The plasticity of the human perceptual systems in processing auditory- and spatial-frequencies

Abstract

A common technique in research on visual and auditory perception is to embed stimuli in noise. The traditional assumption is that, when noisy stimuli are repeatedly presented to an observer, each noise component remains equally unpredictable.
I challenged that assumption by running two experiments - with visual and auditory stimuli, respectively - under different noise conditions, some of which give the visual (auditory) system the opportunity to learn characteristics of the noise - a form of perceptual learning.

The visual stimuli that I used were Gabor patches - spatial patterns that varied sinusoidally in luminance across the image. The visual system is thought to processes spatial patterns through a series of independent channels, each functioning as a band-pass filter centred around a specific spatial-frequency and thus tuned to a sinusoidal pattern of that frequency. The main property to be determined for each selective channel is its width on the spatial-frequency axis. This has been done by using noise masking, i.e. embedding the sinusoid in notched noise - noise from which certain frequency components have been removed - and gradually increasing the width of the notch until no further improvement in threshold is obtained.
The auditory stimuli were defined in analogy with the visual ones.

My prediction was that the estimates of channel widths derived via noise masking will depend on the way in which the noisy stimuli are presented.
My results confirmed that perceptual learning does occur in vision, and to a lesser degree in hearing, thus suggesting that the auditory channels are more "rigid" than those in vision. The sense in which the visual system is deemed to be more "plastic" is discussed in detail.

I conclude that context can influence data obtained in notched-noise masking paradigms, and I discuss the possible implications that this might have in other psychophysical investigations.