Computational Neuroscience and Neuroengineering
Understanding how nervous system works is a key question studied for decades. Despite many efforts it is challenging to investigate all these processes in in vivo or in vitro experiments. For this reason, in silico approaches are desired e.g. to help with development of neuroprostheses. Up to now, electrical cochlear implant (eCI) is the most successful neuroprosthesis restoring hearing to around 1 million people worldwide. Despite that, hearing rehabilitation with eCIs is far from normal hearing due to limited spectral resolution. This limitation is set by a spread of electrical excitation from each electrical contact inside a cochlea that is filled with a conductive fluid leading to channel interaction. Although most of eCI users achieve fair open-set speech perception in the quiet, performance in complex listening tasks (i.e. speech recognition in noisy and/or reverberant environments) and music appreciation may be limited. This makes optical cochlear implant (oCI) a candidate to overcome spectral resolution limitation with the use of light that can be confined in space allowing for a greater number of non-interacting channels.
Currently the group is focused on investigation how auditory system works in order to develop next generation optical CIs on a hardware and a software level. On a software level it involves establishment of novel sound coding strategies that transfer sound into a stimulation of SGNs to enable hearing in deaf. Optical sound coding strategy that capitalises on the new stimulation modality requires fine-grained, fast, and power-efficient real-time sound processing and control of multiple microscale optical emitters. Such algorithm in case of eCI would be tested in human patients. However, oCI as a non-clinical concept requires an in silico framework. For that we develop computational models constrained by experimental data obtained from close collaboration within the Institute for Auditory Neuroscience.