September 2021
Volume 21, Issue 9
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2021
Capacity and Allocation across Sensory and Short-Term Memories
Author Affiliations
  • Shaoying Wang
    University of Denver
  • Srimant P. Tripathy
    University of Bradford
  • Haluk Ogmen
    University of Denver
Journal of Vision September 2021, Vol.21, 1863. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Shaoying Wang, Srimant P. Tripathy, Haluk Ogmen; Capacity and Allocation across Sensory and Short-Term Memories. Journal of Vision 2021;21(9):1863.

      Download citation file:

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

  • Supplements

Human memory consists of Sensory Memory (SM), Short-Term Memory (STM), and Long-Term Memory. SM enables a large capacity but decays rapidly. STM has limited capacity but lasts longer. Previously, it was shown that major bottlenecks for motion processing exist prior to STM (Ogmen et al, 2013. PLoS ONE) and that the contents of SM are allocated exclusively to the current event segment (Tripathy & Ogmen, 2018. Frontiers in Psychology). Here we used mixture modeling to test if the modeling results obtained in Ogmen et al., (2013) also hold for the dataset from Tripathy & Ogmen (2018) in which stimuli contained one to four disks moving in different directions. Each disk changed its direction at the mid-point of its trajectory. The synchronized deviations indicated an event boundary. Observers were instructed to partially or fully report the directions of disks, randomly selected from one of the two event-segments. Experiment 1 varied set-size with cue-delay set to 0 ms. A mixture model consisting of a Gaussian (Memory) and a Uniform (Guessing) distributions found that intake (fraction of target items retained in memory) and precision (1/Standard-Deviation) decreased with increasing set-size. In Experiment 2, set-size was fixed at 3, with a varying cue-delay. Intake dropped gradually and substantially from stimulus encoding to SM and STM, but precision dropped only moderately across memory stages. These results are in agreement with the mixture analysis in Ogmen et al., (2013). We also asked if a mixture model containing two, as opposed to one, Gaussians would capture better the data. Model comparison based on the Akaike Information Criterion favored the two Gaussian model. In summary, we generalized previous findings about memory capacity and bottlenecks using a different data set and propose a mixture model with two Gaussians as a better statistical model than a single Gaussian version.


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.