Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity
Reconstructing natural images and decoding their semantic category from fMRI brain recordings is challenging. Acquiring sufficient pairs of images and their corresponding fMRI responses, which span the huge space of natural images, is prohibitive. We present a novel self-supervised approach that goe...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-07-01
|
| Series: | NeuroImage |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S105381192200249X |
| _version_ | 1828351077780553728 |
|---|---|
| author | Guy Gaziv Roman Beliy Niv Granot Assaf Hoogi Francesca Strappini Tal Golan Michal Irani |
| author_facet | Guy Gaziv Roman Beliy Niv Granot Assaf Hoogi Francesca Strappini Tal Golan Michal Irani |
| author_sort | Guy Gaziv |
| collection | DOAJ |
| description | Reconstructing natural images and decoding their semantic category from fMRI brain recordings is challenging. Acquiring sufficient pairs of images and their corresponding fMRI responses, which span the huge space of natural images, is prohibitive. We present a novel self-supervised approach that goes well beyond the scarce paired data, for achieving both: (i) state-of-the art fMRI-to-image reconstruction, and (ii) first-ever large-scale semantic classification from fMRI responses. By imposing cycle consistency between a pair of deep neural networks (from image-to-fMRI & from fMRI-to-image), we train our image reconstruction network on a large number of “unpaired” natural images (images without fMRI recordings) from many novel semantic categories. This enables to adapt our reconstruction network to a very rich semantic coverage without requiring any explicit semantic supervision. Specifically, we find that combining our self-supervised training with high-level perceptual losses, gives rise to new reconstruction & classification capabilities. In particular, this perceptual training enables to classify well fMRIs of never-before-seen semantic classes, without requiring any class labels during training. This gives rise to: (i) Unprecedented image-reconstruction from fMRI of never-before-seen images (evaluated by image metrics and human testing), and (ii) Large-scale semantic classification of categories that were never-before-seen during network training. Such large-scale (1000-way) semantic classification from fMRI recordings has never been demonstrated before. Finally, we provide evidence for the biological consistency of our learned model. |
| first_indexed | 2024-04-14T01:34:21Z |
| format | Article |
| id | doaj.art-de04c565f2864f8f9116d8fbdf1c052b |
| institution | Directory Open Access Journal |
| issn | 1095-9572 |
| language | English |
| last_indexed | 2024-04-14T01:34:21Z |
| publishDate | 2022-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | NeuroImage |
| spelling | doaj.art-de04c565f2864f8f9116d8fbdf1c052b2022-12-22T02:20:02ZengElsevierNeuroImage1095-95722022-07-01254119121Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain ActivityGuy Gaziv0Roman Beliy1Niv Granot2Assaf Hoogi3Francesca Strappini4Tal Golan5Michal Irani6Corresponding author.; Dept. of Computer Science and Applied Math, Weizmann Institute of Science, Rehovot, IsraelDept. of Computer Science and Applied Math, Weizmann Institute of Science, Rehovot, IsraelDept. of Computer Science and Applied Math, Weizmann Institute of Science, Rehovot, IsraelDept. of Computer Science and Applied Math, Weizmann Institute of Science, Rehovot, IsraelDept. of Neurobiology, Weizmann Institute of Science, Rehovot, IsraelZuckerman Institute, Columbia University, New York, NY USADept. of Computer Science and Applied Math, Weizmann Institute of Science, Rehovot, IsraelReconstructing natural images and decoding their semantic category from fMRI brain recordings is challenging. Acquiring sufficient pairs of images and their corresponding fMRI responses, which span the huge space of natural images, is prohibitive. We present a novel self-supervised approach that goes well beyond the scarce paired data, for achieving both: (i) state-of-the art fMRI-to-image reconstruction, and (ii) first-ever large-scale semantic classification from fMRI responses. By imposing cycle consistency between a pair of deep neural networks (from image-to-fMRI & from fMRI-to-image), we train our image reconstruction network on a large number of “unpaired” natural images (images without fMRI recordings) from many novel semantic categories. This enables to adapt our reconstruction network to a very rich semantic coverage without requiring any explicit semantic supervision. Specifically, we find that combining our self-supervised training with high-level perceptual losses, gives rise to new reconstruction & classification capabilities. In particular, this perceptual training enables to classify well fMRIs of never-before-seen semantic classes, without requiring any class labels during training. This gives rise to: (i) Unprecedented image-reconstruction from fMRI of never-before-seen images (evaluated by image metrics and human testing), and (ii) Large-scale semantic classification of categories that were never-before-seen during network training. Such large-scale (1000-way) semantic classification from fMRI recordings has never been demonstrated before. Finally, we provide evidence for the biological consistency of our learned model.http://www.sciencedirect.com/science/article/pii/S105381192200249XSelf-Supervised learning, Decoding, Encoding, fMRI, Image reconstruction, Classificationvision |
| spellingShingle | Guy Gaziv Roman Beliy Niv Granot Assaf Hoogi Francesca Strappini Tal Golan Michal Irani Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity NeuroImage Self-Supervised learning, Decoding, Encoding, fMRI, Image reconstruction, Classification vision |
| title | Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity |
| title_full | Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity |
| title_fullStr | Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity |
| title_full_unstemmed | Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity |
| title_short | Self-supervised Natural Image Reconstruction and Large-scale Semantic Classification from Brain Activity |
| title_sort | self supervised natural image reconstruction and large scale semantic classification from brain activity |
| topic | Self-Supervised learning, Decoding, Encoding, fMRI, Image reconstruction, Classification vision |
| url | http://www.sciencedirect.com/science/article/pii/S105381192200249X |
| work_keys_str_mv | AT guygaziv selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT romanbeliy selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT nivgranot selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT assafhoogi selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT francescastrappini selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT talgolan selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity AT michalirani selfsupervisednaturalimagereconstructionandlargescalesemanticclassificationfrombrainactivity |