Perception of “hidden” auditory figures

Abstract

Periodic pulse sequences were “hidden” amidst other periodic pulse sequences within auditory pulse trains. The generating rule was to emit a pulse for the occurrence of any one of up to ten prime counts of a 10-kHz pulse train. The absolute detection, the relative discrimination, and the identification of such sequences were examined for trains of up to 10 prime components. Two extreme hypotheses may be identified: a “masking” hypothesis, based upon the greater masking by lower auditory frequencies than by higher auditory frequencies; and an “extraction” hypothesis, in which periodic components are extracted in terms of the number of occurrences of each component. The former suggests that performance should be best with low-frequency information (large counts); the latter suggests that performance should be best with high-frequency (small counts). The data favor the latter hypothesis: performance is most closely related to the rank-ordering of the to-be-examined component within its presented series.