Thomas Lenarz, J. Michael Herrmann

Complex sound encoding schemes in the Inferior Colliculus

One overall goal of the project is to advance auditory midbrain implants and furthermore to develop a stimulation strategy that can be adaptively controlled by the perceptual responses of the patient.

Auditory implants, particularly the cochlear implant [Zeng, 2004], can reestablish hearing for deaf humans through electrical activation of an intact acoustic nerve. Although hearing performance is not yet to the level of normal hearing individuals, many cochlear implant patients receive sufficient information to understand speech without lip-reading cues and are even able to converse over the telephone. For persons with a damaged acoustic nerve, however, the cochlear implant becomes ineffective.The research group of T. Lenarz from Hannover Medical University (MHH) has developed a new central auditory prosthesis, the auditory midbrain implant (AMI) [Lenarz, et al., 2006; Lim, et al. 2007], which is implanted into the inferior colliculus (IC) of the midbrain. Presently the AMI contains a single shank electrode array and utilizes the encoding scheme developed for cochlear implants. Since the IC has a three dimensional structure, it is expected that a 3-D stimulation array would achieve better perception of sounds. However, the spatial and temporal encoding of the human IC are still not well understood to determine the optimal stimulation strategy.In this project the auditory implant group from the MHH and the nonlinear dynamics group  from the Max Planck Institute for Dynamics and Self-organization (MPI DS) are collaborating.At MHH, animal studies are carried out to determine the spatio-temporal mapping of sounds to the neuronal activity in the IC. At MPI DS, statistical analyses of speech signals and neural responses are performed for feature extraction as well as neural modelling of early speech processing. Classification methods, such as clustering and independent component analysis, are applied to the experimental neural data to extract activity patterns in the IC representing certain features of complex acoustic stimuli. We seek to obtain a better functional understanding of the encoding mechanism with the central auditory system, which will serve to improve the currently applied encoding scheme of the AMI. To further optimize the AMI stimulation strategies in individual patients, a feedback loop will be used. In this loop, the perceptions of the AMI-patient to specific electrical stimuli will be compared to the optimal percepts (e.g.,those achieved by a normal hearing person) and the electrical parameters will be varied iteratively until the best possible match is obtained. This would allow in the future to adapt the stimulation scheme for the patients individually,- that can depend on the degree of auditory damage, location of the implant and other patient-specific properties.