Combining sensory information from two separate critical bands

Rickert, Martin Erhard (1987) Combining sensory information from two separate critical bands. Masters thesis, Memorial University of Newfoundland.

[English] PDF (Migrated (PDF/A Conversion) from original format: (application/pdf)) - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (4MB) [English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. (Original Version)

Abstract

This study investigated how subjects combine auditory signals and noise from two separate frequency-selective channels, or critical bands. In the first experiment, listeners were trained to detect a 0.5 kHz sinusoidal tone in a continuous background of noise. Then, without informing listeners, a 1.3 kHz tone was added to the signal. Initially the detectability of the two-tone complex was no better than the detectability of the 0.5 kHz signal, but it improved in subsequent sessions. Following this, the 1.3 kHz signal was removed, and performance dropped to, or below, the original level. These results indicated that human listeners have the ability to use either a single- or multiple-band listening strategy. In order to determine more about how information from widely separated critical bands is combined, two mathematical models were considered. The information integration model (Green, 1956) assumes that signal and noise energy from individual frequency-selective channels is statistically combined prior to decision. By contrast, the decision threshold model (Schafer and Gales, 1949) postulates that independent decisions are made about the information in each channel, and that the outcomes of the decisions are combined according to an overall decision rule. Data from the first experiment are consistent with the information integration model, but are contrary to predictions made using the decision threshold model. In order to further distinguish between these models, a second, more sensitive, experiment was designed. In this experiment the 0.5 kHz, 1.3 kHz, and combined signals were randomly mixed within the same block of trials. Therefore, subjects presumably used a similar listening strategy for each of the signals. Data from this experiment fit the neither the information integration model nor the decision threshold model. Furthermore, comparisons of detection performance between experiments argue against each of these models. In conclusion, while this study shows that neither the information integration model nor the decision threshold model explains how listeners combine information from two critical bands, it demonstrates that human listeners have the ability to switch from using one critical band to using two.

Actions (login required)

View Item