350 Community Drive
Manhasset, NY 11030

Tel: 516-562-2498
E-mail: david1@nshs.edu

The Imaging Core is directed by David Eidelberg, MD, and is situated within the  Functional Brain Imaging Laboratory  of the Feinstein Institute. The laboratory occupies 2,500 square feet on the second floor of the Feinstein Institute. The imaging laboratory is staffed by three biophysicists and two computer scientists/data analysts who support the research activities of investigators in the Center for Neurosciences as well as General Clinical Research Center (GCRC) projects involving imaging methodologies. The close proximity of the GCRC to the positron emission tomography (PET) imaging laboratory and the new high field magnetic resonance imaging (MRI) facility is a significant advantage to GCRC investigators, who have utilized these facilities consistently since the funding was initiated in July 2003.

The Imaging Core provides GCRC investigators with the support they need to design and perform clinical experiments employing state-of-the-art techniques such as PET and magnetic resonance imaging (MRI). The core serves GCRC investigators as well as other investigators with NIH-funded imaging projects.

Core services include support in the areas of study design, data acquisition, and data analysis.

Specific services include:

• Consultation in the design of neuroimaging protocols for patient-oriented research.
• Training in the use of PET and MRI.
• Facilitating access to the PET and MR scanners for GCRC investigators to perform imaging experiments.
• Facilitating image acquisition and data collection. The Core provides support for radiochemical preparation (PET), PET and MRI data acquisition, and image reconstruction and analysis for GCRC protocols.
• Assistance in the processing and analysis of PET data. Data analysis entails the implementation of image functionalization, PET/MRI co-registration algorithms, the implementation of segmentation algorithms, atrophy correction, ROI placement, as well as voxel-based statistical analyses.
• Assistance in the processing and analysis of MR data including creation of delineation criteria for manual ROI measurements and distortion correction and image registration of diffusion tensor images.
• Bioinformatics support for studies that use this core, as well as informal consulting on bioinformatics questions by other investigators. The Imaging Core also provides training in specialized techniques to faculty of the Feinstein, as well as staff of the Clinical Departments of the North Shore-LIJ Health System, postdoctoral students and fellows, and graduate students.
• Training in the use of PET for assessing cerebral metabolism, hemodynamics, and neurotransmitter function, as well as for the study of activation responses during task performance.
• Training in the use of MRI for high-resolution, anatomical imaging including voxel-based morphology (VBM) and diffusion tensor imaging (DTI), as well as functional mapping (BOLD) and perfusion imaging techniques.

[back to the top]

PET-Cyclotron Facility
 
The research PET facility is staffed by one physicist in charge of overall operations, one physicist in charge of PET scanner quality control, a senior technologist who performs daily quality control and maintenance operations, and two certified nuclear medicine technologists. These studies are conducted on a GE Advance PET camera adjacent to the cyclotron. This scanner has 18 detector rings containing a total of 12,096 bismuth germanate crystal detectors. It generates 35 simultaneous slices 4.25 mm apart. The interplane septa is automatically retracted to switch from 2D to 3D mode in under 45 seconds. The machine resolution is approximately 4.2 mm full width half maximum (FWHM) in each direction. The patient aperture is 59 cm, the radial and axial fields-of-view are 55 and 14.5 cm, respectively. The patient bed is mobile vertically and longitudinally and was designed with consideration to both operational utility and patient comfort. A shielded 68Ge/68Ga attenuation source, remotely positioned, minimizes radiation exposure to operating personnel and provides for acquisition of attenuation information. A laser optical system to define reference planes for accurate patient positioning is also provided. The computer system consists of two Sun Ultra 60 workstations, one dedicated to data acquisition and the other to image analysis. The software is designed for ease of use and permits fast interactive generation of time-activity curves and customized image analysis protocols. The peripherals include a 4 mm DAT backup drive, a 2.3 GB read/write removable optical disk for image archival. These workstations are connected via intranet to a computer network comprised of over 30 PC Windows XP workstations with networked storage hard drives.
 
The Cyclotron/Radiochemistry Facility is situated next to the PET laboratory. These facilities are directed by senior radiochemist, Dr. Thomas Chaly, PhD, and two radiochemists, a radiopharmacist, two cyclotron engineers, and two technicians. It consists of a vault containing a 17 MeV cyclotron and power supply room, a radiochemistry laboratory and a class 100,000 clean room for the routine production of 18F-FDG. The clean room meets the requirements of GMP production area and is in compliance with FDA guidelines.
 
The facility currently houses a General Electric PETtrace Cyclotron. The GE PETtrace is an automated compact self-shielded medical cyclotron capable of producing 16.5 MeV protons and 8.4 MeV deuterons. The high-energy cyclotron provides for a high yield [18F-]-fluorodeoxyglucose (FDG) production and also other important radionuclides such as [11C], [13N], and [15O]. A remotely operated semiautomatic processing system for the production of H215O is mounted on the wall of the cyclotron vault. In addition, we have an automated H215O injection system located in the PET suite. The radiochemistry laboratory is equipped with a six foot chemical fume hood, a solvent storage area, liberal bench space, and a small undercounter refrigerator. Additional instrumentation includes a gas chromatography (GC), capable of packed column and capillary column operations with flame ionization detectors and hot wire detectors, and a two-component gradient elution research high pressure liquid chromatograph (HPLC) with radiochemical and electrochemical detection systems. Both the GC and HPLC operate under the control of a single microcomputer. A pneumatic transfer tube and a small-bore (2mm ID) stainless steel gas transfer tube are provided for delivery of radiopharmaceutical preparations to the PET suite from the radiochemical laboratory.
 
A major focus of the PET component of the Imaging Core is the development and implementation of novel spatial covariance methods for the analysis of functional imaging data acquired through GCRC investigations. This network approach utilizes regional data from PET scans to identify disease-specific patterns in blood flow and metabolism data (see e.g., Eckert and Eidelberg, Lancet Neurol 2007; 6(10):926-932). Investigators within the Imaging Core have optimized this computational algorithm to standardize the criteria for pattern generation on a voxel level, and to assess the reproducibility of pattern expression in healthy subjects and patient cohorts (Ma et al., J Cereb Blood Flow Metab 2007;27(3): 597-605). This procedure was critical for the demonstration of network modulation during therapeutic interventions for Parkinson’s disease (Asanuma et al., Brain 2006; 129:2667-78) and for the measurement of network changes with disease progression (Huang et al, Brain 2007; 130(Pt 7):1834-1846). Both these NIH projects (R01 NS 35069 and P50 NS 38370) produced novel findings based upon analytical routines that were developed, validated, and implemented within this GCRC core.
 
The Imaging Core has also supported the development of network markers for preclinical Huntington’s disease (Feigin et al., Brain, 2007; 130(Pt 11):2858-2867) and for early cognitive impairment in Parkinson’s disease (Huang et al., NeuroImage 2007;34(2):714-723; Brain 2007; 130(Pt 7):1834-1846) and Alzheimer’s disease (Huang et al., Neurobiol Aging 2007; 34(2):714-723). Most recently, investigators within the core used these network tools to develop a fully automated imaging-based algorithm for the differential diagnosis of degenerative brain diseases (Spetsieris et al., SPIE 2006; 6144:61445M1-12). This track record attests to the technical innovation of this core in support of imaging research within the GCRC.

[back to the top]

MRI Center
 
The MRI research program is led by Dr. Aziz Ulug and is staffed by a computer programmer, a data manager, and two technicians who support the activities of the neuroscience investigators. MRI research studies are conducted within the Imaging Core on a 1.5 Tesla GE Signa Echo Speed scanner housed in the North Shore University Hospital Department of Radiology, approximately 200 feet from the PET facility. The system is equipped with a standard quadrature head coil and is echo planar capable with a maximum gradient strength of 2.4 Gauss/cm. Phantom studies, including a diffusion sequence, are performed on a weekly basis to insure the integrity of image quality. Monthly/weekly quality assurance procedures are also performed as part of the manufacturer’s preventive maintenance.
 
MRI studies are also conducted on a new 3T research-dedicated system housed in the North Shore University Hospital MRI Center, also next to the PET facility. The system, a GE Signa HDx 3.0T scanner (General Electric, Milwaukee, WI), was installed in 2007. The unit is equipped with twinspeed gradient set capable of maximum gradient strength of 50mT/m, and slew rate of 150 T/m/s. In addition to the standard quadrature head coil, there is an 8-channel head coil for parallel imaging. The scanner is capable of imaging 0.5 mm thick slices in 2D mode and 0.1mm slices in 3D mode. In diffusion tensor imaging studies, the unit has the capability of maximum b-value of 10,000 s/mm2 and 150 different diffusion directions. Phantom studies, including a diffusion sequence, are performed on a weekly basis to insure the integrity of image quality.
 
A major focus of the MRI component of the Imaging Core is the use of DTI to investigate neurologic and neuropsychiatric populations. GCRC investigators have used this specialized technique to map microstructural abnormalities in the white matter pathways of clinically unaffected carriers of the DYT1 dystonia mutation (Carbon et al., Ann Neurol 2004; 56(2):283-286), in adolescents with attention deficit disorder (Ashtari et al., Biol Psychiatry 2005;57(5):448-455), childhood-onset schizophrenia (Kumra et al., J Am Acad Child Adolesc Psychiatry, 2005;44(9):934-941), first-episode schizophrenia (Szeszko et al., Am J Psychiatry, 2005;162(3):602-605) and obsessive compulsive disorder (Szeszko et al., Arch Gen Psychiatry, 2005;62(7):782-790).
 
[back to the top]

Bioinformatics Sub-Core: Three staff members manage the bioinformatics and computing infrastructure of the Imaging Core, including a biophysicist with extensive experience in brain imaging analysis, a senior programmer with expertise in biomedical computing, and a data manager. The members of the Core have developed and implemented a wide variety of innovative solutions to visualize and analyze multi-modality brain images from PET and MRI. In addition to providing technical consultation to investigators using the Imaging Core, the group supports ongoing research through its long-term commitment to integrate database and analytical routines on a single common computing platform.
 
Computing Environment: The computing facility is comprised a large number of PCs under Windows XP operating systems. These computers are equipped with the latest system software and accessories and can access the central database via the local intranet. There are also a series of standard software packages to perform statistical analysis such as SPSS and SAS. A dedicated state-of-the-art high performance Windows cluster system is being purchased for data archival and image analysis. This system will provide a 64-bit high-performance computational resource capable of multi-core processing, data mining and redundancy backup to the level of industrial standards.
 
Analytic techniques: Analysis of PET and MRI brain images is performed on both regional and voxel basis using a series of the state-of-art approaches. This includes in-house software tools (Scanvp) to quantify cerebral blood flow, glucose metabolism, neuroreceptor binding and MRI data, as well as the custom-built spatial covariance mapping methods available on the Imaging Facility website (http://www.feinsteinneuroscience.org). In addition we also routinely use the third party programs such as statistical parametric mapping (SPM) (http://www.fil.ion.ucl.ac.uk). These programs are written in several languages such as Matlab, C, and Visual Basic. We are currently optimizing new multivariate tools including supervised PCA and ICA algorithms. This effort will provide a full battery of solutions for Core users employing multivariate approaches for image analysis.

[back to the top]

Database: The key task is to store brain images acquired from ongoing PET and MRI studies. They are transferred to our laboratory regularly from PET and MRI scanners through the local area network (LAN) of the North Shore-LIJ Health System. The scans are then inspected as part of the quality control program, archived on the computer server and entered in the database along with the corresponding experimental, clinical, and behavioral information. Databases keep track of preprocessed scans and outcome measures of brain images generated from our research projects.
 
A centralized SQL database management environment is used to support all research studies in the Core, using Microsoft SQL Server 2005 and MS Access.

(1) real time updates of the information system, which comprises one of the largest clinical and imaging databases in the world in patients with Parkinson’s disease and related disorders;
(2) creating and maintaining the extensive clinical and scientific database to be generated as part of the ongoing projects; and
(3) expanding the capability of this system by sharing and exchanging data with other laboratories.
 
Data Mining: This entails systematic analysis of the vast amount of brain images in the database to refine and validate relevant image-based biomarkers, including indices of radiotracer binding, regional MRI-based measures, and individual subject expression of disease-related covariance patterns (i.e, metabolic networks). Such markers are being developed to improve the accuracy of differential diagnosis and to objectively assess the efficacy of new therapeutic agents. The Core also is involved in the analysis of large imaging data sets obtained in other patient populations for purposes of biomarker validation/reproducibility and for the characterization of new disease-related patterns. Dr. Eidelberg has over 20 years experience in the imaging field. Dr. Eidelberg’s research focuses on the use of functional brain imaging and network modeling to study the alterations in brain circuitry that occur in movement disorders. His laboratory has been continuously funded by NIH since 1995. He has been the recipient of several awards for his work including the Fred Springer Award of the American Parkinson Disease Foundation (2005). He is the author of over 300 scientific publications and has served on the editorial boards of Neurology (1996-2001) and Movement Disorders (1999-2003), Journal of Nuclear Medicine (1999-Present), Current Opinion in Neurology (2001-Present) and Annals of Neurology (2006-Present).

Facility Members

Name: Thomas Chaly, PhD
Position: Chief Radiologist
Research: Developed methods and synthesis of PET radiotracers. Designs and produces labeled monoamines and receptor ligands for PET imaging of the nervous system.
E-mail: tchaly@nshs.edu

Name: Vijay Dhawan, PhD
Position: Biophysicist
Research: Responsible for all technical issues attendant to the execution and quantification of the PET studies. His research focuses on compartmental modeling and the application of novel PET radiotracers for the study of the nigrostriatal dopamine system.
E-mail: vdhawan@nshs.edu

Name: Yilong Ma, PhD
Position: Imaging Analyst
Research: Responsible for all technical issues attendant to the quantification and analyses of the PET and MRI studies.
E-mail: yma@nshs.edu

Name:   Aziz Ulug, PhD
Position:   MRI research Physicist
Research:   Development of specialized algorithms for data reconstruction. He also oversees quality assurance and technical aspects of MRI research studies.
E-mail: aulug@nshs.edu

Name: Philip R. Szesko, PhD
Position: Associate Investigator
Research: MR imaging research in patients with neuropsychiatric disease. He is responsible for assisting users of the Imaging Core in applications of diffusion tensor imaging and the study of functional correlates of brain pathology.
E-mail: pszesko@nshs.edu

Name: Shichun Peng, PhD
Position: Data Manager
E-mail: speng@nshs.edu

Name: Phoebe Spetsieris, PhD
Position: Systems Analyst
Research: Responsible for systems management and for the development of specific software applications (image display and visualization, data analysis).
E-mail: pspetsie@nshs.edu

Name: Claude Margouleff, CNMT
Position: Nuclear Medicine Technologist
E-mail: claudem@nshs.edu

Name: Ralph Matacchieri, B.S. Physics
Position: Pet Cyclotron, Svc Eng
E-mail: ralphm@nshs.edu