David N. Levin, M.D., Ph.D. - US grants

Long Term Objectives: Since magnetic resonance (MR) imaging, x-ray computerized tomography (CT), and positron emission tomography (PET) provide complementary information about brain anatomy and function, a given patient often undergoes more than one of these procedures. The physician is then presented with a large number (e.g. 100) of cross-sectional images which may have different geometrical orientations and which portray different aspects of brain anatomy and function. He must integrate this huge amount of fragmented information into a coherent mental picture of the brain. The long term objective of this proposal is to develop software for displaying this information in an explicitly integrated fashion. Specifically, spatial integration will be achieved by using computer graphics to create 3-D renditions of brain anatomy or function from each modality. Multimodality integration will be accomplished by fusing these separate 3-D views into a single, comprehensive model of the brain. Interactive software will be developed for exploring and manipulating this 3-D model, so that anatomical and functional information can be examined at any point and from any viewing angle. This software is expected to be useful for medical diagnosis, surgical planning, radiation therapy planning, and medical education. Specific Aims: Preliminary work in the applicant's laboratory has demonstrated the feasibility of achieving the following aims: 1. Techniques will be developed for using MR images to create 3-D renditions of the surface of the brain, surface of the skin, selected internal structures (e.g. lesions), and blood vessels. The accuracy of each type of rendition will be measured experimentally. 2. PET data will be used to create 3-D views of metabolic activity in the brain's cortex. An existing technique for retrospective image registration will be used to fuse the data from MR and PET into a single integrated 3-D model of brain anatomy and function. Clinical tests of the accuracy of the combined display will be performed. 3. Software "switches" will be developed so that the user can view any combination of anatomical features from MR and functional data from PET. Other tools will enable the operator to "roam through" the 3-D model and inspect cross-sectional images at any selected point. Software for surgery simulation will be written so that the user can "rehearse" surgical procedures on the 3-D model.

The last 15 years have seen dramatic advances in the technology for acquisition of brain images. Magnetic resonance imaging (MRI), positron emission tomography (PET), and x-ray computed tomography (CI) produce scores of cross-sectional images which play a crucial role in diagnosis and therapy. Three-Dimensional models of the patient's brain, created from these data, can be used to Perform computer simulations of neurosurgical procedures or radiation treatments. However, these methods of therapy planning have suffered a missing link namely, the absence of a convenient quantitative method of transferring these highly precise treatment plans from images to the patient's body.The PIs have recently developed an interactive localizer which bridges this gap by registering images with the patient's anatomy in a non-invasive, retrospective manner. The further development of this type of frameless stereotaxy will make it possible to perform fully quantitative image-guided neurosurgery and radiation therapy. Such methods may be less time consuming, less expensive, more convenient, and entail less post-procedure morbidity than conventional qualitative techniques. Specific Aims 1. Construction of a mobile localizer which is suitable for use in inpatient hospital rooms, clinics, operating rooms, or radiation treatment suites. 2. Development of applications software for neurosurgical planning and radiotherapy planning. 3. Testing of the neurosurgical localizer on phantoms, volunteers, and brain tumor patients. 4. Testing of the radiotherapy localizer on phantoms, volunteers, and brain tumor patients.

There is substantial evidence that the central nervous system of young children is marked by "plasticity" or the ability to make compensatory changes in its functional architecture. For example, children may redevelop the ability to speak or to walk after damage to the speech or motor areas. Adults are much less likely to respond in this way. Similar phenomena have been studies in animal experiments where it has been possible to crate detailed cortical maps showing which parts of the remaining brain tissue assume the functions of damaged areas. In the last few years, magnetic resonance imaging (MRI) techniques have been developed to detect signals in functionally active parts of the brain. With appropriate computer graphic and image processing techniques, functional MRI (FMRI) signals can be combined with conventional MR images to create integrated 3-D models of cortical structure and function of individual subjects. However, these computational methods need to be automated so that it is practical to apply them to the huge volumes of data produced by fast MRI acquisitions. Our long-term goal is to develop practical computational tools and MRI technology for the creation of 3-D maps of cortical structure and function. We will use this technology to study the functional reorganization of brains of young adults who sustained anatomical brain damage at an early age. Theoretically, such experiments would increase our understanding of the spatial, temporal, and functional limits of human brain plasticity. Practically, this type of information could help neurosurgeons plan the resection of brain lesions so that important functional areas are left intact. The project will be divided into three parts: 1. Image Segmentation Methods. Software will be developed for automatic identification of the brain surface in MR head images. By greatly reducing the labor required to perform this type of image segmentation, this software will promote the routine use of MRI-derived 3-D brain models for many purposes, including neurosurgical planning and multimodality display. 2. Intraoperative Validation of FMRI. It is important to characterize the accuracy of FMRI as a brain-mapping tool. To do this, FMRI will be sued to predict the location of the sensorimotor cortex in patients scheduled to undergo resection of a small brain tumor. Stereotactic surgical methods will be used to compare the FMRI prediction with the location of the sensorimotor cortex determined by electrocorticography. 3. Brain Mapping of Hemiplegic Subjects. FMRI data and image processing will be used to create cortical maps of bilateral sensorimotor areas in normal volunteers and in age-matched hemiplegic subjects with neonatal brain damage. The spatial distribution of functional activity in these tow groups will be compared with the aid of these cortical maps and also by using a common reference frame, derived from a stereotactic brain atlas. Since the FMRI data from all subjects will be mapped into an atlas-defined reference frame, these data can be readily utilized by other investigators.

DESCRIPTION (provided by applicant): The long range goal of this proposal is to develop practical listening systems (speech-to-text systems) for hearing-impaired persons. Although automatic speech recognition (ASR) technology has made steady progress in recent years, existing systems with large vocabularies often require extensive retraining if the acoustic channel is altered because the noise level changes, the speaker's room or position changes, or the signal conduit changes (telephone vs. room speech). The PI has developed a novel non-linear signal-processing method that can be used in the "front end" of any ASR system in order to increase its channel-independence. The new method does a much better job of decreasing the speech signal's channel-dependence than the commonly-used cepstral mean normalization (CMN). Furthermore, its performance is comparable to that of the combination of CMN and spectral subtraction, despite the fact that it does not utilize the noise measurements required by spectral subtraction. The proposal's first aim is to make certain technical changes that will significantly improve the method's performance. The proposal's second aim is to test the following hypothesis: the word error rate of a typical ASR system will be reduced by incorporation of the new channel normalization technology in its front end.