September 2015
Volume 15, Issue 12
Vision Sciences Society Annual Meeting Abstract  |   September 2015
Breaking Visual Working Memory: Cases of In/dependence between Storage and Manipulation Costs
Author Affiliations
  • Hrag Pailian
    The Johns Hopkins University
  • Justin Halberda
    The Johns Hopkins University
Journal of Vision September 2015, Vol.15, 1295. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Hrag Pailian, Justin Halberda; Breaking Visual Working Memory: Cases of In/dependence between Storage and Manipulation Costs. Journal of Vision 2015;15(12):1295. doi:

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

To what extent are storage and manipulation abilities in visual working memory (VWM) in/dependent? We answer this question by investigating whether factors known to influence VWM storage (memory load, object complexity, precision of representations) produce similar vs. different costs in VWM manipulation. To this end, we developed a novel task that is similar to the shell game. Objects were presented briefly and subsequently hidden behind occluders. These occluders either remained stationary (resembling change detection tasks used to measure VWM storage) or swapped positions a number of times. Since swap conditions required observers to update the featural-spatial bindings of object representations, they allowed us to measure manipulation ability and its interaction with factors, such as memory load, information complexity, and precision of representations. In varying memory load (Exp 1a), we found that participants were able to manipulate only 2 objects without incurring costs, despite being able to store up to 3-4 objects. These manipulation costs were not attributable to temporal decay, interference effects, or limited encoding (Exp’s 1b-d). Furthermore, increasing the complexity of these objects (Exp 2) produced additional costs in storage, but not in manipulation. Similarly, the level of precision with which items were stored in VWM was the same for all manipulated representations (Exp 3). Interestingly, the probability with which these manipulated items remained in memory remained unchanged for only 2 items, but decreased as a function of swaps for higher memory loads. These results inform our conceptualization of the VWM system by providing insight into the in/dependence of storage and manipulation abilities. Factors constraining storage abilities limit manipulation, to the extent that they influence the input on which manipulation computations are performed (dependence: Exp’s 2,3). Nonetheless, these factors are separate from those that constrain the manipulation computations themselves (independence: Exp 1).

Meeting abstract presented at VSS 2015


This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.