BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets.

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

Brain microvasculature has a common topology with local differences in geometry that match metabolic load.

The microvasculature underlies the supply networks that support neuronal activity within heterogeneous brain regions. What are common versus heterogeneous aspects of the connectivity, density, and orientation of capillary networks? To address this, we imaged, reconstructed, and analyzed the microvasculature connectome in whole adult mice brains with sub-micrometer resolution. Graph analysis revealed common network topology across the brain that leads to a shared structural robustness against the rarefaction of vessels. Geometrical analysis, based on anatomically accurate reconstructions, uncovered a scaling law that links length density, i.e., the length of vessel per volume, with tissue-to-vessel distances. We then derive a formula that connects regional differences in metabolism to differences in length density and, further, predicts a common value of maximum tissue oxygen tension across the brain. Last, the orientation of capillaries is weakly anisotropic with the exception of a few strongly anisotropic regions; this variation can impact the interpretation of fMRI data.

Reconstruction of 1,000 Projection Neurons Reveals New Cell Types and Organization of Long-Range Connectivity in the Mouse Brain.

Neuronal cell types are the nodes of neural circuits that determine the flow of information within the brain. Neuronal morphology, especially the shape of the axonal arbor, provides an essential descriptor of cell type and reveals how individual neurons route their output across the brain. Despite the importance of morphology, few projection neurons in the mouse brain have been reconstructed in their entirety. Here we present a robust and efficient platform for imaging and reconstructing complete neuronal morphologies, including axonal arbors that span substantial portions of the brain. We used this platform to reconstruct more than 1,000 projection neurons in the motor cortex, thalamus, subiculum, and hypothalamus. Together, the reconstructed neurons constitute more than 85 meters of axonal length and are available in a searchable online database. Axonal shapes revealed previously unknown subtypes of projection neurons and suggest organizational principles of long-range connectivity.

Single-neuron axonal reconstruction: The search for a wiring diagram of the brain.

Reconstruction of the axonal projection patterns of single neurons has been an important tool for understanding both the diversity of cell types in the brain and the logic of information flow between brain regions. Innovative approaches now enable the complete reconstruction of axonal projection patterns of individual neurons with vastly increased throughput. Here, we review how advances in genetic, imaging, and computational techniques have been exploited for axonal reconstruction. We also discuss how new innovations could enable the integration of genetic and physiological information with axonal morphology for producing a census of cell types in the mammalian brain at scale.

Distinct descending motor cortex pathways and their roles in movement.

Activity in the motor cortex predicts movements, seconds before they are initiated. This preparatory activity has been observed across cortical layers, including in descending pyramidal tract neurons in layer 5. A key question is how preparatory activity is maintained without causing movement, and is ultimately converted to a motor command to trigger appropriate movements. Here, using single-cell transcriptional profiling and axonal reconstructions, we identify two types of pyramidal tract neuron. Both types project to several targets in the basal ganglia and brainstem. One type projects to thalamic regions that connect back to motor cortex; populations of these neurons produced early preparatory activity that persisted until the movement was initiated. The second type projects to motor centres in the medulla and mainly produced late preparatory activity and motor commands. These results indicate that two types of motor cortex output neurons have specialized roles in motor control.

Dissociable Structural and Functional Hippocampal Outputs via Distinct Subiculum Cell Classes.

The mammalian hippocampus, comprised of serially connected subfields, participates in diverse behavioral and cognitive functions. It has been postulated that parallel circuitry embedded within hippocampal subfields may underlie such functional diversity. We sought to identify, delineate, and manipulate this putatively parallel architecture in the dorsal subiculum, the primary output subfield of the dorsal hippocampus. Population and single-cell RNA-seq revealed that the subiculum can be divided into two spatially adjacent subregions associated with prominent differences in pyramidal cell gene expression. Pyramidal cells occupying these two regions differed in their long-range inputs, local wiring, projection targets, and electrophysiological properties. Leveraging gene-expression differences across these regions, we use genetically restricted neuronal silencing to show that these regions differentially contribute to spatial working memory. This work provides a coherent molecular-, cellular-, circuit-, and behavioral-level demonstration that the hippocampus embeds structurally and functionally dissociable streams within its serial architecture.

Crowdsourcing the creation of image segmentation algorithms for connectomics.

To stimulate progress in automating the reconstruction of neural circuits, we organized the first international challenge on 2D segmentation of electron microscopic (EM) images of the brain. Participants submitted boundary maps predicted for a test set of images, and were scored based on their agreement with a consensus of human expert annotations. The winning team had no prior experience with EM images, and employed a convolutional network. This "deep learning" approach has since become accepted as a standard for segmentation of EM images. The challenge has continued to accept submissions, and the best so far has resulted from cooperation between two teams. The challenge has probably saturated, as algorithms cannot progress beyond limits set by ambiguities inherent in 2D scoring and the size of the test dataset. Retrospective evaluation of the challenge scoring system reveals that it was not sufficiently robust to variations in the widths of neurite borders. We propose a solution to this problem, which should be useful for a future 3D segmentation challenge.

Thalamocortical input onto layer 5 pyramidal neurons measured using quantitative large-scale array tomography.

The subcellular locations of synapses on pyramidal neurons strongly influences dendritic integration and synaptic plasticity. Despite this, there is little quantitative data on spatial distributions of specific types of synaptic input. Here we use array tomography (AT), a high-resolution optical microscopy method, to examine thalamocortical (TC) input onto layer 5 pyramidal neurons. We first verified the ability of AT to identify synapses using parallel electron microscopic analysis of TC synapses in layer 4. We then use large-scale array tomography (LSAT) to measure TC synapse distribution on L5 pyramidal neurons in a 1.00 × 0.83 × 0.21 mm(3) volume of mouse somatosensory cortex. We found that TC synapses primarily target basal dendrites in layer 5, but also make a considerable input to proximal apical dendrites in L4, consistent with previous work. Our analysis further suggests that TC inputs are biased toward certain branches and, within branches, synapses show significant clustering with an excess of TC synapse nearest neighbors within 5-15 µm compared to a random distribution. Thus, we show that AT is a sensitive and quantitative method to map specific types of synaptic input on the dendrites of entire neurons. We anticipate that this technique will be of wide utility for mapping functionally-relevant anatomical connectivity in neural circuits.

Biological reactivity of nanoparticles: mosaics from optical microscopy videos of giant lipid vesicles.

Emerging fields such as nanomedicine and nanotoxicology, demand new information on the effects of nanoparticles on biological membranes and lipid vesicles are suitable as an experimental model for bio-nano interaction studies. This paper describes image processing algorithms which stitch video sequences into mosaics and recording the shapes of thousands of lipid vesicles, which were used to assess the effect of CoFe(2)O(4) nanoparticles on the population of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine lipid vesicles. The applicability of this methodology for assessing the potential of engineered nanoparticles to affect morphological properties of lipid membranes is discussed.

Principal curves as skeletons of tubular objects: locally characterizing the structures of axons.

Developments in image acquisition technology make high volumes of neuron images available to neuroscientists for analysis. However, manual processing of these images is not practical and is infeasible for larger and larger scale studies. Reliable interpretation and analysis of high volume data requires accurate quantitative measures. This requires analysis algorithms to use mathematical models that inherit the underlying geometry of biological structures in order to extract topological information. In this paper, we first introduce principal curves as a model for the underlying skeleton of axons and branches, then describe a recursive principal curve tracing (RPCT) method to extract this topology information from 3D microscopy imagery. RPCT first finds samples on the one dimensional principal set of the intensity function in space. Then, given an initial direction and location, the algorithm iteratively traces the principal curve in space using our principal curve tracing (PCT) method. Recursive implementation of PCT provides a compact solution for extracting complex tubular structures that exhibit bifurcations.

Motion flow analysis in cell videos using a multi-level clustering method.

Analyzing motion flow of cells is an important task for many biomedical applications. It is a challenging problem due to noise in images and uncontrolled motion of cells. In this study, a method to find regions of organized motion and direction of flow is proposed. Since dense optical flow methods might fail due to homogeneous regions and irregular motion patterns, the technique involves analyzing trajectories of strong corner features. Trajectories are clustered to find dominant flow patterns for different regions of the frame, where a multilevel clustering scheme is followed. Experiments show that the technique gives accurate results for detecting region and direction of flow.

Centerline extraction with principal curve tracing to improve 3D level set esophagus segmentation in CT images.

For radiotherapy planning, contouring of target volume and healthy structures at risk in CT volumes is essential. To automate this process, one of the available segmentation techniques can be used for many thoracic organs except the esophagus, which is very hard to segment due to low contrast. In this work we propose to initialize our previously introduced model based 3D level set esophagus segmentation method with a principal curve tracing (PCT) algorithm, which we adapted to solve the esophagus centerline detection problem. To address challenges due to low intensity contrast, we enhanced the PCT algorithm by learning spatial and intensity priors from a small set of annotated CT volumes. To locate the esophageal wall, the model based 3D level set algorithm including a shape model that represents the variance of esophagus wall around the estimated centerline is utilized. Our results show improvement in esophagus segmentation when initialized by PCT compared to our previous work, where an ad hoc centerline initialization was performed. Unlike previous approaches, this work does not need a very large set of annotated training images and has similar performance.

Towards respiration management in radiation treatment of lung tumors: transferring regions of interest from planning CT to kilovoltage X-ray images.

Tracking of lung tumors is imperative for improved radiotherapy treatment. However, the motion of the thoracic organs makes it a complicated task. 4D CT images acquired prior to treatment provide valuable information regarding the motion of organs and tumor, since it is manually annotated. In order to track tumors using treatment-day X-ray images (kV images), we need to find the correspondence with CT images so that projection of tumor region of interest will provide a good estimate about the position of the tumor on the X-ray image. In this study, we propose a method to estimate the alignment and respiration phase corresponding to X-ray images using 4D CT data. Our approach generates Digitally Reconstructed Radiographs (DRRs) using bilateral filter smoothing and computes rigid registration with kV images since the position and orientation of patient might differ between CT and treatment-day image acquisition processes. Instead of using landmark points, our registration method makes use of Kernel Density Estimation over the edges that are not affected much by respiration. To estimate the phase of X-ray, we apply template matching techniques between the lung regions of X-ray and registered DRRs. Our approach gives accurate results for rigid registration and provides a starting point to track tumors using the X-ray images during the treatment.

Analysis of lipid vesicle populations from microscopy video sequences.

Conducting research on lipid vesicles is very convenient, since they provide a stable and controllable environment for in vitro observations. Their resemblance to biological cell membranes allows biologists to assess hazardous potential of nanoparticles by exposing the vesicles instead of live organisms. When considering behavior of vesicles during incubation with nanoparticles, majority of existing research focus on observing single vesicles only. Our approach provides an ability to observe thousands of lipid vesicles for more representative behavior estimation. We developed an efficient algorithm to transform video sequences acquired with video microscopy into quantitative data. This includes steps required to filter noise, use multiple frames for more precise content presentation, detection of regions of interest, and segmentation of circular and non-primitively shaped vesicles. Presented work is a crucial step towards the creation of an automated computer analysis for lipid vesicles behavior assessment.

Measurement of brain iron distribution in Hallevorden-Spatz syndrome.

PURPOSE: To investigate spatial distribution of iron accumulation in the globus pallidus (GP) in patients with Hallevorden-Spatz syndrome (HSS) using phase imaging. We compared sensitivity of a phase imaging technique to relaxation rate measurement methods (R1,R2,R2*) for iron quantification. MATERIALS AND METHODS: R1, R2, and R2* were measured in GP structure of the brain of eight pantothenate kinase-associated neurodegeneration (PKAN) patients and a healthy volunteer using a 3T magnetic resonance imaging (MRI) scanner. The phase of gradient-echo images was preprocessed to eliminate phase 2pi wrapping and filtered to remove phase background variations. Phase gap across GP structure was used as a metric for iron effects quantification. RESULTS: Among the relaxation rates the most sensitive to iron accumulation was the R2* rate. The R1 and R2 rates demonstrated only small variations in this group of subjects. Up to an order of magnitude phase gap changes were measured between one PKAN patient and an age-matched healthy volunteer. Assuming that phase gap differences scale linearly with iron concentration we estimate that up to 2 mg Fe/g ww accumulates in GP of these patients. CONCLUSION: Our results demonstrate significantly higher sensitivity of the phase measurements for quantitative assessment of iron concentration compared to the relaxation rate measurements. Phase measurements could potentially be used for monitoring a progression and a response to therapy in PKAN.

Towards fish-eye camera based in-home activity assessment.

Indoors localization, activity classification, and behavioral modeling are increasingly important for surveillance applications including independent living and remote health monitoring. In this paper, we study the suitability of fish-eye cameras (high-resolution CCD sensors with very-wide-angle lenses) for the purpose of monitoring people in indoors environments. The results indicate that these sensors are very useful for automatic activity monitoring and people tracking. We identify practical and mathematical problems related to information extraction from these video sequences and identify future directions to solve these issues.