Disruption of visual short-term memory by changing-state auditory stimuli: The role of segmentation

Abstract

Typically, serial recall performance can be disrupted by the presence of an irrelevant stream of background auditory stimulation, but only if the background stream changes over time (the auditory changing-state effect). It was hypothesized that segmentation of the auditory stream is necessary for changing state to be signified. In Experiment 1, continuous random pitch glides failed to disrupt serial recall, but glides interrupted regularly by silence brought about the usual auditory changing-state effect. In Experiment 2, a physically continuous stream of synthesized vowel sounds was found to have disruptive effects. In Experiment 3, the technique of auditory induction showed that preattentive organization rather than critical features of the sound could account for the disruption by glides. With pitch glides, silence plays a preeminent role in the temporal segmentation of the sound stream, but speech contains correlated time-varying changes in frequency and amplitude that make silent intervals superfluous.Irrelevant background speech disrupts serial short-term memory for material presented visually (see D. M. Jones & Morris, 1992, in press, for reviews). A reasonably clear picture of the features of the task and of the speech that elicits this effect is now emerging, as are the general implications for our understanding of attentional factors in memory (see D. M. Jones & Broadbent, 1991). Two views of this disruption are contrasted here. One is that the extent of disruption depends on the degree to which the sound resembles speech. Salame and Baddeley (1989) have proposed that the effects of speech and nonspeech sounds on serial recall could be understood in terms of a filter or detector system that selectively passes into memory sounds that resemble speech. We wish to put forward an alternative model, the changing-state hypothesis, whose pivotal feature is that the sound has to show particular variation over its time course (see D. M. Jones, in press). Our primary concern in the present paper is to refine the notion of change within the auditory changing-state concept; our secondary purpose is to assess the view that sound has to be "speech-like" before it disrupts serial recall. The purpose of the experimental work reported here is to identify cues common to speech and nonspeech stimuli that nevertheless may serve as the basis for disruption. Specifically, interest is attached to the role of segmentation in signifying changing state, and to the ways this may lead to disruption of recall by both speech and Thanks are due Richard Blight for his general help in setting up the equipment. Helen Bowman and Craig Steel ran some of the subjects.

The impact of attention load on the use of statistical information and coarticulation as speech segmentation cues

“…Whereas in that study, visual TP computation was penalized by the concurrent presentation of changing speech material, in the present study speech segmentation afforded by TP computation was penalized by the concurrent presentation of a visual stream (RVSP) of pictures. These results could be in line with the interference-by-process hypothesis (e.g., Jones, 1999;Jones et al, 1993;Marsh et al, 2009), in that both visual sequence and speech material would require seriation, which resulted in interference in processing of the critical material (e.g., the visual sequence in Neath et al, 2009; the speech material in the present study). Accordingly, the key determinant of disruption in spoken word processing could be the extent to which both the irrelevant, distractor information and spoken word processing share similar processes.…”

“…It is in line with previous results showing that statistically driven segmentation can occur even in conditions in which attention is not fully allocated to the AL stream (e.g., Toro et al, 2005; see also Saffran, Newport, Aslin, Tunick, & Barrueco, 1997, who observed ALL with a concurrent drawing task), and with listeners who cannot be formally instructed to attend to the speech stream (human infants-see, e.g., Saffran, Aslin, & Newport, 1996-or nonhuman primates-see, e.g., Hauser, Newport, & Aslin, 2001). More generally, these results converge with evidence showing that the speech signal (or other dynamic and complex auditory stimuli; see, e.g., Jones, Macken, & Murray, 1993) is processed even when participants are asked to ignore it, as demonstrated, for example, by the irrelevant sound effect (or irrelevant speech effect, cf. Colle & Welsh, 1976; for a review, see, e.g., Jones, 1999) observed in short-term memory tasks and more recently in visual statistical learning (Neath, Guérard, Jalbert, Bireta, & Surprenant, 2009).…”

“…It could be the case that the interference-by-process hypothesis of Dylan Jones and colleagues (e.g., Jones, 1999;Jones et al, 1993;Marsh, Hughes, & Jones, 2009) would extend its application beyond working memory to the speech domain.…”

The psychoacoustics of the irrelevant sound effect

“…3.2). What is further noteworthy, is that (in contrast to [27]) even non-interrupted FM appears to follow that pattern generating a small performance decrement (by some 11%) near that frequency.…”

The irrelevant-speech effect and children: theoretical implications of developmental change

“…Following the work of Salamé and Baddeley (1982), evidence that varied speech or nonspeech sounds were more disruptive to memory performance than sounds that remained the same (Jones, Madden, & Miles, 1992) led to the emergence of another model of the effects of irrelevant sounds on performance, the object-orientedepisodic record model (O-OER; Jones, Macken, & Murray, 1993). Two central tenets of the O-OER are the changing-state hypothesis and the equipotentialityhypothesis.…”