{"title":"灵长类动物颞下皮层神经元反应与前馈深度神经网络模型在人脸信息处理中的比较。","authors":"Narihisa Matsumoto, Yoh-Ichi Mototake, Kenji Kawano, Masato Okada, Yasuko Sugase-Miyamoto","doi":"10.1007/s10827-021-00778-5","DOIUrl":null,"url":null,"abstract":"<p><p>Feed-forward deep neural networks have better performance in object categorization tasks than other models of computer vision. To understand the relationship between feed-forward deep networks and the primate brain, we investigated representations of upright and inverted faces in a convolutional deep neural network model and compared them with representations by neurons in the monkey anterior inferior-temporal cortex, area TE. We applied principal component analysis to feature vectors in each model layer to visualize the relationship between the vectors of the upright and inverted faces. The vectors of the upright and inverted monkey faces were more separated through the convolution layers. In the fully-connected layers, the separation among human individuals for upright faces was larger than for inverted faces. The Spearman correlation between each model layer and TE neurons reached a maximum at the fully-connected layers. These results indicate that the processing of faces in the fully-connected layers might resemble the asymmetric representation of upright and inverted faces by the TE neurons. The separation of upright and inverted faces might take place by feed-forward processing in the visual cortex, and separations among human individuals for upright faces, which were larger than those for inverted faces, might occur in area TE.</p>","PeriodicalId":54857,"journal":{"name":"Journal of Computational Neuroscience","volume":null,"pages":null},"PeriodicalIF":1.5000,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/s10827-021-00778-5","citationCount":"4","resultStr":"{\"title\":\"Comparison of neuronal responses in primate inferior-temporal cortex and feed-forward deep neural network model with regard to information processing of faces.\",\"authors\":\"Narihisa Matsumoto, Yoh-Ichi Mototake, Kenji Kawano, Masato Okada, Yasuko Sugase-Miyamoto\",\"doi\":\"10.1007/s10827-021-00778-5\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Feed-forward deep neural networks have better performance in object categorization tasks than other models of computer vision. To understand the relationship between feed-forward deep networks and the primate brain, we investigated representations of upright and inverted faces in a convolutional deep neural network model and compared them with representations by neurons in the monkey anterior inferior-temporal cortex, area TE. We applied principal component analysis to feature vectors in each model layer to visualize the relationship between the vectors of the upright and inverted faces. The vectors of the upright and inverted monkey faces were more separated through the convolution layers. In the fully-connected layers, the separation among human individuals for upright faces was larger than for inverted faces. The Spearman correlation between each model layer and TE neurons reached a maximum at the fully-connected layers. These results indicate that the processing of faces in the fully-connected layers might resemble the asymmetric representation of upright and inverted faces by the TE neurons. The separation of upright and inverted faces might take place by feed-forward processing in the visual cortex, and separations among human individuals for upright faces, which were larger than those for inverted faces, might occur in area TE.</p>\",\"PeriodicalId\":54857,\"journal\":{\"name\":\"Journal of Computational Neuroscience\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.5000,\"publicationDate\":\"2021-08-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://sci-hub-pdf.com/10.1007/s10827-021-00778-5\",\"citationCount\":\"4\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Journal of Computational Neuroscience\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1007/s10827-021-00778-5\",\"RegionNum\":4,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2021/2/17 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q3\",\"JCRName\":\"MATHEMATICAL & COMPUTATIONAL BIOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Computational Neuroscience","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1007/s10827-021-00778-5","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2021/2/17 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"MATHEMATICAL & COMPUTATIONAL BIOLOGY","Score":null,"Total":0}
Comparison of neuronal responses in primate inferior-temporal cortex and feed-forward deep neural network model with regard to information processing of faces.
Feed-forward deep neural networks have better performance in object categorization tasks than other models of computer vision. To understand the relationship between feed-forward deep networks and the primate brain, we investigated representations of upright and inverted faces in a convolutional deep neural network model and compared them with representations by neurons in the monkey anterior inferior-temporal cortex, area TE. We applied principal component analysis to feature vectors in each model layer to visualize the relationship between the vectors of the upright and inverted faces. The vectors of the upright and inverted monkey faces were more separated through the convolution layers. In the fully-connected layers, the separation among human individuals for upright faces was larger than for inverted faces. The Spearman correlation between each model layer and TE neurons reached a maximum at the fully-connected layers. These results indicate that the processing of faces in the fully-connected layers might resemble the asymmetric representation of upright and inverted faces by the TE neurons. The separation of upright and inverted faces might take place by feed-forward processing in the visual cortex, and separations among human individuals for upright faces, which were larger than those for inverted faces, might occur in area TE.
期刊介绍:
The Journal of Computational Neuroscience provides a forum for papers that fit the interface between computational and experimental work in the neurosciences. The Journal of Computational Neuroscience publishes full length original papers, rapid communications and review articles describing theoretical and experimental work relevant to computations in the brain and nervous system. Papers that combine theoretical and experimental work are especially encouraged. Primarily theoretical papers should deal with issues of obvious relevance to biological nervous systems. Experimental papers should have implications for the computational function of the nervous system, and may report results using any of a variety of approaches including anatomy, electrophysiology, biophysics, imaging, and molecular biology. Papers investigating the physiological mechanisms underlying pathologies of the nervous system, or papers that report novel technologies of interest to researchers in computational neuroscience, including advances in neural data analysis methods yielding insights into the function of the nervous system, are also welcomed (in this case, methodological papers should include an application of the new method, exemplifying the insights that it yields).It is anticipated that all levels of analysis from cognitive to cellular will be represented in the Journal of Computational Neuroscience.