The overall goal of this procedure is to image the brain's dopaminergic response to smoking cigarettes. This is accomplished by first imaging a smoker smoking in the pet scanner after the dopamine antagonist tracer 11 C RAC lip pride has been injected. The second step is to reconstruct the list mode data to images taking motion into account, and then smooth the pet images with a hyper filter.
Next, the time activity curves are modeled at the voxel level with LPNT PET and the resulting dopamine curves are retained only for those voxels that are well fit by the model. The final step is to color code the fractional dopamine level at each foxhole over time in significant foxholes and capture the color coded values as a series of dynamic images. Ultimately, it is hoped the LPNT PET generated dopamine movies of cigarette smoking will be used to reveal a unique spatiotemporal pattern of addiction.
The main advantage of our kinetic model over conventional ones like the two tissue compartment model, SRTM or the Logan plot, is that our model contains a time varying parameter, but like SRTM and the Logan plot, it is linear in parameters. The presence of a time varying parameter in the LPNT PET model means that we can describe the uptake or displacement of the PET tracer even when the endogenous neurotransmitter is not at steady state. A clinical aspect over our techniques that cannot be overlooked is the statistical step for us.
The app test, which we use to identify those box locations that require a time bearing models for etiquette descriptions. Although the method is presently applied to addiction and the dopamine system, it could also be applied to other conditions and neurotransmitter systems such as depression and serotonin, for example, provided that we have a pet tracer that can be displaced by changes in level of serotonin. We hope that we can use this technique to identify spatial and temporal patterns of brain activity that are unique to addiction, perhaps even neurochemical signatures that are uniquely female or uniquely male.
We first had the idea for making dopamine movies when our co-author Christian Constantine tried to display time varying patterns of dopamine activity by creating an old fashioned flip book. Visual demonstration of our method is critical because the end product is a movie and it defies demonstration on the static pages of a journal. But because our eyes are good pattern recognition systems, we hope to discover patterns or sequences of dopamine release that no one was even looking for, demonstrating the aspects of our procedure that must be performed by licensed medical personnel will be Elisa Hildago, certified nuclear medicine technician and Cindy Dko, registered nurse from the Yale Pet Center.
To begin acquire a pre pett structural MR scan, such as a 3D MP rage on a separate day from the PET scan. The MR scan will provide an anatomical reference for the pet images. Also arrange for the subject to practice the smoking motion in the PET scanner before the scan.
To begin patient prep, an IV must first be inserted by a registered nurse and ready for a later attachment to the pump that delivers the tracer. Also, prepare the injection pump program the pump with the proper infusion paradigm for the RAC lapide tracer. Next, set up the head motion monitor a fixed reflective spheres to the top of the subject's head here.
The spheres are attached to a rigid cross shaped tool, which is attached to a CRA swim cap. The Vic CCRA system head tracking lasers will pull the position of reflective spheres at a rate of 20 hertz. Finally, to eliminate secondhand smoke from the pet suite.
During the experiment, position the intake of an air filter in front of the scanner and above the subject's head leave room for the subject to bring the cigarette to his or her mouth while smoking. Position the subject in the scanner. Then prior to injection of tracer and PET acquisition, first acquire a nine minute transmission scan.
Then have a certified nuclear medicine technologist administer the tracer. Generally, a team of two technologists initiates tracer administration and pet data acquisition simultaneously administer simple questionnaires orally to the subject immediately prior to and following smoking. The smoker should rate his or her craving satisfaction of craving nicotine high and feelings of aversion on a scale of one to a hundred.
In order to capture the dopamine response to a naturalistic smoking experience, have the subject perform the smoking. Rather than having nicotine or the cigarette administered by study personnel instruct the smoker to smoke at his own pace and smoke his own brand of cigarettes. Subjects should have abstained from smoking since the previous midnight and should smoke two cigarettes in succession.
It generally takes about 10 minutes to complete both cigarettes. Then again, administer the rating scales posts smoking as previously described after the scanning session apply a variant of the spatial filtering method. Highly constrained back projection to all pet images in a frame by frame manner.
The appeal of hyper LR is that it reduces spatial noise without degrading the temporal information at every voxel that we will use to create our dopamine movies. Align the PET to the subject's MR.Data, then apply a stri AAL mask. RACLOPRIDE has sufficient signal to background contrast to be used only in the striatum, a brain area implicated in drug addiction.
Apply a mask of the pre-commercial striatum to all the PET data in the template space. Next, select dopamine response functions that are consistent with possible dopamine responses to the stimulus by selecting a particular set of response functions. When can constrain the shape and timing of the estimated dopamine responses to curves that are expected for the stimulus for smoking?
Expect a unimodal rise and fall of dopamine concentration in a gamma variant shaped curve. In the case of smoking at 45 minutes into the scan, families of response functions with takeoff times of 40 minutes and later should be included. Now fit the LPNT PET model to the PET Time activity curves at each individual voxel in the masked region.
The operational equation of the model is seen here. The integral of the product of the PET time activity curves with each response function becomes a set of linear basis functions be of T that contribute to the model. Because LPNT PET is a linear basis function-based method for fitting the dynamic pet data, it can be implemented to rapidly estimate both kinetic parameters governing action of the tracer and a time profile of relative dopamine concentration change during the scan session at each foxhole.
Next record the map of the weighted sum of squared residuals from the fit of LPNT PET to the data at each foxhole fitting the model at each foxhole produces images of the tracer parameters. R one is the relative flow value. K two is the FLX rate from the free compartment of the target region.
K two A is the apparent FLX rate from the target tissue modeled as one compartment and gamma is the magnitude of the dopamine signal. The weighted sum of squared residuals at each vle can be thought of as an image as well. Now fit the MULTILINE reference tissue model or MRTM to the PET time activity data at each individual voxel in the masked region.
This is a linear model that is identical to LPNT pet, except that it lacks a time varying dopamine term fitting. MRTM to voxel wise data yields estimates of only three parametric images, R one, K two and K two a record the weighted sum of squared residuals map from the fit of MRTM to the data at each voxel as well. Next, create an F map from the sum of squares maps by calculating the F statistic at each voxel in the mask.
The F statistic compares the weighted sum of squared residuals from LPNT PET to the weighted sum of squared residuals from MRTM correcting for differences in degrees of freedom in the respective fits threshold. The FMAP at a value that translates to a probability of P less than 0.05 based on degrees of freedom in the model fits then ize the map to make a new significance mask that retains only those voxels in the striatum whose pet time activity curves are fit better with the LPNT PET than with MRTM. Next perform a morphological opening on the significance mask to eliminate tiny isolated clusters of voxels that we assume to be due to noise.
An isotropic two by two by two voxel kernel should be used to remove isolated groups of voxels with diameters of two voxels or smaller. This results in the final significance mask. Perform the same analysis on data from every experimental condition to be examined for the protocol.
Demonstrated data was acquired and analyzed for each subject in two separate conditions. Smoking and control. Finally compare smoking to control by constructing a composite dopamine movie.
Produce dopamine movies for the same subject in different conditions. For example, baseline or sham task versus smoking. Also produce a composite dopamine movie for one subject for all slices of the striatum for baseline and smoking.
Here we can see the effects of two different hypers spatial filters versus no filtering on the smoothness of the time. Activity data at a single tri AAL voxel, the top and middle rows show images filtered with different kernel sizes while the bottom row shows corresponding time activity curves from the same single voxel location in the left dorsal caudate. Here we see representative examples of dopamine response functions that take off at 40 minutes or 45 minutes post tracer injection.
These were pre-computed for fitting the LPNT PET model to the PET time activity data at each voxel. This figure shows fits of the conventional MRTM and new LPNT PET models respectively to the time activity data from a voxel in the left caudate. The MRTM fit is in blue and the LPNT pet fit is in red.
Here we see parametric images of the weighted sum of squared residuals for MRTM and LP NT pet. The respective WSSR maps are compared via the F statistic to create the F map, which in turn is threshold at the desired probability level to a binary mask and then filtered to produce the final significance mask. This figure shows one coronal slice of the final significance mask for the smoking condition in a single subject compared to the final significance mask for a corresponding subject and slice in the control condition.
This movie shows the frame by frame dopamine level relative to the basil or resting dopamine level. The dopamine levels are encoded in color. Specifically the colors represent the change in dopamine above the basal level as a percent of basal.
Again, the dopamine levels are shown only for voxels. In the final significance mask that exceed the P less than 0.05 significance level. This movie shows the multis slice Multidi dopamine movie for a male smoker.
All slices of the ventral striatum are displayed simultaneously for the smoking condition and baseline condition. Notice the major activation in the right ventral striatum at about the time of the second cigarette. The movie repeats a few times for ease of viewing.
This movie is made from the scan of a female smoker. Notice the major activation in the left dorsal caudate at around the time of the first cigarette. The movie repeats a few times for ease of viewing.
Once mastered the creation of dopamine movies from pet data can be executed in a few days times. We hope to automate more of the procedure, particularly the capture of individual movie frames in the future. Making dopamine movies can be considered a constrained optimization problem.
The constraints on the possible dopamine outcomes are applied through the selection of the initial dopamine response functions. Now that we have created dopamine movies, it remains for us to identify the best similarity and dissimilarity measures so that we can compare the movies from different subjects, say dependent smokers and former smokers. With the requisite measures in place, we can ask what aspect of the brain activity changes or returns to normal.
When a smoker kicks the smoking habit, It is hope that this technique will pave the way for other pet researchers to properly describe and analyze their pet data, which they are acquiring. In the study of other addictive behaviors with a stimulus such as alcohol may produce a brief neurotransmitter response, which does not come to study levels post drug. We hope to use similar experimental procedures using different pet tracers to look at neurochemical changes in the brain with other drugs of abuse.
After watching this video, you should have a good understanding of how to create dopamine movies from pet data, which we have done in order to capture the transient and especially limited response of the brain's dopamine system to smoking cigarettes. Recall that the dopamine movies are created in five steps. Pet experiment, pre-processing at the images, modeling the images is statistical step and the visualization step.
