Mapping Mouse Brain Slice Sequence to a Reference Brain Without 3D Reconstruction

Abstract

Histological brain slices are widely used in neuroscience to study anatomical organization of neural circuits. Since data from many brains are collected, mapping the slices to a reference atlas is often the first step in interpreting results. Most existing methods rely on an initial reconstruction of the volume before registering it to a reference atlas. Because these slices are prone to distortion during sectioning process and often sectioned with nonstandard angles, reconstruction is challenging and often inaccurate. We propose a framework that maps each slice to its corresponding plane in the atlas to build a plane-wise mapping and then perform 2D nonrigid registration to build pixel-wise mapping. We use the L2 norm of the Histogram of Oriented Gradients (HOG) of two patches as the similarity metric for both steps, and a Markov Random Field formulation that incorporates tissue coherency to compute the nonrigid registration. To fix significantly distorted regions that are misshaped or much smaller than the control grids, we trained a context-aggregation network to segment and warp them to their corresponding regions with thin plate spline. We have shown that our method generates results comparable to an expert neuroscientist and is significantly better than reconstruction-first approaches.

Anatomical, Physiological, and Functional Heterogeneity of the Dorsal Raphe Serotonin System

“…To systematically and quantitatively visualize the organization of these output-defined serotonin neurons, we developed an image registration algorithm that allowed us to register all DR-containing histological sections to the Allen Institute reference brain ( Figure 1A, right panels; STAR Methods; Xiong et al, 2018). We then created clusters using the combined data from four brains with the same injection sites ( Figure 1B, right insets).…”

“…We further made improvement based on the data-specific properties of our experimental dataset including segmenting the aqueduct with a convolutional neural network and local warping them with thin plate spline (TPS). For a more detailed description of this procedure see Xiong et al (2018).…”

“…Specifically, serotonin neurons projecting to subcortical areas (PVH, CeA, LHb, and dLGN) tended to localize more in the dorsal DR, especially the dorsal wings ( Figure 1B, top row). By contrast, serotonin neurons that project to the OB and three cortical areas (OFC, PIR, and ENT) preferentially localized in the ventral DR and were rarely found in the dorsal wing ( Figure 1B, bottom row).To systematically and quantitatively visualize the organization of these output-defined serotonin neurons, we developed an image registration algorithm that allowed us to register all DR-containing histological sections to the Allen Institute reference brain ( Figure 1A, right panels; STAR Methods; Xiong et al, 2018). We then created clusters using the combined data from four brains with the same injection sites ( Figure 1B, right insets).…”

A tool for analyzing electrode tracks from slice histology

“…We recommend prioritizing that the transform is accurate nearby the region of interest. An alternative approach could incorporate non-rigid transformations as in Fürth et al, 2018, Song et al, 2018, and Xiong et al, 2018 One feature generally lacking in electrodelocalization toolkits is the use of brain regions' distinct electrophysiological signatures. A future Figure 4.…”

“…Fürth et al, 2018 addresses the latter two problems with automated nonlinear slice registration and visualization tools but expects unangled coronal slices and a manual estimation of the anterior-posterior location. Song et al, 2018 andXiong et al, 2018 demonstrate automated registration of angled slices but do not yet provide an interface for registration and manual fine-tuning. Methods that produce volumetric images of the brain, such as serial-section two-photon imaging and light-sheet imaging of cleared brains, can resolve all three problems (Mayerich et al, 2008, Ragan et al, 2012, Renier et al, 2014, Niedworok et al, 2016, Vandenberghe et al, 2016 but can be impractical due to the extensive setup and expensive equipment required.…”