Past work indicates that attended visual verbal stimuli are automatically encoded together with their spatial location but not vice versa. A new study by Elsley and Parmentier confirms this observation and demonstrates that the automatic binding of the "where" to the "what" remains stable during at least 15 seconds (the longest interval used in the study).
Schematic illustration of trial types in the Verbal Task (VT) and the Spatial Task (ST), with corresponding correct responses. Participants were presented with an array of consonants in distinct locations. In the verbal task participants were instructed to remember the consonants. In the spatial task, they remembered the locations. Following a visual mask and a retention interval of variable duration, they were asked to judge a probe consonant and press a "yes" key if the probe represented one of the to-be-remembered consonants (verbal task) or locations (spatial task), and to press the "no" key otherwise.
Error data (% incorrect) for intact and recombined probes as a function of retention interval in the verbal task (panel A) and spatial task (panel B). Error bars represent one standard error of the mean. The results show that probes made of a consonant presented in its original location (intact probe) were better remembered than probes made of a consonants and locations that formed part of the array but not paired (recombined probes). This binding effect was stable across all retention intervals. It was however absent in the spatial task (where participants attended the locations only).
Reference: Elsley, J. V., & Parmentier, F. B. R. (in press). The asymmetry and temporal dynamics of incidental letter-location bindings in working memory. The Quarterly Journal of Experimental Psychology.