Figure - PMC

Skip to main content

An official website of the United States government

Here's how you know

Here's how you know

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

View full-text article in PMC

. 2019 Oct 30;10:4934. doi: 10.1038/s41467-019-12623-6

Search in PMC
Search in PubMed
View in NLM Catalog
Add to search

© The Author(s) 2019

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

PMC Copyright notice

Fig. 4 — Receptive field reconstructions of two example model units. Two out of the seven model units that were found to significantly predict responses of single face contacts in a leave-one out search (see Supplementary Fig. 12 for the full set). Model units were defined as units which were best correlated with a single iEEG contact responses to all N−1 exemplar sets and significantly predicted the held out responses based on a linear fit. a Model unit found in layer pool5 for set 1. Top left scatter plot depicts the responses of the model unit versus the mean HFA responses (50–500 ms) of the face contact. Red dots correspond to the three exemplars for which receptive field visualizations are presented (see below). Red line is the least-squares linear fit. Top right plot depicts the mean categorical responses of the face-selective contact to which the model unit matched. Bar inset presents the categorical selectivity of the model unit, which responded solely to faces and not to places or patterns. All error bars denote s.e.m. across exemplars. Bottom gallery presents visualizations of the model unit’s receptive field for three example face exemplars: left column shows the original images in a descending order of the model unit’s response; middle column shows the reconstructed receptive field using deconvolution; right column shows the delta between the altered image following activation maximization and the original image, reflecting the added image that maximally amplified the model unit’s response. For visualization purpose, the contrast of each receptive field image was normalized by stretching the active range of RGB values. b Same as panel a, depicting a different model unit detected in layer pool5 for set 1 and its matching face contact. Note the view point selectivity and consistency of the face fragments that are highlighted by the two visualization procedures, and the large expanse of image changes following activation maximization. Copy rights holders: Woodie Allen photo—Colin Swan, license link https://creativecommons.org/licenses/by-sa/2.0/; Christopher Falkenberg photo—Christopher Falkenberg; Gila Gamliel photo—https://en.wikipedia.org/wiki/Gila_Gamliel#/media/File:Gila_Gamliel.jpg, license link https://creativecommons.org/licenses/by-sa/3.0/deed.en; Barack Obama photo—Pete Souza, license link https://creativecommons.org/licenses/by/3.0/deed.en