System for analysis of biological voltage signals

Abstract

System for analyzing biological signals representative of voltage changes, including obtaining an analog biological signal representative of voltage changes, using digital processing software to digitize the biological signal, displaying the processed biological signal in analog form on a display in a time compressed format, wherein an amount of compression for the time compressed format is selected such that graphical patterns are made perceivable on the display that signify an abnormality in the biological signal, and visually analyzing the biological signal on the display to characterize the abnormality.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This application is a continuation of application Ser. No. 10 / 664,889, filed Sep. 22, 2003, now pending, which is a continuation of application Ser. No. 10 / 078,355, filed Feb. 21, 2002, now abandoned, which is a continuation of U.S. Ser. No. 09 / 405,233, filed Sep. 24, 1999, now U.S. Pat. No. 6,370,423 and the application hereby claims priority on U.S. Provisional Application No. 60 / 103,154 filed Oct. 5, 1998, the entire contents of which are hereby incorporated by reference in this application.FIELD OF THE INVENTION [0002] The instant invention relates to improved methods and systems for analysis of dynamic electrocardiograms and other similar waves of biological origin with the purpose of facilitating improved diagnosis of pathological states in human and veterinary medicine. More particularly, the instant invention advantageously uses advances in sound wave technology to improve the recovery, preservation, enhancement and cost effecti...

AI Technical Summary

Benefits of technology

[0073] A primary object of the instant invention is to increase the accuracy and decrease the cost of biologic signal analysis for use in mass screening, clinical practice and research.

[0081] CVAT is different from current forms of biological signal analysis in that it preserves the integrity of the analog signal, enhances dynamic range, the fidelity and resolution of the original signal obtained. All these features lead to better interpretation of the signal using compressed visual patterns, which, in turn, leads to quick and easy identification of abnormalities suggestive of pathologic states. CVAT is based on the application to biological signal analysis of advances made in the software, hardware and electronic technology used to process and analyze sound waves. This is a major departure from current obsolete ways to digitize analog signals, which include the use of extreme lossy digital compression, Fast Fourier Transformation and other mathematical and autocorrelational engineering based algorithms which markedly deteriorate the quantity and quality of the signal to be evaluated.

[0082] A main application of the present invention is to improve the analysis of the Holter electrocardiogram. The invention departs from the current Holter ambulatory electrocardiogram analysis in that it replaces auto-correlational communications engineering techniques and quantification-dependent analysis of the electrocardiogram done with obsolete computer technology which eliminates most of the original signal and distorts the fidelity, resolution and dynamic range of the small fraction kept in the digital file for algorithm driven analysis.

[0083] Instead, CVAT relies on morphologic and pattern evaluation signal analysis complemented with quantification when necessary. The totality of the signal originally recorded is preserved with protection and enhancement of dynamic range, resolution and fidelity of the signal.

[0088] Digital sound processing software and techniques are used for the processing and analysis of biological signals. The inventor has determined that one suitable sound processing software is SOUND FORGE, which is designed for processing digital audio. Other similar software programs (such as, but not limited to seismographic and geologic software) used for wave analysis may also be used in accordance with the present invention. Such software allows various steps to be performed to enhance the signal (without introducing distortion) in the voltage and time domains and enhances pattern visualization and other forms of analysis;

Problems solved by technology

About 50% of those affected do not reach the hospital due to poor recognition of the disease before a cataclysmic, often terminal event has occurred.

Visual analysis is a very time consuming (hence costly) process, which required an operator with intimate knowledge of electrocardiography and cardiology.

For this reason the use of visual analysis has been limited to academic research and it has not been possible to extend its benefit to patient care in the community.

The current methods are unable to reliably detect ambulatory ischemia or risk for potentially lethal arrhythmia.

Such risks are not detectable in a cost-effective manner with prior art techniques.

These shortcomings of the prior art have a significant impact on cardiovascular morbidity and mortality.

By failing to disclose evidence of risks for catastrophic events, current Holter analysis lulls clinicians into the falsehood of absence of evidence misrepresented as evidence of absence of potentially lethal risks.

Consecutive obsolete methodologic steps in current Holter analysis severely diminish the quantity and degrade the quality of the signal encoded in original Holter recording media.

However, today, the only reliable form of Holter analysis is visual scanning of the magnetic tape itself, not the “over reading” of the expunged and distorted digital file which misrepresents the original signal.

Visual analysis by an expert electrocardiographer is a very time consuming method used only by highly motivated experts in research programs.

Due to time and cost involved, visual analysis of the analog signal cannot be applied to clinical practice or mass screening of at risk population with known methods.

Communications engineering paradigms and techniques are best limited to the evaluation of non-biological signals where reproducibility and repetition of waves and other phenomena are the norm.

A major drawback of engineering autocorrelation is that it is sensitive to waveform changes in the time domain (X-axes) and poorly sensitive to changes in the voltage domain (Y-axes).

In current Holter analysis, autocorrelation is wrongly applied to a small sample of degraded biological signal with poor dynamic gain which magnifies the limitations of autocorrelation to recognize voltage changes.

Non-biological techniques used to analyze biological data yield, at best, mediocre results, which become poor when analysis is done using a distorted, minuscule fraction of the original signal recorded.

Speed fluctuation in the 10% range is a signal acquisition problem; the best research efforts have dropped it to 3%, which is still too high for accurate quantitative ECG analysis.

The norm today is to digitize the analog signal by playing back the cassette tapes at speeds as fast as 480 times real time; this is the beginning of major degradation of the analog ECG.

High-speed playback degrades fidelity by limiting frequency response.

Tape stretching due to repeated stopping and starting of the tape is another source of signal degradation.

Current algorithms use elision and omission of vast amounts of the originally recorded ECG signal to achieve extreme, unnecessary and deleterious data compression.

However, extreme digital compression gravely decreases the integrity, fidelity, resolution and most importantly the dynamic range of the stored electrocardiogram or any other signal.

Such creative approach is done after drastic lossy compression has irretrievably discarded more than 90% of the original signal with great loss of integrity, dynamic range, resolution and fidelity.

The continuing use of vastly outmoded computer and signal processing technology impede the use of Dr.

Obsolete and unnecessary compression strategies reduce 24-hours worth of analog Holter data down to a little more than a single megabyte digital file.

When the algorithms for Holter analysis were created, extreme limitations in available memory existed.

Thus, extreme data compression was needed.

Thirty years ago, in the infancy of the computer industry, when silicon chips were as expensive as they were limited in their RAM or ROM capacity, data compression was a necessary evil.

Now that computer memory is as cheap as it is truly vast in capacity, data compression is an undesirable tool mainly used by producers of entertainment and other non-essential computer applications, i.e. whenever loss of data is deemed acceptable for reasons of practicality and / or fast transmission over consumer-level internet connections.

Like all biologic signals, ECG, as audio data are remarkably hard to compress effectively.

All compression routines are known to deteriorate dynamic range, signal quantity and quality.

For 16-bit data, companies like Sony and Philips are spending millions of dollars to develop proprietary schemes that as yet are not fully successful.

Although great strides of innovation are now being made in techniques of data compression, a 350:1 data compression ratio keeping the integrity of the signal is as yet impossible, nor is it necessary.

The problem is that such pre-compression decision regarding ambulatory ECG signal is not and can not be made without rendering compressed Holter files useless except for detection of gross arrhythmia.

Although the algorithm used in MP3 compression is quite advanced, the process still degrades the quality of the original signal in an invariably noticeable (almost ‘trademark’) fashion.

Such degradation, however, lies within an ‘acceptable’ window of loss for the consumer-oriented purposes of the technology, i.e. exchanging recordings of popular songs over the Internet.

One overriding fact remains clear: the application of any inherently omissive data compression strategies to a 24 hr ECG recording prior to any and all analysis of the totality of the signal is wrong.

Holter analysis remains a vastly under addressed technological obsolescence which is an obstacle for detection of risk for lethal events and in doing so puts lives directly at risk.

In addition, the only limits containing further development and refinement of the CVAT process are those temporarily imposed by the ephemeral and upwardly spiraling limits of computer and signal analysis technology.

Several different morphologies of the T wave are associated with non-homogeneous repolarization, a sign of myocardial cell hypoxygenation and risk for lethal arrhythmia.

Correction for presence of Ta (atrial ischemia) is unheard off in the current art, since it is unable to visualize this subtle but important change.

The analytic paradigm and totally obsolete limitations in computer technology imposed this major source of false negative reports.

Current Holter algorithms can not detect ECG signs of abnormal repolarization in a reliable and reproducible manner.

This false isoelectric point and spill over of the Ta negative voltage into the ST segment are common pitfalls that introduce error in ischemia detection by current algorithms.

The prior art taught by conventional Holter monitoring systems cannot retrieve, store, display or analyze high fidelity signals in the microvolt or microsecond range.

Fast magnetic tape play back done without optimizing the dynamic range, scanty sampling, poor quantization and extreme data compression deteriorate and diminish the signal.

All the above contribute to the poor diagnostic performance of current Holter technology for conditions other than gross arrhythmia.

Conventional Holter monitoring and ECG systems cannot detect, preserve or recover signals at or beyond the microvolt or microsecond range.

Undue reliance is placed on a physician over reading of very small depictions of low fidelity greatly deteriorated ECG tracings recovered from the digital file.

This is a basic problem which has to be dealt with even when neural networks, used in research only, select beats to “train” the computer to recognize “normal” beats.

Therefore, the method is used as a last resort, when setting the other parameters does not help, which can occur with patients who have peaked T waves.

Whenever the ST segment shifts up or down due to myocardial ischemia, the T morphology is usually abnormal and not amenable to template classification.

While templates work well for arrhythmia, over reliance in abnormal beat classification using predetermined templates is a reason for the poor performance of computer automated Holter analysis in the diagnosis of conditions other than arrhythmia.

However superimposition of fast played back, scantily sampled, mercilessly compressed, filtered, smoothed and / or Fast Fourier Transformed beats cannot be trusted, since it processes a signal different from that originally encoded in the magnetic tape.

Template detection may be convenient, but applied to a digital file which lacks integrity, dynamic range, fidelity and resolution, it cannot be sensitive or specific nor can it detect abnormalities in microvolt regions such as the PQ, ST segments or the P and T waves.

The sophisticated cardiology community is aware of the current Holter analysis shortcomings; hence, this method is not routinely used as an aid in the diagnosis of highly lethal cardiovascular risks.

This algorithm is widely used in patient care and research and further demonstrates the disadvantages of current Holter processing techniques.

“Taking every third sample provides a limited sampling rate and scaled differential coding provides limited resolution.

After the “decimating compression” it is only benign to say that the algorithm driven file will have poor resolution and fidelity.

This is a grave problem that needs immediate redress.

It is not surprising that the quality of the ECG recovered from current Holter analysis algorithms is too poor to identify anything but arrhythmia with some degree of certainty.

With this algorithm, all the microvolt nuances will certainly be irretrievably lost.

The price paid is extremely poor ECG data unsuitable for recognition of ischemic and other dire electrocardiographic signs with any degree of certainty.

The 12-lead electrocardiogram is not expected or designed to detect transient and unpredictable episodes of myocardial ischemia or arrhythmia since it depicts only 3 of the 100,000 or more heart beats we have in 24 hours.

All methods available today, other than the Holter technique, are unable to detect myocardial ischemia due to transient spastic and / or thrombotic causes of decreased coronary blood flow.

Hence, this grave condition escapes detection unless Holter recordings are done under the fleeting and often difficult to identify forms of daily life stress that induces the attacks in a given patient.

The limiting factor is the current computerized Holter analysis that is unsuitable for detection of anything but gross arrhythmia.

The current art suffers from false negative findings which have dire consequences for patients considered healthy when they are not.

Such visual Holter analysis is time consuming and hence, done only in few research efforts and not cost effective or applicable to daily clinical practice or mass screening.

Ischemia-induced abnormalities are in the microvolt range and are unlikely to stand the decimating affects of current algorithms devoted to minimize file size.

The minor changes introduced by computer algorithms are not sufficient for reliable detection of ischemia or risk for potentially lethal arrhythmia.

The magnitude and morphologic changes of the T wave are additional indicators of ischemia which the current algorithms are unable to detect.

However, silent or symptomatic, ischemia can equally induce arrhythmia, myocardial infarction or sudden death.

All these intensive computational niceties are done on a digital file known to be incomplete and with major fidelity, resolution and dynamic range deficiencies.

Fibrillation occurs when transient triggers impinge upon an electrically unstable heart causing normally organized electrical activity to become disorganized and chaotic.

Complete cardiac dysfunction results and may end in sudden death.

An episode of poor oxygenation of the heart (myocardial ischemia) is probably the most frequent cause of ventricular fibrillation and death.

Microvolt signals are easily obliterated by poor dynamic range, “decimating” compression algorithms, creation of “imaginary” points, etc used by algorithms in the quest for automation and trans-telephonic transmission of minimized Holter files.

Such computer programs have had only limited success in diagnosing pathological conditions which compromise a patient's cardiovascular system.

As a result, many patients have had pathological conditions go undetected.

Images