David A. Rottenberg

Magnetic Resonance Image Tissue Classification Using a Partial Volume Model

We describe a sequence of low-level operations to isolate and classify brain tissue within T1-weighted magnetic resonance images (MRI). Our method first removes nonbrain tissue using a combination of anisotropic diffusion filtering, edge detection, and mathematical morphology. We compensate for image nonuniformities due to magnetic field inhomogeneities by fitting a tricubic B-spline gain field to local estimates of the image nonuniformity spaced throughout the MRI volume. The local estimates are computed by fitting a partial volume tissue measurement model to histograms of neighborhoods about each estimate point. The measurement model uses mean tissue intensity and noise variance values computed from the global image and a multiplicative bias parameter that is estimated for each region during the histogram fit. Voxels in the intensity-normalized image are then classified into six tissue types using a maximum a posteriori classifier. This classifier combines the partial volume tissue measurement model with a Gibbs prior that models the spatial properties of the brain. We validate each stage of our algorithm on real and phantom data. Using data from the 20 normal MRI brain data sets of the Internet Brain Segmentation Repository, our method achieved average kappa indices of kappa = 0.746 +/- 0.114 for gray matter (GM) and kappa = 0.798 +/- 0.089 for white matter (WM) compared to expert labeled data. Our method achieved average kappa indices kappa = 0.893 +/- 0.041 for GM and kappa = 0.928 +/- 0.039 for WM compared to the ground truth labeling on 12 volumes from the Montreal Neurological Institutes BrainWeb phantom.

The Quantitative Evaluation of Functional Neuroimaging Experiments: The NPAIRS Data Analysis Framework

We introduce a data-analysis framework and performance metrics for evaluating and optimizing the interaction between activation tasks, experimental designs, and the methodological choices and tools for data acquisition, preprocessing, data analysis, and extraction of statistical parametric maps (SPMs). Our NPAIRS (nonparametric prediction, activation, influence, and reproducibility resampling) framework provides an alternative to simulations and ROC curves by using real PET and fMRI data sets to examine the relationship between prediction accuracy and the signal-to-noise ratios (SNRs) associated with reproducible SPMs. Using cross-validation resampling we plot training-test set predictions of the experimental design variables (e.g., brain-state labels) versus reproducibility SNR metrics for the associated SPMs. We demonstrate the utility of this framework across the wide range of performance metrics obtained from [(15)O]water PET studies of 12 age- and sex-matched data sets performing different motor tasks (8 subjects/set). For the 12 data sets we apply NPAIRS with both univariate and multivariate data-analysis approaches to: (1) demonstrate that this framework may be used to obtain reproducible SPMs from any data-analysis approach on a common Z-score scale (rSPM[Z]); (2) demonstrate that the histogram of a rSPM[Z] image may be modeled as the sum of a data-analysis-dependent noise distribution and a task-dependent, Gaussian signal distribution that scales monotonically with our reproducibility performance metric; (3) explore the relation between prediction and reproducibility performance metrics with an emphasis on bias-variance tradeoffs for flexible, multivariate models; and (4) measure the broad range of reproducibility SNRs and the significant influence of individual subjects. A companion paper describes learning curves for four of these 12 data sets, which describe an alternative mutual-information prediction metric and NPAIRS reproducibility as a function of training-set sizes from 2 to 18 subjects. We propose the NPAIRS framework as a validation tool for testing and optimizing methodological choices and tools in functional neuroimaging.

Scaled Subprofile Model: A Statistical Approach to the Analysis of Functional Patterns in Positron Emission Tomographic Data

The data obtained from measurements of regional rCMRglu using [18F]fluorodeoxyglucose (FDG)/positron emission tomographic (PET) data contain more structure than can be identified with group mean rCMRglu profiles or regional correlation coefficients. This additional structure is revealed by a novel mathematical-statistical model of regional metabolic interactions that explicitly represents rCMRglu profiles as a combination of region-independent global effects, a group mean pattern and a mosaic of interacting networks. In its application to FDG/PET data, this model removes global subject effects [global scaling factors (GSFs)] and a group mean pattern (profile) so as to maximize statistical power for the detection and simultaneous discovery of all networks of two or more regions that form a significant and consistent linearly covarying pattern. The model approach presented here was applied to the combined rCMRglu data from 12 demented AIDS patients and 18 normal controls: Two significant metabolic covariance pattern descriptors that together accounted for 71 to 96% of the rCMRglu/GSF variation across subjects for 22/28 regions in the AIDS group were extracted. Each descriptor was found to be highly correlated with performance on several neuropsychological tests, providing independent validation of the analysis technique as a means of discovering and describing behaviorally related components of group rCMRglu profiles.

Qualitative and quantitative evaluation of six algorithms for correcting intensity nonuniformity effects.

The desire to correct intensity nonuniformity in magnetic resonance images has led to the proliferation of nonuniformity-correction (NUC) algorithms with different theoretical underpinnings. In order to provide end users with a rational basis for selecting a given algorithm for a specific neuroscientific application, we evaluated the performance of six NUC algorithms. We used simulated and real MRI data volumes, including six repeat scans of the same subject, in order to rank the accuracy, precision, and stability of the nonuniformity corrections. We also compared algorithms using data volumes from different subjects and different (1.5T and 3.0T) MRI scanners in order to relate differences in algorithmic performance to intersubject variability and/or differences in scanner performance. In phantom studies, the correlation of the extracted with the applied nonuniformity was highest in the transaxial (left-to-right) direction and lowest in the axial (top-to-bottom) direction. Two of the six algorithms demonstrated a high degree of stability, as measured by the iterative application of the algorithm to its corrected output. While none of the algorithms performed ideally under all circumstances, locally adaptive methods generally outperformed nonadaptive methods.

A randomized, double-blind, placebo-controlled trial of antidepressants in Parkinson disease

Objective: To evaluate the efficacy and safety of a selective serotonin reuptake inhibitor (SSRI) and a serotonin and norepinephrine reuptake inhibitor (SNRI) in the treatment of depression in Parkinson disease (PD). Methods: A total of 115 subjects with PD were enrolled at 20 sites. Subjects were randomized to receive an SSRI (paroxetine; n = 42), an SNRI (venlafaxine extended release [XR]; n = 34), or placebo (n = 39). Subjects met DSM-IV criteria for a depressive disorder, or operationally defined subsyndromal depression, and scored >12 on the first 17 items of the Hamilton Rating Scale for Depression (HAM-D). Subjects were followed for 12 weeks (6-week dosage adjustment, 6-week maintenance). Maximum daily dosages were 40 mg for paroxetine and 225 mg for venlafaxine XR. The primary outcome measure was change in the HAM-D score from baseline to week 12. Results: Treatment effects (relative to placebo), expressed as mean 12-week reductions in HAM-D score, were 6.2 points (97.5% confidence interval [CI] 2.2 to 10.3, p = 0.0007) in the paroxetine group and 4.2 points (97.5% CI 0.1 to 8.4, p = 0.02) in the venlafaxine XR group. No treatment effects were seen on motor function. Conclusions: Both paroxetine and venlafaxine XR significantly improved depression in subjects with PD. Both medications were generally safe and well tolerated and did not worsen motor function. Classification of Evidence: This study provides Class I evidence that paroxetine and venlafaxine XR are effective in treating depression in patients with PD.

Quantitative comparisons of image registration techniques based on high-resolution MRI of the brain.

Objective A variety of methods for matching intrasubject MRI-MRI, PET-PET, or MRI-PET image pairs have been proposed. Based on the rigid body transformations needed to align pairs of high-resolution MRI scans and/or simulated PET scans (derived from these MRI scans), we obtained general comparisons of four intrasubject image registration techniques: Talairach coordinates, head and hat, equivalent internal points, and ratio image uniformity. In addition, we obtained a comparison of stereotaxic Z frames with a customized head mold for MRI-MRI image pairs. Materials and Methods and Results Each technique was quantitatively evaluated using the mean and maximum voxel registration errors for matched voxel pairs within the brain volumes being registered. Conclusion We conclude that fiducial markers such as stereotaxic Z frames that are not rigidly fixed to a patients skull are inaccurate compared with other registration techniques, Talairach coordinate transformations provide surprisingly good registration, and minimizing the variance of MRI-MRI, PET-PET, or MRI-PET ratio images provides significantly better registration than all other techniques tested. Registration optimization based on measurement of the similarity of spatial distributions of voxel values is superior to techniques that do not use such information.

The Evaluation of Preprocessing Choices in Single-Subject BOLD fMRI Using NPAIRS Performance Metrics

This work proposes an alternative to simulation-based receiver operating characteristic (ROC) analysis for assessment of fMRI data analysis methodologies. Specifically, we apply the rapidly developing nonparametric prediction, activation, influence, and reproducibility resampling (NPAIRS) framework to obtain cross-validation-based model performance estimates of prediction accuracy and global reproducibility for various degrees of model complexity. We rely on the concept of an analysis chain meta-model in which all parameters of the preprocessing steps along with the final statistical model are treated as estimated model parameters. Our ROC analog, then, consists of plotting prediction vs. reproducibility results as curves of model complexity for competing meta-models. Two theoretical underpinnings are crucial to utilizing this new validation technique. First, we explore the relationship between global signal-to-noise and our reproducibility estimates as derived previously. Second, we submit our model complexity curves in the prediction versus reproducibility space as reflecting classic bias-variance tradeoffs. Among the particular analysis chains considered, we found little impact in performance metrics with alignment, some benefit with temporal detrending, and greatest improvement with spatial smoothing.

Optimizing the fMRI data-processing pipeline using prediction and reproducibility performance metrics: I. A preliminary group analysis.

We argue that published results demonstrate that new insights into human brain function may be obscured by poor and/or limited choices in the data-processing pipeline, and review the work on performance metrics for optimizing pipelines: prediction, reproducibility, and related empirical Receiver Operating Characteristic (ROC) curve metrics. Using the NPAIRS split-half resampling framework for estimating prediction/reproducibility metrics (Strother et al., 2002), we illustrate its use by testing the relative importance of selected pipeline components (interpolation, in-plane spatial smoothing, temporal detrending, and between-subject alignment) in a group analysis of BOLD-fMRI scans from 16 subjects performing a block-design, parametric-static-force task. Large-scale brain networks were detected using a multivariate linear discriminant analysis (canonical variates analysis, CVA) that was tuned to fit the data. We found that tuning the CVA model and spatial smoothing were the most important processing parameters. Temporal detrending was essential to remove low-frequency, reproducing time trends; the number of cosine basis functions for detrending was optimized by assuming that separate epochs of baseline scans have constant, equal means, and this assumption was assessed with prediction metrics. Higher-order polynomial warps compared to affine alignment had only a minor impact on the performance metrics. We found that both prediction and reproducibility metrics were required for optimizing the pipeline and give somewhat different results. Moreover, the parameter settings of components in the pipeline interact so that the current practice of reporting the optimization of components tested in relative isolation is unlikely to lead to fully optimized processing pipelines.

The quantitative evaluation of functional neuroimaging experiments: mutual information learning curves.

Learning curves are presented as an unbiased means for evaluating the performance of models for neuroimaging data analysis. The learning curve measures the predictive performance in terms of the generalization or prediction error as a function of the number of independent examples (e.g., subjects) used to determine the parameters in the model. Cross-validation resampling is used to obtain unbiased estimates of a generic multivariate Gaussian classifier, for training set sizes from 2 to 16 subjects. We apply the framework to four different activation experiments, in this case [(15)O]water data sets, although the framework is equally valid for multisubject fMRI studies. We demonstrate how the prediction error can be expressed as the mutual information between the scan and the scan label, measured in units of bits. The mutual information learning curve can be used to evaluate the impact of different methodological choices, e.g., classification label schemes, preprocessing choices. Another application for the learning curve is to examine the model performance using bias/variance considerations enabling the researcher to determine if the model performance is limited by statistical bias or variance. We furthermore present the sensitivity map as a general method for extracting activation maps from statistical models within the probabilistic framework and illustrate relationships between mutual information and pattern reproducibility as derived in the NPAIRS framework described in a companion paper.

Principal Component Analysis and the Scaled Subprofile Model Compared to Intersubject Averaging and Statistical Parametric Mapping: I. “Functional Connectivity” of the Human Motor System Studied with [15O]Water PET

Using [15O]water PET and a previously well studied motor activation task, repetitive finger-to-thumb opposition, we compared the spatial activation patterns produced by (1) global normalization and intersubject averaging of paired-image subtractions, (2) the mean differences of ANCOVA-adjusted voxels in Statistical Parametric Mapping, (3) ANCOVA-adjusted voxels followed by principal component analysis (PCA), (4) ANCOVA-adjustment of mean image volumes (mean over subjects at each time point) followed by F-masking and PCA, and (5) PCA with Scaled Subprofile Model pre- and postprocessing. All data analysis techniques identified large positive focal activations in the contralateral sensorimotor cortex and ipsilateral cerebellar cortex, with varying levels of activation in other parts of the motor system, e.g., supplementary motor area, thalamus, putamen; techniques 1–4 also produced extensive negative areas. The activation signal of interest constitutes a very small fraction of the total nonrandom signal in the original dataset, and the exact choice of data preprocessing steps together with a particular analysis procedure have a significant impact on the identification and relative levels of activated regions. The challenge for the future is to identify those preprocessing algorithms and data analysis models that reproducibly optimize the identification and quantification of higher-order sensorimotor and cognitive responses.