Abstracts of the British Society of Audiology annual conference (incorporating the Experimental and Clinical Short papers meetings)

Abstract

Neurons in the auditory pathway are topographically arranged by their response to sound frequency. This tonotopic organisation mirrors the distribution of receptors in the cochlea, with neurons at the base of the cochlea responding best to high frequencies, and neurons at the apex responding best to low frequencies. Previous functional magnetic resonance imaging (fMRI) results suggest that, in human auditory cortex, high frequencies are represented at the medial end and low frequencies in the antero-lateral end of Heschl ' s gyrus (Formisano et al, 2003; Humphries et al, 2010). While fMRI has been used to investigate the topography of best frequency (BF) in human auditory cortex, the selectivity, or bandwidth of tuning, for frequency of auditory cortical neurons remains unknown. This work was aimed at assessing the suitability of fMRI to measure frequency selectivity in human auditory cortex. Both conventional and adaptation-based fMRI paradigms were used.Blood oxygen level-dependent (BOLD) responses were recorded with high-resolution fMRI (1 ϫ 1 ϫ 1 mm ³ ) at 7 Tesla (Philips Achieva) using a sparse imaging protocol. Stimuli consisted of narrow-band noises (50-Hz bandwidth), centred around nine test frequencies that varied from 500 Hz to 8 kHz in half-octave steps. Frequency tuning curves of activated voxels were derived by assigning each voxel to the frequency to which it responded best for a given proportion of trials, and then averaging the frequency responses of voxels with the same best frequency for the remaining trials. The resulting tuning curves were fi tted with a bell curve to estimate their BF and tuning bandwidth.The voxel frequency tuning curves were remarkably reliable across trials. Voxels in the area of Heschl ' s gyrus, which is the site of primary auditory cortex in humans, tended to show sharper frequency tuning than voxels in non-primary. This would be expected based on neurophysiological data from animals. Furthermore, the tuning bandwidths also appeared to correlate with the local map structure, in that areas with homogenous frequency preference showed sharper tuning than areas with heterogeneous frequency preference. This dependency of frequency selectivity on local map structure may be a result of the spatial smoothness of BOLD effect.