Previous studies have demonstrated the various benefits of multisensory interactions (Chen & Spence,
2010; Mendonça, Santos, & López-Moliner,
2011; Parise, Spence, & Ernst,
2012; Seitz, Kim, & Shams,
2006; Shams & Seitz,
2008; van Atteveldt, Murray, Thut, & Schroeder,
2014), but only a few studies have explored the disruptive influences of such interactions. Exceptions include studies of spatial ventriloquism (McGurk & MacDonald,
1976) and other visual influences on auditory perception (Teramoto, Kobayashi, & Hidaka,
2013), as well as various demonstrations that auditory signals alter visual perception (Hidaka et al.,
2009; Sekuler et al.,
1997; Shams, Kamitani, & Shimojo,
2000; Shipley,
1964; Teramoto, Hidaka, & Sugita,
2010; Zhou et al.,
2007). However, those previous studies focused on relatively low-level audiovisual tasks; none targeted a higher-order cognitive function such as memory. In order to understand how task-irrelevant auditory signals affected short-term memory, our
first experiment investigated how nominally ignored auditory sequences influenced categorization of concurrent visual sequences. Our
second experiment used Hebb's (
1961) repetition effect to examine the buildup of this influence (incidental learning) over many trials and the transfer of information from short-term memory restricted to the stimuli within a single trial, to longer-term memory that spanned multiple trials. As Sperling and Dosher (
1986) have noted, in higher-order tasks, subjects' strategies enjoy an expanded influence on performance. Because our study's task recruited higher order processes such as selective attention and memory, it was able to reveal some features of the strategies that subjects call upon. For example,
Experiment 2 showed that prior experience (namely, musical training), affected subjects' reliance on task-irrelevant information. Such results confirm the usefulness of studying audiovisual interactions in a task that draws upon top-down as well as bottom-up processing.