research directions Cochlear implant G.M.CLARK FREQUENCY CODING ELECTRICAL RATE STIMULATION - PHYSIOLOGY AND PSYCHOPHYSICS Department ofotolaryngology

Transcription

1 Cochlear implant research directions G.M.CLARK COl1gress of. Sydney, Auslra'ia 2-7 March 1997 Department ofotolaryngology The University ofmelbourne, Melbourne (AUS) The Bionic Ear Institute, Melbourne (A US) FREQUENCY CODING Initial cochlear implant research (Clark, 1969) showed that with electrical stimulation of the auditory nerve there is an electroneural "bottle-neck" limiting the flow of information from sound to the central auditory nervous system. This electroneural "bottle-neck" is due to the difficulty in simulating with electrical stimulation the temporal as well as the place coding of frequency. One of the main aims of our research is to improve cochlear implant performance by widening the "bottle-neck" with better simulation of the temporal and place coding of frequency. Temporal coding is considered to be due to a direct relationship between the intervals between action potentials and the period of the sound wave. Temporal coding is thought to apply to low frequencies, but its importance for high frequencies is still not clear. Place coding is due to excitation of specific sites within the cochlea and the central auditory pathways 'so that a frequency scale is preserved anatomically (i.e. the brain is organized tonotopically). ELECTRICAL RATE STIMULATION - PHYSIOLOGY AND PSYCHOPHYSICS With our previous research we have demonstrated that frequency diserimination and pitch perception for electrical stimulation are comparable with acoustic stimulation for low frequencies up to approximately pulsesls (Clark, et.al., 1996) They are however, poorer for higher rates of stimulation. On the other hand, our electrophysiological research has demonstrated that the pattern of interspike response intervals are signific~ntiy different for low rates of CoPyright 1997 by Monduzzi Edilore S.p.A na (Ilaly) 55

2 XVI Wor1d 50 Q) Q) 416 Hz - Olorllinolaryngology TO '0 30 Qj Qj Sydney, Australia.c 20.c 2 7 March 1997 E E 10 Z Z (/) Electric <D <D (/) 400 pulses/s - '0 400 ' c.c E E 200 Z 100 Z ~ (/) 834 Hz Electric 800 pulses/s Figure 1 Interspike interval histograms from primary-like units in the anteroventral cochlear nucleus of the cat. Left top - acoustic stimulation of 416 Hz; Left bottom - electric stimulation at 400 pulsesls (pps). Right top - acoustic stimulation at 334 Hz; Right bottom - electric stimulation at 800 pulsesls (pps) a b Figure 2 A diagramatic repr:sentation of the unit responses in an ensemble of fibres for a low to rrud acousl1c frequency. (a) Nerve action potentials in a population of neurones, (b) Pure tone acoustic stimulus. (Clark, 1995) Sydney. Australia 2 7 MarCh 1997 stimulation, but, become more similar for high rates above 400 pulsesls (Clark, 1996). Fig. l. At low rates the responses to electrical stimulation are deterministic, that is, there is little jitter around the point of excitation, and there is usually a single interval between responses, which is the same as the period of the sound wave. On the other hand, with acoustic stimulation the response is stochastic, and there are multiple intervals which are harmonically related to the period of the sound wave. The question arises why is there an apparent contradiction between the psychophysical and the physiological results? Why is electrical rate discrimination more like sound at low stimulus rates, but the pattern of nerve firing is dissimilar; and why is rate discrimination poor at high stimulus rates when the interspike interval patterns are similar. i Auditory Nerve 56 CODING - TEl\IPORO SPATIAL PATTERNS The discrepancy between the physiological and psychophysical results can be explained by the assumption that a temporo-spatial pattern of intervals, but not the patterns of intervals in a single fibre are required for the temporal coding of sound, and this is not adequately reproduced by electrical stimulation. An example of a temporo-spatial pattern ofresponses in an ensemble of neurons is demonstrated in Fig. 2. In this example, although each neuron does not fire eaeh sine wave, the overall population does. Furthermore, the firing of individual fibres is probably correlated with the activity of neighbouring nerve fibres due to phase variations in the basilar membrane maxima. This is illustrated in Fig. 3. The responses to sound are shown on the left, and it can be seen that the individual fibres fire at various intervals in time when the maximal displacement of the basilar membrane ~. ' I! L 'j' j: Cochlea Figure 3 ~ft: A diagram of variations in the movement ofthe basilarmembrane around the ~~t~l~f the ~aximu~ vibration at two instances in time, and nerve action potentials. ree nelghbounng groups of nerve ftbres in response to phase variation in the stimulus. Right:. I The simulated temporo-spatial pattern of aett'on potentials produced by e ectneal stimuli with multiple pulses per period. 57

3 XVI Wortd Congress 01 Sydney, Australia 2 7 March 1997 is in one direction. It is proposed that the probability of auditory nerve firing is a function of the phase of the basilar membrane travelling wave, and consequently on the spatial separation of the neurons excited by the two maxima. In order to simulate this tempo-spatial pattern a promising method is to produce patterns of electrical pulses with multiple pulses per period. In this way, (Fig. 3), particularly when the initial pulse has a lower amplitude a!)lore central group of fibres is excited, and then with the larger amplitude pulse a wider group of neurons is stimulated. Furthermore, when the second pulse occurs during the refractory period the central group is not excited effectively simulating the pattern seen with sound. We have commenced psychophysical and physiological research to help determine what patterns of stimulation are going to reproduce sound frequencies most effectively, and this will be important for the development of advanced speech processing strategies for cochlear implants. This will be done by using the appropriate patterns for each frequency location. A simple simulation of such a strategy is illustrated in Fig. 4. PLACE CODING Our research has also been aiming to widen the "bottle-neck" by better place coding of frequency. This is being undertaken by siting electrodes closer to the spiral ganglion cells to provide more localised stimulation of groups of auditory Pulse Pattern SPEAK Strategy nerve fibres. The prescnt Nucleus banded electrode altay lies peripherally away from the spiral gangfion cells which are in the central axis or modiolus of the cochlea. Experimental research in our laboratory has shown that moving an array closer to the modiolus will significantly reduce the threshold, and this should provide better opportunities for improving place and temporal coding. PERTh1ODIOLAR-ELECTRODES A number of perimodiolar electrode arrays are being developed to achieve this goal. They are firstly, electrodes that are held straight prior to insertion, then released to allow them to curl around the modiolus as they are further advanccd. They can be held straight with an insertion tool which then has a releasin" mechanism once the array is within the basal tum or alternatively the precurved array may be held straight with a coating of polyvinyl alcohol which dissolves in the cochlea to allow release to occur. An alternative design is to attach a teflon leadwire which helps draw the electrode into the cochlea and then applies pressure as it reaches its point of maximal insertion. This system is also being evaluated to determine whether it is likely to induce trauma and be bjocompatible. A third type of array is one which is bilaminar and made of material which absorbs water and increases in length. Differential expansion causes the electrode to curl. These carriers will be the basis for improved electrode designs for developing both improved place and temporal coding. The perimodiolar pre-curved banded array has been implanted in an initial patient, and it has been demonstrated that it can result in lower thresholds and comfortable listening levels. Research is in progress to determine how best to present additional place information and how to design improved speech speech processing strategies with the appropriate filter band to electrode transfer function. Sydney, Australia 2 7 March 1997 C!l V Filter Bank i~) <t V Frequency C!l ~l, \nsp~ctral Maxima 0., I. '. (j E ~' II 1'-/' 1"'.".,, " <t I I Frequency Q) rj) ctl co Electrical Stimulus Pattern ~ " "-----.l--"--l"- "" il SPEECH ELEMENT ANALYSIS In addition, further research is required to determine the most important speech elements for intelligibility, and extend the initial research carried out by the Melbourne group to extract formants. There are a number of specific elements in speech responsible for intellibibility, recording information and needing extraction and emphasis for transmission through the electroneural "bottle-neck". It is hoped that one or all of the research directions will create a significant leap in patient results, and extend the benefits of the implant to a wider range of people. ~ REFERENCES.\/':tl;\/~lftll CLARK, G.M. Middle Ear and Neural Mechanisms in Hearing and in the Management of Deafness. PhD Thesis. University of Sydney. 4-25, Electrode Array x Q) Cl. <{ ~ Analysis Period ~ Time CLARK, G.M. Electrical Stimulation of the Auditory Nerve: the Coding of Frequency, the Perception of Pitch and the Development of the Cochlear Implant Speech Processing Strategies for Profoundly Deaf People. Clinical and Experimental Pharmacology and Physiology. 23: , F~re4. A diaqtam of thc speech processing strategy with multiple pulses pcr period simui~ing temporal coding at a frcquency specific location in the cochlea. CLARK, G.M. Cochlear Implants: Future Research Directions. Annals of Otology, Rhinology and Laryngology. I04(Suppl. 166):22-27,

4 TONG, Y.c., BLAMEY, P.I., DOWELL, R.c. & CLARK. G.M. Psychophysical Studies Evaluating the Feasibility of a Speech Processing Strategy for a Multiple-Channel Cochlear Implant. Journal of al Society of Arnerica.74:73-80, Sydney, Auslralia 2-7 March 1997

5 Minerva Access is the Institutional Repository of The University of Melbourne Author/s: Clark, Graeme M. Title: Cochlear implant research directions Date: 1997 Citation: Clark, G. M. (1997). Cochlear implant research directions. In Cochlear Implants: Otohinolaryngology, Sydney, N.S.W. Persistent Link: File Description: Cochlear implant research directions Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.