WO1996024906A1

WO1996024906A1 - Audible real-time digitized eeg monitoring

Info

Publication number: WO1996024906A1
Application number: PCT/US1996/001765
Authority: WO
Inventors: Kenneth G. Jordan
Original assignee: Jordan Kenneth G
Priority date: 1995-02-09
Filing date: 1996-02-09
Publication date: 1996-08-15

Abstract

A method and apparatus for processing digitized EEG signals utilizes 'harmonic linking' to generate audible tones representing brainwave activity in the delta, theta, alpha, and beta frequency ranges. Muscle and other artifact are removed from the EEG signals so that seizures can be accurately identified and treated by attendants. The aural system allows attendants to monitor for seizures or other brainwave abnormality while performing other tasks, such as preparing medications or adjusting other monitors. In one embodiment, multi-channel digital electroencephaloaudiograms (EEAGs or EAGs) are produced from the multi-channel analog EEG signals. The present system allows attendants to monitor patients real-time without visual fatigue from watching an EEG display or evaluation of EEG signals containing frequencies above approximately 30 hertz. The stored results are also beneficial in post hoc evaluation and research.

Description

AUDIBLE REAL-TIME DIGITIZED EEG MONITORING

Background of the Invention The present invention relates to a system for real-time monitoring of patients for seizure activity, and further relates to the representation of electroencephalograms (EEGs) in audio format with frequency content limited to brainwave activity that is of interest.

It is now common to monitor human brainwave activity using EEGs, and such monitoring is used in some form both in clinical settings and for psycho-physiological monitoring. That is, EEGs are used by medical personnel to monitor and analyze the brain function of epileptics, victims of head trauma, and other patients susceptible to seizures. Biofeedback has proved to be a popular application of the concept of brainwave and physiological monitoring.

EEGs involve the placement of electrodes or leads at specific locations around the scalp. Arrays of horizontally and longitudinally positioned leads are referred to as montages, and a montage using a common reference is used as a baseline measurement to convert the analog to digitized EEG signals. Typically, a montage switch or selector is used to select the pairs of electrodes at each scalp "site" corresponding to each of the 8-16 channels of the EEG unit. A differential amplifier with a frequency bandwidth in the 0.5-100 hertz (hz) range is conventionally used to pass analog EEG signals from each electrode pair to a display format of the various channels.

The brainwave frequencies are generally classified as delta (0.5-3.0 hz), theta (3.5-7.0 hz), alpha (8.0-13.0 hz), and beta (14.0-25.0 hz), and generally have amplitudes in the 20-80 microvolt range during normal function. Muscle activity is at approximately 30 hz and greater and generally has higher amplitude than brainwave activity.

A significant problem in EEG monitoring is that muscle frequencies or other artifacts are often embedded in the EEG signals and can obscure identification of the delta, theta, alpha, and beta frequencies as well as general brainwave activity. Also, if a scalp electrode is not properly positioned on the patient, additional artifact can be introduced to the EEG signals. In biofeedback, the presence of artifact is usually not critical, since the signals are used merely as relative indications of a psycho-physiological condition, such as stress. The signals are occasionally linked with a music synthesizer, with tones of various pitch and timbre somewhat arbitrarily assigned to the signals. The signal content includes a wide spectrum of frequencies and any resultant tone will depend upon the dominant combination at the time. Thus, this application, while found to be beneficial to some persons in managing psycho-physiological self- awareness, is basically for non-medical applications.

The representation of EEG signals in audible form has also been performed with ambulatory EEG (AEEG) data, wherein a patient collects data over a period of time by wearing various scalp electrodes while ambulatory. The data is recorded using analog EEG multiplexed to up to 8 channels on a cassette tape. The tape is read by a scanner which demultiplexes the data for visual display. The scanner can play the tape at up to 60 times the recording speed, which brings the recorded waveforms up to frequencies in the audible range. However, since the patient is ambulatory during the recording of the EEG, high frequency "noise" can easily hide the brainwave activity. Also, since a substantial amount of data must be recorded in order to be played 60 times faster for an adequate amount of time, AEEG is not useful for real-time monitoring.

Processing and analysis of digitized EEG signals has allowed neurologists to identify and better understand activity in the various brainwave bandwidths, i.e., delta, theta, alpha, and beta frequency ranges. Spectral analysis of the digitized data is often performed using a Fast Fourier Transform (FFT) algorithm. FFT can be used to represent the frequency content of the EEG signal as amplitude (in volts) or power (in watts or proportional to the voltage squared) at specific frequencies.

However, this and other types of EEG data processing, such as AEEG, require interpretation by experienced neurologists, are usually not performed real-time, and include limited sampling intervals of data acquisition. In fact, post hoc comparison of EEG data with later patient diagnosis has revealed that artifact was mistaken for seizure activity, and, more critically, that seizures had been dismissed as artifact or other activity.

In one system, results are averaged from the digitized, spectral data taken during the sample intervals, or epochs. The averaging of the sampled data reduces or eliminates the possibility of accurately identifying activity in the separate delta, theta, alpha, and beta ranges. These "smoothed" results are visually displayed and may utilize a warning tone to indicate a value above a certain threshold. Thus, this system still requires concentration by a skilled attendant in order to interpret the data, and visual fatigue is common.

In another system, the use of spectral analysis is eliminated by repeating an epoch N times within the same length of time, where this N factor is adequate to make the signal audible, and this repeated waveform is output to a listener. In this system, the important low frequency content of the EEG signals is lost through epoch delineation based on a determined dominant frequency. That is, the epochs are based on a threshold crossover (e.g., "zero" volt crossover) of the signal which indicates a dominant frequency, and therefore frequency components below the dominant frequency are lost when the waveform is reproduced N tones for that epoch. Any opportunity to represent the delta, theta, alpha, or beta brainwaves can easily be lost. Thus, most present systems utilizing EEGs require visual monitoring and interpretation by skilled personnel.

These personnel are relatively expensive and still subject to error when artifact is present in the signals. Aural systems, which attempt to make the EEG signals audible, still fail to adequately present the delta, theta, alpha, or beta spectral frequencies, as desired for real-time or post hoc analysis. In view of the foregoing, a need exists for an improved EEG format that overcomes the problems mentioned.

Summary of the Invention The audible real-time digitized EEG monitoring system of the present invention overcomes the aforeπoted disadvantages by "harmonically linking" digitized EEG frequencies below approximately 30 hz to frequencies near 200 hz or greater, bringing them into the audible range. The real time, harmonically linked, aural EEG signal is easier to monitor, since humans can more readily distinguish between close aural frequencies than they can visually distinguish between low/high frequencies in the signal. This further avoids visual fatigue associated with visual monitoring. Selection of bandwidths corresponding to delta, theta, alpha, and beta frequencies simplifies accurate identification of seizure activity. And, post hoc analysis of the recorded data can yield valuable research information on "normal" versus "abnormal" brainwave activity, uncontaminated by artifact. Thus, on going audio monitoring of patients can result in a wealth of new information valuable to neurologists, psychologists, and others, as described below.

The audible real-time digitized EEG monitoring system of the present invention includes a method for generating an audible signal representing a brainwave for real-time monitoring, and comprises the steps of:

(a) establishing a first signal representing a brainwave;

(b) harmonically linking said first signal to a second signal in an audible frequency range, said harmonic linking including (i) transforming said first signal from the time domain to the frequency domain, (ii) selecting frequencies of said first signal that are below approximately 30 hertz, and (iii) converting said selected frequencies to said second signal with frequencies above approximately 200 hertz; and

(c) converting said second signal to an audible tone for real-time monitoring of brainwave activity in the delta, theta, alpha, and beta frequency range.

The audible real-time digitized EEG monitoring system of the present invention further monitors delta, theta, alpha, and beta brainwave activity from EEG signals, wherein the system comprises: a converter for digitizing EEG signals from a plurality of channels to provide digitized EEG waveforms which can be conditioned to frequencies of interest; means for harmonically linking said digitized EEG waveforms to a plurality of signals in the audible frequency range, said harmonic linking including transforming said digitized EEG waveforms from the time domain to the frequency domain and applying a multiplier to frequencies below approximately 30 hertz to generate conditioned frequencies above approximately 200 hertz; and a transduction device for converting said plurality of signals into sounds to be heard by an attendant for detection of abnormal brainwave activity, said device including an amplifier such that said plurality of signals are amplified to sounds approximately at least 20 decibels. An apparatus of the present invention for generating an audible EEG signal consisting of delta, theta, alpha, and beta brainwaves comprises: a digital frequency shifter for harmonically linking a digitized EEG signal to a conditioned signal by producing a frequency shift of the digitized waveform from frequencies below approximately 30 hertz to frequencies in an audible frequency range; and an amplifier for generating tones at an audible sound level from said conditioned signal in order to represent delta, theta, alpha, and beta brainwave activity.

Also, an apparatus of the present invention for generating an audible EEG signal consisting of delta, theta, alpha, and beta brainwaves comprises: a converter for generating a digitized signal from an analog EEG signal; means for harmonically linking said digitized signal to a conditioned signal by producing a frequency shift of the waveform from frequencies below approximately 30 hertz to frequencies in an audible range; and an amplifier for generating tones at an audible sound level from said conditioned signal in order to represent delta, theta, alpha, and beta brainwave activity.

In a first embodiment, the audible real-time digitized EEG monitoring system of the present invention provides a visual display of the digitized, selected brainwave frequencies in the delta, theta, alpha, and beta ranges in the course of processing the harmonically linked signals to audio form. In another embodiment, a visual form of the partly processed signal is generated in parallel with the audio processing. In this embodiment, multi-channel digital electroencephaloaudiograms (EEAGs or EAGs) are produced from the multi-channel analog EEG signals. The visual verification of the aural signal is both possible and desirable when used either by an attendant during real-time monitoring, or by a researcher during later analysis. Also, a wireless headset worn by the attendant would allow him/her to move out of visual range, such as to an adjoining room, while still monitoring the audible signal.

Progressive structural brain lesions, such as caused by an expanding blood clot or tumor, can be identified and further investigated. Similar monitoring for carotid endarterectomy may be performed, where focal cerebral ischemia alters the amplitude and frequency of EEG waveforms.

Other applications of the present invention include adaptation to monitoring of known epileptic patients, ambulatory patient monitoring, or out-patient monitoring via fiberoptic or telephonic link to a remote monitoring station. Post hoc analysis can yield valuable data for medical research on brainwave activity. For example, polysomnographic studies would utilize auditory analysis to discriminate between sleep stages, a process which is presently performed visually or by automated paradigms similar to seizure detection algorithms now used.

Diagnosis and study of schizophrenia, attention deficit disorder, and other psychiatric disturbances can also be performed with the present invention. "Brain music" resulting from the auralization of the EEG may reveal signatures or "fingerprints" of normal versus abnormal brainwave activity. This diagnostic capability can also provide a method for monitoring the effect of various therapies. A further, non-medical application of the resulting brain music of the present invention includes the creation of musical solos, duets, or even symphonies comprised of the brainwaves of one or more "musicians". The music would vary according to alertness, activity, and chronological age of each person involved.

Yet another application of the present invention is its use in developing artificial intelligence (Al) technology for automated seizure detection. That is, neural networks and expert systems could utilize digitized audio signals rather than digitized visual signals. Just as it is easier for the human ear to distinguish between various aural tones, the digitized audio signals would be more easily and effectively employed by the Al systems in identifying seizures automatically.

Further advantages and applications will become apparent to those skilled in the art from the following detailed description and the drawings referenced herein. Brief Description of the Drawings FIGURE 1 is a schematic representing a first embodiment of the audible real-time digitized EEG monitoring system of the present invention, illustrating a single processing flow of the EEG signals to visual representation of the power spectra and to audible tones. FIGURE 2 is a flow diagram of the first embodiment, illustrating the three Phases a, b, and c comprising the single flowpath of the EEG signal processing.

FIGURE 3 is an example of the power spectral information of the present invention, illustrating the original delta, theta, alpha, and beta frequencies (f*) which are shifted to audible frequencies (F-).

FIGURE 4 is a schematic representing a second embodiment of the audible real-time digitized EEG monitoring system of the present invention, illustrating a dual, parallel processing flow of data, visual and aural representation of the EEG signals.

FIGURE 5 is a flow diagram of the second embodiment of FIGURE 4, illustrating the phases comprising the dual flowpath of the EEG signal processing.

Detailed Description of the Invention

A first embodiment of the audible real-time digitized EEG monitoring of the present invention, simply referred to as aural monitoring herein, is illustrated in FIGURE 1, and indicated by the reference numeral 10. A goal of the aural monitoring is to determine when abnormal or seizure activity occurs in any of the four brainwave frequency ranges or bandwidths of interest • namely, delta, theta, alpha, and beta frequencies ranging from 0.5 to 25 hz • at any location in the brain.

The aural monitoring system may be considered to include three phases, flowing from left to right in FIGURES 1 and 4, and described in detail in FIGURES 2 and 5. These phases or processes are generally comprised of a) digitization of multi channel analog EEG signals, b) harmonic linking of the lower frequency content of the digitized EEG to higher frequencies, and c) audio processing of the harmonically linked signals to generate audible tones.

As illustrated in FIGURE 1, an initial phase 12 of the present invention includes establishing analog EEG signals using conventional electrodes 14, placed in the usual paired arrays, or sites, around a patient's scalp 16. This data is typically displayed for monitoring on a multi channel analog EEG unit 18. This format used alone would have to be carefully and continuously scrutinized by an attendant in order to detect seizure activity, which often occurs near 20 hz.

A displayed waveform 20, either analog or digitized, of a single channel corresponding to an individual site is not limited in its frequency content. The higher frequencies in the waveform 20 may dominate or obscure the lower frequencies, and muscle activity near 30 hz can easily be mistaken for a seizure by the attendant. Likewise, a loose electrode lead on the patient 16 may be falsely read by an attendant, even if the frequency is much higher than the beta upper brainwave frequency of 25 hz. Conversely, a true seizure may not be detected among the range of displayed frequencies, or may be falsely read as muscle activity. The harmonic linking of the present invention includes first transforming digitized EEG, time domain data of the selected site or channel 20 to frequency domain data. That is, representation of the digital signal in terms of time and amplitude (voltage) is transformed to representation in terms of the frequencies and corresponding amplitudes contained in the signal. The digitization, frequency transformation, and other procedures can be performed on almost any of today's portable computers 22. Preferably, a UNIX™ based operating system capable of real-time, multi-tasking is utilized.

FFT algorithms, for example, well known to those skilled in the art, provide one method to accomplish the transformation to the frequency domain and allow identification and retention of the lower frequencies. That is, a waveform in the time domain is decomposed to a range or spectrum of frequencies having associated amplitudes. The FFT algorithm identifies these frequencies and amplitudes over specific intervals of time.

Thus, the four brainwave bandwidths can be selected for further processing, analysis, and display described herein. The selection may be performed in the pre programming of the harmonic linking process, or may be an input to the system by the attendant prior to EEG monitoring. The higher frequencies from muscle and other artifact can, if desired, be effectively removed. Alternately, the present invention includes frequencies in ranges other than the aforementioned brainwave bandwidths which can be selected during the harmonic linking process, for analysis and display in other research or patient monitoring.

As illustrated in the middle of FIGURE 1 as an intermediate phase 24, the FFT results may be displayed on a monitor 26 as vertical bars 28. The FFT generates frequency and amplitude/voltage (V) information, which may be displayed as x (abscissa) and y (ordiπate) axis data, respectively. Power spectra are more usually displayed, where the power (represented as V²) is displayed as the y-axis. Squaring the amplitude (V) enhances the differences in the relative intensities or energies between frequencies. For example, three frequencies with V - 0.5, 1.0, and 5.0, would have power V² - 0.25, 1.0, and 25.0, respectively. Thus, the power or energy disparity between the first and third frequencies is amplified. Other graphical displays can be selected, including topographic scalp "maps" which provide regional pictorial representation of frequency spectra of the brain. The type of display may be a fixed feature of an embodiment of the system of the present invention, or may be a user-selected option at the time of monitoring.

In the present invention, increments within each of the four bandwidths may be displayed, or the mean frequency of each bandwidth may be displayed. That is, the power spectra at 1.0 hz for delta, 5.0 hz for theta, 10.0 hz for alpha, and 18.0 hz for beta may be displayed for simplicity. Preferably, the pre-FFT data and the full power spectral data are stored on a hard drive or other storage device, such as a diskette, whether they are displayed or not.

The harmonic linking of the present invention includes further conditioning of these frequencies from the four bandwidths to obtain frequencies which are audible to the average adult. That is, frequencies as low as 0.5 hz are converted to frequencies near or above 200 hz, although, some humans are able to accurately detect frequencies below 100 hz. This is accomplished through a multiplication factor or multiplier (C) applied to each frequency (f*) or band of the spectra. This multiplier may be a constant for all four bandwidths, or it may be made variable to focus on a particular frequency range. The multiplier brings the lowest spectra band to the minimum audible range, and the higher bands to correspondingly higher frequencies without applying a gain change to the amplitudes.

Accordingly, as illustrated to the right in FIGURE 1 as a final phase 30, the harmonically linked or conditioned EEG signals may now be processed to aural form using audio transducers well known to those skilled in the art. Real-time monitoring or later analysis is assisted by the use of a graphic equalizer 32, which may be used to select and control the individual volume (gain or loudness) of the four brainwave bands as played over a speaker

34 or through headphones 36.

Another option which yields even greater mobility for the attendant is the use of earplugs or a wireless headset 37 containing a radiowave receiver. A radiowave transducer would be included in the audio processing of the final phase of the present invention. Thus, the attendant could be out of visual range of the patient and EEG equipment, such as the next room or down the hall, yet still be able to monitor the audio signals.

A spectral edge for the channel, or the frequency below which is contained 95% of the power, may be selected and listened to as well. That is, the spectral edge may be processed along with the frequency bandwidths and chosen for display by the attendant upon initiating the monitoring. The mean spectral frequency and peak power frequency may be similarly processed and can yield valuable information.

Thus, in the first embodiment of the aural monitoring system 10 of the present invention, the processing of the analog EEG signals proceeds from conventional mechanisms 12 to harmonic linking 24 to auralization 30, in a sequential manner. This single flowpath is illustrated in greater detail in FIGURE 2, separated into Phases a, b, and c.

In the first Phase a of FIGURE 2, comprising Steps 110-124, mechanisms well known to those skilled in the art are employed to obtain the digitized EEG signals. Specifically, in Step 110, twenty-one electrodes 14 may be placed according to a standard international 10-20 method on the patient's scalp 16. The horizontal and longitudinal positions are designated as X and Y, respectively, for each electrode pair. Either manual or automatic site selection may be made using a montage or site selector, as indicated in Step 112.

A differential amplifier in Step 114 is used to generate a single signal from each electrode pair at the selected site. The low drift, high gain operation of the differential amplifier generates an output signal which is proportional to the difference between two input signals, and this output signal is an input to one of the 8-16 channels of the multi-channel analog EEG of Step 116. These channels are commonly displayed on the EEG unit 18 for visual monitoring, and these signals are recorded for later analysis, as indicated in Step 118. The differential amplifier and EEG unit 18 may be obtained from a medical instrument manufacturer such as NeuroCoπcepts, Inc. in Madison, Wisconsin.

Phase a in FIGURE 2 continues with Step 120, with an analog-to-digital converter (ADC), well known to those skilled in the art for digitizing the analog EEG signals. The ADC has at least a 10 bit resolution and provides a sample rate of at least approximately 200 hz. Although, preferably a 12 bit resolution with 500 hz sample rate is used. The output of the ADC comprises the multi-channel digital EEG of Step 122, which is stored on a storage device (not shown) and may be also displayed, as indicated in Step 124. Due to the relatively high sample rate in comparison to the brainwave frequencies, the digitized EEG accurately represents the analog EEG. Not shown is the possible use of cable or modem transmission for remote, real-time monitoring of either the analog or digitized EEG signals. That is, the output of Step 116 or 122 may be relayed to a remotely located attendant for further processing of Phases b and c in the aural monitoring system 10 of the present invention, described below.

Continuing to Phase b of FIGURE 2, the digitized data from each channel of Step 122 is transformed in Step 126 by application of the FFT algorithm. Resultant power spectra 28 of the digitized EEG are generated for each channel on a continuous basis. That is, the continuous FFT processing includes transformation from the time domain to the frequency domain, storage of this data onto a recording medium, and display of the power spectra 28 of the brainwave frequencies, as indicated in Step 128. The EEG processing time thus far in Phases a and b is insignificant and further minimized by the selection of frequencies below approximately 30 hz. That is, the required processing time for these lower frequencies is very small and easily handled by the computer 22. Although, today's processors are capable of handling much higher frequencies without requiring significant time delays.

Thus, by the time of Step 130 and the application of the multiplier to the FFT results, only a small time would have elapsed if a seizure had occurred in the patient. The multiplier to produce frequencies in the audible range may be a fixed value, may correspond to the frequency range (delta, theta, etc.) and/or site, or may be varied by the attendant during use. A default multiplier may be supplied in alternate embodiments.

The value of the multiplier is largely dependent upon the spectra and the site, and the application to the selected frequencies is illustrated in FIGURE 3. Frequencies within a range, such as every 0.5 hz between 0.5 and 3.0 hz for delta, may be multiplied and further processed, or a single value in the range may be chosen for frequency shifting by the multiplier and auralization in Phase c. A preferred application is to multiply the mean of each range by 300, thereby obtaining frequencies of 300 hz for delta, 1500 hz for theta, 3000 hz for alpha, and 5400 hz for beta. In alternate embodiments, the multiplier may be chosen by the attendant as a fixed value for all the bandwidths or as a variable within a range of values, and the attendant may choose either frequency increments or the bandwidth mean for application of the multiplier.

As illustrated in a graph 142 on the left of FIGURE 3, the power spectra 28 prior to application of the multiplier shows the differences in power of each frequency, f,. Power is represented as V², in units of millivolts squared (mV²). A graph 144 on the right illustrates shifted frequencies, f'„ which have equivalent power values 28' as before the application of the multiplier. Thus, the energies of each of the spectra 28 are not changed by the harmonic linking to the higher frequencies.

Referring again to FIGURE 2, Step 132 provides the necessary amplification and audio transduction of the conditioned signal from Step 130. Any well known audio transducer 146 will generate the audio signals to be processed in Phase c, described below. For example, a sound card, such as SoundBlaster™, in a microprocessor- linked transducer 146 would provide audiblization of the signals. Similarly, an amplifier 148 of any type well known to those skilled in the art is used to increase the amplitude of the signal to hearing range, or preferably above 20 decibels (approximately a whisper). The final procedures of FIGURE 2 include Step 134, wherein the graphic equalizer 32 is used to separately monitor the frequencies or bands of interest by the attendant, or, later, by the analyst or researcher. That is, the volume-controlled bands of the equalizer 32 correspond to the alpha, beta, delta, theta bandwidths. The audible, chosen frequencies are stored as in Step 136 and may be played through speakers or headphones, as in Steps 138 or 140, respectively. If earplugs 37 are to be used, as in Step 139, a radiowave transducer (not shown) would be used to generate the transmitted radiowaves.

In the first embodiment 10 of the aural monitoring system, the now audible mean frequency of each of the four brainwave frequency ranges may have an increased volume for real-time monitoring. Or, a selected one or two (say, delta and theta) may have their volumes set high and the remaining ranges have their volumes reduced to their lowest levels. In addition, the spectral edge frequency may be processed, brought to audible range, and selected for monitoring by the attendant.

Thus, abnormal or seizure activity may be determined by an abrupt, higher-pitched sound from among the continuous tones being heard. During research or post hoc analysis of the spectral data, a transient waveform, or transitory shift in frequency/amplitude, in the spectral data may be utilized to distinguish abnormalities in the brainwave patterns.

In the first embodiment of FIGURES 1 and 2, an additional alarm signal for breach of the "confidence limits of normalcy" may be included to further aid in recognizing a seizure. The alarm signal, coupled with the visual displays 20, 28 of the EEG and the aural tones of the present invention provides additional verification of abnormal brainwave activity. A second embodiment utilizing two process flowpaths, one visual and the other aural, is illustrated schematically in FIGURE 4 and in greater detail in FIGURE 5, and generally referenced by the numeral 200. As with the first embodiment 10, the audible real-time digitized EEG monitoring can be divided into three phases, with a leftmost or initial phase 202 of FIGURE 4 similar to the initial phase 12 of FIGURE 1. That is, standard methods are employed to obtain multi-channel analog EEG from the patient using a plurality of electrodes 14. In a visual flowpath 204 of FIGURE 4, the processing continues with digitization of the EEG signals. Next, the application of an FFT algorithm produces a visual display of the power spectra 28, without complete harmonic linking (i.e., a multiplier) to higher frequencies. This is a truncation of the flow of FIGURE 1.

In an aural flowpath 206 of FIGURE 4, the multi channel analog EEG is converted to analog audio signals, using the audio transducer 146. The audio signals at this point will be of frequency and amplitude too low to be beneficial. Digitization of these signals using a computer 207 produces what will be termed multi channel digital electroencephalo audiograms (EEAGs or EAGs), which are then harmonically linked, amplified, and monitored by an attendant via speaker 34, headphones 36, or earplugs 37. Thus, a rightmost, or last, phase 208 of the dual flow of FIGURE 4 is similar to the final phase 30 of FIGURE 1.

Referring in detail to FIGURE 5, it is readily observed that the initial and final procedures are indeed similar to Phases a and c of FIGURE 2. Steps 210-216 to generate multi-channel analog EEG signals correspond to Steps 110-116, respectively. Similarly, Steps 218-222 in FIGURE 5 to generate multi channel digital EEG signals correspond to Steps 120-124 of FIGURE 1. Steps 224 and 226 for visually displaying the spectra 28 correspond to Steps 126 and 128. These steps were described above. Thus, the visual flowpath 204 of this embodiment 200 very closely follows the flow 12, 24 of the first embodiment 10 in generating the visual display of the power spectral data 28 for the delta, theta, alpha, and beta frequencies. The aural flowpath 206 of FIGURE 5 branches off from the results of Step 216, wherein the multi-channel analog EEG signals are generated. The audio transducer 146 is utilized in Step 228 to produce the audio signals which are converted by the ADC of Step 230. Multi-channel digital EEAG of Step 232 in the aural flow processing 206 is analogous to the multi-channel digital EEG of Step 220 in the visual flow processing 204.

Next, the EAG signals of the present invention are harmonically linked in the parallel audio processing 206 of FIGURE 5. That is, since audio traπsduction has already occurred, what remains are transformation to the frequency domain, reduction to the desired frequencies for shifting to higher frequencies, and amplification to audible levels. Again, the frequency selection may be pre-programmed or selected at the beginning of monitoring, and the present invention includes frequencies beyond the brainwave bandwidths. Thus, steps 234-238 are similar to Steps 126, 130, and 132 of FIGURE 2. FIGURE 5 Steps 240-246, for volume control for selective listening of the frequencies, are equal to FIGURE 2 Phase c Steps 134-140.

Thus, the audible real-time digitized EEG monitoring system of the present invention processes digitized EEG signals and generates audible tones representing brainwave activity in the delta, theta, alpha, and beta frequency ranges. Muscle and other artifact are removed during the FFT/multiplier processing of the EEG signals so that seizures can be accurately identified and treated by attendants. The present system allows attendants to monitor patients real-time without visual fatigue from continuously watching an EEG display or evaluation of EEG signals containing frequencies above approximately 30 hertz to discern delta, theta, alpha, and beta activity.

The aural system of the present invention allows attendants to monitor for seizures or other brainwave abnormality while performing other tasks, such as preparing medications or adjusting other monitors. By using a wireless headset to receive radiowave transmissions, an attendant could even leave the room while still actively monitoring a patient. The EEG signals may also be remotely transmitted for harmonic linking and aural processing to a monitoring station, such as via telephone lines or fiberoptic cable transmission from a patient's home to a hospital.

The stored results are also beneficial in post hoc patient evaluation and neurological research, such as polysomnographic and psychiatric studies. "Brain music" resulting from the audiblization of the present invention presents a potential, commercial application featuring one or more "musicians". Finally, the digitized audio signals of the present invention can be used for accurate, automatic seizure detection using artificial intelligence technology.

The embodiments illustrated and described above are provided merely to indicate possible embodiments of the audible real-time digitized EEG monitoring of the present invention. Other changes and modifications may be made from the embodiments presented herein by those skilled in the art without departure from the spirit and scope of the invention, as defined by the appended claims.

Claims

What is claimed is:

1. A method for generating an audible signal representing a brainwave for real-time monitoring, comprising the steps of:

(a) establishing a first signal representing a brainwave; (b) harmonically linking said first signal to a second signal in an audible frequency range, said harmonic linking including (i) transforming said first signal from the time domain to the frequency domain, (ii) selecting frequencies of said first signal that are below approximately 30 hertz, and (iii) converting said selected frequencies to said second signal having frequencies above approximately 200 hertz; and (c) converting said second signal to an audible tone for real-time monitoring of brainwave activity in the delta, theta, alpha, and beta frequency range.

2. The method of Claim 1, wherein step (a) includes digitizing said first signal.

3. The method of Claim 1, wherein step (iii) includes applying a multiplier.

4. The method of Claim 3, wherein step (iii) includes adjusting said multiplier.

5. The method of Claim 1, wherein step (c) includes using an amplifier.

6. The method of Claim 1, wherein step (c) includes amplifying said audible tone to approximately at least 20 decibels.

7. The method of Claim 1, further comprising the step of selecting at least one frequency bandwidth for listening wherein said frequency bandwidth corresponds to any of delta, theta, alpha, and beta brainwaves.

8. A method of processing EEG signals from a plurality of brainwaves for real-time monitoring of seizure activity, comprising the steps of:

(a) establishing a plurality of analog EEG signals representing a plurality of brainwaves;

(b) digitizing said plurality of analog EEG signals to generate a first plurality of digitized waveforms; (c) harmonically linking said first plurality of digitized waveforms to a second plurality of digitized waveforms, said harmonic linking including (i) transforming said first plurality of digitized waveforms from the time domain to the frequency domain, (ii) selecting frequencies of said first plurality of digitized waveforms that are below approximately 30 hertz, and (iii) converting said selected frequencies of said first plurality of digitized waveforms to said second plurality of digitized waveforms having frequencies above approximately 200 hertz; and

(d) converting said second plurality of digitized waveforms to generate a plurality of audible signals to monitor seizure activity in the delta, theta, alpha, and beta range.

9. The method of Claim 8, wherein step (iii) includes applying a multiplier.

10. The method of Claim 9, wherein step (iii) includes adjusting said multiplier.

11. The method of Claim 8, wherein step (d) includes aurally transducing and amplifying said second plurality of digitized waveforms.

12. The method of Claim 8, wherein step (d) includes amplifying said plurality of audible signals to approximately at least 20 decibels.

13. The method of Claim 8, further comprising step (e) selecting at least one frequency bandwidth for listening wherein said frequency bandwidth corresponds to any of delta, theta, alpha, and beta brainwave frequency ranges.

14. The method of Claim 13, wherein any of steps (a)-(e) includes a visual display of the result.

15. A system for monitoring delta, theta, alpha, and beta brainwave activity from EEG signals, comprising: a converter for digitizing analog EEG signals from a plurality of channels to provide digitized EEG waveforms which can be conditioned to frequencies of interest; means for harmonically linking said digitized EEG waveforms to a plurality of signals in the audible frequency range, said harmonic linking including transforming said digitized EEG waveforms from the time domain to the frequency domain and applying a multiplier to frequencies below approximately 30 hertz to generate conditioned frequencies above approximately 200 hertz; and a transduction device for converting said plurality of signals into sounds to be heard by an attendant for detection of abnormal brainwave activity, said device including an amplifier such that said plurality of signals are amplified to sounds approximately at least 20 decibels.

16. The system of Claim 15, wherein said multiplier is adjustable.

17. The system of Claim 15, further comprising a graphic equalizer for selecting at least one frequency bandwidth to be listened to by the attendant wherein said frequency bandwidth corresponds to any of delta, theta, alpha, and beta brainwaves.

18. The system of Claim 15, wherein said means for harmonic linking includes a Fast Fourier Transform algorithm for said transformation to the frequency domain.

19. An apparatus for generating an audible EEG signal consisting of delta, theta, alpha, and beta brainwaves, comprising: a converter for generating a digitized signal from an analog EEG signal; means for harmonically linking said digitized signal to a conditioned signal by producing a frequency shift of the waveform from frequencies below approximately 30 hertz to frequencies in an audible frequency range; and an amplifier for generating tones at an audible sound level from said conditioned signal in order to represent delta, theta, alpha, and beta brainwave activity.

20. The apparatus of Claim 19, wherein said converter includes an audio transducer to generate an electroencephaloaudiogram.

21. The apparatus of Claim 19, wherein said means for harmonic linking includes a Fast Fourier Transform algorithm to transform said digitized signal from the time domain to the frequency domain.

22. The apparatus of Claim 19, where said means for harmonic linking includes a multiplier to obtain frequencies of said conditioned signal above approximately 200 hertz.

23. The apparatus of Claim 19, wherein said amplifier includes an audio transducer to make said conditioned signal audible.

24. The apparatus of Claim 19, wherein said amplifier includes a graphic equalizer for selecting among delta, theta, alpha, and beta brainwave frequency ranges.

25. An apparatus for generating an audible EEG signal consisting of delta, theta, alpha, and beta brainwaves, comprising: a digital frequency shifter for harmonically linking a digitized EEG signal to a conditioned signal by producing a frequency shift of the digitized waveform from frequencies below approximately 30 hertz to frequencies in an audible frequency range; and an amplifier for generating tones at an audible sound level from said conditioned signal in order to represent delta, theta, alpha, and beta brainwave activity.

26. The apparatus of Claim 25, wherein said digital frequency shifter for harmonic linking includes a Fast Fourier Transform algorithm to transform the digitized EEG signal from the time domain to the frequency domain.

27. The apparatus of Claim 25, where said digital frequency shifter for harmonic linking includes a multiplier to obtain frequencies of said conditioned signal above approximately 200 hertz.