Statistically unbiased prediction enables accurate denoising of voltage imaging data.
Ontology highlight
ABSTRACT:
SUBMITTER: Eom M
PROVIDER: S-EPMC10555843 | biostudies-literature | 2023 Oct
REPOSITORIES: biostudies-literature
Publications
Statistically unbiased prediction enables accurate denoising of voltage imaging data.
Eom Minho M Han Seungjae S Park Pojeong P Kim Gyuri G Cho Eun-Seo ES Sim Jueun J Lee Kang-Han KH Kim Seonghoon S Tian He H Böhm Urs L UL Lowet Eric E Tseng Hua-An HA Choi Jieun J Lucia Stephani Edwina SE Ryu Seung Hyun SH Rózsa Márton M Chang Sunghoe S Kim Pilhan P Han Xue X Piatkevich Kiryl D KD Choi Myunghwan M Kim Cheol-Hee CH Cohen Adam E AE Chang Jae-Byum JB Yoon Young-Gyu YG
Nature methods 20230918 10
Here we report SUPPORT (statistically unbiased prediction utilizing spatiotemporal information in imaging data), a self-supervised learning method for removing Poisson-Gaussian noise in voltage imaging data. SUPPORT is based on the insight that a pixel value in voltage imaging data is highly dependent on its spatiotemporal neighboring pixels, even when its temporally adjacent frames alone do not provide useful information for statistical prediction. Such dependency is captured and used by a conv ...[more]
Similar Datasets
Project description: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.
| S-EPMC11618097 | biostudies-literature
Project description:Two-photon calcium imaging is now an important tool for in vivo imaging of biological systems. By enabling neuronal population imaging with subcellular resolution, this modality offers an approach for gaining a fundamental understanding of brain anatomy and physiology. Proper analysis of calcium imaging data requires denoising, that is separating the signal from complex physiological noise. To analyze two-photon brain imaging data, we present a signal plus colored noise model in which the signal is represented as harmonic regression and the correlated noise is represented as an order autoregressive process. We provide an efficient cyclic descent algorithm to compute approximate maximum likelihood parameter estimates by combing a weighted least-squares procedure with the Burg algorithm. We use Akaike information criterion to guide selection of the harmonic regression and the autoregressive model orders. Our flexible yet parsimonious modeling approach reliably separates stimulus-evoked fluorescence response from background activity and noise, assesses goodness of fit, and estimates confidence intervals and signal-to-noise ratio. This refined separation leads to appreciably enhanced image contrast for individual cells including clear delineation of subcellular details and network activity. The application of our approach to in vivo imaging data recorded in the ferret primary visual cortex demonstrates that our method yields substantially denoised signal estimates. We also provide a general Volterra series framework for deriving this and other signal plus correlated noise models for imaging. This approach to analyzing two-photon calcium imaging data may be readily adapted to other computational biology problems which apply correlated noise models.
| S-EPMC3110192 | biostudies-literature
Project description:Volumetric functional imaging is widely used for recording neuron activities in vivo, but there exist tradeoffs between the quality of the extracted calcium traces, imaging speed, and laser power. While deep-learning methods have recently been applied to denoise images, their applications to downstream analyses, such as recovering high-SNR calcium traces, have been limited. Further, these methods require temporally-sequential pre-registered data acquired at ultrafast rates. Here, we demonstrate a supervised deep-denoising method to circumvent these tradeoffs for several applications, including whole-brain imaging, large-field-of-view imaging in freely moving animals, and recovering complex neurite structures in C. elegans. Our framework has 30× smaller memory footprint, and is fast in training and inference (50-70 ms); it is highly accurate and generalizable, and further, trained with only small, non-temporally-sequential, independently-acquired training datasets (∼500 pairs of images). We envision that the framework will enable faster and long-term imaging experiments necessary to study neuronal mechanisms of many behaviors.
| S-EPMC9440141 | biostudies-literature
Project description:SignificanceVoltage imaging is a powerful tool for studying the dynamics of neuronal activities in the brain. However, voltage imaging data are fundamentally corrupted by severe Poisson noise in the low-photon regime, which hinders the accurate extraction of neuronal activities. Self-supervised deep learning denoising methods have shown great potential in addressing the challenges in low-photon voltage imaging without the need for ground-truth but usually suffer from the trade-off between spatial and temporal performances.AimWe present DeepVID v2, a self-supervised denoising framework with decoupled spatial and temporal enhancement capability to significantly augment low-photon voltage imaging.ApproachDeepVID v2 is built on our original DeepVID framework, which performs frame-based denoising by utilizing a sequence of frames around the central frame targeted for denoising to leverage temporal information and ensure consistency. Similar to DeepVID, the network further integrates multiple blind pixels in the central frame to enrich the learning of local spatial information. In addition, DeepVID v2 introduces a new spatial prior extraction branch to capture fine structural details to learn high spatial resolution information. Two variants of DeepVID v2 are introduced to meet specific denoising needs: an online version tailored for real-time inference with a limited number of frames and an offline version designed to leverage the full dataset, achieving optimal temporal and spatial performances.ResultsWe demonstrate that DeepVID v2 is able to overcome the trade-off between spatial and temporal performances and achieve superior denoising capability in resolving both high-resolution spatial structures and rapid temporal neuronal activities. We further show that DeepVID v2 can generalize to different imaging conditions, including time-series measurements with various signal-to-noise ratios and extreme low-photon conditions.ConclusionsOur results underscore DeepVID v2 as a promising tool for enhancing voltage imaging. This framework has the potential to generalize to other low-photon imaging modalities and greatly facilitate the study of neuronal activities in the brain.
| S-EPMC11519979 | biostudies-literature
Project description:Diffusion weighted imaging (DWI) with multiple, high b-values is critical for extracting tissue microstructure measurements; however, high b-value DWI images contain high noise levels that can overwhelm the signal of interest and bias microstructural measurements. Here, we propose a simple denoising method that can be applied to any dataset, provided a low-noise, single-subject dataset is acquired using the same DWI sequence. The denoising method uses a one-dimensional convolutional neural network (1D-CNN) and deep learning to learn from a low-noise dataset, voxel-by-voxel. The trained model can then be applied to high-noise datasets from other subjects. We validated the 1D-CNN denoising method by first demonstrating that 1D-CNN denoising resulted in DWI images that were more similar to the noise-free ground truth than comparable denoising methods, e.g., MP-PCA, using simulated DWI data. Using the same DWI acquisition but reconstructed with two common reconstruction methods, i.e. SENSE1 and sum-of-square, to generate a pair of low-noise and high-noise datasets, we then demonstrated that 1D-CNN denoising of high-noise DWI data collected from human subjects showed promising results in three domains: DWI images, diffusion metrics, and tractography. In particular, the denoised images were very similar to a low-noise reference image of that subject, more than the similarity between repeated low-noise images (i.e. computational reproducibility). Finally, we demonstrated the use of the 1D-CNN method in two practical examples to reduce noise from parallel imaging and simultaneous multi-slice acquisition. We conclude that the 1D-CNN denoising method is a simple, effective denoising method for DWI images that overcomes some of the limitations of current state-of-the-art denoising methods, such as the need for a large number of training subjects and the need to account for the rectified noise floor.
| S-EPMC9477507 | biostudies-literature
Project description:We present a modular approach for analyzing calcium imaging recordings of large neuronal ensembles. Our goal is to simultaneously identify the locations of the neurons, demix spatially overlapping components, and denoise and deconvolve the spiking activity from the slow dynamics of the calcium indicator. Our approach relies on a constrained nonnegative matrix factorization that expresses the spatiotemporal fluorescence activity as the product of a spatial matrix that encodes the spatial footprint of each neuron in the optical field and a temporal matrix that characterizes the calcium concentration of each neuron over time. This framework is combined with a novel constrained deconvolution approach that extracts estimates of neural activity from fluorescence traces, to create a spatiotemporal processing algorithm that requires minimal parameter tuning. We demonstrate the general applicability of our method by applying it to in vitro and in vivo multi-neuronal imaging data, whole-brain light-sheet imaging data, and dendritic imaging data.
| S-EPMC4881387 | biostudies-literature
Project description:Prediction of T-cell receptor (TCR) interactions with MHC-peptide complexes remains highly challenging. This challenge is primarily due to three dominant factors: data accuracy, data scarceness, and problem complexity. Here, we showcase that "shallow" convolutional neural network (CNN) architectures are adequate to deal with the problem complexity imposed by the length variations of TCRs. We demonstrate that current public bulk CDR3β-pMHC binding data overall is of low quality and that the development of accurate prediction models is contingent on paired α/β TCR sequence data corresponding to at least 150 distinct pairs for each investigated pMHC. In comparison, models trained on CDR3α or CDR3β data alone demonstrated a variable and pMHC specific relative performance drop. Together these findings support that T-cell specificity is predictable given the availability of accurate and sufficient paired TCR sequence data. NetTCR-2.0 is publicly available at https://services.healthtech.dtu.dk/service.php?NetTCR-2.0 .
| S-EPMC8433451 | biostudies-literature
Project description:The influence of genetics on DNA methylation (DNAme) variation is well documented, yet confounding from population stratification is often unaccounted for in DNAme association studies. Existing approaches have been developed to address confounding by population stratification by directly using DNAme data, but have not been validated in additional human populations or tissues, such as the placenta. Results: To aid future placental DNAme studies in assessing population stratification, we developed an ethnicity classifier, PLANET (placental elastic net DNAme ethnicity classifier), on combined Infinium Human Methylation 450k BeadChip array (HM450k) data from placental samples. We used data from five North American cohorts from private and public repositories (n = 509) and show that PLANET can not only accurately predict (accuracy = 0.9379, kappa = 0.8227) major classes of self-reported ethnicity/race (African: n = 58, Asian: n = 53, Caucasian: n = 389), but can also produce probabilities that are highly correlated with genetic ancestry inferred from genome-wide SNP (>2.5 million SNP) and ancestry informative markers (n=50) data. We found that PLANET’s ethnicity classification relies on 1860 DNAme microarray sites, and over half of these were also linked to nearby genetic polymorphisms (n=955). Lastly, we found our placental-optimized method outperforms existing approaches in assessing population stratification in our placental samples from individuals of Asian, African, and Caucasian ethnicities. Conclusion: PLANET outperforms existing methods and heavily relies on the genetic signal present in DNAme microarray data. PLANET can be used to address population stratification in future placental DNAme association studies, and will be especially useful when ethnicity information is missing and genotyping markers are unavailable.
2020-10-14 | GSE128827 | GEO
Project description:A fundamental challenge in fluorescence microscopy is the photon shot noise arising from the inevitable stochasticity of photon detection. Noise increases measurement uncertainty and limits imaging resolution, speed and sensitivity. To achieve high-sensitivity fluorescence imaging beyond the shot-noise limit, we present DeepCAD-RT, a self-supervised deep learning method for real-time noise suppression. Based on our previous framework DeepCAD, we reduced the number of network parameters by 94%, memory consumption by 27-fold and processing time by a factor of 20, allowing real-time processing on a two-photon microscope. A high imaging signal-to-noise ratio can be acquired with tenfold fewer photons than in standard imaging approaches. We demonstrate the utility of DeepCAD-RT in a series of photon-limited experiments, including in vivo calcium imaging of mice, zebrafish larva and fruit flies, recording of three-dimensional (3D) migration of neutrophils after acute brain injury and imaging of 3D dynamics of cortical ATP release. DeepCAD-RT will facilitate the morphological and functional interrogation of biological dynamics with a minimal photon budget.
| S-EPMC9931589 | biostudies-literature
Project description:BackgroundNanopore metagenomics has been used for infectious disease diagnosis for bacterial pathogens. However, this technology currently lacks comprehensive performance studies in clinical settings for simultaneous detection of bacteria, fungi, and viruses.MethodsWe developed a dual-process of Nanopore sequencing for one sample, with unbiased metagenomics in Meta process and target enrichment in Panel process (Nanopore Meta-Panel process, NanoMP) and prospectively enrolled 450 respiratory specimens from multiple centers. The filter system of pathogen detection was established with machine learning and receiver operator characteristic (ROC) curve to optimize the detection accuracy based on orthogonal test of 21 species. Antimicrobial resistance (AMR) genes were identified based on the Comprehensive Antibiotic Resistance Database (CARD) and single-nucleotide polymorphism matrix.FindingsOur approach showed high sensitivity in Meta process, with 82.9%, 88.7%, and 75.0% for bacteria, fungi (except Aspergillus), and Mycobacterium tuberculosis groups, respectively. Moreover, target amplification improved the sensitivity of virus (>80.0% vs. 39.4%) and Aspergillus (81.8% vs. 42.3%) groups in Panel process compared with Meta process. Overall, NanoMP achieved 80.2% sensitivity and 98.8% specificity compared with the composite reference standard, and we were able to accurately detect AMR genes including blaKPC-2, blaOXA-23 and mecA and distinguish their parent organisms in patients with mixed infections.InterpretationWe combined metagenomic and enriched Nanopore sequencing for one sample in parallel. Our NanoMP approach simultaneously covered bacteria, viruses and fungi in respiratory specimens and demonstrated good diagnostic performance in real clinical settings.FundingNational Key Research and Development Program of China and National Natural Science Foundation of China.
| S-EPMC10652131 | biostudies-literature