NIHLM2015604_SOW.docx

STATEMENT OF WORK

Project Title: Advanced Computational Methods for Anatomical and Functional Brain Data Analysis

INTRODUCTION:

Description of services

Professional services are required to develop and evaluate techniques for analyzing anatomical and functional brain data using deformable shape and appearance volume models (Metamorphs/Active Volume Models), stretching open active contours (SOAX), and advanced classification methods, including deep learning. These methods will be investigated and state-of-the-art tools developed for the segmentation of brain MRI and diffusion imaging data and the analysis of fMRI data, with the aim of supporting research into understanding functional brain circuits and their anatomical correlations. Functional-anatomical atlases will be developed to facilitate comparisons across individuals and for statistical modeling. The required work is projected to be a multi-year effort, with the first year concentrating on feasibility and prototype development. Subsequent two-year work, if justified by first-year results, will concentrate on the development and further evaluation of the prototypes as mature tools that contribute to the wider national research initiative to accurately model functioning of the human brain.

Background

Multidisciplinary research to understand functional brain connections and circuits and their anatomical correlations continues to expand, especially in view of the NIH Brain Initiative. The understanding of the biological underpinning for human behavior and experience, including cognitive states, emotion, perception, communication, and motor control, is expected to play a major role in future disease treatment (including diseases of cognitive decline and psychological disorders), accelerated brain/machine interface development, and perhaps enhancements to sensory perception, memory, and learning.

Toward this goal research and development is required to develop and evaluate tools that contribute to the analysis of anatomical and functional data of the brain by supporting researchers in the discovery and validation of connectivities and circuits within functional neuroimaging data (e.g., fMRI) and in correlating these functional characteristics with the underlying brain anatomy.

Recent methods with possible application to brain segmentation have been developed, but not extensively evaluated, namely the Metamorphs/Active Volume Models [9,35] methods for generic volume segmentation and the SOAX [37] method, which is expected to have application for DTI data segmentation.

Also, the application of pattern recognition/classification methods to the analysis of fMRI data has become a topic of increasing research interest [1-6], including, recently, the application of deep learning methods [7]. Finally, a growing body of research literature is concerned with the fusion of anatomical and functional data. (See [8], for example, for a survey of DTI and fMRI fusion methods.)

Proposed Work

The contractor will acquire functional and anatomical brain data from open source repositories (e.g., the Human Connectome Project) or as otherwise negotiated with the government. The anatomical data will include both conventional MRI anatomical data and Diffusion Tensor Imaging (DTI) data. The functional brain data will include fMRI data.

The contractor will implement and evaluate methods to segment the anatomical brain data using previously-developed algorithms for deformable shape and appearance volume models and stretching open active contours (for the DTI data) and quantitatively compare these methods with methods from the published literature.

The contractor will investigate state-of-the-art methods for analysis of fMRI data, and will implement and evaluate algorithms with the goal of activation regions and connectivities in both task-related and resting state fMRI data. In particular, the contractor will investigate pattern recognition and classification techniques, and their application to fMRI data analysis, with particular emphasis on machine learning methods; these will include, specifically, deep learning and its use in fMRI.

REQUIREMENTS:

REQUIRED TASKS

Task Description: Base Period

Task 1: Acquire human brain MRI and DTI anatomical data, and fMRI functional brain data from an open source repository, or as otherwise negotiated with the government.

Investigate the application of active volume models, MetaMorphs, and SOAX for segmentation of brain anatomy; implement and provide initial experimental evaluations of at least one of these methods.

Investigate the application of pattern recognition/classification methods including, specifically, deep learning, to the analysis of fMRI data. As negotiated with the government, and considering the available fMRI data, implement a deep learning method to identify anatomical activation sites, connectivities, and/or classification of brain states (for example, in task-related data collections where the subject is shown various classes of objects). Provide initial experimental results.

Task 2: Implement and provide refined experimental evaluations for each of the active volume models, MetaMorphs, and SOAX methods for segmentation of brain anatomy.

Provide refined experimental results for the deep learning method application of Task 1 and a comparative analysis of state-of-the-art pattern recognition/classification methods as applied to the analysis of fMRI data.

Deliverables

Partial payments will be authorized to the Contractor upon receipt of deliverables. The deliverables under this statement of work shall be provided by the Contractor according to the following schedule:

Period of Performance:

It is anticipated that the period of performance shall be 12 months from the date of award, with two (2) one-year option periods, if exercised. An award is anticipated to be made on or about September 18, 2015. Award is this requirement is subject to available funding.

Notice of Government Unlimited Rights to Work First Produced Under This Contract

Government rights to work Ufirst produced under this contractU are established by Federal law including, but not limited to, this specific reference: FAR 42.227-14, Rights in Data – General, (b) (1).

Requirement to Notify Government of Proprietary Work Dependencies

It is the intent of the Government that all algorithms and code developed under this contract be for public use and benefit, including, but not limited to, internal use by government institutions, use by researchers or other parties collaborating in government work, and public domain release, at the option of the government. The Contractor is required to notify the Government in writing of any dependencies of the deliverables under this contract on proprietary, copyrighted, or patented work that potentially inhibits, restricts, or requires permission for the dissemination of the deliverables to the public, other governmental agencies or research groups, or to any other parties whatsoever.