Abstract
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 . 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.
Highlights
The hemodynamic response to neural activation has become a vital tool in understanding brain function and pathologies [1]
Due to rapid deterioration of measurement contrast with imaging depth in 2PM, these methods have been unable to demonstrate effective segmentation for vasculature deep beneath the brain surface. We address these limitations and present a computationally efficient framework for 2PM vascular segmentation that allows us to effectively process much larger regions of the mouse brain compared to existing methods at significantly faster computation speed in terms of voxels segmented per second
We propose a novel deep learning method for vascular segmentation of cerebral 2PM angiograms, which overcomes the aforementioned limitations of the existing techniques and demonstrates state-of-the-art segmentation performance
Summary
The hemodynamic response to neural activation has become a vital tool in understanding brain function and pathologies [1]. Mathematical models have been proposed for these neuroimaging methods, which provide valuable insight into the relation between the measured signals, and the underlying physiological parameters, such as cerebral blood flow, oxygen consumption, and rate of metabolism [4,5,6,7]. These mathematical models often require a topological representation of the blood vessels as a graph of BME Frontiers spatially distributed nodes, connected via edges [5, 7]. Our method demonstrates accurate segmentation for significantly deeper vasculature compared to the state-of-the-art
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
