scholarly journals Memory Image Completion: Establishing a task to behaviorally assess pattern completion in humans

Hippocampus ◽  
2018 ◽  
Vol 29 (4) ◽  
pp. 340-351 ◽  
Author(s):  
Paula Vieweg ◽  
Martin Riemer ◽  
David Berron ◽  
Thomas Wolbers
2018 ◽  
Author(s):  
Paula Vieweg ◽  
Martin Riemer ◽  
David Berron ◽  
Thomas Wolbers

AbstractFor memory retrieval, pattern completion is a crucial process that restores memories from partial or degraded cues. Neurocognitive aging models suggest that the aged memory system is biased toward pattern completion, resulting in a behavioral preference for retrieval over encoding of memories. While there are behavioral tasks to assess the encoding side of these memory differences, pattern completion has received less attention in the literature. Here, we built on our previously developed behavioral recognition memory paradigm – the Memory Image Completion task (MIC) – a task to specifically target pattern completion. First, we used the original design with concurrent eye-tracking in order to rule out perceptual confounds that could interact with recognition performance. Second, we developed parallel versions of the task to accommodate test settings in clinical environments or longitudinal studies. The results show that older adults have a deficit in pattern completion ability with a concurrent bias toward pattern completion – a replication of previously found effects. Importantly, eye-tracking data during encoding could not account for age-related performance differences. At retrieval, spatial viewing patterns for both age groups were more driven by stimulus identity than by response choice, but compared to young adults, older adults’ fixation patterns overlapped more between stimuli that they (wrongly) thought had the same identity. This supports the observation that when making errors older adults choose responses perceived as similar to the correct stimulus, which is interpreted as a bias toward pattern completion. Additionally, two shorter versions of the task yielded comparable results, and no general learning effects were observed for repeated testing. Together, we present evidence that the MIC is a reliable behavioral task that targets pattern completion, that is easily and repeatedly applicable, and that is made freely available online.


2021 ◽  
Vol 11 (2) ◽  
pp. 624
Author(s):  
In-su Jo ◽  
Dong-bin Choi ◽  
Young B. Park

Chinese characters in ancient books have many corrupted characters, and there are cases in which objects are mixed in the process of extracting the characters into images. To use this incomplete image as accurate data, we use image completion technology, which removes unnecessary objects and restores corrupted images. In this paper, we propose a variational autoencoder with classification (VAE-C) model. This model is characterized by using classification areas and a class activation map (CAM). Through the classification area, the data distribution is disentangled, and then the node to be adjusted is tracked using CAM. Through the latent variable, with which the determined node value is reduced, an image from which unnecessary objects have been removed is created. The VAE-C model can be utilized not only to eliminate unnecessary objects but also to restore corrupted images. By comparing the performance of removing unnecessary objects with mask regions with convolutional neural networks (Mask R-CNN), one of the prevalent object detection technologies, and also comparing the image restoration performance with the partial convolution model (PConv) and the gated convolution model (GConv), which are image inpainting technologies, our model is proven to perform excellently in terms of removing objects and restoring corrupted areas.


Author(s):  
Iddo Drori ◽  
Daniel Cohen-Or ◽  
Hezy Yeshurun
Keyword(s):  

Optik ◽  
2014 ◽  
Vol 125 (17) ◽  
pp. 4985-4989 ◽  
Author(s):  
Hao Wu ◽  
Zhenjiang Miao

2015 ◽  
Vol 159 ◽  
pp. 157-171 ◽  
Author(s):  
Hao Wu ◽  
Zhenjiang Miao ◽  
Yi Wang ◽  
Jingyue Chen ◽  
Cong Ma ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document