Robust self-supervised denoising of voltage imaging data using CellMincer
Abstract Voltage imaging is a powerful technique for studying neuronal activity, but its effectiveness is often constrained by low signal-to-noise ratios (SNR). Traditional denoising methods, such as matrix factorization, impose rigid assumptions about noise and signal structures, while existing dee...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-12-01
|
| Series: | npj Imaging |
| Online Access: | https://doi.org/10.1038/s44303-024-00055-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849220609644429312 |
|---|---|
| author | Brice Wang Tianle Ma Theresa Chen Trinh Nguyen Ethan Crouse Stephen J. Fleming Alison S. Walker Vera Valakh Ralda Nehme Evan W. Miller Samouil L. Farhi Mehrtash Babadi |
| author_facet | Brice Wang Tianle Ma Theresa Chen Trinh Nguyen Ethan Crouse Stephen J. Fleming Alison S. Walker Vera Valakh Ralda Nehme Evan W. Miller Samouil L. Farhi Mehrtash Babadi |
| author_sort | Brice Wang |
| collection | DOAJ |
| description | Abstract Voltage imaging is a powerful technique for studying neuronal activity, but its effectiveness is often constrained by low signal-to-noise ratios (SNR). Traditional denoising methods, such as matrix factorization, impose rigid assumptions about noise and signal structures, while existing deep learning approaches fail to fully capture the rapid dynamics and complex dependencies inherent in voltage imaging data. Here, we introduce CellMincer, a novel self-supervised deep learning method specifically developed for denoising voltage imaging datasets. CellMincer operates by masking and predicting sparse pixel sets across short temporal windows and conditions the denoiser on precomputed spatiotemporal auto-correlations to effectively model long-range dependencies without large temporal contexts. We developed and utilized a physics-based simulation framework to generate realistic synthetic datasets, enabling rigorous hyperparameter optimization and ablation studies. This approach highlighted the critical role of conditioning on spatiotemporal auto-correlations, resulting in an additional 3-fold SNR gain. Comprehensive benchmarking on both simulated and real datasets, including those validated with patch-clamp electrophysiology (EP), demonstrates CellMincer’s state-of-the-art performance, with substantial noise reduction across the frequency spectrum, enhanced subthreshold event detection, and high-fidelity recovery of EP signals. CellMincer consistently outperforms existing methods in SNR gain (0.5–2.9 dB) and reduces SNR variability by 17–55%. Incorporating CellMincer into standard workflows significantly improves neuronal segmentation, peak detection, and functional phenotype identification, consistently surpassing current methods in both SNR gain and consistency. |
| format | Article |
| id | doaj-art-dbf291f876f244e9bc6260017bca652d |
| institution | Kabale University |
| issn | 2948-197X |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Imaging |
| spelling | doaj-art-dbf291f876f244e9bc6260017bca652d2024-12-08T12:41:05ZengNature Portfolionpj Imaging2948-197X2024-12-012112110.1038/s44303-024-00055-xRobust self-supervised denoising of voltage imaging data using CellMincerBrice Wang0Tianle Ma1Theresa Chen2Trinh Nguyen3Ethan Crouse4Stephen J. Fleming5Alison S. Walker6Vera Valakh7Ralda Nehme8Evan W. Miller9Samouil L. Farhi10Mehrtash Babadi11Data Sciences Platform (DSP), Broad Institute of MIT and HarvardData Sciences Platform (DSP), Broad Institute of MIT and HarvardSpatial Technology Platform (STP), Broad Institute of MIT and HarvardSpatial Technology Platform (STP), Broad Institute of MIT and HarvardStanley Center for Psychiatric Research at the Broad Institute of MIT and HarvardData Sciences Platform (DSP), Broad Institute of MIT and HarvardDepartments of Molecular & Cell Biology and Chemistry and Helen Wills Neuroscience Institute, UC BerkeleySpatial Technology Platform (STP), Broad Institute of MIT and HarvardStanley Center for Psychiatric Research at the Broad Institute of MIT and HarvardDepartments of Molecular & Cell Biology and Chemistry and Helen Wills Neuroscience Institute, UC BerkeleySpatial Technology Platform (STP), Broad Institute of MIT and HarvardData Sciences Platform (DSP), Broad Institute of MIT and HarvardAbstract Voltage imaging is a powerful technique for studying neuronal activity, but its effectiveness is often constrained by low signal-to-noise ratios (SNR). Traditional denoising methods, such as matrix factorization, impose rigid assumptions about noise and signal structures, while existing deep learning approaches fail to fully capture the rapid dynamics and complex dependencies inherent in voltage imaging data. Here, we introduce CellMincer, a novel self-supervised deep learning method specifically developed for denoising voltage imaging datasets. CellMincer operates by masking and predicting sparse pixel sets across short temporal windows and conditions the denoiser on precomputed spatiotemporal auto-correlations to effectively model long-range dependencies without large temporal contexts. We developed and utilized a physics-based simulation framework to generate realistic synthetic datasets, enabling rigorous hyperparameter optimization and ablation studies. This approach highlighted the critical role of conditioning on spatiotemporal auto-correlations, resulting in an additional 3-fold SNR gain. Comprehensive benchmarking on both simulated and real datasets, including those validated with patch-clamp electrophysiology (EP), demonstrates CellMincer’s state-of-the-art performance, with substantial noise reduction across the frequency spectrum, enhanced subthreshold event detection, and high-fidelity recovery of EP signals. CellMincer consistently outperforms existing methods in SNR gain (0.5–2.9 dB) and reduces SNR variability by 17–55%. Incorporating CellMincer into standard workflows significantly improves neuronal segmentation, peak detection, and functional phenotype identification, consistently surpassing current methods in both SNR gain and consistency.https://doi.org/10.1038/s44303-024-00055-x |
| spellingShingle | Brice Wang Tianle Ma Theresa Chen Trinh Nguyen Ethan Crouse Stephen J. Fleming Alison S. Walker Vera Valakh Ralda Nehme Evan W. Miller Samouil L. Farhi Mehrtash Babadi Robust self-supervised denoising of voltage imaging data using CellMincer npj Imaging |
| title | Robust self-supervised denoising of voltage imaging data using CellMincer |
| title_full | Robust self-supervised denoising of voltage imaging data using CellMincer |
| title_fullStr | Robust self-supervised denoising of voltage imaging data using CellMincer |
| title_full_unstemmed | Robust self-supervised denoising of voltage imaging data using CellMincer |
| title_short | Robust self-supervised denoising of voltage imaging data using CellMincer |
| title_sort | robust self supervised denoising of voltage imaging data using cellmincer |
| url | https://doi.org/10.1038/s44303-024-00055-x |
| work_keys_str_mv | AT bricewang robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT tianlema robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT theresachen robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT trinhnguyen robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT ethancrouse robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT stephenjfleming robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT alisonswalker robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT veravalakh robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT raldanehme robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT evanwmiller robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT samouillfarhi robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer AT mehrtashbabadi robustselfsuperviseddenoisingofvoltageimagingdatausingcellmincer |