Project 2: High Resolution Digital Atlases

There are four specific aims:

1. Expand the atlases of the monkey brains sectioned in the three standard planes by the addition of images derived from other microscopic and molecular biological techniques, in order to show the localization of gene expression for the major neurotransmitters and their receptors, with quantitative data of their relative levels.
2. Prepare a comparable atlas of the human thalamus and adjacent regions, emphasizing quantitative morphometric data and comparing these in brains from normals and from subjects suffering from schizophrenia or manic depressive illness.
3. Develop visualization tools for using these atlases, particularly involving the storage of the large amounts of data, for interactive 3-D visualizations, for exporting them in a form readily usable by other scientists, and for extending databases of Neuroscience information associated with them; develop an atlas browser (NeuArt) and multi-resolution technology for interactive and remote data visualization.
4. Provide image data, visualization tools and data about morphological and chemical features (metadata) to be used by the Program as a whole in the preparation of a  highly detailed database of the structure, connectivity and function of the thalamus, using the lateral geniculate nucleus as a model.  This will be constructed by Project 1 in close collaboration with Projects 2-4. The particular visualization techniques that Project 2 will develop and the Program as a whole support, will include slicing three-dimensional image data with interactive re-sectioning tools, generating transparent volume images (volume renderings), superpositioning image data retrieved from diverse data sources, and semi-automatic feature definition/matching mechanisms.

Contributions of Project 2 to the program as a whole

The contributions of this project to the program as a whole are: it provides high resolution cytoarchitectonic brain images and the tools for manipulating them; it provides the images of the primate lateral geniculate nucleus to the database being developed in Project 1; it provides lexical data about lateral geniculate structure, chemistry and cell types to the database.

Specific contributions of Project 2

The contributions specific to this project are: it is developing the highest resolution brain atlases of monkey and human brains yet available; it is developing a browser for incorporating image-based data held in other formats into the atlas; taking advantage of new developments in visualization techniques, it developing interactive visualization tools for generating derived image data and for manipulating the very large image files in remote and distributed settings using high speed computer networks.

In addition to the sample images of the monkey brain available in the above link, we have made images of cat and pigeon  brains available by directory browsing.  These images later will be incorporated into the images available through the atlas viewer; we present them now for interested parties.

Project status

1. Used the improved browser and input tools to populate the monkey atlas database with images, outlines, and necessary metadata (outlines of nuclei and laminae etc.) and to facilitate interleaving of other image material.
2. Built input tool subunits to: (i) parse Adobe Illustrator files into images, cell groups, and names. (ii) generate indexing polygons/bounding boxes from vector-based Illustrator files. (iii) simplify polygons to meet storage requirements of the DBMS.
3. Connected the atlas viewer with the terminology server (jTerm) developed in Project 1. This connection enables a user to interact with the server by: (i) opening a tool that accepts user textual input and then acts on it. (ii) clicking on a region of interest in the atlas image. This retrieves location and textual information from the atlas database and inserts it into the terminology tool.
4. Constructed a language acceptor for the Terminology Query Language (TQL) developed in Project 1. This is the first step toward building a  compiler/interpreter for the language. NeuArt when opened by the viewer operates in a 2-pass paradigm. Pass 1 permits a textual query such as show the results of localization of GABAA receptor mRNAs and the application returns text about the relevant experiments. Pass 2 permits selection of a subset of  data e.g. for a particular GABAA receptor subunit. These can then be viewed as full images presented as an overlay of the currently displayed index atlas level. A spatial query manager permits new atlas levels with the data for this mRNA to be selected.  Interactive tools permit the display to be changed to reduce less relevant content. The spatial query manager enables highly relevant regions such as the lateral geniculate nucleus to be outlined, information such as quantitative mRNA levels found in it to be retrieved and queried against other mRNAs or against connectional or cellular studies. Modifications and enhancements of the textual database, made with Project 1, permit entry to be either visual or textual.
5. The annotator API in NeuArt has been enhanced with the capability to perform queries of and display results from XANAT (Project 4).
6. Digitization of the Berman and Jones Atlas.  We digitized the plates of this cat forebrain atlas and added it to the similarly prepared stereotaxic atlases of Karten (Project 3). The files have been written to CDs for distribution.
7. Processing of High Resolution Histological Images.   The visualization team developed data reduction algorithms for dealing with the high resolution images of sections from the monkey brains, including: large-scale visualization , large-scale volume rendering, and interactive visualization . A project with Arthur Toga's HBP includes  segmentation, automated ice removal and cavity detection in human brain cryosections.
8. Volumetric  Studies on the Human Thalamus in Schizophrenia and Depression. Thalami from 6 pairs of brains from schizophrenics and controls and from 7 pairs of depressed patients and controls, matched for age, sex and autolysis time have been sectioned and stained with routine stains and by in situ hybridization to localize gene expression for mRNAs associated with the GABA and glutamate neurotransmitter systems. A control gene expression atlas has been prepared and preliminary quantitative data are available for the schizophrenics . Using the optical disector on sections from the schizophrenic series, we found a consistent, >30% reduction in neurons restricted to the parvocellular division of the mediodorsal nucleus (MD) in schizophrenics in comparison with controls . This division of MD projects to dorsolateral prefrontal cortex which has been implicated in the pathophysiology of schizophrenia in imaging, behavioral  and gene expression studies.

Personnel

Principal Investigator: Edward G. Jones, UC Davis
Co-PI: Bernd Hamann, UC Davis
Co-PI: Kenneth I. Joy, UC Davis
Co-PI: Nelson L. Max, UC Davis
Co-PI: Kwan-Liu Ma, UC Davis