Functional MRI Core Facility

Bandettini, Peter

Abstract

Space Utilization: The Functional Magnetic Resonance Imaging Facility (FMRIF) currently occupies approximately 5000 sq ft of space in Building 10, divided between the B1level scanner bays, control rooms and electronics/machine rooms for 3TA/3TB, 3TC, and the Siemens 7T-Classic, (about 1800 sq ft, 1100 sq ft and 1300 sq ft respectively) and office space within the Nuclear Magnetic Resonance (NMR) center. On the first floor are the Functional MRI Facility and the Section on Functional Imaging Methods suites (approximately 800 sq ft total) for office space and shared conference space for all staff employed full-time by the facility. Staff: The FMRIF staff (currently 14 full- time positions) consists of: the facility director, four staff scientists to keep the scanners running, six MRI technologists, an information technology specialist, a programmer, and an administrative laboratory manger. Investigators: The functional MRI facility supports the research of over 30 Principal Investigators translating to over 300 researchers overall. Over 70 research protocols are active and making use of FMRIF scanners. Each scanner has scheduled operating hours of 105 hours per week. Papers published using the core: A strong measure of the utility of a core facility is the quantity and quality of scientific papers published by investigators using the facility. We have kept careful records of papers published and their corresponding citations, such that we have been able to create a core facility h-index. Since its inception in 2000 until August 2019, a total of 1,159 peer-reviewed publications from intramural investigators have used data acquired in the FMRIF core facility. The total is distributed among 658 papers from NIMH, 329 papers from NINDS, and 172 from the other institutes. These papers have been cited a total of 128,839 times for a combined h-index of 175. In other words, 175 papers using the FMRIF have been cited at least 175 times. Core Projects of the Staff Scientists: Staff Scientists working on projects are designated as: Sean Marrett (SM), Vinai Roopchansingh (VR), Andy Derbyshire (AD), and Linqing Li (LL). Other staff are Jan Varada (JV) and Roark Maccado (RM). These projects span training, subject interface, IT infrastructure, pulse sequences, processing, and hardware development. Training: We continue to provide support for high-resolution FMRI - working with individual groups to implement 1mm-1.6mm resolution protocols using the CMRR-MB sequence. On the FMRIF-7T, this is about 75-80% of research FMRI scans. The FMRIF also added support for novel methods for ultra-high-resolution on the FMRIF-7T, including VASO and VAPER. SM organized a small training workshop on non-BOLD VASO contrast for acquiring sub-millimeter layer-specific FMRI images. This session was videoed. Training text and scripting was provided by Dr. Laurentius Huber (SFIM). The workshop focused on practical techniques for improving image quality by optimizing acquisition and reconstruction parameters. Final integration of the most up-to-date computer console for the FMRIF-7T that increased storage and processing capacity, enabling easier custom pulse sequence development and raw data migration using a high-capacity portable disk subsystem. SM worked on safety testing and integration of a new version of a 32-channel surface array coil organized into 2 flexible 16-channel modules . This coil was custom built for FMRIF via a specialty company. This is a collaboration with Joe Murphy-Boesch of NINDS who designed the interface logic to integrate this coil with his custom head-transmit coil and also with Dr. Huber (SFIM). PAB organized a summer neuroimaging course centered on current issues and controversies. This consisted of 24 lectures (2 per week) from June to Sept. Subject Interface: SM installed a DLP projector for the 7T as well as standardized high-speed eye-tracking solutions across all our scanners. A custom eye-tracker and screen assembly is in development for the FMRIF-7T in collaboration with the Section on Instrumentation and Dr. Eli Merriam. The FMRIF improved audio stimulus delivery with integration and training for an active noise cancelling system that is now available on all FMRIF scanners. This new system also allows for robust collection of speech data from patients concurrently with fMRI exams (e.g. for Dr. Sara Inati, NINDS). Finally, SM has worked closely with Dr. Peter Molfese (Center for Multimodal Neuroimaging (CMN)/SFIM ) to support concurrent high-density Electroencephalography (EEG) data collection on 3TB and 3TD (with Dr. Jen Evans/Zarate group). IT infrastructure: A prototype pipeline for deploying automated MRI quality control (MRIQC) and a pilot Django-based system for secure archiving, retrieval and better meta-data searching of FMRIF image data was developed by JV. A replacement and upgrade of the storage area network that is the backbone of the FMRIF computing and data archiving network is underway (RM). This will expand the current storage capacity from 200TB to 384 TB and increase access speed. VR upgraded internal networks on all FMRIF scanners to stream raw data using 10-gigabit connections. Pulse sequences: (all pulse sequence testing used Protocol number 93-M-0170, NCT00001360) LL has been developing an imaging technique DANTE prepared EPI for quantitative mapping of cerebral blood volume (CBV) and cerebral blood flow (CBF) of brain activity. Results were presented at the International Society for Magnetic Resonance Imaging in Medicine (ISMRM) 2019. LL and Yuhui Chai (SFIM) developed the use of DANTE (Delay Alternating with Nutation for Tailored Excitation) pulse trains combined with 3D-EPI to acquire an integrated VASO and perfusion (VAPER) contrast. LL in collaboration with the Spectroscopy core developed a method for quantitative measurement of brain neurotransmitter glutamate. AD has maintained a virtual-machine based implementations of the Siemens IDEA Pulse Sequence Development environments currently including VB17A, VD13A, VD13D, VE11C, VE11U, VE11K and VE12U. Processing An updated neuro-feedback API in AFNI, which allowed AFNI to communicate with PsychoPy, thereby allowing its use in neuro-feedback experiments, was developed by VR and collaborators. AFNIs real-time plugin was updated by VR working with Dr Gonzalez-Castillo from the Section on Functional Imaging Methods SFIM, and Richard Reynolds, from the Scientific and Statistical Computing Core) to provide researchers the ability to use multi-echo FMRI data for real-time neuro-feedback studies. This is currently being used by Dr Ramot (from Dr Martins Section Cognitive Neuropsychology in NIMH) on the FMRIFs 3TB scanner. JAD has developed methods to recover lost/hidden scan parameters from legacy and current GE DICOM image data in particular the identification of blip-up versus blip-down EPI trajectories. This is helpful for re-processing legacy/historical fMRI data from image databases using newer image processing algorithms and pipelines. Working with John Rogers-Lee (then in the Data Science and Sharing Team, latterly AFNI group) these methods have been incorporated into the dcm2niix software package widely used by the neuroimaging community. VR and AD helped to develop and implement new cross-vendor sharable raw data format called ISMRMRD. Hardware AD has initiated the system integration effort for the head gradient coil that is being sourced by NIBIB (Carlo Pierpaoli is PI) is from Stanford University and is being built at the University of Western Ontario. In use, the head gradient will offer significantly higher gradient performance: amplitude 120mT/m and slew rate 1200mT/m/ms.