Acoustic simulation of cochlear implants in reverberation - Essay Example

ACOUSTIC SIMULATION OF COCHLEAR IMPLANTS IN REVERBERATION Acoustic Simulation of Cochlear Implants in Reverberation Introduction The cochlear implant is a minute electronic device that aids a profoundly deaf person or a person with hearing difficulties to sense a sound signal. This implant is also referred to as the bionic ear. The device works by stimulating the auditory nerves located inside the cochlea with an electric pulse. The implant consists of two portions. One part of the device sits behind the human ear while the second part is implanted under the skin. The device does not amplify the sound but stimulates any working auditory nerves (NIDCD, 2007). History of cochlear Alesandro Volta who developed the battery, placed metal rods and when he connected them to a 50 volts circuit he experienced a jolting noise, later French- Algerian surgeons, Andr Djourno and Eyries reported that a patient would hear sounds when electrodes were placed on the patient nerves exposed during an operation. In 1957 both French surgeons developed a clinical cochlear implant. This device was inefficient and only provided the patient with the rhythm. In 1961, William house and Doyle James developed a five wire device. This device used five electrodes each applied with the same single signal. The speech signal was modulated to travel at 16 KHz. The device was implanted and functioned better than the French surgeon's equipment. In 1984 the device was approved by food and drug administration (FAD) in the USA. In 1964 some patients were implanted with six channel devices but the recipients did not understand the speeches, the equipment was developed by Blair Simons of Stanford University. Parallel developments of the same device were being conducted at the University of California, the group comprising of Robin Michelson and Michael Merzenich made major medical implants. In 1976 Pialoux, Chouard and McLeod tested an implant device that enabled patients to hear and perceive at least half of the spoken words. In the 1990, miniaturization of the cochlear device dominated the device research. By the year, 2006 most children were implanted with miniature devices that were put behind the ear. On October 2005, 3 people were implanted with Tiki device; this device did not have external components. This device is however not commercially available. Developments have also focused on providing devices that can be implanted on both ears. These devices aid in hearing in a noisy environment. Current development enables the device to be fitted on babies five months old (Deafblind, 2008). Components of the cochlear implant; The device has to major components; the external part and the internal part External part (A) Microphone. This device picks the sound from the external environment. Sound waves vibrate a membrane that converts the mechanical energy to electric pluses. (Desloge, et al 1997). (B) Speech processor This device comprises of filters that process the signals detected by filtering off the noise giving priority to the audible sounds. The filtered audible signals are then sent to the transmitter. (C) Transmitter The processed signals from the speech processors are received by the transmitter. The transmitter comprises of a coil held by a magnet placed behind the ear. The purpose of the transmitter is to transmit signals by electromagnetic induction to an internal receiver. Internal parts These parts are implanted inside the skin. (A) receiver/stimulator This device receives the signals from the transmitter and converts them to electric pulses. The pulses are then transmitted to the electrodes using cables implanted internally. The receiver is placed in the bone below the skin. (B) The electrodes This comprises of 22 electrodes that are wound round the cochlea, they send the impulses to the nerves in scala tympani. The signals are then sent to the brain via the auditory nerve system. PROCESSING OF THE AUDITORY SIGNALS Sound comprising of compression and rare functions is received by the microphone. The sound waves are processed to determine which electrode should receive that sound (Georgiou & Toumazou, 2005). The sound must be processed to remove any noise and give priority to the audible sound. The strategy used to filter the sound is the Fast Fourier Transforms. The transforms divides the received signals into different frequency bands. An algorithm helps in selecting the strongest output from the filters. The number of outputs is determined by the number of electrodes and other factors. Most of the strategies emphasize on the transmission of spectral aspects of speech and coarse temporal information. Great research is being undertaken to determine to incorporate fine timing aspects to the system (Hersh & Johnson, 2003). Feature extraction strategies are used in to process the vowels. Each vowel is given a fundamental frequency and format. The fundamental frequency is the lowest frequency peaks while the format represents peaks with higher frequencies. The format for the fundamental frequency and formats are unique for every vowel sound. The processing algorithms help in the identification of the vowels and emphasize on its feature. The strategies emphasize on the transmission aspects of the speech. Depending on the different manufacturers, different strategies are used to aid in the processing of the speech (Cheeran, et al, 2004) Some of the processing strategies used are ACE which is a strategy used by Cochlear company. In this strategy, the number of maxima from the available maxima in the sound is selected. Another method for processing strategy is the CIS, SAS, and HIRes, these strategies are used by Advanced Bionics Company. The strategies rely on stimulating the full spectrum. The choice of the processing strategies is vital. Patient's understand speech using 4 electrodes but are faced with some difficulties when trying to understand music which requires fine tuning structure stimulation. Some strategies allow the processing of the sound using fine structure presentation. These strategies are availed by advanced bionics and Medel strategies. Fine structure presentation uses the Hilbert Transform to process the signal path. This is in contrast with other strategies that use short time Fourier transforms (Neuteboom, et al, 1997). Transmission After processing the speech signal, it is transmitted to via the radio link to the internal device. The radio link does not require the use of physical cable and connection that reduce the pain that would be experienced if such connections were used. The use of physical wires has the limitation of inducing some infections on the patient. The transmitter is attached to the receiver through the use of a magnet that holds it through the skin. Receiver The receiver is a sophisticated device that allows the translation of the processed sound information. The receiver gets direction from the speech processor through magnetic induction sent from the transmitter element. The receiver also gets the power to run it function through transmission. The receiver processes the sound information received and controls the electric current being sent to the electrodes in the cochlea. The receiver is usually embedded in the skull behind the ear. The electrodes The electrodes are manufactured from platinum or a highly conductive material. The material must not rust or cause infection on the patient. Platinum is known for being one of the best materials as it resists nearly all types of corrosion and is also a good conductor. The electrode array is made of silicone rubber. The electrode arrays are connected to the internal receiver at one end and the other end is surgically inserted into the cochlea. The cochlea is wound around the auditory nerve. When electric pulses are routed to it, an electric field is generated and the fibers of the auditory nerve are stimulated. The depth of inserting the electrodes varies and greatly influences the performance of the device. Typical insertions reach up to a depth of 25 mm corresponding to a frequency of 400-600 HZ. The distance between the electrodes is about 2.5 mm. this varies between different implant manufacturers. Speech processing This is the component of the cochlea device that aids in the transformation of the sound picked by microphone to electronic signals. The generated signals can be transmitted to the internal receiver. To do its functions, it has to be programmed. The programs are stored in a processor. The sound signals are coded and then sent to the receiver and finally to the cochlea. Two types of processors are usually available, behind the ear processor commonly refereed to as BTE which is a minute processor worn on the ear and has a microphone. This type of processor is used by the adults and older children. The other type of processor is the body processor that is used by a small child whose ear cannot support the hearing device. The device s kept on the body and wires extend to microphone earpiece to connect it to the processor. Programming the device is done by an audiologist. Each implant is coded for different users. The audiologist sets the minimum and maximum electrode current based n the level of loudness of the speech to the user. The audiologist also selects the best speech processor for different users (Deiss & Huang, 2003). Typical simulation of the cochlea implant The filter bank is used for filtering the input signals. It consists of several filters. The signal received is half wave rectified and low pass filtered. (Shannon, 1995). This enables the extraction of envelope for each channel. The extracted envelopes are used to modulate the carrier signal. (University of Granada, 2004). The carrier wave for a sinusoidal wave resembles a sine wave. The output is then filtered to band limit every channel. There is a minimum channel interaction with non overlapping filters (Dorman, 1997). The final signal strength is generated by the summation of all channels (Widrow, 2001). Noise Vocoder The green wood map defines the centre frequencies and the cutoff frequencies. Diagram 1: adapted from (Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California.) Tone generation The method used is the Gaussian envelope tone vocoder. Amplitude modulation on the carrier wave is done by a Gaussian envelope. The carrier signal maybe pure tone or band pass limited, this can be expressed by the following equation Equation From the equation controls the envelope duration which results in a tone pulse with good spectral aspects. Rectangular and trapezoidal envelopes use the spatter or sinusoidal modulations with the frequency content being localized around the carrier frequency, a single enveloped tone pulse of 5 KHz is shown below. The duration of the envelope is about 1ms. Drawing 2 (adapted from: Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California.) To carry out the CI stimulation, three parameters are used. Rate of simulation: this is the repetition rate for the Gaussian envelope tone train. Stimulation place; this is usually manipulated by the carrier tone frequency. Electric field spread: the electric field spread is inversely proportional to the temporal envelope duration Diagram showing vocoder speech (adapted from: Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California.) Diagram 4 showing vocoder speeches (adapted from: Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California.) Consonant and vowel recognition The recognition of consonant and vowels is examined as a function of the number of channels. Noise based simulations and sine vocoder results to overestimation of the CI performance Diagram 5 (adapted from Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California.) Amplitude modulation detection and discrimination The sensitivity of the modulation frequency decreases with the increase in modulation frequency. With the normal hearing, as the modulation frequency increases the normal listeners shows reduced performance. This is as illustrated in the graph below; (Durant, 2004) Diagram 6; adapted from: Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California. As the modulation is increased. The demodulation discrimination worsens. The noise level is 1 inch broadband and a Gaussian of 1K and the amplitude modulated tone of 2 KHz. This is illustrated in the diagram above. Importance of acoustic simulation in CI research The acoustic simulations are important in cochlea implant research as the help in the advancement of the CI devices. Simulation tests reveal many parameters that help in the improvement of the CI devices. An example is the determination of the number of electrodes to be used. In this processes the patients are inserted with the different electrodes and the simulation test conducted. After conducting the test, the patient who responded best to audio speech determines the number of electrodes to be implanted. The research is also extended to the distance between the electrodes. This too has an effect on the effectiveness of the CI implant. By conduction the simulation test and drawing the electrodograms it is possible to determine the effects of varying the lengths of the electrodes. A simulation diagram also helps to reveal the depth of the insertion of the electrodes, different electrode insertions produces differing results that can be analyzed by using acoustic simulations (University of Granada. 2004). Secondly simulation helps in determining different programming techniques and speech processing techniques that produce the best results. Different simulations are run using different methods and the simulation charts reveal the best method for speech processing. Acoustic simulation using cochlear implants aids in determining the performance of CI measured against the performance with normal hearing persons. This aids in developing CI equipments, speech processing devices and other hearing aid that are up to standard and closely resemble the real life situations (Gao et al, 2003) References Cheeran, A. N., Pandev, P. C. (2004). Speech processing for hearing aids for moderate bilateral sensorineural hearing loss. IEEE International Conference on Acoustics, Speech and Signal Processing. Dorman, M, Loizou. P and Rainey, D. (1997). "Speech intelligibility as a function of the number of channels of stimulation for signal processors using sine-wave and noise-band outputs," Journal of Acoustical Society of America, 102(4), 2403-2411 Deafblind.2008. Cochlear Implants. [Online]. Available at http://www.deafblind.com/cochlear.html Accessed 20 June 2009 Deiss, A., & Huang, Q. (2003). A low-power 200-MHz receiver for wireless hearing aid devices Rana, R. S., & Garg, H. K. (2004). A novel architectural concept for hearing aid devices," PAT04-022/ICS-GHIC004, Institute of Microelectronics, Singapore Hersh, M. A., & Johnson, M. A. (2003). Assistive technology for the hearing-impaired, deaf and deaf blind. New York: Springer publisher Durant, E. A., Wakefield, G. H., van Tasell, D. J., & Rickert, M. E. (2004). Efficient perceptual tuning of hearing aids with genetic algorithms. IEEE Transactions on Speech and Audio Processing. Desloge, J. G., Robinowitz, W. M., & Zurek, P. M. (1997). Microphone-array hearing aids with binaural output. I. Fixed-processing systems. IEEE Transactions on Speech and Audio Processing. Georgiou, J., & Toumazou, C. (2005). A 126-W Cochlear Chip for a Totally Implantable System. IEEE Journal of Solid-State Circuits. Gao, R., Basseas, S., Bargiotas, D. T., & Tsoukalas, L. H. (2003) Next-generation hearing prosthetics. Robotics & Automation Magazine, IEEE. McDermott, H. (1998). A programmable sound processor for advanced hearing aid research. IEEE Transactions on Rehabilitation Engineering. Neuteboom, H., Kup, B. M. J., & Jassens, M. (1997). A DSP-based hearing instrument IC. IEEE Journal of Solid-State Circuits. NIDCD.2007. Cochlear Implants. NIDCD Information Clearinghouse. [Online] available at http://www.nidcd.nih.gov/health/hearing/coch.asp accessed 23 June 2009. Shannon, Zeng, Wygonski, Kamath, and Ekelid (1995). Simulations: Acoustic Simulations of Cochlear Implants. [Online] available at http://www.healthaffairs.uci.edu/hesp/Simulations/simulationsmain.htm Accessed 21 June 2009 Thomas L, Jeff C & Zeng, F (2006), Acoustic Simulation of Cochlear Implants. Hearing and Speech Research Laboratory California: University of California. University of Granada. 2004. Cochlear Implant Simulation: How has "Cochlear Implant Simulation" been developed [Online]. Available at http://www.ugr.es/atv/web_ci_SIM/en/seccion_4_en.htm Accessed 20 June 2009 University of Granada. 2004. Cochlear Implant Simulation. What is not "Cochlear Implant Simulation" [Online]. Available at http://www.ugr.es/atv/web_ci_SIM/en/seccion_5_en.htm Accessed 20 June 2009 University of Granada. 2004. Cochlear Implant Simulation: How does "Cochlear Implant Simulation" work [Online]. Available at http://www.ugr.es/atv/web_ci_SIM/en/seccion_3_en.htm Accessed 20 June 2009 University of Granada. 2004. Cochlear Implant Simulation: Sound perception through a cochlear implant [Online]. Available at http://www.ugr.es/atv/web_ci_SIM/en/seccion_2_en.htm Accessed 20 June 2009 University of Granada. 2004. Cochlear Implant Simulation: Software installation [Online]. Available at http://www.ugr.es/atv/web_ci_SIM/en/installation.htm Accessed 20 June 2009 Widrow, B. (2001). A microphone array for hearing aids. IEEE Circuits and Systems Magazine. Read More