{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/using-deep-learning-to-reveal-the-neural-code","title":"Using deep learning to reveal the neural code for images in primary visual cortex","arxiv_id":"1706.06208","date":"2017-06-19","proceeding":null,"authors":["William F. Kindel","Elijah D. Christensen","Joel Zylberberg"],"abstract":"Primary visual cortex (V1) is the first stage of cortical image processing,\nand a major effort in systems neuroscience is devoted to understanding how it\nencodes information about visual stimuli. Within V1, many neurons respond\nselectively to edges of a given preferred orientation: these are known as\nsimple or complex cells, and they are well-studied. Other neurons respond to\nlocalized center-surround image features. Still others respond selectively to\ncertain image stimuli, but the specific features that excite them are unknown.\nMoreover, even for the simple and complex cells-- the best-understood V1\nneurons-- it is challenging to predict how they will respond to natural image\nstimuli. Thus, there are important gaps in our understanding of how V1 encodes\nimages. To fill this gap, we train deep convolutional neural networks to\npredict the firing rates of V1 neurons in response to natural image stimuli,\nand find that 15% of these neurons are within 10% of their theoretical limit of\npredictability. For these well predicted neurons, we invert the predictor\nnetwork to identify the image features (receptive fields) that cause the V1\nneurons to spike. In addition to those with previously-characterized receptive\nfields (Gabor wavelet and center-surround), we identify neurons that respond\npredictably to higher-level textural image features that are not localized to\nany particular region of the image.","url_abs":"http://arxiv.org/abs/1706.06208v1","url_pdf":"http://arxiv.org/pdf/1706.06208v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"using-deep-learning-to-reveal-the-neural-code","repo_url":"https://github.com/jzlab/v1_predictor","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}