Author(s): Senan Toros Ufuk, Kohlrausch Armin, Jelfs Sam
Summary:
Short-term memory disruption ascribed to the nature of sound has been studied under the paradigm of "``irrelevant sound,'' where test participants try to perform cognitive tasks in the presence of background noise. By analyzing the test scores for different acoustic stimuli in relation to a reference condition from such experiments, the "irrelevant sound (speech) effect" (ISE) can be quantified. The ISE is thought to result from the interference of two parallel ordering processes; one for the deliberate processing of visually presented items which the subject needs to memorize, and one for the involuntarily processing of the acoustically presented irrelevant sound. This interference is observed as a decrease in memory performance and is often explained by the changing state hypothesis: The background sound must be separable into tokens where successive tokens are perceptually different. We analyze a spectral predictor, called frequency domain correlation coefficient (FDCC), for its ability to predict this interference process. The definition of the FDCC follows the idea of the changing-state hypothesis: The sound is divided into tokens and the change between successive states is quantified by computing the correlation coefficient between the power spectra of the successive tokens. Previous research has shown that the FDCC was able to predict the serial-recall results to a certain extent but also demonstrated limitations. The present study investigates those limitations and attempts to improve the parameter's prediction accuracy by integrating a peak detection phase into the token segmentation stage of the algorithm. The improved metric is evaluated by employing a large set of irrelevant sound stimuli from the ISE literature and comparing the experimental results with prediction values of this spectral parameter.
Name: Mr Toros Ufuk Senan
e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Country: Netherlands