1,721,004 research outputs found
Beyond mind reading: advanced machine learning techniques for FMRI data analysis
Get PDFThe advent of functional Magnetic Resonance Imaging (fMRI) has significantly improved the knowledge about the neural correlates of perceptual and cognitive processes. The aim of this thesis is to discuss the characteristics of different approaches for fMRI data analysis, from the conventional mass univariate analysis (General Linear Model - GLM), to the multivariate analysis (i.e., data-driven and pattern based methods), and propose a novel, advanced method (Functional ANOVA Models of Gaussian Kernels - FAM-GK) for the analysis of fMRI data acquired in the context of fast event-related experiments. FAM-GK is an embedded method for voxel selection and is able to capture the nonlinear spatio-temporal dynamics of the BOLD signals by performing nonlinear estimation of the experimental conditions. The impact of crucial aspects concerning the use of pattern recognition methods on the fMRI data analysis, such as voxel selection, the choice of classifier and tuning parameters, the cross-validation techniques, are investigated and discussed by analysing the results obtained in four neuroimaging case studies.
In a first study, we explore the robustness of nonlinear Support Vector regression (SVR), combined with a filter approach for voxel selection, in the case of an extremely complex regression problem, in which we had to predict the subjective experience of participants immersed in a virtual reality environment.
In a second study, we face the problem of voxel selection combined with the choice of the best classifier, and we propose a methodology based on genetic algorithms and nonlinear support vector machine (GA-SVM) efficiently combined in a wrapper approach.
In a third study we compare three pattern recognition techniques (i.e., linear SVM, nonlinear SVM, and FAM-GK) for investigating the neural correlates of the representation of numerical and non-numerical ordered sequences (i.e., numbers and letters) in the horizontal segment of the Intraparietal Sulcus (hIPS). The FAM-GK method significantly outperformed the other two classifiers. The results show a partial overlapping of the two representation systems suggesting the existence of neural substrates in hIPS codifying the cardinal and the ordinal dimensions of numbers and letters in a partially independent way.
Finally, in the last preliminary study, we tested the same three pattern recognition methods on fMRI data acquired in the context of a fast event-related experiment. The FAM-GK method shows a very high performance, whereas the other classifiers fail to achieve an acceptable classification performance.L’avvento della tecnica di Risonanza Magnetica funzionale (fMRI) ha notevolmente migliorato le conoscenze sui correlati neurali sottostanti i processi cognitivi. Obiettivo di questa tesi è stato quello di illustrare e discutere criticamente le caratteristiche dei diversi approcci per l’analisi dei dati fMRI, dai metodi convenzionali di analisi univariata (General Linear Model - GLM) ai metodi di analisi multivariata (metodi data-driven e di pattern recognition), proponendo una nuova tecnica avanzata (Functional ANOVA Models of Gaussian Kernels - FAM-GK) per l’analisi di dati fMRI acquisiti con paradigmi sperimentali fast event-related. FAM-GK è un metodo embedded per la selezione dei voxels, che è in grado di catturare le dinamiche non lineari spazio-temporali del segnale BOLD, effettuando stime non lineari delle condizioni sperimentali. L’impatto degli aspetti critici riguardanti l’uso di tecniche di pattern recognition sull’analisi di dati fMRI, tra cui la selezione dei voxels, la scelta del classificatore e dei suoi parametri di apprendimento, le tecniche di cross-validation, sono valutati e discussi analizzando i risultati ottenuti in quattro casi di studio.
In un primo studio, abbiamo indagato la robustezza di Support Vector regression (SVR) non lineare, integrato con un approccio di tipo filter per la selezione dei voxels, in un caso di un problema di regressione estremamente complesso, in cui dovevamo predire l’esperienza soggettiva di alcuni partecipanti immersi in un ambiente di realtà virtuale.
In un secondo studio, abbiamo affrontato il problema della selezione dei voxels integrato con la scelta del miglior classificatore, proponendo un metodo basato sugli algoritmi genetici e SVM non lineare (GA-SVM) in un approccio di tipo wrapper.
In un terzo studio, abbiamo confrontato tre metodi di pattern recognition (SVM lineare, SVM non lineare e FAM-GK) per indagare i correlati neurali della rappresentazione di sequenze ordinate numeriche e non-numeriche (numeri e lettere) a livello del segmento orizzontale del solco intraparitale (hIPS). Le prestazioni di classificazione di FAM-GK sono risultate essere significativamente superiori rispetto a quelle degli alti due classificatori. I risultati hanno mostrato una parziale sovrapposizione dei due sistemi di rappresentazione, suggerendo l’esistenza di substrati neurali nelle regioni hIPS che codificano le dimensioni cardinale e ordinale dei numeri e delle lettere in modo parzialmente indipendente.
Infine, nel quarto studio preliminare, abbiamo testato e confrontato gli stessi tre classificatori su dati fMRI acquisiti durante un esperimento fast event-related. FAM-GK ha mostrato delle prestazioni di classificazione piuttosto elevate, mentre le prestazioni degli altri due classificatori sono risultate essere di poco superiori al caso
The spatial representation of numerical and non-numerical ordered sequences: Insights from a random generation task
No full text
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Get PDF“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Get PDFWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Discrimination and Ordinal Judgments of Temporal Durations at 3 Months
No full textThis study presents the first evidence that 3-month-old infants success in a timing matching task and in an ordinal timing task, when numerical information are controlled. Three- month-old infants discriminated a brief temporal durations that differ by a 1:3 ratio, relying solely on temporal information. Moreover, at 3 months of age infants were able to discriminate between a monotonic and a non-monotonic time-based series, when numerical and temporal information are inconsistent .These findings strength the hypothesis that a magnitude representational system for temporal quantities is operating very early in the ontogenetic development
Dispelling the Myths Behind First-author Citation Counts
Get PDFWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
- …
