Richard E. Greenblatt - US grants

Magnetoencephalography (MEG), the measurement of magnetic fields generated by electric currents in the brain, promises to be a safe, sensitive, non-invasive procedure for the measurement and localization of normal and pathological brain function. The main objective of this proposal is to develop and implemnt better source localization algorithms. Single current dipole localization procedures will be extended and validated. An interactive graphical display will be developed which will combine dipole fit, time series, magnetic field distribution and head shape data. A source localization software tools library will be designed. During Phase II, the library will be implemented and MEG data will be combined with structural images from MRI and/or CT. This project has the potential for significantly improving existing source localization software. The results will find application in the functional study of the normal brain, improved localization of epileptic foci and other neurological disorders, neuropharm- acological studies, and in the development of new sensor technology for neuromagnetometry.

Magnetoencephalography (MEG) is the measurement of extracranial magnetic fields generated by intracellular electric currents in the brain. It offers the possibility to locate discrete source of normal and pathological electrical activity with three-dimensional accuracy of a few millimeters. Current applications include epileptogenic foci localization, evoked response studies, and cognitive processing research. MEG promises to become an important noninvasive spatial imaging modality that reveals dynamic electrical function of the brain. This project addresses a major source of uncertainty in locating the volume of brain tissue responsible for the measured fields, namely the difficulty in accurately positioning the MEG sensor with respect to the skull. The current positioning systems are too imprecise and so slow and difficult to use that they limit the feasibility of MEG for many of the most promising applications where hour-long measurement sessions will not be tolerated. All aspects of the problem are addressed. These include a gantry to position and support the MEG senor, a combined patient support table and chair, means to accurately read the sensor position relative to the head, and a text fixture (phantom) to verify system performance.

Software and hardware will be developed to visuaalize the estimated extent and time course of activity within the brain from an array of EEG electrodes placed on the scalp. This requires the development, implementation, and validation of signal processing algorithms to identify appropriate signal components and estimate the source locations and time course. Preliminary results suggest the feasibility of the approach, which uses novel linear estimation techniques, not currently available for the analysis of electrical recordings. Methods will be verified using phantoms (phase I) and physiological data (phase II). Potential applications of these methods include both research and eventual clinical uses, such as the localization and lateralization of language function, monitoring of stroke and psychiatric disorders, and the estimation of cortical activity during sensory evoked responses and cognitive activity.

Software will be developed to visualize the estimated extent and time course of activity within the brain from an array of MEG sensors placed over the head. The software will include both a Windows-based turnkey package for analysis and display, and a toolkit/subroutine library to aid researchers in developing their own analysis applications. This requires the implementation and validation of algorithms to identify multiple sources and estimate their locations and time course. Preliminary results suggest the feasibility of the approach, which uses novel estimation techniques, not currently available for the analysis of magnetic recordings. Methods will be verified using phantoms and physiological data. Potential applications of these methods include both research and eventual clinical uses, such as the localization and lateralization of language function, monitoring of stroke and psychiatric disorders, and the estimation of cortical activity during sensory evoked responses and cognitive activity.

Multimodal functional brain imaging software will be developed to estimate and visualize the estimated spatial extent and time course of brain activity by combining information from magnetic resonance imaging (MRI) with electroencephalography (EEG) and/or magnetoencephalography (MEG). Structural information from MRI will be combined with extracranial EEG and/or MEG measurements through algorithms developed to segment the MR images and to represent scalp, skull, and brain boundaries as computational objects. This structural information may then be used to improve the spatial accuracy and resolution of existing EEG and MEG source estimation algorithms, while supporting millisecond temporal resolution. The software will comprise a PC/Windows-based program suite for analysis and display. The methods will be verified both with simulated data and with physiological data. The algorithms and software may be used to study both normal brain function, such as measurements in cognitive neuroscience which may be studied with evoked response/event related potentials or spontaneous EEG, and in diseases of the brain, such as epilepsy, where precise spatial and temporal resolution may be of value for diagnosis and presurgical evaluation. PROPOSED COMMERCIAL APPLICATION The techniques which we propose are a non-invasive, non-radiological and relatively low cost addition to existing EEG, MEG and MRI systems, and provides information which is not currently available from these systems independently. The resulting software will have direct application in clinical and cognitive neuroscience research. If clinical value is demonstrated, systems based on this methodology may find applications in the areas of psychiatry, neurology and psychology.

DESCRIPTION (Adapted from Applicant's Abstract): Functional imaging techniques are important to brain researchers and clinicians alike because many phenomena cannot be observed by anatomical techniques alone. Among functional imaging methods, only magneto- and electro- encephalography (MEG, EEF, or jointly M/EEG) can noninvasively resolve events with a millisecond time scale. Statistical tools for M/EEG functional brain imaging software will be developed to estimate and visualize the spatial extent and time course of brain activity. Algorithms will be developed for the incorporation of a priori information into source estimation, and for estimating the uncertainty of the estimates. These tools will permit the use of information derived from anatomy, physiology, and other functional imaging modalities (such as fMRI and PET) to be combined with M/EEG data to improve the robustness, reliability, and objectivity of the M/EEG analysis. The algorithms will be incorporated into prototype software, and the software validated with both simulated and experimental data. The software will comprise a PC/Windows-based program suite for analysis and display. The algorithms and resulting software may be used to study both normal brain function, such as measurements in cognitive neuroscience which may be studied with evoked response/event related potentials or spontaneous EEF, and in diseases of the brain, such as the epilepsies, where precise spatial and temporal resolution may be of value for diagnosis and presurgical evaluation.

Near infrared (NIR) optical imaging may provide the basis for a safe, noninvasive, transportable and relatively inexpensive device for functional brain imaging. Absorbtion by human tissue of light in the red and NIR wavelength regions is low enough that diffusively transmitted light can be collected from regions centimeters below the skin surface. The transmitted light may be used to obtain spatiotemporal information about the brain s optical properties, especially those properties that are known to be modified by the metabolic state of the brain (such as blood oxygenation) and by neuronal activity. The long-term objective of this project is the development of analysis and imaging software which will function as part of an instrument for NIR functional imaging, and support integration with other imaging modalities, such as MRI. During Phase I, prototype analysis and visualization software will be developed and tested. In addition, mathematical models of photon transport through tissue will be implemented as computer software, using finite element techniques. These calculations will be compared with results obtained from experimental measurements on a calibrated phantom. The successful completion of this preparatory work will lay the basis for the development and testing of image reconstruction software during Phase II. The result will be a novel and innovative product which has the potential to facilitate the development of basic and preclinical research in NIR imaging, with the possible future development into part of a clinical diagnostic imaging product. PROPOSED COMMERCIAL APPLICATION Research interest in NIR imaging is expanding and the proposed software may facilitate this growth by providing a useful product where none now exists. The resulting software will have direct application in preclinical and cogntive neuroscience research. If clinical value is demonstrated, systems based on this methodology may find applications in the areas of neurology and neurosurgery.

Functional imaging techniques are important to brain researchers and clinicians alike because many phenomena cannot be observed by anatomical techniques alone. Among functional imaging methods, only magneto- and electro-encephalography (MEG, EEG, or jointly M/EEG) can noninvasively resolve events with a millisecond time scale. Statistical tools for M/BEG functional brain imaging software will be developed to estimate and visualize the spatial extent and time course of brain activity. Algorithms will be developed for the incorporation of a priori information into source estimation, and for estimating the uncertainty of the estimates. These tools will permit the use of information derived from anatomy, physiology, and other functional imaging modalities (such as fMR1 and PET) to be combined with M/EEG data to improve the robustness, reliability, and objectivity of the M/EEG analysis. The algorithms will be incorporated into prototype software, and the software validated with both simulated and experimental data. The software will comprise a PC/Windows-based program suite for analysis and display. The algorithms and resulting software may be used to study both normal brain function, such as measurements in cognitive neuroscience which may be studied with evoked response/event related potentials or spontaneous EEG, and in diseases of the brain, such as the epilepsies, where precise spatial and temporal resolution may be of value for diagnosis and presurgical evaluation. PROPOSED COMMERCIAL APPLICATIONS: The techniques which we propose are a non-invasive, non-radiological and relatively low cost addition to existing EEG, MEG and MRI systems, and provides information which is not currently available from these systems independently. The resulting software will have direct application in clinical and cognitive neuroscience research. If clinical value is demonstrated, systems based on this methodology may find applications in the areas of psychiatry, neurology and psychology.

Functional imaging techniques are important to brain researchers and clinicians alike because many phenomena cannot he observed by anatomical techniques alone. Among functional imaging methods, only magneto- and electro-encephalography (MEG, EEG, or jointly M/EEG) can noninvasively resolve events with a millisecond time scale. The aim of this work is the development, validation, and commercialization of a detailed, realistic model of electrical current flow in the head, the currents originating from sources within the brain, and measured electrically and/or magnetically on or near the surface of the scalp. 3D finite element models of individual heads will be constructed with the aid of magnetic resonance imaging (MRI). The finite element method provides the needed flexibility for modeling an asymmetric, anisotropic, and inhomogeneous medium like the head. The algorithms will he incorporated into prototype software, and the software validated with both simulated and experimental data. The software will comprise a realistic modeling module of our EMSE suite, our PC/Windows-based program suite for analysis and display. The algorithms and resulting software may be used to study both normal brain function, such as measurements in cognitive neuroscience which may he studied with evoked response/event related potentials or spontaneous EEG, and in diseases of the brain, such as the epilepsies, where precise spatial and temporal resolution may be of value for diagnosis and presurgical evaluation. PROPOSED COMMERCIAL APPLICATIONS: The technique which we propose is a non-invasive, non-radiological and relatively low cost addition to existing EEG, MEG and MRI systems, and provides information which is not currently available from these systems independently. The resulting software will have direct application in clinical and cognitive neuroscience research. If clinical value is demonstrated, systems based on this methodology may find applications in the areas of psychiatry, neurology and psychology.

DESCRIPTION (provided by applicant): The epilepsies are a family of disorders of brain dynamics. Epilepsy patients are susceptible to seizures (changes in sensation, awareness, or behavior caused by brief electrical disturbances in the brain). Often, the electrical disturbances originate from a "focus" or "epileptogenic zone" in part of the brain, from which the disturbances are propagated to other parts of the brain. Epilepsy affects approximately 2.5 million persons in the United States, and over 50 million persons worldwide, and 150,000 to 200,000 new cases occur annually in the U.S. Although most epilepsy patients can control their seizures with the use of antiepileptic drugs, between 20% to 25% of patients cannot bring their seizures under control using drug therapy. Many patients with pharmacologically intractable seizures can eliminate their disability largely or completely by neurosurgical intervention, which typically involves resection of tissue in the epileptogenic zone to prevent the spread of electrical disturbances. Candidates for epilepsy surgery are generally evaluated first with scalp EEG. In approximately 20% of cases, however, intracranial telemetry (iEEG) with brain surface (ElectroCorticoGraphy, or ECoG) and/or depth electrodes is required to determine the site(s) of seizure onset, but determination of the seizure onset zone is often ambiguous. We propose to develop software tools that will aid the clinician in the identification of the epileptogenic zone from iEEG data, through the multimodal integration of iEEG data with structural and functional imaging data, visualization of iEEG data, both in space and time, and analysis of iEEG data. During Phase I, we designed, implemented, and validated software to integrate the visualization of ECoG electrode locations on sMRI data, on a subject-specific basis, and visualize ECoG activity patterns in both space and time. During Phase II, we will extend these results to include depth electrode data. We will also develop, implement and verify statistical signal processing tools to aid in the identification and localization of the seizure onset zone. These methods will be validated by comparison with clinical findings. A principal design objective is the creation of software that can be used by a trained technician to produce a reliable visualization with less than 1 hour of preparation time. PUBLIC HEALTH RELEVANCE: The work described in this proposal addresses the need to develop improved software tools for the visualization and analysis of intracranial electroencephalographic (iEEG) measurements obtained from the surface (electrocorticography, or ECoG) or from the depths of the human brain. If successful, these tools will advance the clinical practice of surgical management for epilepsy, as well as providing an enabling technology for basic and clinical human brain research.

[unreadable] DESCRIPTION (provided by applicant): The epilepsies are a family of disorders of brain dynamics. Patients are susceptible to seizures (changes in sensation, awareness, or behavior caused by brief electrical disturbances in the brain). Often, the electrical disturbances originate from an "epileptogenic zone" in part of the brain, from which the disturbances are propagated to other parts of the brain. Epilepsy affects approximately 2.5 million persons in the United States, over 50 million persons worldwide, and 150,000 to 200,000 new cases occur annually in the U.S. Although most epilepsy patients can control their seizures with the use of antiepileptic drugs, about 20% of patients cannot bring their seizures under control using drug therapy. Many patients with pharmacologically intractable seizures can eliminate their disability largely or completely by neurosurgical intervention, which typically involves resection of tissue in the epileptogenic zone to prevent the spread of electrical disturbances. Candidates for epilepsy surgery are generally evaluated with scalp electroencephalography (EEC) telemetry, and, increasingly, with magnetoencephalography (MEG). We propose to develop computer software that will aid clinicians in the diagnosis and treatment of medically intractable epilepsy, using data obtained from structural magnetic resonance imaging (sMRI), MEG, and EEG. Our goal is to develop and validate improved methods for the non-invasive localization of the epileptogenic zone through the automated analysis of spike-like electrical activity recorded in the interval between seizures (interictal spikes). During Phase I, we improved the sensitivity, specificity, and computational efficiency of our automatic interictal spike identification and classification software. The software was evaluated, using MEG and sMRI data obtained from epilepsy surgery patients. During Phase II, we will add additional features for spike propagation analysis, extend the validation to EEG data, and incorporate the resulting algorithms into our commercial EMSE Suite software. [unreadable] [unreadable]