| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Departments of Radiology (D.W.M., K.R.M.) and Neurological Surgery (K.R.M., J.R.M., H.R.W.), University of Washington, 1959 NE Pacific St, Box 357115, Seattle, WA 98195. From the 1999 RSNA scientific assembly. Received December 30, 1999; revision requested February 28, 2000; revision received May 15; accepted June 1. D.W.M. supported by an RSNA Research Resident Grant from the RSNA Research and Education Foundation. Work also supported by National Institutes of Health grant NS21076. Address correspondence to D.W.M.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
Index terms: Animals, 10.121412, 10.121415, 10.121416 • Brain, MR, 10.121412, 10.121415, 10.121416 • Experimental study, 10.121412, 10.121415, 10.121416 • Magnetic resonance (MR), echo planar, 10.121416 • Magnetic resonance (MR), functional imaging, 10.121412, 10.121415, 10.121416
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
2.0 T) imagers and slower gradient-echo pulse sequences. The limitations of these models include the fact that they are not using human subject echo-planar imaging techniques and that they require specialized small-bore research magnets. Human functional MR imaging is typically performed by using a 1.5-T whole-body imager with single-shot gradient-echo echo-planar pulse sequences. To develop and test new human functional MR imaging methods, it will become increasingly important to have a small-animal model of functional MR imaging that uses techniques similar to those used in humans. Using a previously described rat model of electric stimulation of the forepaw to activate the contralateral sensory cortex (5,1), we determined parameters for BOLD functional MR imaging in a 1.5-T whole-body MR imager by using a single-shot echo-planar imaging sequence to see if a reliable rat functional MR imaging model could be developed with techniques similar to those used in human studies (6).
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
A 22-gauge intraperitoneal catheter was then placed through a small, anterior midline abdominal incision. After this, tracheotomy was performed, and the trachea was cannulated by using a 14-gauge catheter. Mechanical ventilation was administered by using a rodent respirator. The rats were then paralyzed with intravenous administration of 1 mg of D-tubocurarine chloride (Abbott Laboratories, North Chicago, Ill) per kilogram of body weight. The inspired gas mixture of oxygen and room air was adjusted to produce a fraction of inspired oxygen of 60%. The halothane anesthesia was gradually discontinued over the course of 15 minutes and replaced with an intraperitoneally administered mixture of 
-chloralose (Part C8091; Sigma-Aldrich, St Louis, Mo) and urethane (Part U2500; Sigma-Aldrich) in a 1:10 (-chloralose:urethane, by weight) ratio. The initial dose of this mixture was 30/300 mg/kg (-chloralose/urethane).
Subcutaneous needle electrodes made from 30-gauge copper transformer wire were placed in the radial and ulnar aspects of the left forepaw. The animal was then placed prone in an MR-compatible stereotactic head holder constructed from polycarbonate (Dupont, Wilmington, Del) and acetyl resin (Delrin; Dupont) to minimize added susceptibility artifacts. Functional imaging did not begin until at least 1 hour after discontinuation of halothane anesthesia to ensure adequate anesthetic washout.
On the basis of arterial blood gas analysis, changes to ventilator settings and fraction of inspired oxygen were made to maintain pH, PaCO2, and PaO2 within the physiologic range. Body temperature was maintained by using a circulating water warming blanket (Gaymar Industries, Orchard Park, NY). Supplemental intravenous doses of D-tubocurarine chloride (1 mg/kg) were administered every 30 minutes for the duration of the experiment. Supplemental doses of an -chloralose/urethane mixture (15/150–20/200 mg/kg) were administered intraperitoneally approximately every 60 minutes for the duration of the experiment. At the conclusion of each experiment, each animal was killed while anesthetized, according to the protocol approved by the institutional animal care committee.
Functional MR Imaging
All experiments were performed with a 1.5-T whole-body MR imager (Horizon Echo Speed; GE Medical Systems, Milwaukee, Wis) in a radio-frequency-shielded room. A 2.5-cm-diameter transmit-receive surface coil designed in our laboratory (Hayes C, unpublished data, 1997) was used for all measurements.
Functional imaging was performed by using a single-shot multisection two-dimensional standard gradient-echo echo-planar pulse sequence with the following parameters: 2,000/40, 50, or 60 (repetition time msec/echo time msec [effective]); number of signals acquired, one; flip angle, 90°; field of view, 7 x 4; matrix, 70 x 40; echo train length, 40; linear phase encode ordering; and 1-mm section thickness. The animal was placed prone in the imager with the 7-cm field of view axis and the frequency acquisition axis oriented in the left-right direction.
Both fat saturation and ramp sampling were enabled. Fat saturation forces the imager to use a narrow-band, sync-type radio-frequency pulse that produces a sharper section profile than the default spatial-spectral radio-frequency pulse, which allows submillimeter section thickness with the standard two-dimensional gradient-echo echo-planar pulse sequence. Fat saturation also eliminates a checkerboardlike artifact seen with the spatial-spectral pulse, which is thought to be related to off-axis fat excitation. With ramp sampling, an adaptive wideband sampling technique is used to allow shorter effective echo times while still collecting full echoes. With ramp sampling enabled, the maximal receiver bandwidth was 65 kHz. These parameters resulted in cubic 1-mm3 voxels.
Ten different section locations in the rat brain were imaged during each functional MR imaging measurement. One set of 10 sections was acquired during each repetition time, with individual sections acquired sequentially in an interleaved fashion—odd sections first, then even. By using this strategy, each section was stimulated once during each repetition time, while the MR imager generated a radio-frequency pulse and received an echo every repetition time divided by 10 (ie, every 200 msec).
Each functional MR imaging measurement consisted of 51 phases (ie, 51 sets of 10 sections), lasted 102 seconds, and generated 510 images. There was a rest period of at least 10 minutes between functional MR imaging measurements in animals undergoing multiple functional MR imaging measurements.
Electric forepaw stimulation was accomplished with 5-V, 0.3-msec, 3-Hz pulses administered by means of a pulse generator (SD9 Stimulator; Grass Instruments, Quincy, Mass). The maximal current delivered during each pulse was 0.75–1.00 mA. A boxcar pattern of forepaw stimulation was used as follows: off, on, off, on, and off for 11, 10, 10, 10, and 10 phases, respectively, for a total of 51 phases. The pulse generator was controlled by using a personal computer (PowerMac 8100/100; Apple Computer, Cupertino, Calif) with a commercially available software package (PSYSCOPE; New Micros, Dallas, Tex) (7) via a hardware signal interface (PSYSCOPE Button Box; New Micros). By using a 40-msec monostable multivibrator (74123; Texas Instruments, Dallas, Tex), the scope trigger output from the integrated pulse generator, or IPG, module of the MR imager was shaped into a pulse compatible with transistor-transistor logic, or TTL, which was then buffered with a 50-
line driver (74S140, Texas Instruments) and passed to an optically isolated input of the hardware signal interface. This allowed the software to count each image in the functional MR imaging measurement as it was recorded and use this information to enable and disable the pulse generator at precise predetermined times during each functional MR imaging acquisition.
Magnet Shim
We found that manually shimming the magnet, particularly along the z axis, was critical for BOLD activations to be observed. After initial imaging sequences to achieve three-axis localization within the magnet bore, an automated shim was performed with the shim volume limited to the sensory cortex. Regions near the paranasal sinuses and mastoid air cells were specifically excluded from this volume. The results of the automated shim were then used as a starting point for a manual gradient shim. We found it possible to achieve an adequate shim by adjusting only the first-order x, y, and z gradients. Once shimming had been completed, the gradient shim and the animal’s position within the magnet remained unchanged for the duration of the experiment.
Data Processing
Rapid preliminary image analysis (FUNCTOOLS; GE Medical Systems) was performed with a UNIX-based workstation during the experiment (D.W.M.) to verify the presence of BOLD activations. After completion of the experiment, images were extracted into digital imaging and communications in medicine, or DICOM, format, and a detailed analysis was performed (MEDx; Sensor Systems, Sterling, Va) with a LINUX-based computer. z scores were generated for each voxel by using an unpaired Student t test referenced to a boxcar waveform that matched the pattern of forepaw stimulation with a 4-second hemodynamic delay. Activation maps were generated by performing thresholding of these z maps, with activation defined as present in any voxel with a z score greater than or equal to 3. A cluster analysis was used to identify clusters of activated voxels. Clusters were analyzed for the number of active voxels and the peak BOLD response observed in a single activated voxel.
These parameters were entered into a spreadsheet program for final analysis (performed by D.W.M. and reviewed by K.R.M.). Activation maps were overlaid on high-spatial-resolution anatomic images by using the preliminary image analysis software package and then exported (PHOTOSHOP; Adobe Software, Seattle, Wash) for final formatting and printing.
Volume of Activated Sensory Cortex during Forepaw Stimulation
In eight different animals, we performed a single BOLD measurement to estimate the volume of sensory cortex responding to forepaw stimulation. Effective echo time was 50 msec in all cases. The number of activated voxels for each animal was used as an estimate of the activated cortex volume.
Determination of Optimum Echo Time
In three of these eight animals, we also performed BOLD measurements while varying effective echo time from 40 to 60 msec in 10-msec increments. A single BOLD measurement was made for each effective echo time in each animal. The number of activated voxels and the BOLD response amplitude in the maximally responding activated voxel were used for analysis.
Signal-to-Noise Ratio Calculations
Signal-to-noise ratio (SNR) was defined as follows: SNR = 20 · log10[RMS(signal)/RMS(noise)], where RMS is the root-mean-square of the indicated time-varying value. The root-mean-square signal intensity was estimated by calculating the root-mean-square of the signal intensity within a region of interest containing only the sensory cortex over the course of a BOLD measurement. Root-mean-square noise intensity was estimated by calculating the root-mean-square signal intensity within a region of interest containing only air over the course of a BOLD measurement.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
|
|
|
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
We observed a localized focus of activation, centered 1–2 mm anterior and 4 mm lateral to the bregma, in response to forepaw stimulation, which is consistent with the known location of the forepaw somatosensory cortex (8,9). We determined the activation volume to be 11 mm3 ± 4, which is similar to the findings of Hyder et al (1) and Gyngell et al (3) but somewhat larger than that observed by Hsu et al (4). Differences between the activation volumes reported by these groups are likely to represent differences in a combination of factors, including forepaw electrode placement, stimulation paradigm, imaging pulse sequence, animal weight and age, and statistical threshold for activation.
We observed maximal BOLD activation with a 4-second hemodynamic delay incorporated into our boxcar reference waveform. This 4-second delay is consistent with the temporal course of deoxyhemoglobin levels in response to visual stimulation reported by Malonek et al (10). Given the more rapid time course of sensory evoked potentials (approximately 50 msec) and sensory cortex arteriolar vasodilatation (approximately 1.5 seconds) in response to peripheral nerve stimulation (5,11), our results are consistent with the hypothesis that the BOLD response is primarily related to changes in deoxyhemoglobin levels associated with neural activation.
The primary advantage of imaging rats with a whole-body 1.5-T MR imager and a standard human echo-planar imaging sequence is the potential to use this system as a tool for testing and developing human functional MR imaging techniques. New pulse sequence parameters or unique pulse sequence designs could first be worked out in this small-animal system and then rapidly applied to human studies.
An additional advantage of the described technique is that it permits small-animal functional MR imaging research studies to be conducted on a clinical 1.5-T MR platform with only minor hardware and software modifications. These experiments were previously performed only with small-bore high-field-strength systems.
Signal-to-noise ratio is the primary limitation of small-animal functional MR imaging at 1.5 T. With careful coil design and pulse sequence parameter selection, we were able to achieve a high signal-to-noise ratio that permitted resolution of cubic 1-mm3 voxels within the sensory cortex. This resolution should be sufficient to observe activations in many rat brain structures, such as the visual cortex, forepaw and hind paw sensory cortex, or olfactory cortex.
Manual shimming is essential to achieve good results. Whole-body imagers are optimized to image large structures by using a large field of view. Rat functional MR imaging measurements require a small field of view and a small-geometry radio-frequency coil. However, the small field of view and small-geometry radio-frequency coil provide insufficient signal for a traditional automated shim. We found it sufficient to adjust only the first-order shims while optimizing the height and width of the water resonance peak.
In conclusion, it is possible to develop a robust small-animal model of BOLD functional MR imaging by using single-shot echo-planar imaging techniques with a 1.5-T whole-body imager. Since this model uses techniques commonly used for human functional MR imaging studies, it should prove useful for testing and developing new human functional MR imaging techniques, as well as for enabling basic research into the neurovascular physiology underlying the BOLD response by using a 1.5-T clinical MR system.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, D.W.M.; study concepts, all authors; study design, D.W.M., K.R.M.; definition of intellectual content, D.W.M., K.R.M.; literature research, D.W.M., K.R.M.; experimental studies, D.W.M.; data acquisition and analysis, D.W.M.; statistical analysis, D.W.M.; manuscript preparation and editing, D.W.M.; manuscript review, all authors; manuscript final version approval, D.W.M., K.R.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D.W. Morton, B. Keogh, K. Lim, and K.R. Maravilla Functional Brain Imaging Using a Long Intravenous Half-Life Gadolinium-Based Contrast Agent AJNR Am. J. Neuroradiol., August 1, 2006; 27(7): 1467 - 1471. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Morton, K. R. Maravilla, J. R. Meno, and H. R. Winn Systemic Theophylline Augments the Blood Oxygen Level--Dependent Response to Forepaw Stimulation in Rats AJNR Am. J. Neuroradiol., April 1, 2002; 23(4): 588 - 593. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
