Full text
Computational modelling of the auditory system to unravel tinnitus mechanisms oTinnitus is the perception of sound without an external source. The underlying mechanisms remain difficult to explain due to its heterogeneity. oComputational modelling provides a way to test hypothesized mechanisms by linking cellular dynamics to network-level activity. oProject goal: develop a biophysically inspired and experimentally validated model of the auditory pathway of both healthy controls and tinnitus patients. oFirst step: modelling of the auditory brainstem. Auditory Brainstem Response (ABR) oNon-invasive recording via EEG, elicited by brief acoustic stimuli. oMost prominent peaks of the ABR are: •Wave I →generated by spiral ganglion cells (SGC) •Wave III →generated by cochlear nucleus (CN) bushy cells •Wave V →generated by inferior colliculus (IC) neurons oA biophysical dipole model of the ABR connects neural simulations to scalp-recorded signals, enabling direct evaluation of tinnitus-related mechanisms. INTRODUCTION Towards a biophysical dipole model for ABR generators WAVES, DEPT. OF INFORMATION TECHNOLOGY, GHENT UNIVERSITY Sarah Vandepitte, Thomas Tarnaud, Emmeric Tanghe, Sarah Verhulst Conversion from sound wave to neuron activity The first stage of the model converts acoustic input into neuronal membrane potential signals, using the Verhulst et al. (2018) [1] auditory periphery framework. Neural activity propagation along the auditory nerve and brainstem To simulate the ABR at the network level, auditory neurons are modelled with biophysical and morphologically realistic properties, using the NetPyNE package and NEURON simulator. Spiral ganglion cells are implemented with a multicompartment Hodgkin–Huxley framework (Figure 3). This realistic cellular implementation forms the basis for dipole calculations and ABR waveform generation. METHODS sarah.vandepitt[email protected] Tonotopy and phase locking of the SGC Figure 4 illustrates the response of the auditory pathway model to pure-tone stimuli of different frequencies. oThe input signal is decomposed into centre frequencies (CFs), each processed by a parallel ascending pathway. The figure shows CFs of 500 Hz, 1 kHz, 2 kHz, 4 kHz, and 8 kHz; the full model comprises 201 CF channels spanning the entire tonotopic axis. oTonotopy: neurons respond most strongly when CF matches stimulus frequency. oPhase locking: at low and mid frequencies, firing patterns synchronize with input waveform. oCurrent limitation: modelled cells show reduced temporal precision, unable to fully track the input frequencies. ABR wave I Figure 5 shows the simulated dipole response of the spiral ganglion cell population to a 1 ms click stimulus at 80 dB. oThe broadband stimulus evokes highly synchronous firing across the SGCs. oSummed transmembrane currents generate a strong dipole moment, corresponding to the biophysical correlate of ABR wave I. oFor comparison with experimental ABR recordings, the dipole must be projected to the scalp via forward modelling within the NetPyNE framework. RESULTS CN bushy cells and IC neurons as ABR wave III and V generators oIncorporating biophysical models of the CN bushy cell and IC principal cell types improves the physiological accuracy of ABR simulations. oMulticompartment model of the CN bushy cells (Figure 6). oThe SGC →CN connection is called the Endbulb of Held synapse and enables bushy cells to preserve the timing and synchrony of the SGC input. oThis implementation will link cellular and synaptic mechanisms to specific ABR components (waves III and V)., aiding in the interpretation of how auditory processing –and conditions such as tinnitus –manifests in scalp-recorded ABRs. NEXT STEPS Figure 4: simulation results of the peripheral and SGC model for pure tone inputs. Note that the outliers in the SGC dipole plots are artifacts that still need to be removed. Figure 2: mean ABR waveform of a young normal hearing cohort, measured by Pauline Devolder from the Hearing Technology lab. Figure 3: morphology of the multicompartmental SGCs. Adapted from [2] Figure 6: morphology of the CN bushy cell. Adapted from [3]. Figure 5: SGC population dipole for a 1 ms click of 80 dB. Note that the outliers are artifacts that still need to be removed. [1] S. Verhulst et al. (2018). Hearing research, 360, 55–75. [2] J. Woo et al. (2009). IEEE transactions on bio-medical engineering, 56(9), 2177–2180. [3] P. B. Manis & L. Campagnola (2018). Hearing research, 360, 76–91. REFERENCES Figure 1: Schematic overview of the auditory pathway. AN: auditory nerve, CN: cochlear nucleus, IC: inferior colliculus, TRN: thalamic reticular nucleus, SOC: superior olivary complex.
