In cases of profound cochlear dysfunction, electrical cochlear implants (eCIs) offer partial restoration of hearing. Nevertheless, the auditory experience with eCIs remains imperfect due to the broad excitation of auditory nerve neurons by electrical signals that limits the perceptual channels available to users. Optogenetics offers a breakthrough by enabling the stimulation of the auditory nerve by an optical cochlear implant (oCI). Light, unlike electrical signals, can be precisely confined in space, resulting in reduced spread of excitation and improved frequency selectivity. However, optogenetic stimulation of the auditory nerve differs from electrical stimulation in several key aspects, including lower temporal fidelity and higher frequency resolution. Consequently, sound coding strategies used to deliver electrical currents in eCIs cannot be directly applied to oCIs.The current project endeavors to devise tailored strategies for delivering light to the optogenetically modified auditory nerve via an optical cochlear implant (oCIs). Specifically, our approach involves training convolutional neural networks to predict responses in the Inferior Colliculus to both optical and acoustical stimulation. Subsequently, we leverage these models to perform cross-modal optimization of the optical sound coding strategy, aiming to evoke responses in the Inferior Colliculus that closely resemble those produced by auditory stimuli. Inferior colliculus responses are recorded from anesthetized adult gerbils implanted with a blue creeLED implant featuring 5-10 channels, using a 32 channel Neuronexus probe. Animals were injected with AAVs carrying Channelrhodopsin transgenes into the round window postnatally (p8).