Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results
Deep neural networks have been shown to predict neural responses in higher visual cortex. The mapping from the model to a neuron in the brain occurs through a linear combination of many units in the model, leaving open the question of whether there also exists a correspondence at the level of indivi...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Technical Report |
| Language: | en_US |
| Published: |
Center for Brains, Minds and Machines (CBMM)
2018
|
| Online Access: | http://hdl.handle.net/1721.1/118847 |
| _version_ | 1826189380243423232 |
|---|---|
| author | Arend, Luke Han, Yena Schrimpf, Martin Bashivan, Pouya Kar, Kohitij Poggio, Tomaso DiCarlo, James J. Boix, Xavier |
| author_facet | Arend, Luke Han, Yena Schrimpf, Martin Bashivan, Pouya Kar, Kohitij Poggio, Tomaso DiCarlo, James J. Boix, Xavier |
| author_sort | Arend, Luke |
| collection | MIT |
| description | Deep neural networks have been shown to predict neural responses in higher visual cortex. The mapping from the model to a neuron in the brain occurs through a linear combination of many units in the model, leaving open the question of whether there also exists a correspondence at the level of individual neurons. Here we show that there exist many one-to-one mappings between single units in a deep neural network model and neurons in the brain. We show that this correspondence at the single- unit level is ubiquitous among state-of-the-art deep neural networks, and grows more pronounced for models with higher performance on a large-scale visual recognition task. Comparing matched populations—in the brain and in a model—we demonstrate a further correspondence at the level of the population code: stimulus category can be partially decoded from real neural responses using a classifier trained purely on a matched population of artificial units in a model. This provides a new point of investigation for phenomena which require fine-grained mappings between deep neural networks and the brain. |
| first_indexed | 2024-09-23T08:14:08Z |
| format | Technical Report |
| id | mit-1721.1/118847 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T08:14:08Z |
| publishDate | 2018 |
| publisher | Center for Brains, Minds and Machines (CBMM) |
| record_format | dspace |
| spelling | mit-1721.1/1188472019-04-09T17:28:58Z Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results Arend, Luke Han, Yena Schrimpf, Martin Bashivan, Pouya Kar, Kohitij Poggio, Tomaso DiCarlo, James J. Boix, Xavier Deep neural networks have been shown to predict neural responses in higher visual cortex. The mapping from the model to a neuron in the brain occurs through a linear combination of many units in the model, leaving open the question of whether there also exists a correspondence at the level of individual neurons. Here we show that there exist many one-to-one mappings between single units in a deep neural network model and neurons in the brain. We show that this correspondence at the single- unit level is ubiquitous among state-of-the-art deep neural networks, and grows more pronounced for models with higher performance on a large-scale visual recognition task. Comparing matched populations—in the brain and in a model—we demonstrate a further correspondence at the level of the population code: stimulus category can be partially decoded from real neural responses using a classifier trained purely on a matched population of artificial units in a model. This provides a new point of investigation for phenomena which require fine-grained mappings between deep neural networks and the brain. This material is based upon work supported by the Center for Brains, Minds and Machines (CBMM), funded by NSF STC award CCF-1231216. 2018-11-02T18:33:39Z 2018-11-02T18:33:39Z 2018-11-02 Technical Report Working Paper Other http://hdl.handle.net/1721.1/118847 en_US CBMM Memo Series;093 application/pdf Center for Brains, Minds and Machines (CBMM) |
| spellingShingle | Arend, Luke Han, Yena Schrimpf, Martin Bashivan, Pouya Kar, Kohitij Poggio, Tomaso DiCarlo, James J. Boix, Xavier Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title | Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title_full | Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title_fullStr | Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title_full_unstemmed | Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title_short | Single units in a deep neural network functionally correspond with neurons in the brain: preliminary results |
| title_sort | single units in a deep neural network functionally correspond with neurons in the brain preliminary results |
| url | http://hdl.handle.net/1721.1/118847 |
| work_keys_str_mv | AT arendluke singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT hanyena singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT schrimpfmartin singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT bashivanpouya singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT karkohitij singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT poggiotomaso singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT dicarlojamesj singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults AT boixxavier singleunitsinadeepneuralnetworkfunctionallycorrespondwithneuronsinthebrainpreliminaryresults |