Modulation discrimination interference and comodulation masking release as a function of the number and spectral placement of narrow-band noise modulators.

Abstract

The discrimination of the depth of amplitude modulation (AM) of a target carrier can be adversely affected by the presence of other modulated carriers (flankers), an effect called modulation discrimination interference (MDI). Conversely, when the task is to detect a sinusoidal signal added to a modulated carrier of the same frequency, the presence of comodulated flankers (with the same modulation as the target carrier) can improve performance (comodulation masking release: CMR). This paper examines how CMR and MDI are influenced by the number and placement of the flankers. Flankers were spaced at 2-ERB intervals from the target, and their number was manipulated by adding flankers that were more remote from the target. Narrow-band noise (centered at 10 Hz) was used as the modulator for all carriers. In experiment 1, the MDI task required detection of a decrease in modulation depth (m) of the target from a reference value of 0.7. Unmodulated flankers did not affect thresholds. Modulated flankers (m = 0.15) produced MDI, which increased as more flankers were added. MDI was generally slightly greater for comodulated than for noncomodulated flankers. The CMR task required detection of a sinusoidal signal added (in 90 degrees phase) to the target carrier (m = 0.7). Unmodulated flankers or flankers that were noncomodulated with the target had no effect on threshold. Comodulated flankers (m = 0.7) gave CMR, which generally increased as more flankers were added. This effect was greater for flankers below the target frequency than for flankers above. Experiment 2 examined MDI using similar stimuli to experiment 1, except that the task was to detect an increase in m from a reference value of 0.1. Even unmodulated flankers elevated thresholds, and this effect increased as more flankers were added, especially on the high-frequency side of the target. Modulated flankers (m = 0.2) produced additional increases in threshold, and modulation thresholds generally increased as more flankers were added. However, the amount of MDI (the difference between thresholds with modulated and unmodulated flankers) did not vary systematically with the number of flankers and did not show consistent differences between comodulated and noncomodulated flankers. The results are interpreted in terms of the mechanisms underlying MDI and CMR, and especially perceptual grouping processes.