ELECTROENCEPHALIC NEUROFEEDBACK APPARATUS AND METHOD.

TECHNICAL FIELD

The invention relates in general to the reception and processing of bioelectrical signals from an organism. Furthermore, the invention relates to an improved method and apparatus for receiving bioelectrical signals, processing the signals and signaling back to the organism so as to allow the organism to either inhibit or facilitate a selected frequency.

BACKGROUND ART

A number of different feedback-type methods and apparatus are known dating back to as early as 1960. Early studies by several researchers focused on bioelectrical feedback on persons suffering from hemiplegia, i.e., paralysis of one lateral half of the body resulting from injury to the motor centers of the brain.

In I960, A. A. Marinacci and M. Horande investigated neurofeed back with respect to left-sided hemiplegia. As reported in "Electromyogram in Nueromuscular Re-education", Bulletin of the Los Angeles Neurologic Society. 25: 57-71, I960, they inserted needle electrodes into the involved left arm muscles, and could find no voluntary nerve impulses. Electrodes were inserted into the normal right deltoid to

feedback. The electrodes were then inserted into the paralyzed left deltoid muscle. The patient was able to generate from 10 to 15 percent motor action potential in a location from which there had been no previous detectable activity. The same procedure was utilized successfully at other muscle sites.

In 1964, J. M. Andrews reported on study utilizing a patient group of hemiplegics who had electromyogram EMG electrodes inserted in the involved tricep muscles as reported in "Neuromuscular Re-education of the Hemiplegic with the Aid of the Electromyograph," Archives of Physical

Medicine and Rehabilitation. 45: 530-532, 1964. Auditory feedback was provided as the subjects tried to generate sound and movement. A five-minute trial period was allowed, and seventeen out of the twenty patients showed an increase in motor action potentials.

In 1973, H. E. Johnson and W. E. Garton reported on ten hemiplegic patients, who utilized EMG practices as an aid in total rehabilitation rather than just the return of voluntary movement as in the Andrews study as discussed in

"Muscle Re-education in Hemiplegia by use of Electromyograph

Device", Archives of Physical Medicine and Rehabilitation.

54: 320-325, 1973. Five out of ten subjects had enough improvement to eliminate leg bracing on the involved side.

In 1974, J. Brudny and others used EMG feedback to treat a group of thirty-six patients, thirteen of whom had hemiparesis. Brudny, J. Korein, J. , Levidow, L. Grynbaum,

B.B., Lieberman, A., and Friedman, L. W. , "Sensory Feedback

Disorders of Voluntary Movement," Neurology. 24: 925-932,

1974. In this study surface electrodes were used instead of inserted needle electrodes. In two individuals there was no change. In one patient there was relief from muscle spasticity. In six patients function of the extremity was re-established, and in four cases prehension became possible.

Also, in 1974, D. Swaan, P. C. . Van ieringer and S. D. Fokkema explored EMG feedback of seven patients, four of whom were hemiplegic. "Auditory Electromyographic Feedback Therapy to Inhibit Undesired Motor Activity," Archives of Physical Medicine and Rehabilitation. 57:9-11, 1974. The subjects taught to inhibit the peroneus longus muscle while contracting their quadricep muscle. Conventional rehabilitation methods were used to suppress the undesirable hyperactivity of the peroneus longus muscles along with the feedback. No justification was given for reinforcement of the quadriceps and inhibition of the peroneus longus. In 1975, J. V. Basmajian, C. G. Kukulka, M. G. Narayan and K. Takebe compared EMG biofeedback plus physical therapy with the results of standard rehabilitation procedures in cases of ankle dorsiflexion paralysis after stroke repeated in "Biofeedback Treatment of a Foot-Drop After Stroke Compared With Standard Rehabilitation Techniques: Effects on Voluntary Control and Strength," Archives of Physical Medicine and Rehabilitation. 56: 231-236, 1975. The authors claimed that an increase in both strength and range of motion in the biofeedback group was twice as great as the

ac eveme s o e exerc se con ro group. e wo group of patients were not variably matched. When biofeedback was added to physical therapy, the mixed variables were not controlled. In 1976, L. P. Taylor and B. Bongar described the use of electromyometry feedback for the treatment of cerebrovascular lesion patients in Clinical Applications in Biofeedback Therapy. Psychology Press, Los Angeles, California, 1976. Patients were taught to inhibit one set of muscles while simultaneously facilitating others. For example, inhibition of thumb flexion was attempted while thumb extension was facilitated.

In 1979, F. Keefe and K. Trombly utilized EMG feedback to aid a hemiplegic patient judge limb position without being able to see the limb in "Impaired Kinesthetic Sensation: Can EMG Feedback Help?" Presented at the Proceedings of Biofeedback Society of America, Tenth Annual Meeting, February, 1979 in San Diego, California. The patient participated in an A-B-A-B withdrawal design to evaluate the effects of EMG biofeedback on accurate limb positioning. EMG feedback with audio feedback produced improvement in performance relative to baseline. Withdrawal of feedback produced a decrement in performance, and when EMG feedback was re-instituted, performance once again improved. The patient was able to generalize the EMG feedback training to improved functional use of the arm.

In 1979, R. Koheil, et al, at the Ontario Crippled Children Centre, developed a Joint Position Trainer to provide precise feedback of limb position to three

. , . , , ., , . ,

G., "Joint Position Training for Hyperextension of the Knee in Stroke Patients: Preliminary Results", presented at the Proceedings of the Biofeedback Society of America, Tenth Annual Meeting, February, 1979 in San Diego, California. The Joint Position Trainer provided feedback of position rather than of muscle activity, and incorporated a goniometer attached to a leg cuff with auditory feedback of knee joint angle. Two of the three patients developed improved gait with increased control of knee hyperextension.

The results of the techniques involved had limited results because the difficulty of recognizing particular frequencies generated which could not be readily determined, nor could the subject have the ability to control these frequencies.

Other research efforts were conducted specifically upon those bioelectrical signals emanating from the brain. One of the earlier works was written in 1966 by T. Mullholland and C. R. Evans who described the use of alpha waves (approximately 7.5 - 11.5 Hertz) emanating from the brain to drive a feedback signal that could be perceived by the test subject and induce relaxation. Mulholland, T. , and Evans, C.R. , Nature. 211: 1278, 1966. The alpha waves could be controlled to some degree by the test subject's recognition of a tone or light when alpha waves were produced. Similarly, differentiation of particular frequencies of bioelectrical signal prevented the test subject from readily acknowledging and either inhibiting or facilitating particular frequencies.

Other electroencephalograph (EEG) feedback dev ces are described in publications by Spunda, J. and Radii-Weiss, T. , "A Simple Device for Measuring the Instaneous Frequency of the Dominant EEG Activity", Electroencephalographic Clinical Neurophvsiolog _f 32: 434, 1972. This device converted EEG frequencies into voltage levels for analysis using bandpass analysis. A series of wave form generators were activated by the flip-flop at each positively directed zero point of the filtered signal resulting in a voltage level corresponding to the frequency.

Another EEG feedback device was described in Hicks, R.G., and Angner, E., "Instrumental evaluation of EEG Time Relationships", Psvchophvsiology. 6:44, 1970. This device analyzed minute time displacements of EEG waves from cortical waves using peak detection and a type of logic as a feedback device.

Also, in Boudrot, R. , "An Alpha Detection and Feedback Control System", Psvchophysiology. 9:467, 1972, a feedback device picked up alpha waves and provided auditory and visual stimulus feedback to the patient. Then, in Pfeifer, E.A. , and ϋsselmann, C., "A Versatile Amplitude Analyzer for EEG Signals to Provide Feedback Stimuli to the Subject", Med. Biol. Eng. 8: 309, 1970, a feedback device analyzed the amplitude and provided feedback cues to subjects in studies of EEG modification. It allowed for usage with bandpass analysis incorporating logic components and a display.

Additionally, U.S. Patent No. 3,837,331 to Sidney A. Ross, issued September 24, 1974, entitled "System and Method

describes an apparatus and method for determining particular frequencies of a bioelectric signal which is analog by nature. The Ross device like the other devices require the use of band pass analysis or other techniques to filter out particular frequencies to study a particular frequency of interest.

In band pass analysis, analog filters analyze how much frequency is produced in a given period of time, or a frequency in relationship to time and voltage. The apparatus required includes a precision attenuator, an active band pass filter, a rectifying means and an integrating means in addition to those components normally utilized in bioelectrical feedback devices. Furthermore, band pass analysis or power spectral analysis is necessary to isolate the particular frequency of interest. Such analysis is typically performed on a large computer requiring special analytical skills and extensive computing time. Power spectral analysis looks at the variance of a bioelectrical signal or the covariance between one or more signal channels. The signal is broken into different frequency bands in relationship to the power density which is then analyzed using a fourier series program.

Not only is the above approach burdensome and time consuming, but also inaccurate. The frequency results are often distorted because the analytical approach used is based on exponential and logrithic analysis. Determining the actual frequency desired to inhibit or facilitate is at best haphazard due to the margins for error in the above

approac .

Also, the above referenced devices lack suitable means for displaying and recording the changing frequencies under study for subsequent review and manipulation for purposes of analysis. But, most importantly, these EEG feedback devices lacked the ability to establish selected limits or thresholds in which to gauge the progress of a test subject or reward the test subject once the test subject learned to inhibit or facilitate a particular frequency of interest. The above-mentioned deficiencies are overcome by the present invention. There is a great interest in the neuropsychology and neurophysiology community for such a device which overcomes these deficiencies. Researchers and practitioners should recognize the value immediately of the present invention. Furthermore, persons suffering from nervous disorders, induced by trauma, drug use or cogenital aberration can greatly benefit from the present invention. The present invention operates as a diagnostic tool, as well as, a means for curing nervous disorders or abnormalities in the body, particularly the brain.

DISCLOSURE OF THE INVENTION

The invention includes two embodiments of a method and two embodiments of an apparatus for practicing the methods, respectively.

The first method is an improved method of detecting and displaying analog bioelectrical frequencies in an organism's or person's body comprising the steps of detecting an analog bioelectrical signal at a selected location in the person's

, , the analog bioelectrical signal to digital signals representing particular frequencies, and selecting a particular digital signal of interest and displaying and/or recording the particular signal of interest in the form of a continuous waveform over time.

The apparatus used to perform the above method comprises a receptor means for attachment to the selected location of the person or organism for receiving an analog bioelectrical signal, an amplifier for amplifying the analog bioelectrical signal received by the receptor means and associated therewith, and an analog to digital converter for converting the analog bioelectrical signal to digital signals representing corresponding electrical frequencies. The converter receives the amplified analog bioelectrical signal from the amplifier.

A computer is used for integrating amplitude of one of the digital signals over a predetermined duration and dividing a resulting value by voltage to determine a change in voltage with respect to time duration, and thereby determine electrical frequencies emitted by the person or organism. The computer then converts the resulting value divided by voltage to a format which can be plotted over time. A display monitor for displaying the prescribed format of the corresponding electrical frequencies can be incorporated.

Furthermore, a magnetic disk recorder for recording the resulting value divided by voltage over time is used to record the information for subsequent review or analysis.

The second embo ment of the method o t e nven on involves the first embodiment of the method described above. Additional steps include sending a signal to the person when the particular digital signal of interest falls within a selected amplitude threshold for a predetermined duration, and causing the person to mentally concentrate so as affect the amplitude of the particular digital signal of interest. This allows the person to affect a certain corresponding analog bioelectrical frequency emitted at the selected location of the person's body.

Subsequent to the step of establishing an amplitude for the particular digital signal, an additional step may include integrating the -amplitude of the selected digital signal with respect to duration and dividing by the amplitude threshold voltage. This yields a determination of whether voltage of the selected digital signal is changing, and thereby whether the corresponding analog bioelectrical frequency is being inhibited or accentuated.

Also, another step may be added by displaying the selected digital signal as a continuous waveform and displaying the voltage thresholds as horizontal lines associated with the continuous waveform to facilitate the determination of whether the corresponding analog bioelectrical frequency is being inhibited or facilitated. Similarly, a step of recording may be added where the digital signals are recorded to a magnetic medium for review and data manipulation.

The step of continuously displaying the particular digital signal of interest as a waveform changing with

threshold voltage is displayed as a straight line.

A step of suppressing exteraneous signals unrelated to the analog bioelectrical signal of interest may be used to suppress unwanted noise.

The apparatus used in the second embodiment is similar to that apparatus described above. The step of detecting an analog bioelectrical signal is achieved by using two electrodes positioned at the selected location of the person's body to be affected. The step of amplifying is achieved by using a signal amplifier associated with the electrodes and which recognizes voltage potential between the electrodes. The step of converting is achieved by receiving signals representing voltage variations over time and using an analog to digital converter to convert voltage variations into digital pulses.

The step of selecting is achieved using a numerical analyzer to select digital pulses corresponding to particular analog bioelectrical frequencies. The step of sending is achieved by use of a light or sound to the person when the analog bioelectrical frequency recognized between the electrodes falls within a predetermined range for a predetermined duration. The step of displaying is achieved using a waveform scroller for displaying particular digital signals as a continuous wave over time and using a monitor for displaying the continuous wave over time as plotted by the waveform scroller. The step of integrating is achieved using a computer which can compute integrals.

Other aspects and advantages of the present invention

w ll become apparent from the follow ng descr pt on o e preferred embodiments , taken in conj unction with the accompanying drawings , which illustrate, by way of example, the principles of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a functional block diagram of a first preferred embodiment of the improved method of the present invention; FIG. 2 is a functional block diagram of a second preferred embodiment of the improved method of the present invention;

FIG. 3 is a top right perspective view of a first preferred embodiment of the improved apparatus of the present invention shown associated with a test subject; and FIG. 4 is a top right perspective view of a second preferred embodiment of the improved apparatus of the present invention shown associated with a test subject.

BEST MODE FOR CARRYING OUT THE INVENTION

As shown in the drawings, wherein like numerals represent like elements, an organism 5 can be monitored using a receptor means 7 as shown by the schematics of FIGS. 1-4. The receptor means 7 comprises at least two electrodes 9 of a ferrous materials shown attached to a rabbit 11 in FIG. 3 and a man 13 in FIG. 4. The electrodes 9 can be placed on any portion of the body where bioelectric signals may be of interest. As shown in FIGS. 3 and 4 the bioelectrical signals of the head 15 of the rabbit 11 and

two electrodes 9 can be sensed over time and transmitted by wires 19 to an amplifying means or amplifier 21. One amplifier used is the morendocino microcomputer EEG amplifier or equivalent. The amplifier 21 is of a kind that can amplify analog signals such as bioelectrical signals produced by living organisms. Amplification is necessary because bioelectrical signals are typically vary faint and cannot be readily analyzed by electrical means. Such amplifiers are commonly known in the art of electroencephelgraphic applications.

Once the analog bioelectrical signal is sufficiently amplified, the signal can be digitized or demodulated by an analog to digital converter means 23 so as to produce discrete digital signals which correspond to the frequencies inherent in the analog bioelectric signal. Some signal filtering may be necessary to properly process the analog bioelectric signal prior to demodulation. Demodulation of the analog bioelectric signal can be achieved by using a DASH-16 analog/digital input card 25. The input card 25 is manufactured by Metrabyte Corp. , and is a high speed multifunction analog/digital I/O expansion board for a personal computer.

The DASH-16 uses an industry standard (HI-674A) 12 bit successive approximation converter with a 12 microsecond conversion time giving a maximum throughput rate of 60KHz in D.M.A. mode. The channel input configuration is switch selectable on the board, providing a choice between 16 single ended channels or 8 differential channels with 90dB

common mo e re ec on an - v common mo e range.

Analog to digital conversions may be initiated in any one of 3 ways, by software command, by internal programmable interval timer or by direct external trigger to the analog to digital. At the end of the analog to digital (A/D) conversion, it is possible to transfer the data by any of 3 ways, by program transfer, by interrupt or by D.M.A. All operating modes are selected by a control register on the DASH-16 and are also supported by its accompanying utility software.

High input impedance ranges of +lv, +2v, +5v & +10v unipolar and +/-0.5v, +/-ϊv _r +/-2.5V, +/" ^{5v &} +/~10 ^v bipolar are switch selectable. These ranges are common to all channels and are controlled by the gain of the input instrumentation amplifier. Other ranges may be realised with a single user installed resistor. All inputs are multiplexed through a low drift, fast settling instrumentation amplifier/sample-hold combination and the channel input configuration is switch selectable to operate as either 16 single ended or 8 differential channels.

A 3 channel programmable interval timer (Intel 8254) provides trigger pulses for the A/D at any rate from 250 KHz to 1 pulse/hr. 2 channels are operated in fixed divider configuration from an internal 1MHz xtal clock (optional 10MHz jumper selectable on DASH-16F) . The third channel is uncommitted and provides a gated 16 bit binary counter that can be used for event or pulse counting, delayed triggering, and in conjunction with the other channels for frequency and period measurement.

. converters may be operated with a fixed -5v reference available from the DASH-16 board to give a 0 - +5v output. Alternatively an external D.C. or A.C. reference may be used to give different output ranges or programmable attenuator action on an A.C. signal. D/A's are double- buffered to provide instantaneous single step update.

A -5v (+/-0.05v) precision reference voltage output is derived from the A/D converter reference. Typical uses are providing a D.C. reference input for the D/A converters and offsets and bridge excitation to user supplied input circuits.

Digital I/O consists a 4 bits of TTL/DTL compatible digital output and 4 bits of digital input. Apart from being addressed as individual I/O ports, some of the digital inputs do double duty in some modes as A/0 trigger and counter gate control inputs.

The analog/digital input card 25 is incorporated within a modified personal computer 27 which can be of an IBM PC or XT type sufficient to interface with the analog/digital input card 25.

The digital signals resulting can be further analyzed with respect to certain analog bioelectric frequencies sensed. A selector means 29 allows sufficient manipulation of the digital signals so as to separate particular digital signals which correspond to the particular bioelectric analog frequencies to be monitored. This function can also be achieved by using the DASH-16 card.

e compu er compr ses a numer ca processor card 29 with an expanded memory card 30 which among other things allows the selection of particular digital signals of interest corresponding to the bioelectric frequency to be studied. Such a selection can be performed by digital separation using certain digital filters commonly known in the art of digital filtering. Many of these techniques can be implemented by use of the DASH-16 card.

Furthermore, a computing means 31 is used to integrate the amplitude of the particular digital signal over a pre¬ determined duration. The resulting value is divided by microvoltage to determine a change in voltage with respect to time duration, and thereby determine the bioelectric frequencies emitted by the person or animal. This calculation is achieved by the computer 27 driven by particular algorithms which are herein disclosed and addressed in Appendix A attached hereto. Appendix A is a list of software variables and their interrelationships. The computer 27 with a keyboard 28 then converts the resulting values divided by micro voltage to a format which can be plotted over time. Threshold values can be entered by the keyboard 28 and established such that when the amplitude of the resulting signal is within a certain amplitude range (the bioelectric signal is within a particular frequency) a signaling means signals to the person an auditory, visual, or sensible signal indicating that bioelectric frequencies are within the pre-determined ranges. An amplitude sensing means 35 is used which may be a simple algorithm preset or varied by a technician or the subject coordinating the

The signaling means 33 is simply a light box or sound box 37 that emits a sound or light or series of sounds or lights which indicate to the subject that the bioelectric frequencies received by the monitoring electrodes 9 are within a preset frequency range. Of course, the signaling means can also be any other type of stimuli that can be sensed by the subject. One approach considered is the use of a video type game displaying animations which can be controlled by way of controlling the brainwave frequencies.

Once the subject is aware that bioelectric frequencies within a certain range can be sensed, a feedback phenomenon is possible. The subject can be trained to focus mentally upon obtaining the stimuli from the signaling means 33 and thereby, alter the bioelectric frequencies produced at the location of the electrodes 9. Bioelectric frequencies of a predetermined kind can either be facilitated and produced more readily or inhibited once the subject can be alerted to whether those frequencies produced are within the preset range.

Furthermore, a suppression means 39 or artifact suppression device can be used to suppress unwanted signals that would normally trigger the signaling means 33. Unwanted signals include "noise" attributable to bioelectric activity in the subject's muscles. A bipolar hookup with an ear reference may be used incorporating baseline wire 41 connected to the subject's ear or other part of the body which provides a signal to the suppressing means 33 which indicates that a comparable signal received from the

e ec ro es s ou e suppresse . e s function of the suppressed means 39 are commonly known in the art and are not further herein discussed. Suffice it to say that extraneous bioelectric frequencies not of interest emanating from a location not around the electrodes 9 can be suppressed. As shown in FIG. 4, the requisite suppression circuity can be found within the amplifier 21 or within the computer 27.

A display means 43 or monitor 45 can be used in conjunction with a waveform scroller display card 45 and a graphics display card 46 to display the demodulated and processed data resulting from the processed bioelectric signal. The graphics display card is a Hercules monochrome card, but a number of IBM compatible cards can be used. The waveform scroller card 45 is commonly known in the art of computer graphics for displaying data points over time on a monitor.

The particular frequency of interest is displayed as a sinosoidal wave 48 which changes amplitude depending upon whether the particular frequency of interest is being inhibited or facilitated. More particularly, the waveform scroller card 47 processes the data received by the computing means 31 and displays it on the monitor 45. The actual data displayed are the values of the integration of the digital signals of interest over time divided by micro- voltage representing the particular frequencies of interest making up the bioelectric signal. Upper and lower threshold levels can be established and displayed on the display monitor 45 as horizontal lines 49. The horizontal lines 49

particular bioelectric frequencies. When the amplitude of the sinosoidal wave exceeds the limits imposed by the horizontal lines 49, the signaling means 33 or light or sound box 37 send a particular signal to the subject to indicate that a particular frequency is not being suppressed or inhibited. When the sinosoidal wave 48 is within the horizontal lines 49 the subject is sent a different signal indicating that the particular bioelectric frequency of interest is being inhibited at least within the microvolt levels defined by the horizontal lines 49. Upon practice by the subject, the horizontal lines 49 can be brought closer together representing a requirement that even less of a particular bioelectric frequency must be produced to achieve a reward response by the signaling means. Of course, the use or nonuse of lights or reward systems can be modified.

A timing means 51 comprising timing circuitry commonly known in the art can be used to time the required duration that the level of bioelectric frequency production must stay within proscribed limits before a reward signal is given. The duration can be lessened once the subject has improved his ability to inhibit the production of certain bioelectric frequencies. Variability of the duration is a feature which allows greater clinical and therapeutic customization of the invention depending on the particular subject encountered and the level of skill developed by the subject.

A recording means 53 can be used to record to a magnetic medium or hard disk 55, or floppy disk drive 57 to record on a floppy disk (not shown) . If a hard disk 55 is

use , a ar s con ro er car s necessary a commonly known in the art of magnetic recording medium. The therapy session results displayed on the display monitor 45 can be stored in a format to be displayed and compared with past sessions or to be compared with future sessions. A large variety of software tools commonly known in the art of statistical evaluation to compare the subject's progress in inhibiting or facilitating particular bioelectric frequencies can be used. Most helpful is compressing the waveform produced over the entire session to a format which can be viewed in its entirety on a screen or page of paper. Such analysis is helpful to see the subject's ability to inhibited or facilitate a particular frequency over the period of the entire therapy session. This information is not only helpful to the clinician, but also to the subject since the cognitive effect on the subject may enhance his or her ability to further facilitate or inhibit certain bioelectric frequencies.

Although feedback is primarily provided to the subject by the signaling means 33, the display means can be used by the subject to monitor his own progress in facilitating or inhibiting certain bioelectric frequencies.

Also, as shown in FIGS. 1 and 3, the invention can further be used without a feedback or signaling means 33 simply to accurately and easily monitor certain selected bioelectric frequencies from an organism 5 or rabbit 11 as shown. It should be indicated also, that a number of different EEG type channels can be feed into the invention for purposes of monitoring various areas provided the proper

the art of EEG monitoring and recording. A large number of cables, plugs, and jacks necessary for proper operation of the invention are not herein described as those accessory items are commonly known in the art of data processing with a microcomputer.

The software to be used operates with respect to the parameters as described below and as embodied in Appendix A, as well as with the hardware described. Those skilled in the art will find the following considerations helpful.

Data collection by the software allows wrap-around in its storage, so there can be no meaning assigned to > or < in comparison of the pointers. Only == or != are useful comparisons. Data collection proceeds continuously whenever the system is enabled. The pointer "ad_ptr" always points to the next storage location to be used. This pointer is incremented by the interrupt handler, and is never changed by others except at the start of data collection, when it is set to the start of raw data memory.

Data filtering also proceeds continuously under most circumstances. The pointer filtered always points to the next datum to be applied to the filters. It is incremented by the collar of the filter programs, and is never changed by others except at the start of data collection or the start of data review.

Plotting may be started, stopped, reversed, restarted, etc. Two pointers are involved:

(1) "plotted", which points to the next datum to be

p o e ; an

(2) "plotptr", which is updated by interrupt service to indicate that it is time to plot another point when plotting at normal speed. These pointers are incremented or decremented according to plotting direction: "int runn" is +1 for forward plotting, -1 for reverse, and 0 for for stopped. Thus adding "runn" to the pointer changes it appropriately.

There are a number of points to considering in constructing the program: setting up a patient file, connection to the patient, starting the program, and running the program.

Setting UP a Patient File

An example of a patient file is shown below. New files may be created and edited with a test editor which can incorporated with the system. The example file contains notes about the file contents. When a new patient file is being set up these notes can be erased from the new file.

Example

c:/adinp/adcoef 20 100 Raw data trace scale and artifact suppression level 20 30 Middle trace scale and threshold 20 20 Bottom trace scale and threshold

The patient's name must appear as the first line of this file. adcoef names a file where filter coefficients are to be found. It MUST appear as the second line of this file. scale and artifact suppression level must be in line 3. scale and threshold for the middle trace must be in line 4. scale and threshold for the bottom trace must be in line 5.

Any other notes may be placed later in this file, and the file can then be read by the DOS command:

type patient

Notes can be added by use of any word processor that produces clean ASCII files, but not such as Wordstar.

Starting the Program

If the computer is off, open the floppy disk drive 57 and turn the power on. The computer 27 will go through an elaborate self-testing program; then it will look to drive A: to see if it can read a disk there. If the door is open, it cannot read from drive A: so it will read the hard disk, drive C: . It will go through some more testing and loading, and finally it will show the DOS "prompt" on the screen. C>

Insert a patient's existing disk, or create a new one,

n r ve A: an c ose e oor. Type EEG a e ey oar and follow it with the "Enter" key.

The program can be loaded from the hard disk 55, and a file called "patient" on the floppy disk can be read. If that file is not found, the program will quit with a message stating: "Cannot find patient file."

If the necessary file is there, a display will indicate the same. The display shows the name of the patient file on line 12. This normally comes from disk drive A, so it will show: "a:patient opened.", but the system can also be set up to take the file from drive B or C, or a subdirectory of C.

The name of the patient appears on line 14 of the display, and below line 14 is scaling information and the name, size and date of the last file saved on the disk. Then, the display shows the name of the new file that will be created if you save data resulting from current session, and indicates how much more data can be placed on the disk. The program then asks whether the information shown is correct. A separate disk for each patient, and often several disks for one patient is advisable.

If you answer yes by typing "Y", the computer 27 will go on to load more information and display a menu. If you answer "N" (or any other key) the program will allow you to change the disk or switch to a different drive or to quit.

Running the Program

The numbers representing "raw input", "low pass

and "bot", have multiple scaling factors applied either inherently or by intention. Below are these relationships. Inputs in can be defined as follows: "raw" = completely raw input to the amplifier, measured in microvolts.

"sig" = signal value after amplification, analog/digital conversion, conversion to 16 bit twos complement notation and low pass filtering. "Sig" is in arbitrary units, and sig' numbers will generally be larger than the corresponding number of microvolts . Sig is stored in memory and is applied to the 4-7 Hz and to the 12-15 Hz filters, which are referred to here as the middle and bottom filters. Outputs can be defined as follows:

"top" = deflection of the top trace in response to "sig". "mid" = deflection of the middle trace in response to the output of the middle filter when a steady state in-band signal of amplitude sign is applied. "bot" = deflection of the bottom trace in response to the output of the bottom filter when a steady state in-band signal of amplitude sign is applied.

Scaling is important and described below. Because the middle and bottom filters may be changed and scaling may be changed, and the new filters and scales may be applied in review of existing data, it is convenient to perform scaling as part of the filtering process. Therefore, "mid" and "bot" are generated directly from the filters, ready for display, so we can also define:

m = ou pu o e m e er w s g as an npu .

"bot" = output of the bottom filter with "sig" as an input.

It is undesirable to alter the values of the original data by scaling. Therefore, the output of the low pass filter, "sig", is separately scaled immediately before plotting, converting it from microvolts to a trace deflection measured in pixels.

Scaling factors must be properly set. Scaling occurs, in part as a side effect of other functions, such as signal amplification, analog to digital conversion and filtering, and in part because the system operator chooses to set a scale for convenience in viewing.

The detailed system design is concerned with trace deflection measured in pixels, but the system operator is more comfortable thinking about a scale in terms of microvolts per centimeter, or some other grid unit. Therefore, a relationship is established as follows: "grid" - the number of pixels corresponding to the scale chosen by the user. A likely value is 20, so that if the user selects a scale of 50 microvolts, then a signal deflection of 50 microvolts will generate a plot deflection of 20 pixels.

The computer 27 provides for input of three scaling values from the patient file or by manual operation during a run. (The operator presses the ^% F4' key and then enters three numbers.) These values are identified as: "iscale[top]", "iscale[mid]", "iscale[bot] ⁿ which are signal levels in microvolts representing a deflection of ^v grid' pixels on the display monitor 45.

, generates a smaller picture. The relationship is that a trace on the monitor 45 for example of two grid units in height represents a signal of 2 "iscale" microvolts. Other arbitrary scaling factors can be used. In order to achieve the desired scaling on the display monitor 45, in the face of other scaling factors that are not fully known or understood, a set of factors (one each for "top", "mid" and "bot") is chosen experimentally that will achieve the desired result. These values are related to the amplifier 21, the selector means 29 and possibly some other factors, but all are independent of the operator and the patient, so they are stored in a filter coefficient file and are not subject to change by the operator. They relate the arbitrary ^x sig' units stored in memory to true microvolt measurements, and are called (in the computer program) "fullscal[trac]". Therefore:

microvolts = sig / fullseal[trace]

If the scale chosen by the operator is equal to the grid size, then one pixel will correspond to one microvolt. If the operator chooses a different scaling, it must be applied as well. sig grid top = * fullscalftop] iscale[top] Several factors for the top trace are combined into a signal factor. For convenience this factor is stored in the

coe icien array. coeff[top][16] = grid / ( iscale[top] * fullscal[top] )

so that the multiplication in the plotting program is:

top = sig * coeff[top] [16] with "sig" in arbitrary units and "top" in pixels

The "fullscal[trace]" values for the middle and bottom traces contain the same relationship of *sig' units to microvolts and also compensate for gain of the filters. As above, the grid size and the operator selected scaling are combined with "fullscal[trace]" into a single factor. For convenience this is stored in the filter coefficient array as:

coeff[trace] [16] = grid / ( iscale[trace] * fullscaltrace]

These factors are applied to the datum (in sig' units) at the filter inputs so that the filter output is directly in pixels.

Thresholds must be established for the system. Each of the traces has an associated threshold value which leads to a decision: clamping of the filters in the case of the top trace, reward for the middle trace and inhibition of the reward for the bottom trace. Therefore, "thresh[top]",

"thresh[mid]", "thresh[bot]" and decision values in microvolts, i.e., input from patient file

.

For the top trace, the actual decision is based on the signal level stored in the memory in the arbitrary sig' units. Therefore, the top trace's microvolt threshold is converted to these units by:

sigthres - threshftop] * fullscal[top]

For the other traces the decision is based on the filter outputs, which are already in pixels, so the microvolt thresholds are converted to pixels by:

m.thres = thresh[mid] * grid / iscale[mid] b.thres = threshfbot] * grid / iscalefbot]

A similar conversion is made on the top trace threshold for the purpose of graphic display of the threshold.

topthres = threshftop] * grid / iscale[top]

These calculations of thresholds and scaling factors are made in the function ^λ chscale() ' which are part of the module thrhold.c' .

Reward logic of the program is critical to proper operation. The digital filters in "philtre.asm" generate two values: "bottom.led" and "middle.led", representing the integrated amplitude of the filtered signal divided by the threshold set for that signal. (Thus a number > 1.0 means that the signal is greater than the threshold.) Indicator

l ghts of the s gnal ng means 33 and a counter (not shown) depend on these two values. A counter can be used to keep track of the number of times the subject has been rewarded over the test period. Three indicator lights 38 may be provided in addition to the counter: they cure referred to as red, yellow and green. Red is the inhibit indicator controlled by the undesired frequency through the 5 Hz filter and the variable "middle.led". Yellow is the enhance indicator controlled by the desired frequency. The green light signals that a reward is being posted to the counter.

The red light is turned on if "middle,led" is greater than 1.0, indicating that the 5 Hz signal exceeds the threshold. Actually, "middle.led" must become greater than 1.0 by some fixed increment to turn the red light on, and less than 1.0 by the same amount to turn the indicator off. At exactly 1.0 or very close to that value the indicator does not change. This will reduce flickering of the indicator. The yellow light is continuously variable. Its brilliance is proportional to:

15 Hz Integral - Threshold

Threshold

which is equal to "bottom,led" = -1.0.

inhibit signal. Both lights, however, will go off if a muscle or eye blink artifact is detected, because the inputs to the filters are clamped for one second upon detection of the artifact as part of the suppression means 39.

The green light signals a reward. When the yellow light is on continuously for 0.5 second during which the red light is off, then the green light is turned on. An audible signal is given, and the counter is incremented. The green light stays on for 0.5 second and is then, off until the next reward. The program allows time setting an interval after one reward before the next one can start to be earned, so that the frequency of rewards can be limited if desired.

Obviously, the approach described herein can be greatly modified by the clinician or the subject

It should be appreciated for the foregoing description that the present invention provides an improved, more simplistic, more accurate, and less time intensive monitoring and/or control of particular bioelectric frequencies of organisms. The apparatus of the invention can be assembled from components readily available and easily assembled. Furthermore, diagnostic evaluation and patient use is greatly improved. Furthermore, the use of power spectral analysis and analog active band pass signal filtering can be eliminated.

Although the present invention has been described in detail with reference only to the presently-preferred embodiments, it will be appreciated by those of ordinary skill in the art that various modifications can be made

w ou epar ng rom e nven on. ccor ng y, e invention is limited only by the following claims.

APPENDIX A

Page 1

File DASH 16. ASM

•ov c):.Cbp+fa3 (interval in m croseconds .D

CALL 1

PRETIMER adinpx: pop es pop ds pop d pop si pop bp

RET

PROC

ENAB. I

•ov dx.ADSTAT (Mriting anything to this address out dx.a I clears and reenables the IRQ 3

•ov al,OBSh (Enable interrupts using lOll 010 out 21h,al (IRQ fc-diβkette IRQ 3-AD IRQ 1-

RET

HOLD LI

PROC H0LD_L1

3

Page 2

F le DASHltr.ASM

out 21h,al βti

Page

F le DASH16.ASM

shl bx.l (Make word address of t mov a ,ex | | | xchg a.:,tstartCbx_l (prev ous measurement sub ax,ex (time difference in clo sub dx.dx (multiply by 1/1.19316

MOV ex,54925 (to get time in m crose mul ex

MOV timdifCb>:3,dx

MOV ax.dx

READ2 pop bp

RET

SETLAB

PROC SETLAB

(Prepare for A/D conversion in selected channel

(Set MUX control register and bx,7 (A/D channel number (0-7 on

MOV a>: , b>: mov cl.4 shl ax,cl ^' (high channel into high nibb or ax.bx (same channel into low nibb mov d>:,MXHL out dx,al

(Set DASH16 control register

{ 1NTE Interrupt Level X DMA Start on Time

I 1 o 0 1 1

MOV al,0B3h mov dκ.ADCNTL out dx.al sub ax,ax (set DA outputs to 0 MOV dx,DA0L0 out dx,al inc dx (DAOHI out dx.al inc dx (DA1L0 out dx,al inc dx (DA1HI out dx.al

RET

PRETIMER

PROC PRETIMER

CALL

-

Page 4

File DASH16.ASM

27

Page

File DASH 14. ASM

REWARD LOGIC \ if .ledreg M» GREEN) /♦ if already giving re < beepcnt— ^■ ( I if I beepcnt-«0 >

< beeper off ( ledreg«ϋ ( >. > if < emledpt > ' /♦ i* bad integral > th { rewcnt * rewtim ( /* restart counting tow ledreg I' RED ( els >e ledreg no re Id } if ( retry 1 /* to limit frequency o

{ retry— ( J rewcnt — rewtim ( outp >ut pulswide to Itimer O ( /• yellow if < 'bledpt > /» f good integral > t

{ rewcnt— ( | if < rewcnt ■■ 0 )

< ledreg 1- GREEN ( beeper on ( beepcnt _* beepti ( rewcnt — rewtim { retry - delaytry ( } > reward: mov bx,offset ledreg

AND

Jump if V,

dashe.:. mov dx. DSTAT icH ear the source of the int out dx,al ( any data written to ADSTAT mov al,20h (clear the interrupt out 20h,al. I t mov al,byte ptr imasl: (reenable IRQ 3

FAKER

push dx mov ax,1phase add ax.1delta

FA ER2

FAKER2 shr *x,cl (make whole part into word a and ax.OFFFEh

**0

Page 1

DASHINT File

PROC DASHINT push ax push bx push ex push dx push si push di push bp push ds push es eld

(DEBUGGING in al,21h

MOV ah,al or al,B (disable IRQ 3 out 21h.al mov bx,seg dgroup mov ds,bx mov byte ptr i ask,ah mov dx,ADSTAT in al.dx (read status

DASHINTO

DASHI T6 DASHINTO

DASHINT1 then mark DASHINT1 f status

Ψ

Page 2

DASHINT File

Jump if

DASH I NT __ zero

CMP- FAKE and O

Jump if

DASHINT2 zero

CALL

FAKE*

Jump

DASHI T3 Unconditional DASHI T2 mov dx.ADLO (base address of DASH1 and in al,dx (read converted data mov bl,al mov dx.ADHI in al,dx mov bh,al mov ax.bx

(CONVERT RAW DATUM Ptύrt 12 BIT BIASED ZERO TO TW add bκ,ΘOOOh (make raw datum two's comple mov cl,4 HI sar bx.cl (right juβtify and sign exte

(arrive at _.-_sh.nt3 with ax » raw analog value.

DASHINT3

(COPY I f jT TO OUTPUT mov dx.DAlLO out dx,al mov dx.DAlHI mov al , ah out dx.al

(STORE TWO'S COMPLEMέNT VALUE, marled as ew dat xor bx.BOOOh (mark high bits 01 or 10 mov es,ad_seg___:θ 03-l mov byte ptr es : Cd ( tore low byte in 64k mo ,1! bl v es,ad_segsC23 || mov byte ptr es:Cdi3,bh (store high byte in 64k (FINISHED WITH NEW DATA

(mark buffer full

Z.

Page 3

DASHINT File

DASHINT6

DASHINT6

DASHINT7

DASHI T7

Page 4

DASHINT F le

REWAR 1

REWAR 1

(green and red both off (beeper off

REWARD1

(pointer to middle led

REWARD:

(bad integral > thresho (time required for a re

REWARDS REWARD2 mov di,offset retry (no reward during retry inte

CMP

CDI3 and

0

Jump if

REWARD3 zero

(at present not p πlanning to limit retry rate, bu (allows a delay to be entered as delaytry dec Cdi

REWARDS mov bx, pulswide Ji (:calculated by pulstinte mov dx,DAlLO out dx.al mov al , ah mov ax.bx

^■ m

Page 5

DASHINT File

REWARD3B

REWARD3A allowed pulse if tn

REWARD3A _ι mov dx.CNTO (addresε timer to drive ye out d ,al mov al,ah out dx,al

REWARD3B mov sii.bledpt (pointer to bottom led from

CMP

CSI3 and

O

Jump if EWARD: zero

DEC REWCNT

Jump i

REWARDS not zero

CMP

CDI3 and

O

Jump if

REWARD4 not zero or ledreg.SREENCN nT (give reward with green led in al,61h (read PPI or al,3 (beeper on out 61h,al I)) mov ax,beeptιm (time to beep mov beepcnt, ax mov ax, delaytry mov Cdll.ax ^' (delaytry -> retry

H≤

Page 6

DASHINT File

REWARD4 mov ax,rewtim (time to try again mov rewcnt,ax EWARD: mov ax,ledreg (low byte 2'green, 4»count xor ah.REDSREEN (OAh means display now, 00 1 mov ledreg.ax (so led's are pulsed on alte

(interrupts, rather than on full time or ah, COUNT (but don't flicker the count and al , ah mov dx,DIGIO out dx.al I

C

Page 7

DASHINT File mov ax,sιnetb>:3 cwd mov cx,fakamp (scale up amplitude imul ex mov bx,dx (bx * high word of the produ mov a , dx mov ex ,4 shl ax.cl add ax , B OOh (ax — left justified, bias z pop dx

mov es.ax mov si.BOTLI add si.m_.rknt mov ax.marker

,al

pop es

RET

INT

Page 6

DASHINT File OBH

RET

RESETCNT

PROC RESETCNT mov al.RESET mov dx.DIGIO out dx,al sub ax, -M mov score,ax mov word ptr longtime,ax mov word ptr longtιme£23,a>:

RET