Skip to content. | Skip to navigation

Personal tools
You are here: Home Forum Using data sets good neurons in pvc-8

good neurons in pvc-8

Up to Using data sets

good neurons in pvc-8

Posted by Yimeng Zhang at May 10. 2018


The pvc-8's associated paper, "Flexible Gating of Contextual Influences in Natural Vision.", only analyzes 207 or 126 neurons out of all the available data. I tried to follow the paper's Online Methods in order to know which 207 or 126 neurons were used in the paper, without much progress.

In particular, for the 207 subset, Online Methods claims that "1) The RF was centered on the stimulus... 2) The neuron responded above its spontaneous rate to at least 4 small natural images.". Neurons that satisfy condition 1) are explicitly flagged in the data; however, I couldn't implement condition 2) correctly to get those 207 neurons. Can the authors provide some demo script on this issue? I tried two interpretations of a neuron's "spontaneous rate", one calculated for each image (there are 540 different spontaneous response rates, one for each natural image) and the other calculated as the average of the 540 rates mentioned above. These rates were always calculated from last 50 bins of `train_resp_blk`.


Powered by Ploneboard
Document Actions