INVESTIGATION ON NEURAL RESPONSES RELATED TO THE LOCALIZATION OF NATURAL SOUNDS

Spatial hearing allows the localization of sounds in complex acoustic environments. There is considerable evidence that this neural system rapidly adapts to changes in sensory inputs and behavioral goals. However, the mechanisms underlying this context-dependent coding are not well understood. In fact, previous studies on sound localization have mainly focused on the perception of simple artificial sounds, such as white-noise or pure tone bursts. In addition, previous research has generally investigated the localization of sounds in the frontal hemicircle while ignoring rear sources. However, their localization is evolutionary relevant and may show different neural coding, given the inherent lack of visual information. Here we present a pilot electroencephalography (EEG) study to identify robust indices of sound localization from participants listening to a short natural sound from eight source positions on the horizontal plane. We discuss a procedure to perform a within-subject classification of the perceived sound direction. Preliminary results suggest a pool of discriminative subject-specific temporal and topographical features correlated with the characteristics of the acoustic event. Our preliminary analysis has identified temporal and topographical features that are sensitive to spatial localization, leading to significant decoding of sounds direction for individual subjects. This pilot study adds to the literature a methodological approach that will lead to the objective classification of natural sounds location from EEG responses.