Purchase this article with an account.
Mingliang Gong, Yuming Xuan, Xiaolan Fu; The effect of consistency on scene short-term memory. Journal of Vision 2011;11(11):1133. doi: 10.1167/11.11.1133.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The term scene consistency refers to the probability that a given object naturally appears in a given context. For inconsistent object-context pairs this probability is low (e.g. a bird that flies underwater). Categorization and naming is faster and more accurate for consistent pairs, while fixation and detection of orientation changes is faster for inconsistent pairs. The present study examines the role of scene consistency in scene short-term memory in a change detection paradigm. Experiment 1 demonstrates that changes are detected more accurately when scene consistency changes from probable to improbable (or from improbable to probable) than when consistency does not change. However, a consistency-change can also be interpreted as a between-category change, which might be more easily detected than within-category changes. To test this alternative explanation, we presented objects and backgrounds separately in experiments 2 and 3. Results showed that participants detected within-category changes and between-category changes equally well when objects and backgrounds were presented in different frames (experiment 2). However, they detected between-category changes better than within-category changes when objects and backgrounds were presented in the same frame (experiment 3). In experiments 4 and 5, we conceptually replicated experiments 1 and 3 with monochrome stimuli instead of chromatic ones. In summary, this study demonstrates that scene consistency might play a role in scene short-term memory. However, the effects of scene consistency- and category-changes on change detection performance were not fully disentangled. Future research might differentiate between these two possible interpretations of our results.
This PDF is available to Subscribers Only