JPH04502876A - Multi-peak audio processor - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] Multi-peak audio processor Technical field The present invention is a pulsatile multichannel cochlear implant system for people with complete or profound hearing loss. Regarding.

Background of the invention Pulsating type multi-channel pot bovine inplant system is commonly used in pot bovine inplant, external audio pro processor, and an external headset. The cochlear endoplant is placed inside the pot. An electrical stimulation pulse is emitted to an electrode array (for example, 22 electrodes). audio process The sensor and headset transmit information and power to the Bogyu endophyte.

Is the audio processor from the microphone in the headset or an alternative source? receives an incoming acoustic signal from a computer and extracts specific acoustic parameters from this signal. put out These acoustic parameters are used to determine electrical stimulation parameters They are the transmitting coil in the headset and the receiving coil forming part of the implant. encoded and transmitted via.

Many people with profound hearing loss believe that the reason for their hearing loss is the conversion of acoustic signals into nerve impulses. Absence or damage of hair cells in pot beef. These people are therefore exposed to acoustic stimulation It is impossible to generate nerve stimulation from sound in a normal manner no matter how large the Therefore, they cannot benefit from regular hearing aids. The pot cow inplant system is Provides electrical stimulation directly to acoustic nerve fibers, leading to sound perception in the brain.

Electrical power inside the pot connected to the outside world via a cable and a connector attached to the patient's skull. From the polar inplant to the external computer via radio frequency power and data link To achieve this objective, A large number of methods have been described in the past.

The invention described herein provides a method for using externally powered wearable audio processors. Inside the pot connected to a multi-channel implanted stimulator unit that receives power and data. A mental object that contains multi-channel electrodes implanted within it and has a known sound processing strategy. Based on physical phenomena and personalized through the use of diagnostic and programming units. It is particularly suitable for use in prosthetic devices tailored for each patient. the An example of such a prosthesis is the G.I. Also illustrated and described in US Pat. No. 4,532,930 to Rosby et al. It is.

To best understand the invention, some of the physiology and anatomy of the human hearing organ will be explained. It is necessary to know the characteristics of the audio signal and have knowledge of the characteristics of the audio signal. In addition, electricity Auditory sensations induced by emotional stimuli occur in people with normal hearing due to acoustic stimuli. It is necessary to discuss the psychophysics of electrical stimulation of the auditory system. It is essential. In people with normal hearing, the sound impinges on the eardrum as shown in Figure 1, and then amplification and acoustic input to a piston or membrane called a non-round window connected to the pot chamber. through bones called ossicles that act as levers to provide impedance matching. communicated.

The pot chamber is approximately 35mm long extended, and along most of its length divided by partitions. This partition is called the basement membrane. The lower chamber is the tympanic cavity called floors. An opening at the far end of the pot chamber communicates between its upper and lower halves. Cheating. The pot is filled with a liquid that is about twice as viscous as water. The scala tympani is responsible for the fluid. with another piston or membrane called a round window which acts to pick up the displacement. ing.

When the non-round window is driven acoustically through the auditory ossicles, the basilar membrane is displaced by Due to its mechanical nature, the basement membrane is located at the distal or apical end of the whorl. It vibrates at a maximum at a low frequency at a point, and at its base or near the oval window it vibrates at a high frequency. It vibrates. The displacement of the basement membrane is caused by the formation of hair cells, which are located in special structures on the basement membrane. stimulate a population of cells. The movement of these hair cells is caused by the second nerve or auditory nerve. Generates an electrical discharge in the fibers. In this way, the capillaries closest to the round window (bottom edge of the m cow) Nerve fibers from the vesicle carry information about high-frequency sounds, and fibers closer to the apex The nerve fibers carry information about low frequency sounds, which are the dot pick machines of the nerve fibers in the pot. It is called structure.

There are many causes of hearing loss, and there are generally two types. Conductive hearing loss is all over the place The normal mechanical path of sound to the hair cells of the ear is obstructed, for example by damage to the auditory ossicles. Occurs when Conductive hearing loss amplifies the sound so that the acoustic signal reaches all the way into the pot. Often helped by the use of hearing aids. Some types of conductive hearing loss are It can be alleviated through medical treatment.

Sensorineural hearing loss results from damage to the hair cells or nerve fibers within the cauldron. This type Nine hearing aids will not bring about any improvement in a patient with I.P. sound energy god This is because the mechanism that converts it into transimpulses is damaged. This loss of functionality is partially This is because the auditory nerve can be directly stimulated.

The system described here, and some other pot inplant systems in the prior art. In this method, stimulating electrodes are surgically placed into the scala tympani, close to the basilar membrane, and a The electrical current that passes through it causes nerve stimulation in the nerve fibers.

The human vocal system consists of numerous resonant cavities, namely the oral cavity and the nasal cavity. is excited by air passing through the glottis or vocal cords, causing them to vibrate. vibration The numbers are heard as the pitch of the speaker's voice and range between approximately 100 and 400 Hz. Change. The pitch of female speakers is generally higher than that of male speakers.

Listeners can, for example, distinguish between explanations and questions, separate sentences in continuous discourse, and intonation of a sentence, allowing you to detect which parts are particularly emphasized. It is the pitch of this human vocalization that gives the effect. What is the difference between this and signal amplification? Provides loose prosody information.

Speech is the movement of the speaker who excites the vocal cords and uses the tongue, lips and jaw to produce different sounds. created by the manipulation of an acoustic cavity by motion. Some sounds reach the excited vocal cords These are called voiced sounds. Other sounds are teeth that produce voiceless sounds. and by other means, such as the passage of air between the tongue and the tongue. “Z” like this Although the sounds are voiced, the sound "S" is unvoiced, and the sound "B" is voiced. "P" is a voiceless sound, etc.

Audio signals can be analyzed in several ways. One useful analysis technique is Spectral analysis, whereby the audio signal is analyzed into the frequency domain, and The spectrum can be thought of as amplitude (and relative) versus frequency when the cavity of the zero-voice system is excited. When a large number of spectral peaks occur, and these spectral peaks Frequency and relative amplitude change over time.

The number of spectral peaks ranges between about 3 and 5, and the formants and Called. These formants range from the lowest frequency formation, usually called F1, to the highest frequency formation. High frequency formations are numbered and the voice pitch is usually referred to as FO.

The characteristic sound of various vowels is that the vocalization changes the frequency and relative strength of these formants. created by changing the shape of the oral cavity and nasal cavity, which has a changing effect Ru.

Specifically, the second formant (F2) is important for conveying vowel information. It turned out that. For example, the vowels “oO” and “ee” are produced with the same vocalization of the vocal cords. Yes, but it will sound different because of the different second formant characteristics.

Of course, there are various different sounds in speech, and the way they are generated is complex.

However, for purposes of understanding the present invention here, there are two main types of pronunciation: voiced. and unvoiced, and that the time course of formant frequency and amplitude is It is enough to remember that you have most of the understanding of the signals.

The term "psychophysics" as used here refers to the use of electrical stimulation of the auditory nerve in patients. Refers to the study of evoked perception. at speeds between 100 and 400 pulses per second. For a stimulus, a noise that changes with the stimulus speed is perceived. this is that It is a distinguishable sensation that, by variation, can convey a melody to the patient.

By stimulating the electrodes at a rate proportional to speech pitch (FO), prosodic information is transmitted to the patient. It is possible to convey to This idea will be the only way to convey information. It has been used in some pot inplant systems and can be implemented with a single electrode.

Formant information contains most of the intelligibility of a speech signal, so it is important to It is even more important to communicate this to others. Psychophysical experiments have shown that far An auditory signal stimulating one end produces a low-frequency sensation, and a signal stimulating a nearby end produces a low-frequency sensation. A similar phenomenon can be seen with electrode stimulation, which produces high-frequency sensations. has been discovered. induced by electrical stimulation at different locations inside the pot. Perception is not based on pitch as it is, but rather on knowledge that changes in “clarity” or “unintelligibility.” Reported by patients as causing sensations. However, the frequency between the electrodes Differences in number perception can be determined by changing the formants or spectral peaks by electrode selection or panning. It is possible to code by selecting the stimulation site in the middle.

Psychophysical experiments show that the perceived It turns out that the loudness of the sound has a larger dynamic range than the dynamic range of the stimulus itself. It was. For example, an electrical stimulus with a dynamic range of 220 dB has a dark value or is barely perceptible. Perceptions can range from none to pain darkness. In people with normal hearing, sound perception The dynamic range of is on the order of 100 dB.

Additionally, psychophysical experiments have shown that the pitch of sound perception due to electrical stimulation depends on the frequency of the stimulation. Depending on the number, however, the perceived pitch may not be the same as the stimulus frequency. It's been found. In particular, the maximum pin that can be perceived by varying stimulation rates alone is on the order of l kHz, and stimulation at rates above this highest level does not result in an increase in the perceived frequency or pitch of the sound. In addition, inside the pot For electrical stimulation, the perceived pitch depends on electrode position. multi-electrode system , the perception due to stimulation at one electrode is the same as the perception due to simultaneous stimulation of adjacent electrodes. It's not unrelated. Also, the perceived quality of pitch, i.e. “clarity” and loudness. cannot vary independently of stimulation rate, electrode position, and stimulation amplitude.

Some prior art pot inplant systems are proportional to the energy of a specific frequency band. It is designed to stimulate many electrodes at the same time, but this is because the stimulation electrodes in close proximity This is done without considering the perception caused by the stimulation current. As a result, the channel There is an interaction between and the magnitude is influenced by this.

Until now, auditory nerve fibers have been studied using electrodes partially or adjacent to the structure in the pot. Numerous attempts have been made to provide useful hearing through electrical stimulation of the fibers. single pair A system using electrodes is disclosed in Michelson U.S. Pat. No. 3,751,60. No. 5 and Bartz, US Pat. No. 3,752,939.

In these systems, an external audio processing unit converts acoustic input through the skin. to a signal suitable for transmission to the implanted receiver/stimulator unit. Ru. These devices apply continuously varying stimuli to a pair of electrodes that stimulate a population of auditory nerve fibers. stimulates at least a portion of the body, and produces audible anger and relaxation.

The stimulus signals generated from a given acoustic input are Although each has shown some degree of effectiveness, performance varies by system. The design of these systems is experimental. The cause of this variation has not been developed and is not based on detailed psychophysical observations. could not be determined. Therefore, it was not possible to reduce the fluctuation.

Another approach is to stimulate groups of nerve fibers according to the frequency spectrum of the acoustic signal. The method is to use the dot pick tissue in the hot pot.

A system using this technology is US Pat. No. 4.207.441 to Ricard. No. 3,449,753 of Doyle, No. 4.06 of K15s iah No. 3.048, Hochmair et al. No. 4,284,856, No. 4,357 ．． No. 497.

The system described by K15siah converts acoustic signals into individual auditory spectra. analyzer to separate into a number of frequency components with predetermined frequency ranges within Use a set of log filters. These analog signals are the analog signals they represent. It is converted into a digital pulse signal with a pulse rate equal to the frequency of the log signal, and the corresponding Digital signals stimulate the part of the auditory nerve that normally carries information in the same frequency range used for. Stimulation is achieved by placing electrodes at intervals within the pot. .

The K15siah system works at the limits of the normal hearing frequency range, i.e. up to 10kHz. It utilizes electrical stimulation at a rate of 0.05 and independent actuation of each electrode. which god The maximum number of excitations of the warp fibers is also limited to 1 or 2 kHz by physiological mechanisms. And since there is only a small perceptual difference at electrical pulse rates of 8007 or higher, the instructions experience suggests that it would be inappropriate to stimulate at a given rate. different sound sources that produce considerable uncontrolled loudness fluctuations depending on It does not take into account the interaction between the stimulation electrodes caused by the electrodes used. Also, this The stem contains a percutaneous connector with an associated risk of infection.

The system proposed by Doyle allows the number of stimulations to be arbitrary for any fiber group. fibers to respond to sequential stimulation. It is multiple Utilizing transfer channels, each channel provides simple combined power/de-energization to a bipolar pair of electrodes. send a data signal. Time multiplexing similar to that described below that is later used by Rtcard. In the fashion voltage source stimulation is used and proposed adjacent pairs of electrodes An independent stimulus of would produce similar uncontrolled loudness fluctuations. difference In addition, the need for a number of transmission links equal to the number of electrode pairs reduces this requirement for more than a few electrodes. Prohibit the use of type systems.

The system proposed by Ricard uses filters to analyze acoustic signals. bank and provides time-multiplexed output to the electrode set implanted in the pot. A single wireless link to send both power and data to the implanted receiver/stimulator Use links. Single-phase voltage stimulation is used, with - electrodes connected to the voltage source at a time. and the rest are connected to a common ground line. A small slit between the electrodes within the scala tympani. Attempts were made to isolate the stimulation currents from each other by placing recone strips.

Monophasic voltage stimulation was used, and the electrodes returned to a common reference level after delivering each stimulation. Therefore, the electrical storage property of the electrode/electrolyte interface remains for several hundred cycles after the driving voltage returns to zero. A certain amount of current will flow in microseconds. This reduces the net potential transfer. (and hence electrode corrosion), but this potential return phase now coincides with the next stimulus. overlap in time. Any spatial overlap of these stimuli does not control loudness. fluctuations will occur.

In the Hochmair et al. patent, multiple carrier signals are in the auditory frequency band. modulated by a pulse corresponding to the signal. The carrier signal is likely to receive the transmitted signal. and is transmitted to a receiver with an independent channel for demodulation. detected pulses are selectively delivered into the pan to stimulate areas with the desired frequency response. is applied to an electrode on the inplant in the pot, which is an electrode placed on the pot. The pulse is the auditory zone the frequency corresponding to the signal frequency within the auditory band, and the amplitude corresponding to the signal amplitude within the auditory band. It has a ruth width.

No. 4,267,410 to Forster et al. Utilizes biphasic current stimulation in time and offers good instantaneous control of both stimulation and return phases. Describe the system provided. However, fixed pulse durations may Forbid changes in this parameter necessitated by physical variations. Furthermore, this The data transmission system described in the system provides constant pulse rate stimulation. Severely limits the number of available pulse rates.

U.S. Pat. No. 4,593,696 to Hochmair et al. an analog signal is applied to an electrode implanted in the patient, and at least one pulse is applied to the electrode implanted in the patient; A system is described in which a signal is applied to an implanted electrode. Analog signals are voice signals. and the pulse signal has different frequencies such as formant frequency and pinch frequency. Provide specific audio characteristics.

Patrick et al., US Pat. No. 4,515,158, discloses that a set of currents is implanted. A system for applying voltage to selected electrodes in an array of electrodes is described. incoming sound the voice signal is processed to generate an electrical input corresponding to the received voice signal; An electrical signal representative of the acoustic characteristics of the audio signal is then generated from the input signal. Professional grammable means acquire and store data from the electrical signal and to the electrode array. Establishing the set of electrical stimuli to be applied and the generation of the audio signal for voiced pronunciation. The rate derived from the vocal frequency and the rate that is independent of the vocal frequency for unvoiced pronunciation. A command signal is created to control the sequential application of pulse stimulation to the electrodes. It will be done.

The prior art, over which the present invention provides an improvement, is perhaps best understood by Crosby et al., supra. U.S. Patent No. 4,532,930 entitled ``Pot Inplant System for Sensory Prosthesis'' Best exemplified by No.

The subject matter of the Crosby et al. patent is incorporated herein by reference.

The Crosby et al. patent includes multiple platinum ring electrodes in a silicone carrier! pole row We describe a pot bovine implantation system in which a pot bovine is implanted in the ear pot bovine.

The electrode array is a multichannel receiver/stimulator containing semiconductor integrated circuits and other components. unit 7), which is implanted in the patient adjacent to the ear. receiver /The stimulation unit is an inductive link with an external audio processor that can be worn by the patient. receives data information and power through the tuned coil. audio pro sensor - releases data from Erasable Bucogrammable Read Only Memory (EFROM) It includes integrated circuits and various components arranged or mapped to E PROM was determined by testing the patient and his implanted stimulator/poles. Programmed to suit each patient's electrical stimulation perception. This test diagnostic and programming interface unit connected to the audio processor. It is implemented using a processing unit (DPU).

Crosby et al.'s system includes voiced, unvoiced homophones, and efficiency information. various audio programs, including amplitude compression of the dominant spectral peak and audio pitch. Allows the use of processing strategies. The voice processing strategy employed is based on known elaborations. Based on divine physical phenomena and through the use of diagnostic and programming units is tailored to each individual patient. Biphasic pulses in various operating modes The various combinations of electrodes are supplied by switch-controlled current sinks. day The transmission of data depends on the chosen electrode, electrode mode shape, stimulation current, and amplitude burst period. by a series of separate data verses representing the amplitude determined by the interval.

Each patient will have a different perception resulting from the electrical stimulation in the pot. In particular, The stimulus intensity required to induce an auditory perception of the same loudness varies from patient to patient and from one patient to another. may differ from electrode to electrode. The patient also knows the pitch changes from electrode to electrode. There are differences in the ability to understand.

Speech processors adapt to differences in psychophysical perception between patients and Compensate for differences between electrodes. Audio pro- gramming takes into account the individual patient's psychophysical response. The sensor determines the stimulus level, electrode frequency boundaries, and other parameters that evoke the appropriate auditory perception. encode acoustic information about the parameters. Determine such stimulus parameters from the acoustic signal. The psychophysical information used to determine the speech process is called a map. stored in RAM inside the server. Audiologists can diagnose and program systems A map is created and fine-tuned for each patient using a system (DPS). DPS is Administer appropriate tests, provide controlled stimulation, and review and review test results. used to record and record information.

This multi-electrode pot bovine prosthesis has been used successfully by patients with significant hearing loss for many years. and is part of the daily life of many people in various countries around the world. . The implanted portion of this prosthesis is designed to reduce the overall thickness of the device and Designs such as those made to include implanted magnets to eliminate the need for dosing Relatively unchanged except for the in-line changes.

External audio processors have undergone significant changes from those in the early stages of prosthetics . The vocal coding modalities used by early patients can be divided into three types into implant users. Provided acoustic features. These are the amplitudes provided as the current level of electrical stimulation and the fundamental frequency or audio pitch provided as the rate of the pulsatile stimulation, and the This is the second formant frequency provided by the position of the extreme electrode pair. this code For adults with significant late-onset speech hearing loss, the mode of speech hearing loss (FOF2) has a substantial effect on their speech perception. Provide sufficient information to demonstrate improvement.

This initial encoding style is similar to later encoding styles when additional spectral information is provided. There was a natural progression to the ceremony. In this style, the voice represents the first formant. A second stimulating electrode pair was added.

This new modality (FOFIF2) addresses all areas of speech perception for adult patients. showed improved properties.

The success of audio processors using the FOFIF2 style over the past few years also However, a large number of problems were identified.0 For example, people behave better in quiet conditions. Patients may have significant problems when moderate levels of background grout noise are present. Ru. Also, FOFIF2 codes frequencies up to about 3,500Hz, but only Many phonemes and environmental sounds have a high proportion of their energy above this range. , making them inaudible to the implant user in some cases.

According to one aspect of the invention, an incoming audio signal is processed by a speech feature extractor and a speech feature extractor. and their passbands are different from each other, and at least one of them is the first pass band of the audio signal. 2-formant or multiple bands at frequencies higher than the normal range of frequency peaks. Improved pulsation for in-pot prostheses simultaneously provided to the end-pass filter system is provided. The energy within these passbands is at the base of the electrode array. control the amplitude of the electrical stimulation of the corresponding number of fixed electrode bears adjacent to the edge of the Provides additional information about high frequency sound at appropriate locations within the pot. Like Or 2,000 to 2800Hz.

2.800 to 4,0007 and 4,000 to 8. In the range of 0007 Three additional bandpass filters are employed.

The overall stimulation rate remains at FO (fundamental frequency or voice pitch), but only Previously used FOFIs generate only two types of pulses per audio pitch period. Compared to the F2 strategy, four electrical stimulation pulses for each glottal pulse were also Occur. For voiced sounds, there are 2,000 first and second formants. 2,800 and 2,800 to 4.0001 (representing energy in the z range) Provided with additional stimulation pulses. For unvoiced phonemes, 4°000Hz Still other pulses are provided representing energies above, but within this frequency range. Since no energy is present, no stimulus is provided to the first formant.

Stimuli occur at a random pulse rate of approximately 260Hz, which is similar to the previous sound. It is approximately twice that used for the voice coding scheme.

According to another aspect of the invention, a A tissue stimulating multi-channel electrode array that can be implanted in a patient suitable for stimulation. the audio spectrum signal received from the microphone to create the desired signal. A method is provided for processing. The method includes obtaining approximately 280)f from the audio signal. Select the first dominant frequency peak of the frequency band between approximately 1,000 and Z , and the top of the electrode array according to the spectral information contained in the first peak. stimulating at least one electrode in the area; and stimulating at least one electrode in the area; Select the second dominant frequency peak of the frequency hand between 0.07 and approximately 800& , and the base of the electrode array according to the spectral information contained in the second peak. stimulating at least one electrode in the audio signal area; extracting spectral information in at least one region of the spectrum; and at least one predetermined area in the electrode array according to the spectral information obtained. stimulating electrodes, said predetermined electrodes stimulating said base of said electrode array; It is characterized by being located within the space area.

Preferably, additional preselected electrodes are also obtained from said audio signal. 2,000 to 2.00 each.　800Hz, 2°8000 to 4,000 Spectral energies in the audio frequency range above 0.07 and 4.000Hz Stimulated using ghee.

According to another aspect of the invention, an incoming acoustic signal received by a microphone Multi-spectral peaks that extract a certain number, for example, 5 types of spectral peaks from a sample. (MPEAK) Improved for pot prostheses that employ coding tracks. Audio bromo nosar is provided. This audio processor uses this information as a pulse train. and send it to the selected electrodes of the inplant in the pot. 1st Forman) (Fl) Spectral peak (280-1000Hz) and second formant (F2) spectrum vector peaks (800-4000Hz) are coded and labeled as top and base, respectively. is provided to the ground electrode. Fl and F2 electrode selection follows the dot topic structure in the pan. cormorant. High frequency spectral information is sent to electrodes closer to the base, while lower frequency electrodes It is sent to an electrode closer to the apex. 2000-28007, 2800-40007 , and spectral energies in the range of 4000 Hz and above are encoded, Provided to three fixed electrodes. The fundamental or voicing frequency (FO) is the stimulus during the voicing period. Determine the pulse rate of the pulse, and the pseudorandom aperiodic rate of the silent period stimulus. Determine the pulse rate. The amplitude of the acoustic signal in the five bands determines the stimulation intensity do.

Concave dish Ω This specification sets out the claims particularly pointing out and distinctly claiming the subject matter of the invention. The present invention will be better understood from the following description when taken in conjunction with the accompanying drawings. I believe it will happen. In the drawings, Figures 1A and 18 are internal anatomical diagrams of the human ear. and a cross-sectional view of the inside of the pot.

FIG. 2 is a block diagram of the overall pot inplant system of the present invention.

FIG. 3 shows this system, including parts that can be implanted and parts that can be worn by the patient. FIG.

Figures 3A and 3B show the side and side views of the implantable parts of this system, respectively. FIG.

FIG. 4 is a graph of current versus time showing two sets of current waveforms used in the present invention. .

FIG. 5 shows the electrode base for voiced speech using the multi-peak coding strategy of the present invention. It is a graph which shows an example of the sequential stimulation pattern of (a).

FIG. 6 shows the electrode base for unvoiced speech using the multi-peak coding strategy of the present invention. It is a graph which shows an example of the sequential stimulation pattern of (a).

FIG. 7 shows a collection of various stationary phonemes using the multi-peak coding strategy of the present invention. 3 is a chart showing an example of an electrical stimulation pattern.

FIG. 8 is a graph showing typical loudness growth for the audio processor of the present invention. It is.

FIG. 9 shows the microphone of a pulsating type, multi-channel pot inplant system according to the present invention. FIG. 2 is a block diagram of the voice processor and audio processor portion.

5 ships for The pot inplant system of the present invention shown in FIG. 2 includes several components.

The electrode array 1 is implanted in the pot. Electrode row 1 is a flexible silicone plastic carrier and a number of rings or bands of molded platinum. preferably total There are 32 platinum bands. The 22 bands at the tip are active electrodes and weld to them with connected wires. The 10-electrode band at the base is used for stiffening. , and act as an adjunct to surgical insertion. In a typical example, an electrode ring is approximately 0.05 mm thick and 0.3 mm wide, and 0.6 mm at the proximal end. It has an outer diameter from mm to 0.4 mm at the tip. The diameter of the ring is from the tip of the row to 1. It changes smoothly so as to taper over 0 mm. The ring is at the tip of the electrode row The exposed outside of the ring is spaced 0.75mm on center over 25mm. All of the area is used as active electrode area. Silicone material is Dow Corning The company's MDX4-4210 is sufficient.

22 t-pole wires run from electrode row 1 to receiver/stimulation unit) (R3U) 3 through cable 2. The invention described is limited to this electrode array design. using a number of alternative electrode designs such as those described in the prior art. can be used. R3U3 is attached to R5U from an external source and the skin Information and power are received through a tuned receiver coil 5 located just below the skin. R3U also provides electrical stimulation pulses to the electrode array 1. power and which electrode and which The stimulation intensity data is processed using an external multi-peak audio processor (MSP). ) 7, transmitted across the skin using an inductive link 6 operating at radio frequency. It will be done. In normal operation, the MSP emits sound from a conveniently worn microphone 8. Pick up the acoustic stimulus and determine the stimulation electrode, rate and amplitude from the signal Extract information used to.

Because each patient's response to electrical stimulation is different, each patient's MSP can be determined by his or her It is necessary to adapt it to your own needs. Therefore, MSP is suitable for each whistleblower. It has a RAM programmed as follows.

The patient's response to electrical stimulation was significantly reduced for some time after implantation of R5U using the patient's MSP. and the results of these tests are Used to set MSP to . This connects MSP to connectors and cables. to the diagnostic programming interface unit (PPI) 10 via the It is done by connecting. DPI is itself a cable and connector 11 through the Diagnostic and Programming Unit 7) (DPU) called 12. Connected to a general purpose computer.

An external view of the system used by the patient is shown in Figures 3.3A and 3B. ing. The electrode array 20 is flexible and is attached to the basement membrane separating the scala tympani from the rest of the pot. When inserted along, it fits into the shape in the pot (Figures 1A and 18). electric The pole array is connected to the RSU 22 by a silicone coated cable 21. cable 21 to provide additional stress relief to prevent wire breakage within the cable. Specially designed. The information and power receiving coils are embedded electronics in R3U22. A single turn of multi-strand platinum wire 23 which is a transformer connected to the circuit. Ru.

Externally worn transmitting coil 24 is supported adjacent to each of coils 23 and 24, for example. by a cooperating magnet or to hold the coil to the user's head. 24 with fixings (not shown) or with adhesive tape inside R3U. It is fixed against the head above the site of the implant 22. The coil 24 is connected to the coil cable 2 6 and a hearing aid microphone 27 to an audio processor 29 .

The hearing aid microphone 27 is worn in the ear closest to the implant site, and the microphone Audio data from connector 27 is transferred to MSP 29 via three wire cable 28. connected to. Transmission data is the same three wire cables 28 and coil cable 26 from the MSP 29 to the coil 24. This three wire structure , entitled “Three-Wire System for Pot Inplant Processors”, 1989 Christopher N. Daly, U.S. Patent Application No. 40, filed September 7, 2013 4.230. This application is assigned to the assignee of the present invention and This is incorporated herein by reference. Worn over a tie clasp or attached to the user's clothing. Alternative microphone configurations are possible, including attached microphones.

Coiled cable 26 and three wire cable 28 have removable connectors. -32. Attached to microphone 27 and MSP29 by 33 and 34. It will be done. The MSP29 uses a commonly available battery (e.g. Powered by a single AA size battery. TV, radio, or high quality Plug-in to allow connection of an external audio signal source, such as from a crophon. An input jack 31 is provided.

Referring to Figure 4, the pulses used to electrically stimulate the pot are biphasic. Ru. That is, it has a period of negative current stimulation followed by equal periods of positive current of equal amplitude. These two periods (known as φ1 and φ2) are short periods of no stimulation. separated by φ1 and φ2 are in the range of 12 to 400 microseconds ( 200 microseconds), and the interruption interval is typically about 50 microseconds. It is a cross second. Amplitudes of φ1 and φ2, their durations and times of interruption intervals The interval is based on the information decoded from the signal transmitted by the audio processor (Figure 3). Therefore, it is determined. The actual values of these parameters depend on the patient's psychophysical tests. As a result of the cormorant. Reversing the polarity of φl and φ2 ensures that there is no net DC component in the stimulus. It is important because This is because long-term DC excitation can lead to electrode corrosion and cardiac pace. It is thought to be more important in pot inplants than in manufacturers. this is a pot Medium stimulators stimulate nerve fibers, whereas cardiac pacemakers stimulate the heart muscle. This is because it is designed to Nerve tissue is further damaged by electrical stimulation. Because of this, pot-implanted systems are subject to more stringent safety requirements than cardiac pacemakers. It is thought that it should be designed using all coefficients. This system provides both stimulus The same stimulus source is designed to be used for both phases. This two-phase The loops are created simply by reversing the connections to the electrodes. Therefore, extremely good load B can be obtained, and as long as the durations of φl and φ2 are equal, they offer a high level of safety. provide

The stimulation circuit is preferably configured as a constant current source.

This is compared to a constant voltage source, even if the electrode impedance changes ( (as is often observed), the current delivered to the electrode has a wide range of electrode impingement. - has the benefit of not changing over the dance. Current is several microamperes to 2 mA, producing a very wide range of sound loudness perception and to allow for wide variation between patients.

The stimulus generation circuit in R5U3 (Figure 2) is preferably configured in one of two modes. It is designed to operate in the same way. The first mode is “Depending on frequency or restrictions” When carrying out the circuit design, it is necessary to use the two selected electrodes (for convenience, A and A). and B). In phase φ1, the current is supplied by A. and is reduced by B. In phase φ2, the current is supplied by B. supplied and reduced by A, the other electrodes play no role in stimulation. R3U 3 is preferably configured such that any pair of electrodes can be selected for bipolar stimulation. It will be done. Therefore, there is great flexibility in the selection of stimulation strategies.

It should be understood that only these two specific stimulation modes were selected. but However, this does not exclude other stimulation modes. For example, multipole or distributed grounding The system may also be used when all electrodes do not act as a distributed ground. 9. With appropriate modification of the receiver/stimulator , current source, current sink, or inactive during either stimulation phase. so you can choose anytime.

The primary objective of the present invention is to provide improved voice communication for people suffering from significant hearing loss. The goal is to provide However, improved voice communication In addition to providing environmental sounds, such as telephones, doors, alarm sirens, doorbells, etc. It is also important to convey the sounds that are a part of life. The system described so far is based on This is essentially the patent of Crosby et al. that has been referenced and incorporated as a reference. Ru. In the Crosby et al. patent, the second formant F2 is used to explain the speech signal. The first formant F1 contains much of the naturalness of the signal. Despite this, it contributes only a small amount to comprehension.

Crosby et al. show that the third and higher fermants are as small as the second formant. Observed that it does not carry much information. They also stimulate a large number of electrodes simultaneously Given the limited knowledge at the time regarding interactions between electrodes during The method uses appropriate electrodes or parts in the pot to provide the most important formant information. He was angry that it was meant to encode the second formant in the second place. such stimulation The amplitude of is derived from the amplitude of the second formant.

The Crosby et al. system also provides prosodic information in the form of pulse rate. So Systems such as compress the stimulation rate to the range of 100-2507. this range Maximum pitch discrimination is obtained from the stimulus pulse rate at .

An additional factor adopted by Crosby is that above the current acoustic stimulus level 10 to 20 dB is used to determine the stimulation amplitude. . This compresses the entire acoustic loudness range into a small range of usable electrical stimulation. Instead, use only the top part. Thus, the signal amplitude of Crosby et al. The width is fully represented by a 5-pin binary code that provides only 30dB of dynamic range. be done.

In summary, the speech processing strategy of Crosby et al. The main spectral peaks within a range of approximately 4000 Hz are used to encode the electrode position. use.

2. The amplitude of the main spectrum I-L-beak used to encode the electrode position is the stimulus used to determine the amplitude.

3. Audio pitch (FO) is compressed and used to determine stimulation rate. It will be done.

Due to unvoiced and environmental sounds, the system of CrOsby et al. still generates a stimulus. However, the stimulation rate and electrode position depend on the raw nature of the acoustic signal. will be determined by For example, for the sibilant consonant (“S”), the stimulus rate is fairly fast but not constant, and the stimulated electrodes are at high frequencies. These are electrodes that secretly create wave perception.

A second speech processing strategy useful for certain patients was adopted by Crosby et al. It is used. The second strategy is that the electrode position is determined by the formant frequency. It is similar to the above in that respect. However, the stimulation rate is in the first form. and the stimulation amplitude is equal to the acoustic signal at the time of the F1 peak. is determined for the peak value of . This means that the stimulation rate is faster and the patient This has the benefit of inducing a more natural-sounding discourse perception for the user. In addition, F The l signal is amplitude modulated, and the FO relation is temporarily good, and the patient hears rhymes. It also perceives FO or vocal pitch, which is useful for conveying rhythmic information.

One of the other speech processing strategies considered by Crosby et al. The patient is stimulated at a rate of Fl, but the stimulation is FO relay The idea is to pattern the stimuli so that they are gated at the same time.

Over recent years, Crosby et al.'s FO, Fl, F2 speech processing code Despite the success of voice processors that use coded formats, such voice processors A number of issues still remain regarding the use of processor coding styles. instructions before As shown, patients who perform well in quiet conditions have moderate levels of banking. You can have significant problems when round noise is present. Furthermore, F.O.

Since the Fl, F2 style encodes frequencies up to about 4000 Hz, and phonemes and since environmental sounds have a high proportion of their energy above this range , such phonemes and environmental sounds are in some cases inaudible to inplant users.

According to the invention, the audio signal has a second frequency peak or formant of the audio signal. Band pass filter with multiple, e.g. three bands within or above the normal range of F2 a pulsating motion system, such as that disclosed in the Crosby et al. A multi-channel pot endoprosthesis is provided. Audio encoding style disclosed herein is called multispectral peak coding strategy (MPEAK). MPEAK is designed to provide additional high-frequency information that aids in the perception of speech and environmental sounds. It is.

MPEAK coding strategy extracts Fl and F2 spectral peaks and More apical and more for stimulation using encoded and extracted frequencies Evaluate to select the electrode bear near the base. Each selected electrode has a basic circumference. It is stimulated at a pulse rate equal to the wavenumber FO. In addition to Fl and F2, Three high frequency bands of spectral information are extracted. Band 3 (2000-28 007), Band 4 (2800-4000 ministers), and Band 5 (4000) (above z), the fixed electrodes of electrode row I (Fig. 2), such as the seventh and fourth and a first electrode, respectively.

The first, fourth and seventh electrodes are selected as opaque electrodes for high frequency bands. Ru. They are sufficiently far apart that most patients experience stimulation at these three locations. This is because it is possible to discriminate between These default assignments can be reprogrammed as needed. It should be noted that

If three high-frequency bands are assigned to the three base-most electrodes in the MAP If so, many patients will not find the additional high frequency information useful. this Patients often do not show good position-pitch discrimination between adjacent base-biased electrodes. It is from. Additionally, the overall pitch perception obtained from electrical stimulation is too high. It would be nice.

The following table ■ shows various formants employed in the speech coding style of the present invention. Indicates the frequency range of

280-10007 F1 800-40007 F2 2000-2800 & Hand 3- Electrode 72800-4000Hz Band 4- Electrode 4400〇7 or more Band 5-Electrode 1 If the input signal is voiced, it has a periodic fundamental frequency.

The electrode pairs selected from the evaluation of Fl, F2 and bands 3 and 4 are equal to FO. stimulated sequentially at a high rate. The electrode pair closest to the base is stimulated first; Then, as shown in FIG. 5, electrode pairs closer to the top are gradually stimulated. Figure 5 shows , band 5 is not provided. This frequency band should be ignored for voiced sounds. This is because it only contains information that can be obtained.

If the input energy is unvoiced, the F1 band (280-10001-Iz ) is practically zero. Therefore, it extracts more than 40,007 pieces of information. Replaced by the frequency band to be emitted. In this situation, F2 and band 3 Electrode pairs selected from evaluations 34 and 5 receive pulsatile stimulation. The stimulation rate is It is non-periodic and fluctuates between 200 and 300 Hz. Figure 6 shows that the stimulus is 2 illustrates a sequential stimulation pattern for an unvoiced sound, proceeding from the top to the top.

In this way, the MPEAK coding strategy extracts five spectral peaks and However, only four spectral peaks are present for which stimulus sequence. However, it can be understood that it is merely encoded.

Figure 7 shows the various steady-state phonemes when using the MPEAK coding strategy. Figure 2 illustrates the pattern of electrode stimulation for. The primary function of the MAP is to identify the main spectral peaks. the frequency of the arc (Fl and F2) into electrode selection. This feature In practice, the electrodes are numbered sequentially starting from the front window in the pan. Electrode 1 is It is the electrode closest to the base, and the electrode 22 is also the electrode closest to the top of the electrode array.

Stimulation of different electrodes usually produces a pinch sensation that expresses the dot-topic mechanism in the pot. . Electrode 22 induces the lowest pitch perception, or dullest sound. Electrode 1 is the highest Evoke pitch perception, or the sharpest sound.

To allocate the frequency range for Fl and F2 spectral peaks to the total number of electrodes , the default mapping algorithm reduces the total number of available electrodes to approximately Divide into a 1:2 ratio. Approximately 1/3 of the electrodes are therefore allocated to the F1 frequency range.

These are the top electrodes, and they have a frequency of 280 to IOOHz. will cover a range of numbers. The remaining two-thirds of the electrodes cover the F2 frequency range (800~ 4000Hz). The highest frequency range covering frequencies from 280 to 1000Hz Electrodes closer to the top are assigned proportionally equal frequency bands. F2 estimate The corresponding frequency range is assigned to the remaining near-base electrodes and logarithmically divided into equal frequency bands. This frequency distribution is proportional/logarithmic (lin/lo g) called spacing.

A second optional mapping algorithm (not shown) maps the entire frequency range to Divide into logarithmically equal frequency bands for Fl and F2 (log/log g spacing). Compared to Iin/log spacing, this is 1000H Electrodes assigned frequency boundaries below z result in a relatively wide frequency band vinegar. Because of the wide frequency band for these electrodes, many vowels are will stimulate the poles and therefore make it difficult to distinguish between these vowels.

The Fl/F2]in/log function of the defaulting algorithm is This is preferred because it provides better position resolution in the F1 range than the Iog function.

In addition, this algorithm is suitable for vowels and Provides consonant discrimination.

The mapping portion of the DPS program allows flexibility in assigning frequency bands to electrodes. To tolerate. If fewer electrodes are included in the MAP, fewer electrodes are included. A wide frequency band is automatically assigned by the computer and the overall frequency range will be covered. In addition, the computer-generated spectrum of frequency bands It is possible to override the Upper frequency limit for any frequency range By changing the number limit, it can be assigned to any electrode.

The table below shows two results using 20 electrode pairs and the MPEAK coding strategy. The default bounds (tin/log) are shown for the map created in phase+1 mode.

(Margin below) :LL Lin/Log frequency boundary and for 20 electrodes in BP+1 mode and electrode allocation for three types of high frequency bands - Sato1 defeat boundary - 1LJ3m -Eng-1--Yo-l- 426] 1 2904 13595 or more For electrode near band 3 The table below shows the same results using only 14 electrodes and the MPEAK coding strategy. Indicates the default boundary of the mode.

(Margin below) L in / L o g period for 14 electrodes in lL BP+1 mode Electrode assignment for wavenumber boundaries and three types of high frequency bands - 凰波臀 boundary 1 one! -4--ball-1--yo-l- 1013603, 587 43428 or more Electrode = 7 hands 3 The amplitude of the electrical stimulation is divided into five frequency bands (Fl, F2 and bands 3.4 and 3.4). and 5) from the amplitude of the incoming acoustic signal in each mouth. However, electricity Since the poles have different darkness values (T) and maximum allowable loudness (C) levels, the audio The processor stabs each electrode separately based on the amplitude of the incoming signal within each band. The intensity level must be determined.

MSP (Figure 2) is a non-linear sound system that converts acoustic signals into electrical stimulation patterns. Contains a size growth algorithm. First, MSPs should refer to Figure 8. In order to better understand the acoustic signal, it can be expressed as a digital straight line with values from 0 to 150. Convert to scale. This digital scale (,!!!) and C) determine the actual amount of charge delivered to the electrode.

Signals whose amplitude level is coded as 1 generate a stimulus at the T level. Will. Signals whose amplitude level is coded 150 are at C level. will cause irritation.

Referring to Figure 9, the micro-hole of the pulsating type multi-channel inplant system. A block diagram of the system and part of the audio processor is shown. System 100 is a sound Pick up the voice and pass it through automatic gain control amplifier III to voice feature extractor 112 provides an electrical audio signal to the Audio feature extractor 112 analyzes the signal; 1 and 2 identified as Fl, AI, F2 and A2 in FIG. Provides a digital output corresponding to two formant frequencies and amplitudes.

The speech feature extractor 112 also detects and outputs the speech pitch FO and extracts the patient's precision. Using MAP114, which contains information about the divine physical test results, Translating pitch information into a pattern of electrical stimulation on two electrodes that are stimulated sequentially The encoder 113 is started. The data so translated is Receiver/stimulator unit R3U3 implanted by coil 115 (Figure 2) sent to.

Three bandpass filters 116, 117 and 118 are also connected to the microphone 110. receiving an audio signal from before it is applied to the audio feature extractor 112; Then, convert the signal into three components of different frequencies, that is, 2000 to 28007 signals. 1z signal to band 3, 2800-4000) 1z signal to band 4, and 4000-8 000Hz signal into band 5. Signals from bands 3.4 and 5 are coder 113, and the mapping of these signals is performed by the first and twenty-seventh encoder. of the resulting pattern of electrical stimulation to the appropriate electrodes Translation is done as previously discussed.

Automatic gain control amplifier 111 controls the signal fed to filters 116 and 117. used to control the amplitude of Filter II8 is the unvoiced part of the audio signal. Its amplitude is never large and therefore the signal is self-contained. Does not require dynamic gain control. Therefore, amplifier 119 has a built-in automatic gain control function. does not have

In summary, the psychophysical measurements made using the DPS software were information to translate the acoustic input into patient-specific stimulation parameters. Ru. The dark value (T) and maximum (C) level of the electrical stimulation reservoir are measured for each electrode pair. be done. These values are stored in the MAP. They can be applied to any given electrode pair Determine the relationship between incoming acoustic signal amplitude and stimulation level for.

Inside the audio processor, RAM stores a set of numerical tables collectively called MAP. Ru. This MAP contains stimulation parameters for Fl, F2 and bands 3 to 5. Determine the amplitude estimate. Coding of stimulus parameters is a sequence of separate steps. Follow the instructions. The steps can be summarized as follows.

1. The first formant frequency (Fl) is the main spectrum in the range of 280 to 1000 Hz. It is converted to a number based on Kutbek.

2. The Fl number should be stimulated together with one of the MAP tables to represent the first formant. used to determine the electrode. This mode determines which electrodes do not work. Ru.

3. The second formant frequency (F2) is the main frequency within the range of 800-4000Hz. It is converted into a number based on the spectral peaks.

4. The F2 number should be stimulated together with one of the MAP tables to represent the second formant. used to determine the electrode. This mode determines which electrodes do not work. Ru.

5, the amplitude estimates for bands 3, 4 and 5 are to the three default electrodes 7, 4 and 1 for or selected when the MAP is created. assigned to other electrodes that can be determined.

6. The amplitude of the acoustic signal in each frequency band is converted to a number between 0 and 150. release The level of stimulation that will be delivered is determined by the acoustic amplitude (in the range 0-150) in step 2. ．． M related to the stimulation level for the particular electrode selected in 4 and 5. It is determined by referring to a set of AP tables.

7. The data is further encoded in the audio processor and sent to the receiver/stimulator. It is transmitted to the device. The latter encodes the data and sends the stimulation to the appropriate electrodes. Ru. Stimulus pulses are applied at a rate equal to FO during voiced periods and at a rate equal to FO during voiceless periods. and runs within the F1 formant (typically 200 to 300 Hz) Provided at dumb aperiodic rates.

From the above description, it can be seen that the multispectral peak speech coding style of the present invention is similar to the prior art. Provides all of the information available in the FOFIF2 format, but with three high frequencies It will be clear that a bandpass filter provides additional information. this Their filters are 2000-28007, 2800-40007 and 4000 Covers a frequency range of ~80007. Energy within these ranges is control the amplitude of electrical stimulation of three fixed electrode pairs at the base end of the this Therefore, additional information about high-frequency sounds is provided to the tonoprotic in the pot at the appropriate location. Served.

The overall stimulation rate remains as FO (Fundamental Frequency or Vocal Pitch), but only However, in the mode of the present invention, four electrical stimulation pulses are associated with each glottal pulse. It occurs. This is a prior art technique in which only two types of pulses are generated per audio pitch period. compared with the FOFIF2 strategy. In this new style, for voiced sounds Two pulses representing the first and second formants are still provided at 20 Additional values representing energy within the ranges 00-28007 and 2800-40007 A stimulation pulse is generated.

For unvoiced phonemes, still other pulses representing energies above 4000 Hz are proposed. However, since there is no energy in that frequency range, it is the first formant. No stimulus is provided. The stimulus is about twice that used in the previous strategy. Occurs at a random pulse rate of about 2607.

Furthermore, from the foregoing discussion, it can be seen that the present invention addresses various problems associated with previous pot inplant systems. It would be obvious to provide an in-pot inplant system that overcomes this problem. According to the present invention, multiple The use of a spectral peak speech coding strategy provides users with this inplanted system with Significantly improved audio even in the presence of moderate levels of background Provide recognition. In addition, the present invention provides improved phoneme and environmental sound recognition. Served.

While specific embodiments of the invention have been illustrated and described, those skilled in the art will be able to understand its broadest aspects. It is obvious that various changes and modifications can be made without departing from the invention. , it is therefore the purpose of the appended claims to cover such changes and modifications as do not fall within the true spirit of the invention. The objective is to cover the area as belonging to the scope.

Copy and translation of written amendment) Submission form nμ % - 17m m □ 2v Fu 5 hate; Shizuku ζ, WJ I □ Rit Hummus May 1991 7 days

Claims (15)

[Claims] 1. Patients suitable for placement in the cochlea from the apex area of the cochlea to the pace area of the cochlea a tissue stimulating multi-channel electrode array that can be implanted into the patient; an implantable multi-channel stimulator and an electrical stimulation device that transmits sound signals to the stimulator; A programmable audio processor worn externally to the patient for processing into intense signals a multichannel cochlear prosthesis comprising about 280 Based on the extraction of main peaks in the range of Hz to about 1000 Hz, determine first formant spectral information of the formant, and determine the first formant spectral information of the formant. stimulating at least one electrode in the top area of the electrode array according to torque information; and means for Based on major peak extraction in the range of about 800 Hz to about 4000 Hz, the determining second formant spectral information in a sound signal; at least one electrode in the base area of said electrode array according to spectral information of a means for stimulating extracting spectral information in at least one region of the spectrum of the sound signal; at least one of the electrode arrays according to the extracted spectral information. with at least one high frequency band filter for stimulating selected electrodes; , the preselected electrode is within the base area of the electrode array; The multichannel cochlear prosthesis characterized by: 2. extracting spectral information in a corresponding number of regions of the sound signal; said for stimulating at least a corresponding number of said preselected electrodes in said All of the preselected electrodes include a plurality of high frequency band filters. The multichannel cochlear prosthesis of claim 1 in said base area of said electrode array. 3. The electrical stimulation consists of pulses delivered at a pulse rate that depends on the pitch of the sound signal. 3. The multichannel cochlear prosthesis of claim 2, wherein the voltage is applied to said electrodes in the form of a constant current. 4. 5. The pulse rate is in a range between about 80 Hz and about 400 Hz. Term 3 multichannel cochlear prosthesis. 5. At least three of the radio frequency bands, each with a preselected electrode, Multichannel cochlear prosthesis according to any one of claims 1 to 4, comprising a filter. . 6. The first high frequency band filter has frequencies of approximately 2000Hz and 2800Hz. Spectral information is extracted from the sound signal in the frequency range between The second filter filters out sounds in the frequency range between about 2800Hz and about 4000Hz. Extracting spectral information from the signal, the third of the high frequency band filters is about 4 extracts spectral information from signals in the frequency range between 000Hz and approximately 8000Hz. 6. The multi-channel cochlear prosthesis of claim 5. 7. The electrodes in said electrode array are continuous starting from their base end and extending to their top end. It can be considered that the first, second and third high frequency waves are numbered. The amplitude estimate obtained from the spectral information extracted from the band filter is the amplitude from said first filter is applied to a number of preselected electrodes; The estimate is applied to a higher numbered electrode than the amplitude estimate from the second filter. and the amplitude estimate from the second filter is equal to the amplitude estimate from the third filter. 7. The multichannel cochlear prosthesis of claim 6, wherein the voltage is applied to a higher number of electrodes than the estimate. 8. The electrode array includes approximately 22 electrodes therein, and the base area of the electrode array is located at the front. The top area includes about 2/3 of the electrodes in the electrode array, and the top region includes about 1/2 of the electrodes in the electrode array. 8. The multichannel cochlear prosthesis of claim 7, comprising: /3. 9. Spectra extracted from the first, second and third high frequency band filters The amplitude estimates obtained from the information are the 7th, 4th and 1st 9. The multichannel cochlear prosthesis of claim 8, wherein each voltage is applied to an electrode. 10. In the case of voiced sound signals, the electrodes selected to stimulate are based on the information obtained from the first and second filters, and the The electrodes are stimulated sequentially at a rate based on the pitch of the sound signal, with the electrodes being stimulated most closely to the base. 7. The electrodes at the top are stimulated first, followed by stimulation of the electrodes progressively closer to the top. multichannel cochlear prosthesis. 11. In the case of unvoiced sound signals, the electrode selected to stimulate the second formant , based on information obtained from the first, second and third filters; The poles are at non-periodic rates in the range of formant 0 to formant F1. stimulated sequentially, with the electrode closest to the base being stimulated first, then progressively 8. The multichannel cochlear prosthesis of claim 7, which continues to stimulate the apical electrode. 12. the electrodes at a non-periodic rate within the range of about 200 Hz to about 300 Hz; 8. The multichannel cochlear prosthesis of claim 7 which is stimulated. 13. The aperiodic rate may be in the range of about 200Hz to about 300Hz. Multichannel cochlear prosthesis according to claim 11. 14. suitable for placement within the cochlea from the apical area of the cochlea to the base area of the cochlea Generates signals to stimulate a tissue-stimulating multichannel electrode array that can be implanted in the patient It is a method of processing audio spectral signals received from a microphone. Thus, the method includes the method of extracting the Select a first dominant frequency peak of the audio signal and include the frequency peak included in the first peak. at least one electrode in the top area of said electrode array according to spectral information provided stimulating said audio from a frequency band between about 280 Hz and about 4000 Hz. a second dominant frequency peak of the audio signal and included in said second peak; at least one electrode in the base area of said electrode array according to spectral information provided stimulating at least one region of the spectrum of the audio signal; extracting spectral information and distributing the information in the electrode array according to the extracted spectral information; stimulating at least one predetermined electrode of the The preselected electrode is located at a base area of the electrode array. Said method. 15. 2000 to 2800Hz, 2800 to 4000Hz, and and obtained from said audio signal in the audio frequency range of 4000 Hz and above. Additional preselected electrodes are stimulated using the spectral energy generated. 15. A method of processing an audio spectral signal according to claim 14.