Query-Guided Visual Analysis of Large Volumetric Neuroscience Data
IEEE SciVis 2013
||Jeff W. Lichtman
|Harvard University, KAUST
This paper presents ConnectomeExplorer, an application for the interactive exploration and query-guided visual analysis of large volumetric electron microscopy (EM) data sets in connectomics research. Our system incorporates a knowledge-based query algebra that supports the interactive specification of dynamically evaluated queries, which enable neuroscientists to pose and answer domain-specific questions in an intuitive manner. Queries are built step by step in a visual query builder, building more complex queries from combinations of simpler queries. Our application is based on a scalable volume visualization framework that scales to multiple volumes of several teravoxels each, enabling the concurrent visualization and querying of the original EM volume, additional segmentation volumes, neuronal connectivity, and additional meta data comprising a variety of neuronal data attributes. We evaluate our application on a data set of roughly one terabyte of EM data and 750 GB of segmentation data, containing over 4,000 segmented structures and 1,000 synapses. We demonstrate typical use-case scenarios of our collaborators in neuroscience, where our system has enabled them to answer specific scientific questions using interactive querying and analysis on the full-size data for the first time.
enables the interactive visual analysis of large data volumes in connectomics research, integrating 3D volume data of brain tissue and segmented objects, connectivity of cells and neurites, and additional meta data. A powerful query algebra allows neuroscientists to pose domain-specific questions in an intuitive manner, and to interactively analyze the results. We show a teravoxel volume via (left) a tree widget of segmented objects, a 3D volume view, a connectivity graph of neurites (axons and dendrites), a visual query builder for dynamic query specification, and a statistics view at the bottom; (right) visualization of a dynamically specified region of interest (top: all neuronal objects in a cylindrical region; bottom: only the spines of the red dendrite).
Our system consists of two main parts: data-driven modules (left, light gray) are triggered by image acquisition, while visualization/user-driven modules (right, yellow) are active at run time. All generated data are stored in the visualization archive. At run time, the user initializes data requests either by specifying a dynamic query or by directly interacting with the visualization. A visual query builder allows users to intuitively specify queries which are translated into a powerful query algebra and evaluated. Results are shown in different linked visualizations.
A user can dynamically create queries by either using the visual query builder or by direct interaction in the provided views via picking or selection. User input is represented and evaluated via a query algebra, and all result sets can be immediately visualized.
Example query tree.
Each user-defined query corresponds to a tree, where the leaves are objects or sets, and the internal nodes are predicates or operators. Each node also corresponds to its result set. Here, the left sub-tree returns all axons connected to dendrite 1 (see connectivity graph in left inset); the right sub-tree evaluates a region of interest. The final result set (tree root) is the set intersection of both.
Dynamic queries and linked views.
All views allow direct user interactions (e.g., picking, selection) that can be used as input for new queries or for exploration of the data set. New queries are specified in the visual query builder and automatically update all views after query evaluation.
Example analysis session.
a) The user starts with a manual inspection of the segmentation, before looking at a single dendrite (in red) and its connected axons. b) Looking at all connections that the red dendrite makes, and querying the axon it makes most synapses with (blue axon). c) Looking at all connections of the blue axon and their strengths, grouped by dendrite, viewed in a histogram. d) Comparing all synapses of the red dendrite based on specific properties: the size of dendritic spines vs. the vesicle count of axons/boutons were analyzed using a scatter plot.