Unifying Amplitude and Phase Analysis: A Compositional Data Approach to Functional Multivariate Mixed-Effects Modeling of Mandarin Chinese.

Mandarin Chinese is characterized by being a tonal language; the pitch (or F0) of its utterances carries considerable linguistic information. However, speech samples from different individuals are subject to changes in amplitude and phase, which must be accounted for in any analysis that attempts to provide a linguistically meaningful description of the language. A joint model for amplitude, phase, and duration is presented, which combines elements from functional data analysis, compositional data analysis, and linear mixed effects models. By decomposing functions via a functional principal component analysis, and connecting registration functions to compositional data analysis, a joint multivariate mixed effect model can be formulated, which gives insights into the relationship between the different modes of variation as well as their dependence on linguistic and nonlinguistic covariates. The model is applied to the COSPRO-1 dataset, a comprehensive database of spoken Taiwanese Mandarin, containing approximately 50,000 phonetically diverse sample F0 contours (syllables), and reveals that phonetic information is jointly carried by both amplitude and phase variation. Supplementary materials for this article are available online.

Multi-resolution Bayesian regression in PET dynamic studies using wavelets.

In the kinetic analysis of dynamic PET data, one usually posits that the variation of the data through one dimension, time, can be described by a mathematical model encapsulating the relevant physiological features of the radioactive tracer. In this work, we posit that the remaining dimension, space, can also be modeled as a physiological feature, and we introduce this concept into a new computational procedure for the production of parametric maps. An organ and, in the instance considered here, the brain presents similarities in the physiological properties of its elements across scales: computationally, this similarity can be implemented in two stages. Firstly, a multi-scale decomposition of the dynamic frames is created through the wavelet transform. Secondly, kinetic analysis is performed in wavelet space and the kinetic parameters estimated at low resolution are used as priors to inform estimates at higher resolutions. Kinetic analysis in the above scheme is achieved by extension of the Patlak analysis through Bayesian linear regression that retains the simplicity and speed of the original procedure. Application to artificial and real data (FDG and FDOPA) demonstrates the ability of the procedure to reduce remarkably the variance of parametric maps (up to 4-fold reduction) without introducing sizeable bias. Significance of the methodology and extension of the procedure to other data (fMRI) and models are discussed.

Dynamic functional changes associated with cognitive skill learning of an adapted version of the Tower of London task.

In this study, we used a modified version of the Tower of London (TOL) planning task, in conjunction with positron emission tomography, to examine the neural substrates mediating cognitive skill learning. Twelve healthy, right-handed participants solved a total of 90 different TOL problems. They were scanned on four occasions during the fast learning stage as well as in a condition designed to control for internally guided movements. Practice of the TOL task resulted in a significant decrease in planning, execution, and total time taken to solve the problems. Consistent with the results of previous studies, early performance of the TOL task was associated with increased blood flow activity in the dorsolateral prefrontal, orbitofrontal, and parietal cortices on the left, as well as in the caudate nucleus, cerebellum, and premotor cortex, bilaterally. Interestingly, however, activity in the left caudate nucleus was maintained from the beginning to the end of the learning process, suggesting that this structure plays a role in this type of cognitive skill learning. In addition, correlational analyses revealed that improved performance on the TOL task was associated with a significant decrease of activity in the medial orbitofrontal and frontopolar cortices over the course of learning, areas thought to be involved in decision making, guessing, and monitoring of feedback information. In sum, the results lend further support to the idea that the learning of cognitive skills requiring planning and working memory capacities is mediated through a fronto-striatal network.

Estimating the delay of the fMRI response.

We propose a fast, efficient, general, simple, valid, and robust method of estimating and making inference about the delay of the fMRI response modeled as a temporal shift of the hemodynamic response function (HRF). We estimate the shift unbiasedly using two optimally chosen basis functions for a spectrum of time shifted HRFs. This is done at every voxel, to create an image of estimated delays and their standard deviations. This can be used to compare delays for the same stimulus at different voxels, or for different stimuli at the same voxel. Our method is compared to other alternatives and validated on an fMRI data set from an experiment in pain perception.

Reduced dopamine D1 receptor binding in the ventral striatum of cigarette smokers.

Several drugs of abuse, including nicotine, are thought to exert their reinforcing effects through actions on the mesolimbic dopamine system. Animal and human studies suggest that chronic administration of addictive drugs may lead to impaired dopamine neurotransmission in the nucleus accumbens. We measured D1 receptor density in 11 smokers and 18 nonsmokers using positron emission tomography and the D1 receptor ligand [11C]SCH 23390. Ten of the smokers were scanned twice, once after overnight abstinence from cigarettes, and once while smoking at their usual rate, to account for possible acute effects of cigarette smoking on D1 receptor binding. In addition, eight control subjects were scanned twice to assess the reproducibility of the method. We used compartmental modeling to measure [11C]SCH 23390 binding potential, a measure of D1 receptor density. There were no differences in binding between abstinent and nonabstinent scans in smokers or in the two scans in controls. However, there was a significant reduction in [11C]SCH 23390 binding potential in smokers compared to nonsmokers in the striatum, most prominently in the ventral striatum. This suggests that there is a reduction in dopamine D1 receptor density in the ventral striatum of human cigarette smokers relative to nonsmokers, which implies that the postsynaptic mesolimbic dopamine system may be chronically underactive in smokers, either as an antecedent or consequence of addiction to cigarettes. Such a hypodopaminergic state may play an important role in sustaining nicotine-seeking behavior. Alternatively, an inherited reduction in dopamine receptors in the striatum may be associated with an increased risk of addictive behavior.

Statistical modeling of positron emission tomography images in wavelet space.

A new method is introduced for the analysis of multiple studies measured with emission tomography. Traditional models of statistical analysis (ANOVA, ANCOVA and other linear models) are applied not directly on images but on their correspondent wavelet transforms. Maps of model effects estimated from these models are filtered using a thresholding procedure based on a simple Bonferroni correction and then reconstructed. This procedure inherently represents a complete modeling approach and therefore obtains estimates of the effects of interest (condition effect, difference between conditions, covariate of interest, and so on) under the specified statistical risk. By performing the statistical modeling step in wavelet space. the procedure allows the direct estimation of the error for each wavelet coefficient; hence, the local noise characteristics are accounted for in the subsequent filtering. The method was validated by use of a null dataset and then applied to typical examples of neuroimaging studies to highlight conceptual and practical differences from existing statistical parametric mapping approaches.

A statistical method for the analysis of positron emission tomography neuroreceptor ligand data.

A method for voxel by voxel statistical inference of PET radioligand receptor studies is presented. This method is aimed at detecting differences in radioligand binding between baseline and activation scans. It uses nonlinear least squares theory to estimate the ligand-receptor model parameters and utilizes the residuals to calculate their associated variance. The approach both increases the degrees of freedom for statistical testing and produces more accurate estimates of the standard deviation of the parameters. This technique is applicable to any ligand with a validated compartmental model, whether reversibly or irreversibly bound. The method was investigated and compared with a simple voxel-wise t test. Both simulated and real PET data for the dopamine D(1) receptor ligand [(11)C]SCH 23390 were used to assess the method. The assumptions implicit in the residuals methods were validated. The residuals method was found to be more sensitive than a simple t test, while not producing false-positive results. In addition, we showed that this method reliably differentiates changes in radioligand binding from the effects of changes in cerebral blood flow.

Modeling dynamic PET-SPECT studies in the wavelet domain.

This work develops a theoretical framework and corresponding algorithms for the modeling of dynamic PET-SPECT studies both in time and space. The problem of estimating the spatial dimension is solved by applying the wavelet transform to each scan of the dynamic sequence and then performing the kinetic modeling and statistical analysis in the wavelet domain. On reconstruction through the inverse wavelet transform, one obtains parametric images that are consistent estimates of the spatial patterns of the kinetic parameter of interest. The theoretical setup allows the use of linear techniques currently used in PET-SPECT for kinetic analysis. The method is applied to artificial and real data sets. The application to dynamic PET-SPECT studies was performed both for validation purposes, when the spatial patterns are known, and for illustration of the advantages offered by the technique in case of tracers with an unknown pattern of distribution.