{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T20:25:46Z","timestamp":1784060746472,"version":"3.55.0"},"reference-count":61,"publisher":"National Academy of Sciences","issue":"48","license":[{"start":{"date-parts":[[2021,3,1]],"date-time":"2021-03-01T00:00:00Z","timestamp":1614556800000},"content-version":"vor","delay-in-days":181,"URL":"https:\/\/www.pnas.org\/site\/aboutpnas\/licenses.xhtml"}],"funder":[{"DOI":"10.13039\/100006502","name":"Defense Sciences Office, DARPA","doi-asserted-by":"publisher","award":["FA8750-18- 413 C000"],"award-info":[{"award-number":["FA8750-18- 413 C000"]}],"id":[{"id":"10.13039\/100006502","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1524817"],"award-info":[{"award-number":["1524817"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1447476"],"award-info":[{"award-number":["1447476"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001839","name":"University Grants Committee","doi-asserted-by":"publisher","award":["24206219"],"award-info":[{"award-number":["24206219"]}],"id":[{"id":"10.13039\/501100001839","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2020,12]]},"abstract":"<jats:p>Deep neural networks excel at finding hierarchical representations that solve complex tasks over large datasets. How can we humans understand these learned representations? In this work, we present network dissection, an analytic framework to systematically identify the semantics of individual hidden units within image classification and image generation networks. First, we analyze a convolutional neural network (CNN) trained on scene classification and discover units that match a diverse set of object concepts. We find evidence that the network has learned many object classes that play crucial roles in classifying scene classes. Second, we use a similar analytic method to analyze a generative adversarial network (GAN) model trained to generate scenes. By analyzing changes made when small sets of units are activated or deactivated, we find that objects can be added and removed from the output scenes while adapting to the context. Finally, we apply our analytic framework to understanding adversarial attacks and to semantic image editing.<\/jats:p>","DOI":"10.1073\/pnas.1907375117","type":"journal-article","created":{"date-parts":[[2020,9,1]],"date-time":"2020-09-01T16:45:27Z","timestamp":1598978727000},"page":"30071-30078","update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":285,"title":["Understanding the role of individual units in a deep neural network"],"prefix":"10.1073","volume":"117","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1744-6765","authenticated-orcid":false,"given":"David","family":"Bau","sequence":"first","affiliation":[{"name":"Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun-Yan","family":"Zhu","sequence":"additional","affiliation":[{"name":"Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139;"},{"name":"Adobe Research, Adobe Inc., San Jose, CA 95110;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hendrik","family":"Strobelt","sequence":"additional","affiliation":[{"name":"Massachusetts Institute of Technology\u2013International Business Machines (IBM) Watson Artificial Intelligence Laboratory, Cambridge, MA 02142;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Agata","family":"Lapedriza","sequence":"additional","affiliation":[{"name":"Media Lab, Massachusetts Institute of Technology, Cambridge, MA 02139;"},{"name":"Estudis d\u2019Inform\u00e0tica, Multim\u00e8dia i Telecomunicaci\u00f3, Universitat Oberta de Catalunya, 08018 Barcelona, Spain;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4030-0684","authenticated-orcid":false,"given":"Bolei","family":"Zhou","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Antonio","family":"Torralba","sequence":"additional","affiliation":[{"name":"Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139;"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"341","published-online":{"date-parts":[[2020,9]]},"reference":[{"key":"e_1_3_4_1_2","unstructured":"B. Zhou A. Khosla A. Lapedriza A. Oliva A. Torralba Object detectors emerge in deep scene CNNs. arXiv:1412.6856 (22 December 2014)."},{"key":"e_1_3_4_2_2","first-page":"818","volume-title":"European Conference on Computer Vision","author":"Zeiler M. D.","year":"2014","unstructured":"M. D. Zeiler, R. Fergus, \u201cVisualizing and understanding convolutional networks\u201d in European Conference on Computer Vision (Springer, Berlin, Germany, 2014), pp. 818\u2013833."},{"key":"e_1_3_4_3_2","first-page":"5188","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Mahendran A.","year":"2015","unstructured":"A. Mahendran, A. Vedaldi, \u201cUnderstanding deep image representations by inverting them\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2015), pp. 5188\u20135196."},{"key":"e_1_3_4_4_2","doi-asserted-by":"publisher","DOI":"10.23915\/distill.00010"},{"key":"e_1_3_4_5_2","unstructured":"A. Bau Identifying and controlling important neurons in neural machine translation. https:\/\/openreview.net\/pdf?id=H1z-PsR5KX. Accessed 24 August 2020."},{"key":"e_1_3_4_6_2","unstructured":"A. Karpathy J. Johnson L. Fei-Fei Visualizing and understanding recurrent networks. arXiv:1506.02078 (5 June 2015)."},{"key":"e_1_3_4_7_2","unstructured":"A. Radford R. Jozefowicz I. Sutskever Learning to generate reviews and discovering sentiment. arXiv:1704.01444 (6 April 2017)."},{"key":"e_1_3_4_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2013.50"},{"key":"e_1_3_4_9_2","doi-asserted-by":"crossref","unstructured":"B. Zhou D. Bau A. Oliva A. Torralba Interpreting deep visual representations via network dissection. arXiv:1711.05611 (26 June 2018).","DOI":"10.1167\/18.10.1244"},{"key":"e_1_3_4_10_2","unstructured":"D. Bau Gan dissection: Visualizing and understanding generative adversarial networks. https:\/\/openreview.net\/pdf?id=Hyg_X2C5FX. Accessed 24 August 2020."},{"key":"e_1_3_4_11_2","first-page":"255","volume-title":"The Handbook of Brain Theory and Neural Networks","author":"LeCun Y.","year":"1995","unstructured":"Y. LeCun, Y. Bengio, \u201cConvolutional networks for images, speech, and time series\u201d in The Handbook of Brain Theory and Neural Networks, M. A. Arbib, Ed. (MIT Press, Cambridge, MA, 1995), vol. 3361, pp. 255\u2013258."},{"key":"e_1_3_4_12_2","first-page":"1097","volume-title":"Advances in Neural Information Processing Systems","author":"Krizhevsky A.","year":"2012","unstructured":"A. Krizhevsky, I. Sutskever, G. E. Hinton, \u201cImagenet classification with deep convolutional neural networks\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2012), pp. 1097\u20131105."},{"key":"e_1_3_4_13_2","unstructured":"K. Simonyan A. Zisserman Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556 (4 September 2014)."},{"key":"e_1_3_4_14_2","first-page":"2672","volume-title":"Advances in Neural Information Processing Systems","author":"Goodfellow I.","year":"2014","unstructured":"I. Goodfellow , \u201cGenerative adversarial nets\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2014), pp. 2672\u20132680."},{"key":"e_1_3_4_15_2","first-page":"3156","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Vinyals O.","year":"2015","unstructured":"O. Vinyals, A. Toshev, S. Bengio, D. Erhan, \u201cShow and tell: A neural image caption generator\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2015), pp. 3156\u20133164."},{"key":"e_1_3_4_16_2","first-page":"770","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"He K.","year":"2016","unstructured":"K. He, X. Zhang, S. Ren, J. Sun, \u201cDeep residual learning for image recognition\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2016), pp. 770\u2013778."},{"key":"e_1_3_4_17_2","first-page":"1125","volume-title":"Proceedings of the IEEE conference on Computer Vision and Pattern Recognition","author":"Isola P.","year":"2017","unstructured":"P. Isola, J. Y. Zhu, T. Zhou, A. A. Efros, \u201cImage-to-image translation with conditional adversarial networks\u201d in Proceedings of the IEEE conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2017), pp. 1125\u20131134."},{"key":"e_1_3_4_18_2","first-page":"2223","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Zhu J. Y.","year":"2017","unstructured":"J. Y. Zhu, T. Park, P. Isola, A. A. Efros, \u201cUnpaired image-to-image translation using cycle-consistent adversarial networks\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2017), pp. 2223\u20132232."},{"key":"e_1_3_4_19_2","unstructured":"T. Karras T. Aila S. Laine J. Lehtinen Progressive growing of GANs for improved quality stability and variation. https:\/\/openreview.net\/pdf?id=Hk99zCeAb. Accessed 24 August 2020."},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0130140"},{"key":"e_1_3_4_21_2","first-page":"117","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Zhou T.","year":"2016","unstructured":"T. Zhou, P. Krahenbuhl, M. Aubry, Q. Huang, A. A. Efros, \u201cLearning dense correspondence via 3D-guided cycle consistency\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2016), pp. 117\u2013126."},{"key":"e_1_3_4_22_2","first-page":"3429","volume-title":"International Conference on Computer Vision","author":"Fong R. C.","year":"2017","unstructured":"R. C. Fong, A. Vedaldi, \u201cInterpretable explanations of black boxes by meaningful perturbation\u201d in International Conference on Computer Vision (IEEE, New York, NY, 2017), pp. 3429\u20133437."},{"key":"e_1_3_4_23_2","first-page":"4765","volume-title":"Advances in Neural Information Processing Systems","author":"Lundberg S. M.","year":"2017","unstructured":"S. M. Lundberg, S. I. Lee, \u201cA unified approach to interpreting model predictions\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2017), pp. 4765\u20134774."},{"key":"e_1_3_4_24_2","first-page":"618","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Selvaraju R. R.","year":"2017","unstructured":"R. R. Selvaraju , \u201cGrad-cam: Visual explanations from deep networks via gradient-based localization\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2017), pp. 618\u2013626."},{"key":"e_1_3_4_25_2","first-page":"3319","volume-title":"Proceedings of the 34th International Conference on Machine Learning","author":"Sundararajan M.","year":"2017","unstructured":"M. Sundararajan, A. Taly, Q. Yan, \u201cAxiomatic attribution for deep networks\u201d in Proceedings of the 34th International Conference on Machine Learning (JMLR, 2017), vol. 70, pp. 3319\u20133328."},{"key":"e_1_3_4_26_2","unstructured":"D. Smilkov N. Thorat B. Kim F. Vi\u00e9gas M. Wattenberg Smoothgrad: Removing noise by adding noise. arXiv:1706.03825 (12 June 2017)."},{"key":"e_1_3_4_27_2","volume-title":"British Machine Vision Conference","author":"Petsiuk V.","year":"2018","unstructured":"V. Petsiuk, A. Das, K. Saenko, \u201cRise: Randomized input sampling for explanation of black-box models\u201d in British Machine Vision Conference (BMVA Press, Malvern, UK, 2018)."},{"key":"e_1_3_4_28_2","doi-asserted-by":"crossref","first-page":"1135","DOI":"10.1145\/2939672.2939778","volume-title":"Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","author":"Ribeiro M. T.","year":"2016","unstructured":"M. T. Ribeiro, S. Singh, C. Guestrin, \u201cWhy should I trust you?: Explaining the predictions of any classifier\u201d in Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (ACM, New York, NY, 2016), pp. 1135\u20131144."},{"key":"e_1_3_4_29_2","unstructured":"B. Kim J. Gilmer F. Viegas U. Erlingsson M. Wattenberg Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV). arXiv:1711.11279 (7 June 2018)."},{"key":"e_1_3_4_30_2","unstructured":"A. Koul A. Fern S. Greydanus Learning finite state representations of recurrent policy networks. https:\/\/openreview.net\/pdf?id=S1gOpsCctm. Accessed 24 August 2020."},{"key":"e_1_3_4_31_2","first-page":"3","volume-title":"European Conference on Computer Vision","author":"Hendricks L. A.","year":"2016","unstructured":"L. A. Hendricks , \u201cGenerating visual explanations\u201d in European Conference on Computer Vision (Springer, Berlin, Germany, 2016), pp. 3\u201319."},{"key":"e_1_3_4_32_2","first-page":"487","volume-title":"Advances in Neural Information Processing Systems","author":"Zhou B.","year":"2014","unstructured":"B. Zhou, A. Lapedriza, J. Xiao, A. Torralba, A. Oliva, \u201cLearning deep features for scene recognition using places database\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2014), pp. 487\u2013495."},{"key":"e_1_3_4_33_2","first-page":"418","volume-title":"Proceedings of the European Conference on Computer Vision","author":"Xiao T.","year":"2018","unstructured":"T. Xiao, Y. Liu, B. Zhou, Y. Jiang, J. Sun, \u201cUnified perceptual parsing for scene understanding\u201d in Proceedings of the European Conference on Computer Vision (Springer, Berlin, Germany, 2018), pp. 418\u2013434."},{"key":"e_1_3_4_34_2","unstructured":"R. Geirhos Imagenet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. arXiv:1811.12231 (14 January 2019)."},{"key":"e_1_3_4_35_2","first-page":"248","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Deng J.","year":"2009","unstructured":"J. Deng , \u201cImagenet: A large-scale hierarchical image database\u201d in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (IEEE, New York, NY, 2009), pp. 248\u2013255."},{"key":"e_1_3_4_36_2","first-page":"2074","volume-title":"Advances in Neural Information Processing Systems","author":"Wen W.","year":"2016","unstructured":"W. Wen, C. Wu, Y. Wang, Y. Chen, H. Li, \u201cLearning structured sparsity in deep neural networks\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2016), pp. 2074\u20132082."},{"key":"e_1_3_4_37_2","unstructured":"H. Li A. Kadav I. Durdanovic H. Samet H. P. Graf Pruning filters for efficient convnets. https:\/\/openreview.net\/pdf?id=rJqFGTslg. Accessed 24 August 2020."},{"key":"e_1_3_4_38_2","unstructured":"A. S. Morcos D. G. Barrett N. C. Rabinowitz M. Botvinick On the importance of single directions for generalization. arXiv:1803.06959 (22 May 2018)."},{"key":"e_1_3_4_39_2","unstructured":"B. Zhou Y. Sun D. Bau A. Torralba Revisiting the importance of individual units in CNNs via ablation. arXiv:1806.02891 (7 June 2018)."},{"key":"e_1_3_4_40_2","unstructured":"A. Radford L. Metz S. Chintala Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv:1511.06434 (19 November 2015)."},{"key":"e_1_3_4_41_2","first-page":"597","volume-title":"European Conference on Computer Vision","author":"Zhu J. Y.","year":"2016","unstructured":"J. Y. Zhu, P. Kr\u00e4henb\u00fchl, E. Shechtman, A. A. Efros, \u201cGenerative visual manipulation on the natural image manifold\u201d in European Conference on Computer Vision (Springer, Berlin, Germany, 2016), pp. 597\u2013613."},{"key":"e_1_3_4_42_2","unstructured":"F. Yu LSUN: Construction of a large-scale image dataset using deep learning with humans in the loop. arXiv:1506.03365 (4 June 2016)."},{"key":"e_1_3_4_43_2","unstructured":"C. Szegedy Intriguing properties of neural networks. arXiv:1312.6199 (21 December 2013)."},{"key":"e_1_3_4_44_2","unstructured":"I. J. Goodfellow J. Shlens C. Szegedy Explaining and harnessing adversarial examples. arXiv:1412.6572 (20 December 2014)."},{"key":"e_1_3_4_45_2","first-page":"39","volume-title":"IEEE Symposium on Security and Privacy (SP)","author":"Carlini N.","year":"2017","unstructured":"N. Carlini, D. Wagner, \u201cTowards evaluating the robustness of neural networks in 2017\u201d in IEEE Symposium on Security and Privacy (SP) (IEEE, 2017), pp. 39\u201357."},{"key":"e_1_3_4_46_2","unstructured":"A. Madry A. Makelov L. Schmidt D. Tsipras A. Vladu Towards deep learning models resistant to adversarial attacks. https:\/\/openreview.net\/pdf?id=rJzIBfZAb. Accessed 24 August 2020."},{"key":"e_1_3_4_47_2","unstructured":"J. Rauber W. Brendel M. Bethge Foolbox: A python toolbox to benchmark the robustness of machine learning models. arXiv:1707.04131 (20 March 2018)."},{"key":"e_1_3_4_48_2","unstructured":"A. Ilyas Adversarial examples are not bugs they are features. arXiv:1905.02175 (12 August 2019)."},{"key":"e_1_3_4_49_2","first-page":"2172","volume-title":"Advances in Neural Information Processing Systems","author":"Chen X.","year":"2016","unstructured":"X. Chen , \u201cInfogan: Interpretable representation learning by information maximizing generative adversarial nets\u201d in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2016), pp. 2172\u20132180."},{"key":"e_1_3_4_50_2","unstructured":"I. Higgins \u03b2-vae: Learning basic visual concepts with a constrained variational framework. https:\/\/openreview.net\/pdf?id=Sy2fzU9gl. Accessed 24 August 2020."},{"key":"e_1_3_4_51_2","doi-asserted-by":"crossref","unstructured":"Q. Zhang Y. Nian Wu S. C. Zhu Interpretable convolutional neural networks. arXiv:1710.00935 (2 October 2017).","DOI":"10.1109\/CVPR.2018.00920"},{"key":"e_1_3_4_52_2","first-page":"1947","article-title":"Emergence of invariance and disentanglement in deep representations","volume":"19","author":"Achille A.","year":"2018","unstructured":"A. Achille, S. Soatto, Emergence of invariance and disentanglement in deep representations. JMLR 19, 1947\u20131980 (2018).","journal-title":"JMLR"},{"key":"e_1_3_4_53_2","doi-asserted-by":"crossref","unstructured":"B. Zhou \u201cScene parsing through ade20k dataset\u201d in Computer Vision and Pattern Recognition (IEEE 2017).","DOI":"10.1109\/CVPR.2017.544"},{"key":"e_1_3_4_54_2","unstructured":"B. Zhou A. Lapedriza A. Khosla A. Oliva A. Torralba Places365-Standard. http:\/\/places2.csail.mit.edu\/download.html. Accessed 1 August 2020."},{"key":"e_1_3_4_55_2","unstructured":"J. Deng ImageNet ILSVRC 2012 data set. http:\/\/www.image-net.org\/download-images. Accessed 1 August 2020."},{"key":"e_1_3_4_56_2","unstructured":"F. Yu LSUN: Construction of a large-scale image dataset using deep learning with humans in the loop. https:\/\/www.yf.io\/p\/lsun. Accessed 1 August 2020."},{"key":"e_1_3_4_57_2","unstructured":"B. Zhou A. Lapedriza A. Khosla A. Oliva A. Torralba Data from \u201cPre-trained CNN models on Places365-Standard.\u201d GitHub. https:\/\/github.com\/CSAILVision\/places365. Accessed 1 August 2020."},{"key":"e_1_3_4_58_2","unstructured":"T. Karras T. Aila S. Laine J. Lehtinen Progressive growing of GANs. GitHub. https:\/\/github.com\/tkarras\/progressive_growing_of_gans. Accessed 1 August 2020."},{"key":"e_1_3_4_59_2","unstructured":"D. Bau Code for understanding the role of individual units in a deep neural network. Github. https:\/\/github.com\/davidbau\/dissect. Deposited 24 August 2020."},{"key":"e_1_3_4_60_2","unstructured":"B. Zhou ADE20K full dataset. https:\/\/groups.csail.mit.edu\/vision\/datasets\/ADE20K\/. Accessed 1 August 2020."},{"key":"e_1_3_4_61_2","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2009.2019809"}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/www.pnas.org\/syndication\/doi\/10.1073\/pnas.1907375117","content-type":"unspecified","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.1907375117","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,4,13]],"date-time":"2022-04-13T05:21:54Z","timestamp":1649827314000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/full\/10.1073\/pnas.1907375117"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,9]]},"references-count":61,"journal-issue":{"issue":"48","published-print":{"date-parts":[[2020,12]]}},"alternative-id":["10.1073\/pnas.1907375117"],"URL":"https:\/\/doi.org\/10.1073\/pnas.1907375117","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,9]]},"assertion":[{"value":"2020-09-01","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}