Machine learning analysis of whole mouse brain vasculature

Hierarchical imaging and computational analysis of three-dimensional vascular network architecture in the entire postnatal and adult mouse brain

“…For example, here we have an isotropic voxel size of 0.65 µm, compared to 1.625 µm x 1.625 µm x 3 µm achieved through multi-dye vessel staining, tissue clearing, and 3D LSFM in 145 . For analysis, anisotropic voxels are often resampled to isotropic voxel size corresponding to the lowest resolution component, as was also done in 145 , resulting in isotropic voxel size of 3 µm for convolutional neural network vessel segmentation and subsequent analysis of the vasculature of the whole mouse brain 145 . Micro CT imaging with use of intravascular contrast agents is used both in-vivo and ex-vivo and made it possible to image the vasculature in whole organs in 3 dimensions with increasing resolution 107,147–150 . Ex-vivo, vascular corrosion casts have shown to be an excellent method to investigate brain vasculature in particular 40,104,106,107,109,110,120 .…”

“…Similar morphological classification is more difficult and tedious with other, classical (automated) histology-based methods due to dye color variations, staining artefacts, and distortion of the true vessel diameter by perivascular cells 138–140 . Histology-based methods including recently published LSFM based methods 141–145 are also less precise in addressing inner vessel diameters. Hierarchical imaging and computational reconstruction of the 3D vascular network via global vascular network morphometry and local vascular network topology allow accurate, complete and quantitative analysis and description of the 3D vessel structure in the postnatal- and the adult mouse brain at the network level. These computational 3D analyses result in vessel parameters (Figures 5–9, and Extended Data Figures 3–9) with direct biological relevance instead of single measures such as ‘percentage of section area occupied by vessels’, that are obtained in classical 2D-section analysis methods. For instance, maximum intensity projections of tissue slices are often used, even though they lead to an overestimation of the vessel density, as the vascular volume fractions of multiple optical sections are summed up, an error that adds up with increasing section thickness.…”

Machine learning algorithms to predict 30-day readmission in patients with stroke: a prospective cohort study

“…When all models were constructed using a given set of clinical inputs, the ANN model was clearly superior to other forecasting models. Additionally, unlike previous works in which the analyses were performed using a dataset for a single medical center, our study used prospective and longitudinal data from multiple medical institutions, which provides a more accurate depiction of current treatment for patients with stroke [10][11][12][13]. Additionally, unlike single-center series studies, our use of registry data provides more accurately depicts stroke treatment in large populations.…”

“…Although many models for predicting outcomes of stroke treatments have been proposed in recent years, models for predicting 30-day readmission have had major shortcomings: (1) recently proposed forecasting models have shown lower prediction accuracy compared to conventional models [4][5][6][7][8][9], (2) proposed forecasting models require use of health insurance claims data, which would not be available in a real-time clinical setting [8,10], (3) predictions of 30-day readmission do not consider post-acute care (PAC) program, patient attributes, clinical attributes and functional status score before rehabilitation [11][12][13][14][15]. In the current study, the best predictors of hospital readmission within 30 days after stroke were identi ed using arti cial neural network (ANN), K nearest neighbor (KNN), support vector machine (SVM), naive Bayes classi er (NBC) and Cox regression (COX) models.…”

Human Developmental Cell Atlas: milestones achieved and the roadmap ahead

“…A 3D imaging pipeline applied to embryonic/fetal human organs and even whole human embryos [49][50][51] has proven its value in mapping cells during certain developmental stages. Progress is being made to increase the multiplex capacity of this approach and use of artificial intelligence/machine learning algorithms to overcome data analytical challenges, as was recently deployed to study whole-organismal vasculature following tissue clearing 52,53 . 61 .…”