RESUMO
The field of fMRI data analysis is rapidly growing in sophistication, particularly in the domain of multivariate pattern classification. However, the interaction between the properties of the analytical model and the parameters of the BOLD signal (e.g. signal magnitude, temporal variance and functional connectivity) is still an open problem. We addressed this problem by evaluating a set of pattern classification algorithms on simulated and experimental block-design fMRI data. The set of classifiers consisted of linear and quadratic discriminants, linear support vector machine, and linear and nonlinear Gaussian naive Bayes classifiers. For linear discriminant, we used two methods of regularization: principal component analysis, and ridge regularization. The classifiers were used (1) to classify the volumes according to the behavioral task that was performed by the subject, and (2) to construct spatial maps that indicated the relative contribution of each voxel to classification. Our evaluation metrics were: (1) accuracy of out-of-sample classification and (2) reproducibility of spatial maps. In simulated data sets, we performed an additional evaluation of spatial maps with ROC analysis. We varied the magnitude, temporal variance and connectivity of simulated fMRI signal and identified the optimal classifier for each simulated environment. Overall, the best performers were linear and quadratic discriminants (operating on principal components of the data matrix) and, in some rare situations, a nonlinear Gaussian naïve Bayes classifier. The results from the simulated data were supported by within-subject analysis of experimental fMRI data, collected in a study of aging. This is the first study that systematically characterizes interactions between analysis model and signal parameters (such as magnitude, variance and correlation) on the performance of pattern classifiers for fMRI.
Assuntos
Envelhecimento/fisiologia , Mapeamento Encefálico/métodos , Córtex Cerebral/fisiologia , Cognição/fisiologia , Imageamento por Ressonância Magnética/métodos , Rede Nervosa/fisiologia , Reconhecimento Automatizado de Padrão/métodos , Adulto , Idoso , Idoso de 80 Anos ou mais , Inteligência Artificial , Simulação por Computador , Interpretação Estatística de Dados , Humanos , Aumento da Imagem/métodos , Interpretação de Imagem Assistida por Computador/métodos , Masculino , Pessoa de Meia-Idade , Modelos Neurológicos , Modelos Estatísticos , Reprodutibilidade dos Testes , Sensibilidade e Especificidade , Análise Espaço-Temporal , Adulto JovemRESUMO
PURPOSE: We present a new morphometric measure of trabecular bone microarchitecture, called mean node strength (NdStr), which is part of a newly developed approach called long range node-strut analysis. Our general aim is to describe and quantify the apparent "latticelike" microarchitecture of the trabecular bone network. METHODS: Similar in some ways to the topological node-strut analysis introduced by Garrahan et al. [J. Microsc. 142, 341-349 (1986)], our method is distinguished by an emphasis on long-range trabecular connectivity. Thus, while the topological classification of a pixel (after skeletonization) as a node, strut, or terminus, can be determined from the 3 × 3 neighborhood of that pixel, our method, which does not involve skeletonization, takes into account a much larger neighborhood. In addition, rather than giving a discrete classification of each pixel as a node, strut, or terminus, our method produces a continuous variable, node strength. The node strength is averaged over a region of interest to produce the mean node strength of the region. RESULTS: We have applied our long range node-strut analysis to a set of 26 high-resolution peripheral quantitative computed tomography (pQCT) axial images of human proximal tibiae acquired 17 mm below the tibial plateau. We found that NdStr has a strong positive correlation with trabecular volumetric bone mineral density (BMD). After an exponential transformation, we obtain a Pearson's correlation coefficient of r = 0.97. Qualitative comparison of images with similar BMD but with very different NdStr values suggests that the latter measure has successfully quantified the prevalence of the "latticelike" microarchitecture apparent in the image. Moreover, we found a strong correlation (r = 0.62) between NdStr and the conventional node-terminus ratio (Nd∕Tm) of Garrahan et al. The Nd∕Tm ratios were computed using traditional histomorphometry performed on bone biopsies obtained at the same location as the pQCT scans. CONCLUSIONS: The newly introduced morphometric measure allows a quantitative assessment of the long-range connectivity of trabecular bone. One advantage of this method is that it is based on pQCT images that can be obtained noninvasively from patients, i.e., without having to obtain a bone biopsy from the patient.
Assuntos
Tíbia/citologia , Tíbia/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos , Idoso , Idoso de 80 Anos ou mais , Densidade Óssea , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Tíbia/fisiologiaRESUMO
We compare 10 methods of classifying fMRI volumes by applying them to data from a longitudinal study of stroke recovery: adaptive Fisher's linear and quadratic discriminant; gaussian naive Bayes; support vector machines with linear, quadratic, and radial basis function (RBF) kernels; logistic regression; two novel methods based on pairs of restricted Boltzmann machines (RBM); and K-nearest neighbors. All methods were tested on three binary classification tasks, and their out-of-sample classification accuracies are compared. The relative performance of the methods varies considerably across subjects and classification tasks. The best overall performers were adaptive quadratic discriminant, support vector machines with RBF kernels, and generatively trained pairs of RBMs.
Assuntos
Interpretação de Imagem Assistida por Computador/métodos , Imageamento por Ressonância Magnética , Reconhecimento Automatizado de Padrão/métodos , Acidente Vascular Cerebral/patologia , Algoritmos , HumanosRESUMO
Historically, research databases have existed in isolation with no practical avenue for sharing or pooling medical data into high dimensional datasets that can be efficiently compared across databases. To address this challenge, the Ontario Brain Institute's "Brain-CODE" is a large-scale neuroinformatics platform designed to support the collection, storage, federation, sharing and analysis of different data types across several brain disorders, as a means to understand common underlying causes of brain dysfunction and develop novel approaches to treatment. By providing researchers access to aggregated datasets that they otherwise could not obtain independently, Brain-CODE incentivizes data sharing and collaboration and facilitates analyses both within and across disorders and across a wide array of data types, including clinical, neuroimaging and molecular. The Brain-CODE system architecture provides the technical capabilities to support (1) consolidated data management to securely capture, monitor and curate data, (2) privacy and security best-practices, and (3) interoperable and extensible systems that support harmonization, integration, and query across diverse data modalities and linkages to external data sources. Brain-CODE currently supports collaborative research networks focused on various brain conditions, including neurodevelopmental disorders, cerebral palsy, neurodegenerative diseases, epilepsy and mood disorders. These programs are generating large volumes of data that are integrated within Brain-CODE to support scientific inquiry and analytics across multiple brain disorders and modalities. By providing access to very large datasets on patients with different brain disorders and enabling linkages to provincial, national and international databases, Brain-CODE will help to generate new hypotheses about the biological bases of brain disorders, and ultimately promote new discoveries to improve patient care.
RESUMO
BACKGROUND: Performance variability in individuals with aphasia is typically regarded as a nuisance factor complicating assessment and treatment. OBJECTIVE: We present the alternative hypothesis that intraindividual variability represents a fundamental characteristic of an individual's functioning and an important biomarker for therapeutic selection and prognosis. METHODS: A total of 19 individuals with chronic aphasia participated in a 6-week trial of imitation-based speech therapy. We assessed improvement both on overall language functioning and repetition ability. Furthermore, we determined which pretreatment variables best predicted improvement on the repetition test. RESULTS: Significant gains were made on the Western Aphasia Battery-Revised (WAB) Aphasia Quotient, Cortical Quotient, and 2 subtests as well as on a separate repetition test. Using stepwise regression, we found that pretreatment intraindividual variability was the only predictor of improvement in performance on the repetition test, with greater pretreatment variability predicting greater improvement. Furthermore, the degree of reduction in this variability over the course of treatment was positively correlated with the degree of improvement. CONCLUSIONS: Intraindividual variability may be indicative of potential for improvement on a given task, with more uniform performance suggesting functioning at or near peak potential.
Assuntos
Afasia/reabilitação , Fonoterapia/métodos , Resultado do Tratamento , Adulto , Idoso , Afasia/etiologia , Feminino , Humanos , Modelos Lineares , Masculino , Pessoa de Meia-Idade , Valor Preditivo dos Testes , Acidente Vascular Cerebral/complicações , Reabilitação do Acidente Vascular CerebralRESUMO
We present a new framework for diffeomorphic image registration which supports natural interpretations of spatially-varying metrics. This framework is based on left-invariant diffeomorphic metrics (LIDM) and is closely related to the now standard large deformation diffeomorphic metric mapping (LDDMM). We discuss the relationship between LIDM and LDDMM and introduce a computationally convenient class of spatially-varying metrics appropriate for both frameworks. Finally, we demonstrate the effectiveness of our method on a 2D toy example and on the 40 3D brain images of the LPBA40 dataset.