As near-infrared spectroscopy(NIRS)broadens its application area to diferent age and diseasegroups,motion artifacts in the NIRS signal due to subject movement is becoming an importantchallenge.Motion artifacts general...As near-infrared spectroscopy(NIRS)broadens its application area to diferent age and diseasegroups,motion artifacts in the NIRS signal due to subject movement is becoming an importantchallenge.Motion artifacts generally produce signal fiuctuations that are larger than physio-logical NIRS signals,thus it is cruciai to corect for them before obtaining an cstimate ofstimulusevoked hemodynamic responses.,There are various methods for correction such as principlecomponent analy sis(P CA),wavelet-based filt ering and spline int erpolation.Here,we introduce anew approach to motion artifact correction,targeted principle component analysis(PCA),which incorporates a PCA filter only on the segments of data identified as motion artifacts.Itis expected that this will overcome the issues of filtering desired signals that plagues standardPCA fitering of entire data sets.We compared the new approach with the most efiective motionartifact correction algorithms on a set of data acquired simultaneously with a collodion-fixedprobe(low motion artifact content)and a standard Velcro probe(high motion artifact content).Our results show that tPCA gives statistically better results in recovering hemodynamic responsefunction(HRF)as compared to wavelet-based fltering and spline interpolation for the Velcroprobe.It resulis in a significant reduction in mean-squauared'error(MSE)and significant en-hancement in Pearson's correlation coeficient to the true HRF,The collodion-fixed fiber probewith no motion correction performed better than the Velcro probe corrected for motion artifactsin terms of MSE and Pearson's correlation coefficient.Thus,if the experimental study permits,the use of a collodion-fixed fiber probe may be desirable.I the use of a collodion-fixed probe is notfeasible,then we suggest the use of tP CA in the processing of motion artifact contaminated data.展开更多
Objective and Impact Statement.Segmentation of blood vessels from two-photon microscopy(2PM)angiograms of brains has important applications in hemodynamic analysis and disease diagnosis.Here,we develop a generalizable...Objective and Impact Statement.Segmentation of blood vessels from two-photon microscopy(2PM)angiograms of brains has important applications in hemodynamic analysis and disease diagnosis.Here,we develop a generalizable deep learning technique for accurate 2PM vascular segmentation of sizable regions in mouse brains acquired from multiple 2PM setups.The technique is computationally efficient,thus ideal for large-scale neurovascular analysis.Introduction.Vascular segmentation from 2PM angiograms is an important first step in hemodynamic modeling of brain vasculature.Existing segmentation methods based on deep learning either lack the ability to generalize to data from different imaging systems or are computationally infeasible for large-scale angiograms.In this work,we overcome both these limitations by a method that is generalizable to various imaging systems and is able to segment large-scale angiograms.Methods.We employ a computationally efficient deep learning framework with a loss function that incorporates a balanced binary-cross-entropy loss and total variation regularization on the network’s output.Its effectiveness is demonstrated on experimentally acquired in vivo angiograms from mouse brains of dimensions up to 808×808×702μm.Results.To demonstrate the superior generalizability of our framework,we train on data from only one 2PM microscope and demonstrate high-quality segmentation on data from a different microscope without any network tuning.Overall,our method demonstrates 10×faster computation in terms of voxels-segmented-per-second and 3×larger depth compared to the state-of-the-art.Conclusion.Our work provides a generalizable and computationally efficient anatomical modeling framework for brain vasculature,which consists of deep learning-based vascular segmentation followed by graphing.It paves the way for future modeling and analysis of hemodynamic response at much greater scales that were inaccessible before.展开更多
A major challenge in neuroscience is visualizing the structure of the human brain at different scales.Traditional histology reveals micro-and meso-scale brain features but suffers from staining variability,tissue dama...A major challenge in neuroscience is visualizing the structure of the human brain at different scales.Traditional histology reveals micro-and meso-scale brain features but suffers from staining variability,tissue damage,and distortion,which impedes accurate 3D reconstructions.The emerging label-free serial sectioning optical coherence tomography(S-OCT)technique offers uniform 3D imaging capability across samples but has poor histological interpretability despite its sensitivity to cortical features.Here,we present a novel 3D imaging framework that combines S-OCT with a deep-learning digital staining(DS)model.This enhanced imaging modality integrates high-throughput 3D imaging,low sample variability and high interpretability,making it suitable for 3D histology studies.We develop a novel semi-supervised learning technique to facilitate DS model training on weakly paired images for translating S-OCT to Gallyas silver staining.We demonstrate DS on various human cerebral cortex samples,achieving consistent staining quality and enhancing contrast across cortical layer boundaries.Additionally,we show that Ds preserves geometry in 3D on cubic-centimeter tissue blocks,allowing for visualization of meso-scale vessel networks in the white matter.We believe that our technique has the potential for high-throughput,multiscale imaging of brain tissues and may facilitate studies of brain structures.展开更多
文摘As near-infrared spectroscopy(NIRS)broadens its application area to diferent age and diseasegroups,motion artifacts in the NIRS signal due to subject movement is becoming an importantchallenge.Motion artifacts generally produce signal fiuctuations that are larger than physio-logical NIRS signals,thus it is cruciai to corect for them before obtaining an cstimate ofstimulusevoked hemodynamic responses.,There are various methods for correction such as principlecomponent analy sis(P CA),wavelet-based filt ering and spline int erpolation.Here,we introduce anew approach to motion artifact correction,targeted principle component analysis(PCA),which incorporates a PCA filter only on the segments of data identified as motion artifacts.Itis expected that this will overcome the issues of filtering desired signals that plagues standardPCA fitering of entire data sets.We compared the new approach with the most efiective motionartifact correction algorithms on a set of data acquired simultaneously with a collodion-fixedprobe(low motion artifact content)and a standard Velcro probe(high motion artifact content).Our results show that tPCA gives statistically better results in recovering hemodynamic responsefunction(HRF)as compared to wavelet-based fltering and spline interpolation for the Velcroprobe.It resulis in a significant reduction in mean-squauared'error(MSE)and significant en-hancement in Pearson's correlation coeficient to the true HRF,The collodion-fixed fiber probewith no motion correction performed better than the Velcro probe corrected for motion artifactsin terms of MSE and Pearson's correlation coefficient.Thus,if the experimental study permits,the use of a collodion-fixed fiber probe may be desirable.I the use of a collodion-fixed probe is notfeasible,then we suggest the use of tP CA in the processing of motion artifact contaminated data.
文摘Objective and Impact Statement.Segmentation of blood vessels from two-photon microscopy(2PM)angiograms of brains has important applications in hemodynamic analysis and disease diagnosis.Here,we develop a generalizable deep learning technique for accurate 2PM vascular segmentation of sizable regions in mouse brains acquired from multiple 2PM setups.The technique is computationally efficient,thus ideal for large-scale neurovascular analysis.Introduction.Vascular segmentation from 2PM angiograms is an important first step in hemodynamic modeling of brain vasculature.Existing segmentation methods based on deep learning either lack the ability to generalize to data from different imaging systems or are computationally infeasible for large-scale angiograms.In this work,we overcome both these limitations by a method that is generalizable to various imaging systems and is able to segment large-scale angiograms.Methods.We employ a computationally efficient deep learning framework with a loss function that incorporates a balanced binary-cross-entropy loss and total variation regularization on the network’s output.Its effectiveness is demonstrated on experimentally acquired in vivo angiograms from mouse brains of dimensions up to 808×808×702μm.Results.To demonstrate the superior generalizability of our framework,we train on data from only one 2PM microscope and demonstrate high-quality segmentation on data from a different microscope without any network tuning.Overall,our method demonstrates 10×faster computation in terms of voxels-segmented-per-second and 3×larger depth compared to the state-of-the-art.Conclusion.Our work provides a generalizable and computationally efficient anatomical modeling framework for brain vasculature,which consists of deep learning-based vascular segmentation followed by graphing.It paves the way for future modeling and analysis of hemodynamic response at much greater scales that were inaccessible before.
基金funding support from:National Institutes of Health grant R01 EY032163(S.C.,L.T)National Institutes of Health grant U54 NS115266(A.C.M.)+8 种基金National Institutes of Health grant P30 AG072978(A.C.M.)National Institutes of Health grant U19 AG068753(A.C.M.)National Institutes of Health grant R01 NS125307(D.L.R.)National Institutes of Health grant RO1 AG075727(D.L.R.and IJ.B)National Institutes of Health grant RF1 AG062831(D.L.R.)Boston University Kilachand Fund Award(I.J.B.,D.L.R.,D.A.B.,and L.T.)National Institutes of Health grant R00 EB023993(HW)National Institutes of Health grant R01 NS128843(H.W.)National Institutes of Health grant U01 MH117023(S.C.,A.N.,D.A.B.)。
文摘A major challenge in neuroscience is visualizing the structure of the human brain at different scales.Traditional histology reveals micro-and meso-scale brain features but suffers from staining variability,tissue damage,and distortion,which impedes accurate 3D reconstructions.The emerging label-free serial sectioning optical coherence tomography(S-OCT)technique offers uniform 3D imaging capability across samples but has poor histological interpretability despite its sensitivity to cortical features.Here,we present a novel 3D imaging framework that combines S-OCT with a deep-learning digital staining(DS)model.This enhanced imaging modality integrates high-throughput 3D imaging,low sample variability and high interpretability,making it suitable for 3D histology studies.We develop a novel semi-supervised learning technique to facilitate DS model training on weakly paired images for translating S-OCT to Gallyas silver staining.We demonstrate DS on various human cerebral cortex samples,achieving consistent staining quality and enhancing contrast across cortical layer boundaries.Additionally,we show that Ds preserves geometry in 3D on cubic-centimeter tissue blocks,allowing for visualization of meso-scale vessel networks in the white matter.We believe that our technique has the potential for high-throughput,multiscale imaging of brain tissues and may facilitate studies of brain structures.
