Human cochlear nerve model

The main purpose of the project is to gather new detailed anatomic and morphometric information about the afferent part of the human cochlear nerve (spiral ganglion) and to use this data in a new computational model for analyzing input output relations for both the acoustically and the electrically stimulated ear. Moreover, the pathways of the spiral ganglion cells will be investigated. The results are of considerable interest to the cochlear implant research in order to overcome shortcomings like speech understanding in a noisy environment. The neural elements responsible for signal transport from the cochlea to the brain show essential anatomic differences between man and mammalian species, leading to several non tested and unproved hypotheses concerning functional consequences. Currently the signaling strategies for cochlea prostheses are mainly based on single fiber recordings in cat. Due to missing fundamental investigations on the neural elements of the human cochlea neurons, even the latest signal processing strategies for cochlear implant users are de facto designed for cats instead of humans. Most striking for the uniqueness of the human spiral ganglion cells is the soma region, typically with (i) poor insulation by myelin and (ii) clusters containing several cell bodies. A goal of the project is to expand to a large extent our knowledge about the relevant ultra-structure. Using existing equipment and material (temporal bones from several individuals and a selection of inner ears from primates) allows to quantify human peculiarities. A comparison with primates should give a clue about the development of the unique morphological features in humans. As an alternative to single fiber recordings which are not available for humans, we will develop a general biophysical model for cochlear neurons including cluster properties in order to understand the functional consequences of neuron variability seen in mammals and humans by analyzing the influence of relevant geometric and electric parameters. With this model we will analyze how the three main auditory nerve coding principles (i-iii) in man are related with the specific human anatomy and morphometry and, additionally, we will include a discussion about consequences for cochlear implants. In more detail, the model analysis is concerned with (i) tonotopic organization: a shift in frequency mapping concerning distal fiber endings originating in the organ of Corti (almost 2 cochlear turns) and soma regions in Rosenthal canal (1 turns), (ii) temporal fine structure of the neural code: a) for short interspike intervals in the distal axon there is a possible loss of signals when passing the soma region and b) in case of electrical stimulation: a confusing bimodal distribution of delay times as a consequence of more than one spike initiation regions along a single cochlear neuron (iii) spontaneous spiking is a supporting mechanism for acoustic signals just above the hearing threshold resulting in a mix of spikes with and without temporal information about the acoustic input; clusters of several somas may act as filter elements as a first step in neural code processing.

The main purpose of the project was to gather new detailed anatomic and morphometric information about the afferent part of the human cochlear nerve (spiral ganglion) and to use this data in a new computational model for analyzing input output relations for both the acoustically and the electrically stimulated ear. Moreover, the pathways of the spiral ganglion cells were investigated. The results are of considerable interest to the cochlear implant research in order to overcome shortcomings like speech understanding in a noisy environment. The neural elements responsible for signal transport from the cochlea to the brain show essential anatomic differences between man and mammalian species, leading to several non-tested and unproved hypotheses concerning functional consequences. Currently the signaling strategies for cochlea prostheses are mainly based on single fiber recordings in cat. Due to missing fundamental investigations on the neural elements of the human cochlea neurons, even the latest signal processing strategies for cochlear implant users are de facto designed for cats instead of humans. Most striking for the uniqueness of the human spiral ganglion cells is the soma region, typically with poor insulation by myelin and clusters containing several cell bodies. Our analysis showed that in human (i) the cell somas are smaller than reported earlier, (ii) the central axon diameter is always nearly just 2 times the peripheral ones and (iii) there are more neural clusters than expected. These three observations are contrary to findings in cats. Furthermore, neural pathway tracing showed spiraled shapes for the apical neurons implying loss of the tonotopically related excitation for low frequency components when cochlear implants substitute the natural neural firing pattern. This effect is explained by a short distance of a target neuron of the auditory nerve to several electrode contacts when the implant is active in the upper cochlear turn.Furthermore, we analyzed the action potential (spike) conductance along the auditory nerve, that is, from the inner hair cell (auditory receptor cell) to the next processing region called cochlear nucleus in the brainstem. Evaluation of our biophysical excitation model demonstrated that the synaptic currents from the inner hair cells are about 15 times stronger than needed. Wasting this synaptic energy boosts spike initiation, which guarantees the rapid transmission of temporal fine structure of auditory signals. Auditory brainstem response measurements are used as clinical tool in order to test the auditory pathway in patients. In spite of its importance it is not clear which neural elements are the generators of the peaks in the recorded signal. Our computer modeling study give evidence that peak I of the human auditory brainstem response results from the somatic regions of type I cells in the auditory nerve.