{"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/toward-goal-driven-neural-network-models-for","title":"Toward Goal-Driven Neural Network Models for the Rodent Whisker-Trigeminal System","arxiv_id":"1706.07555","date":"2017-06-23","proceeding":"NeurIPS 2017 12","authors":["Chengxu Zhuang","Jonas Kubilius","Mitra Hartmann","Daniel Yamins"],"abstract":"In large part, rodents see the world through their whiskers, a powerful\ntactile sense enabled by a series of brain areas that form the\nwhisker-trigeminal system. Raw sensory data arrives in the form of mechanical\ninput to the exquisitely sensitive, actively-controllable whisker array, and is\nprocessed through a sequence of neural circuits, eventually arriving in\ncortical regions that communicate with decision-making and memory areas.\nAlthough a long history of experimental studies has characterized many aspects\nof these processing stages, the computational operations of the\nwhisker-trigeminal system remain largely unknown. In the present work, we take\na goal-driven deep neural network (DNN) approach to modeling these\ncomputations. First, we construct a biophysically-realistic model of the rat\nwhisker array. We then generate a large dataset of whisker sweeps across a wide\nvariety of 3D objects in highly-varying poses, angles, and speeds. Next, we\ntrain DNNs from several distinct architectural families to solve a shape\nrecognition task in this dataset. Each architectural family represents a\nstructurally-distinct hypothesis for processing in the whisker-trigeminal\nsystem, corresponding to different ways in which spatial and temporal\ninformation can be integrated. We find that most networks perform poorly on the\nchallenging shape recognition task, but that specific architectures from\nseveral families can achieve reasonable performance levels. Finally, we show\nthat Representational Dissimilarity Matrices (RDMs), a tool for comparing\npopulation codes between neural systems, can separate these higher-performing\nnetworks with data of a type that could plausibly be collected in a\nneurophysiological or imaging experiment. Our results are a proof-of-concept\nthat goal-driven DNN networks of the whisker-trigeminal system are potentially\nwithin reach.","url_abs":"http://arxiv.org/abs/1706.07555v1","url_pdf":"http://arxiv.org/pdf/1706.07555v1.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":"toward-goal-driven-neural-network-models-for","repo_url":"https://github.com/neuroailab/whisker_model","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"decision-making","task_name":"Decision Making"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1706.07555","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}