Auditory attentional bandwidth: effect of level and frequency range.

This study investigated the relations between theoretical auditory filters and "attentional filters" observed when measuring the detectability in noise of tones of expected or unexpected (probe) frequency. The effect of the level of the noise and the frequency range (narrow or wide) of the probes was assessed. For each 2IFC trial, a tone of the expected target frequency was presented as a cue preceding the two observation intervals--the signal was more often the expected than an unexpected frequency. For the narrow frequency range of probe frequencies, increasing the masker level resulted in a broadening of the probe-signal contour (percent correct as a function of probe frequency); however, detection was significantly better only for probe frequencies farthest from the target frequency. For the wide range of probe frequencies, the percentage of correct detections depended less systematically on masker level; however, detection was better and the attentional band wider than in the narrow frequency range. The results suggest that attentional focusing does not simply reflect auditory filtering: A more central and adaptive process may operate from the outputs of adjacent or more distant auditory filters.