, more particularly, to novel systems and methods for pattern recognition and data interpretation relative to monitoring and categorizing patterns for predictably quantifying and evaluating systems of an observed entity as they react to stimuli such as, for example, drug profiling.
2. The Background Art
The nervous system is a complex network of tissue for carrying and transmitting signals from one part of a body to another. The nervous system can be divided into the central nervous system and the peripheral nervous system. The central nervous system comprises the brain and spinal cord, to which sensory impulses are transmitted and from which motor impulses pass out, and which coordinates the activity of the entire nervous system. The peripheral nervous system incorporates the remainder of the nerve elements of the body. The autonomic nervous system governs involuntary actions and consists of the sympathetic and the parasympathetic nervous systems.
The impulses (currents) of the nervous system may be measured and charted for the purposes of study and evaluation. For example, electroencephalograms (EEG) can be used to detect and record the impulses of the brain (brain waves). Electrocardiograms (EKG) can be used to gather and chart the impulses and currents of the heart (e.g., changes of electrical potential occurring during the heartbeat). Magnetoencephalographs (MEG) can also be used to measure and chart the changing magnetic fields of the brain. An analysis of the impulses of a particular system or organ (e.g., brain, heart or the like) may provide information as to how the particular observed entity (i.e., human or animal) is performing or reacting to stimuli.
A variety of techniques or strategies of analysis have been developed by those skilled in the art to amplify, analyze and interpret EEG, EKG and MEG waveforms. As appreciated, each of these analysis techniques, however, has its own advantages and disadvantages. Specifically, strategies of analysis used by those skilled in the art may range from the early techniques of spectral analysis and multiple-trial waveform averaging, to various transforms, time-frequency distributions, spatial filtering methods, to one or more of the newer approaches of neural networks, fuzzy logic systems and integrated neurofuzzy systems.
One of the disadvantages of prior art spectral analysis techniques is that they are generally limited to the analysis of a single channel or the comparison of two channels at a time. In addition, spectral analysis usually relies on human inspection of the generated waveforms, whereas in frequency representations, time domain information is implicit or hidden.
Multiple-trial waveform averaging is a widely used analysis technique method that uses summing and averaging over many trials to amplify evoked and event-related signals while reducing background noise. While useful for certain applications, averaging techniques have several significant drawbacks. For example, large quantities of information may be lost in the averaging process as only those signals that are robustly time-locked to a stimulus or response are able to survive the summation over multiple-trials. Another serious disadvantage is that the averaging process only provides a comparison between groups of trials rather than between individual trials themselves. Additionally, the need to first record multiple trials before a reliable evoked potential (EP) can be obtained tends to reduce the utility of signal averaging for real-time applications.
Alternative analysis approaches have been developed by those skilled in the art in an attempt to overcome many of the limitations of multiple-trial waveform averaging. These prior art techniques or methods of analysis may include, for example: (1) Fourier Transforms, (2) Hilbert Transforms, (3) Wavelet Transforms, (4) Short-Time Fourier Transforms, (5) Wigner Functions, (6) Generalized Time-Frequency Distributions and other joint time-frequency distributions. While valuable for certain indications, these alternative prior art approaches are usually accomplished using only a single channel. Therefore, there is no spatial information and, accordingly, inter-channel relationships are often missed. Moreover, these prior art alternate approaches have not been integrated with computerized condition discrimination. Like spectral analysis techniques, these alternate prior art approaches rely on human visual inspection of the generated waveforms in concluding findings, which produces a review process fraught with the potential of observer error.
Spatial filtering methods have also been investigated which include: (1) Principal Component Analysis, (2) Singular Value Decomposition and (3) Eigenvalue Analysis. These prior art filtering methods tend to ignore frequency, and often temporal information as well. Additionally, these prior art spatial analysis techniques must be applied to averaged evoked potentials or the noise level is prohibitive. The foregoing prior art spatial filtering methods therefore are not typically useful for single-trial analysis.
Additional analysis techniques and methodology have been developed by those skilled in the art which take advantage of recent increases in computer processing power. Neural networks have been developed to discover discriminant information. The traditional neural network approaches, however, generally take a long time to program and learn, are difficult to train and tend to focus on local minima to the detriment of other more important areas. Moreover, most of these analysis techniques are limited by a lack of integration with time, frequency and spatial analysis techniques.
Although the forgoing analysis techniques and methods of wave signal processing have provided useful concepts and are valuable in their application in particular areas, due to their inherent narrow ranges of applicability, these prior methods of analysis have provided a fragmentary approach to brain signal evaluation. To this end, the prior art methodologies for evaluating waveforms have significant weaknesses and limitations, and none seem to meet the goals of rapid accurate analysis of all pertinent characteristics of a wide variety of single-trial waveforms. What is needed, therefore, is an integrated waveform analysis method capable of extracting useful information from highly complex and irregular waveforms such as EEG, EKG and MEG data.
BRIEF SUMMARY AND OBJECTS OF THE INVENTION
In view of the foregoing, it is a primary object of the present invention to provide novel systems and methods for signal processing, pattern recognition and data interpretation by means of observing the affects of a particular state or event on an observed entity.
It is also an object of the present invention to provide a method for improved drug modeling for evaluating the benefits of drugs and side-effect predication in relation to an observed entity (e.g., human or animal).
It is a further object of the present invention to provide an improved method for drug fingerprinting.
Additionally, it is an object of the present invention to provide novel systems and methods for measuring the effect of a particular event or state on the cognitive skills, motor ability, sensation, perception and the like of an observed entity (e.g., human or animal).
It is still a further object of the present invention to provide novel systems and methods for one or more of the following: (1) determining whether a drug successfully crosses the blood-brain barrier; (2) determining whether a drug alters brain function; (3) determining whether a drug modifies cardiovascular activity; (4) determining of dose response relationships by analyzing the effect of a range of doses on EKG or EEG; (5) measuring drug-induced brain activity patterns indicating the presence of particular side effects inducing drowsiness, nausea, headaches, dizziness or cognitive impairment; (6) identifying the effect of a neurological drug on the electrical activity of the brain and hea
Bright Ideas, L.L.C.
Davis George B.
Pate & Pierce & Baird
LandOfFree
Drug profiling apparatus and method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Drug profiling apparatus and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Drug profiling apparatus and method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3283464