You are here: Home Research


Overview of VML research projects

The Visualization and Modeling Laboratory research projects integrate machine learning, visualization, statistical analysis and advanced data-handling to understand large scale data sets. We focus on building tools that are usable not only in our lab, but for other researchers. Here is a brief overview of some of our projects:


CTAGGER is a MATLAB/Java toolbox that implements a user-friendly, semi-structured and expandable strategy for event annotation in dynamic brain imaging and other time-series. To facilitate common labeling and comparison across data collections and laboratories, an individual’s annotation can be collected in a common community database to encourage annotation reuse. The tools allow the creation of multiple event overlays to facilitate the reuse and combination of brain imaging data for multiple analyses.


Davis (Data Viewing System) is a general-purpose data viewer designed for the simultaneous display of a large number of dynamic data sets. Davis was inspired by the need to explore computational models of the cerebral cortex. These systems are distinguished by complex dynamic elements interconnected in irregular patterns. Neuroscientists study the detailed behavior of individual elements and how these elements interact to achieve cortical function. Davis is written in Java and can be run from a browser or as a standalone application. Davis visualizations can be applied to any collection of space-time data sets, and the Davis infrastructure allows visualizations to be added easily.


DETECT is a MATLAB Toolbox for detecting and identifying events in long multi-channel time series, such as the analysis of electroencephalography (EEG) signals. DETECT can be used to detect multiple types of events and returns labels and time indices where the events occur, allowing for the further analysis of events.


EEGVIS is a MATLAB toolbox that allows users to quickly explore multi-channel EEG and other large array-based data sets using multi-scale drill-down techniques. Customizable summary views reveal potentially interesting sections of data, which users can explore further by clicking to examine using detailed viewing components. The viewer and a companion browser are built on our MoBBED framework, which has a library of modular viewing components that can be mixed and matched to best reveal structure. Users can easily create new viewers for their specific data without any programming during the exploration process. These viewers automatically support pan, zoom, resizing of individual components, and cursor exploration. The toolbox can be used directly in MATLAB at any stage in a processing pipeline, as a plug in for EEGLAB, or as a standalone precompiled application without MATLAB running.


MOBBED is a lightweight, easy-to-use, extensible toolkit that allows users to incorporate a computational database into their normal MATLAB workflow. Although capable of storing quite general types of annotated data, MOBBED is particularly oriented to multichannel time series such as EEG that have event streams overlaid with sensor data. A database provides several advantages not available to users who process one dataset at a time from the local file system.  In addition to archiving primary data in a central place to save space and avoid inconsistencies, such a database allows users to manage, search, and retrieve events across multiple datasets without reading the entire dataset. The database also provides infrastructure for handling more complex event patterns that include environmental and contextual conditions.  The database can also be used as a cache for expensive intermediate results that are reused in such activities as cross-validation of machine learning algorithms.


SIDECACHE is a framework for deploying web services that must be updated periodically.  SIDECACHE supports several types of services including proxy access with rate control, local caching, and automatic web service updates. Web services that access busy sites such as NCBI are subject to rate and access restrictions. The SIDECACHE proxy service allows developers to take advantage of local caching to reduce traffic to other sites, while keeping access rates within restrictions. Many biological web services rely on information that is frequently updated. Deployers of these services must then manually redownload the needed information and rerun deployment algorithms. SIDECACHE rebuildable services allow the developer to specify an update schedule at the time of deployment and SIDECACHE handles the necessary synchronization and updating.


Sidekick is a web-based biological decision-making framework that helps you explore relationships among genes. As its moniker reflects, Sidekick is designed to be an assistant that eases analysis burdens (both computational and book-keeping) and helps you organize both the results and your belief in the quality of the results. Before going to the wet-lab, you can use Sidekick with an initial list of genes or concepts of interest to focus possible directions for research by quickly exploring interactions, orthologies, and enrichment. After experiments or an analysis have produced results, you can further explore and develop explanations for the meaning of these results.