Efeito da coesão semântica sobre a discriminação de itens emocionais em testes de memória de reconhecimento de palavras

Highly emotional itens are best remembered in emotional memory tasks than neutral items. An example of emotional item that benefits declarative memory processes are the taboo words. These words undergo from a conventional prohibition, imposed by tradition or custom. Literature suggests that the strongest recollection these words is due to emotional arousal, as well as, the fact that they form a cohesive semantic group, which is a positive additive effect. However, studies with semantic lists show that cohesion can have a negative effect of interference, impairing memory. We analyzed, in two experiments, the effect of arousal and semantic cohesion of taboo words on recognition tests, comparing with into two other word categories: semantically related and without emotional arousal (semantic category) and neutral, with low semantic relation (objects). Our results indicate that cohesion has interfered whith the performance of the test by increasing the number of false alarms. This effect was strongly observed in the semantic category of words in both experiments, but also in the neutral and taboo words, when both were explicitly considered as semantic categories through the instruction of the test in Experiment 2. Despite the impairment induced by semantic cohesion in both experiments, the taboo words were more discriminated than others, and this result agrees with the indication of the emotional arousal as the main factor for the best recollection of emotional items in memory tests

Reconhecimento de palavras, nomeação de figuras e de palavras impressas em surdos implantados pré-linguais

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Desenvolvimento de arquitetura para sistemas de reconhecimento automático de voz baseados em modelos ocultos de Markov

Este trabalho foi realizado dentro da área de reconhecimento automático de voz (RAV). Atualmente, a maioria dos sistemas de RAV é baseada nos modelos ocultos de Markov (HMMs) [GOM 99] [GOM 99b], quer utilizando-os exclusivamente, quer utilizando-os em conjunto com outras técnicas e constituindo sistemas híbridos. A abordagem estatística dos HMMs tem mostrado ser uma das mais poderosas ferramentas disponíveis para a modelagem acústica e temporal do sinal de voz. A melhora da taxa de reconhecimento exige algoritmos mais complexos [RAV 96]. O aumento do tamanho do vocabulário ou do número de locutores exige um processamento computacional adicional. Certas aplicações, como a verificação de locutor ou o reconhecimento de diálogo podem exigir processamento em tempo real [DOD 85] [MAM 96]. Outras aplicações tais como brinquedos ou máquinas portáveis ainda podem agregar o requisito de portabilidade, e de baixo consumo, além de um sistema fisicamente compacto. Tais necessidades exigem uma solução em hardware. O presente trabalho propõe a implementação de um sistema de RAV utilizando hardware baseado em FPGAs (Field Programmable Gate Arrays) e otimizando os algoritmos que se utilizam no RAV. Foi feito um estudo dos sistemas de RAV e das técnicas que a maioria dos sistemas utiliza em cada etapa que os conforma. Deu-se especial ênfase aos Modelos Ocultos de Markov, seus algoritmos de cálculo de probabilidades, de treinamento e de decodificação de estados, e sua aplicação nos sistemas de RAV. Foi realizado um estudo comparativo dos sistemas em hardware, produzidos por outros centros de pesquisa, identificando algumas das suas características mais relevantes. Foi implementado um modelo de software, descrito neste trabalho, utilizado para validar os algoritmos de RAV e auxiliar na especificação em hardware. Um conjunto de funções digitais implementadas em FPGA, necessárias para o desenvolvimento de sistemas de RAV é descrito. Foram realizadas algumas modificações nos algoritmos de RAV para facilitar a implementação digital dos mesmos. A conexão, entre as funções digitais projetadas, para a implementação de um sistema de reconhecimento de palavras isoladas é aqui apresentado. A implementação em FPGA da etapa de pré-processamento, que inclui a pré-ênfase, janelamento e extração de características, e a implementação da etapa de reconhecimento são apresentadas finalmente neste trabalho.

Reconhecimento de voz através de unidades menores do que a palavra, utilizando Wavelet Packet e SVM, em uma nova estrutura hierárquica de decisão

The automatic speech recognition by machine has been the target of researchers in the past five decades. In this period have been numerous advances, such as in the field of recognition of isolated words (commands), which has very high rates of recognition, currently. However, we are still far from developing a system that could have a performance similar to the human being (automatic continuous speech recognition). One of the great challenges of searches for continuous speech recognition is the large amount of pattern. The modern languages such as English, French, Spanish and Portuguese have approximately 500,000 words or patterns to be identified. The purpose of this study is to use smaller units than the word such as phonemes, syllables and difones units as the basis for the speech recognition, aiming to recognize any words without necessarily using them. The main goal is to reduce the restriction imposed by the excessive amount of patterns. In order to validate this proposal, the system was tested in the isolated word recognition in dependent-case. The phonemes characteristics of the Brazil s Portuguese language were used to developed the hierarchy decision system. These decisions are made through the use of neural networks SVM (Support Vector Machines). The main speech features used were obtained from the Wavelet Packet Transform. The descriptors MFCC (Mel-Frequency Cepstral Coefficient) are also used in this work. It was concluded that the method proposed in this work, showed good results in the steps of recognition of vowels, consonants (syllables) and words when compared with other existing methods in literature

Memória de reconhecimento em indivíduos adultos com mais de 45 anos

As we grow old, there are many cognitive processes which decline in the human brain. One of them is the memory, a function that allows retention and posterior use of knowledge learned during the life, understood as a result of multiple systems highly organized and spread in several neural regions. This work aimed to evaluate the recognition memory in adults over 45 years old through words and pictures recognition tasks and the use of two codification or learning conditions (same distracters and different distracters). Twelve individuals were studied (6 men and 6 women) aged between 45 and 88 years old and with similar demographic characteristics. They presented better performance on picture tasks rather than word tasks. Better results were also verified when the codification context had different distracters, which significantly reflected in a long term principally in elderly individuals. The results reached suggest that the codification context influenced the lists of pictures and words learning, mainly for the elderly ones, when compared to adults, and that these results can be related to the phenomena involved with the recognition memory, the recollection and familiarity

Estampa temporal na memória de reconhecimento em estudantes universitários

Several studies have shown that there is a circadian modulation of explicit memory. This modulation can occur independently in each one of the mnemonic processes. The aim of this study was to evaluate the influence of time of training on short-term memory (STM) and long-term memory (LTM), using a recognition task. Moreover, a possible circadian modulation in retrieving was investigated when this process matched the acquisition hour (time stamp). The chronotype variable was also considered. Fifty-seven undergraduate students aging between 18 and 25 years (21,72 ± 2,14; 28 ♂) participated in this study. In the training phase (acquisition) the subjects heard a ten word list. Following this, they answered a recognition test to evaluate STM and one week later they answered a recognition test to evaluate LTM. In each chronotype, the subjects were divided in groups according to the training hour, part of them in the morning and the other in the afternoon. One week later some of the subjects in each group underwent LTM testing in the morning and others in the afternoon. When the subjects performances were analyzed together, independently of the chronotypes, a training hour effect was found in the LTM. The subjects trained in the afternoon had better performance. No time of day effect was found in the STM and in retrieving from the LTM. However, the morning types who were trained and tested in the same hour had a better performance in the LTM when compared to morning types trained and tested in different hours. This effect did not occur when the other chronotypes were analyzed. The circadian modulation seems to occur at least in two different ways. First, there is a circadian modulation in the acquisition/consolidation processes, with a better performance occuring in the afternoon. Secondly, there is a modulation in the retrieval mnemonic process, called time stamp phenomenon. This phenomenon, that occurred in the morning types, is showed for the first time in humans

Reconhecimento de voz para aplicações em automação implementado em FPGA

In many movies of scientific fiction, machines were capable of speaking with humans. However mankind is still far away of getting those types of machines, like the famous character C3PO of Star Wars. During the last six decades the automatic speech recognition systems have been the target of many studies. Throughout these years many technics were developed to be used in applications of both software and hardware. There are many types of automatic speech recognition system, among which the one used in this work were the isolated word and independent of the speaker system, using Hidden Markov Models as the recognition system. The goals of this work is to project and synthesize the first two steps of the speech recognition system, the steps are: the speech signal acquisition and the pre-processing of the signal. Both steps were developed in a reprogrammable component named FPGA, using the VHDL hardware description language, owing to the high performance of this component and the flexibility of the language. In this work it is presented all the theory of digital signal processing, as Fast Fourier Transforms and digital filters and also all the theory of speech recognition using Hidden Markov Models and LPC processor. It is also presented all the results obtained for each one of the blocks synthesized e verified in hardware

Mining biomedical information from scientific literature

The rapid evolution and proliferation of a world-wide computerized network, the Internet, resulted in an overwhelming and constantly growing amount of publicly available data and information, a fact that was also verified in biomedicine. However, the lack of structure of textual data inhibits its direct processing by computational solutions. Information extraction is the task of text mining that intends to automatically collect information from unstructured text data sources. The goal of the work described in this thesis was to build innovative solutions for biomedical information extraction from scientific literature, through the development of simple software artifacts for developers and biocurators, delivering more accurate, usable and faster results. We started by tackling named entity recognition - a crucial initial task - with the development of Gimli, a machine-learning-based solution that follows an incremental approach to optimize extracted linguistic characteristics for each concept type. Afterwards, Totum was built to harmonize concept names provided by heterogeneous systems, delivering a robust solution with improved performance results. Such approach takes advantage of heterogenous corpora to deliver cross-corpus harmonization that is not constrained to specific characteristics. Since previous solutions do not provide links to knowledge bases, Neji was built to streamline the development of complex and custom solutions for biomedical concept name recognition and normalization. This was achieved through a modular and flexible framework focused on speed and performance, integrating a large amount of processing modules optimized for the biomedical domain. To offer on-demand heterogenous biomedical concept identification, we developed BeCAS, a web application, service and widget. We also tackled relation mining by developing TrigNER, a machine-learning-based solution for biomedical event trigger recognition, which applies an automatic algorithm to obtain the best linguistic features and model parameters for each event type. Finally, in order to assist biocurators, Egas was developed to support rapid, interactive and real-time collaborative curation of biomedical documents, through manual and automatic in-line annotation of concepts and relations. Overall, the research work presented in this thesis contributed to a more accurate update of current biomedical knowledge bases, towards improved hypothesis generation and knowledge discovery.

An embodied approach to language: the role of bodily feedback in discourse processing

Theories of embodied cognition argue that language processing arises not from amodal symbols that redescribe sensorimotor and affective experiences, but from partial simulations (reenactments) of modality-specific states. Recent findings on processing of words and sentences support such a stance emphasizing that the role of the body in the domain of language comprehension should not be overlooked or dismissed. The present research was conducted to extend prior work in two important ways. First, the role of simulation was tested with connected discourse rather than words or sentences presented in isolation. Second, both “online” and “offline” measures of discourse comprehension were taken. In Experiments 1 and 2 participants’ facial postures were manipulated to show that preparing the body for processing of emotion-congruent information improves discourse comprehension. In Experiment 3 the direction of body posture was manipulated to show that implicit properties of simulations, such as spatial dimension or location, are at least somewhat involved in processing of large language segments such as discourse. Finally, in Experiments 4 and 5 participants’ body movement and body posture were manipulated to show that even understanding of language describing metaphorical actions physically impossible to perform involves constructing a sensorimotor simulation of the described event. The major result was that compatibility between embodiment and language strongly modulated performance effectiveness in experiments on simulation of emotion and metaphorical action. The effect of simulation on comprehension of discourse implying spatial dimension was fragile. These findings support an embodied simulation account of cognition suggesting that sensorimotor and affective states are at least partially implicated in “online” and “offline” discourse comprehension.

Dificuldades de aprendizagem na leitura e na escrita: o contributo de jogos eletrónicos na motivação e na aprendizagem

Projeto de Intervenção apresentado à Escola Superior de Educação de Lisboa para obtenção do grau de Mestre em Educação Especial, Especialização em Problemas de Aprendizagem e Comportamento

Normas de terminación para la palabra final de oraciones en español para niños mexicanos

Las normas de terminación de oraciones son un recurso valioso para los investigadores interesados en el estudio de los efectos que tiene un contexto sobre los procesos de reconocimiento de palabras, y la comprensión del lenguaje en general. Este estudio presenta las normas de terminación para la palabra final de 278 oraciones en español para niños mexicanos, obtenidas a través del método de cloze, en una muestra de 420 estudiantes (226 niños y 194 niñas) de 9 a 12 años de edad del 4º al 6º grado de primarias públicas del Estado de México. La tarea de los niños consistió en leer cada una de las oraciones y completarlas con una palabra al cierre. Se calculó la probabilidad de cierre de cada palabra respondida para cada uno de los contextos, de acuerdo a las respuestas de los niños. Aunque el corpus incluyó oraciones con diferentes restricciones contextuales, la respuesta de los sujetos mostró un mayor número de oraciones con alta probabilidad de cierre. Este material puede ser utilizado en la investigación sobre la comprensión del lenguaje tanto auditivo como escrito en niños.

A memória de curto prazo e a síndrome de down: a relação entre contextos de desenvolvimento

Down syndrome (DS) is one of the most frequent causes of intellectual disability, affecting one in every 600 to 1000 live births. Studies have demonstrated that people with DS have a lower capacity for short-term memory (STM) and working memory (WM), which affects their capability to learn new words and to follow spoken instructions, specially when they involve multiple information or consecutive orders/orientations. It seems that the basis of the learning process, as it happens with language and mathematics comprehension and reasoning, relies in the STM and WM systems. Individuals with DS are increasingly included in mainstream education, and yet, very few researches have been conducted to investigate the influence of memory development and the type of enrollment (regular school and special school). This study investigated the relationship between the type of school enrollment with the performance on STM tests and also, the relationship of this performance with early stimulation (ES). The tests used in the first research were the digit span, free recall, word recognition and subtests of the Wechsler Intelligence Scale for Children Third Edition (WISC-III). Individuals enrolled in the regular schools group had higher scores on the digit span test and the subtests of the WISC-III. In the free recall and recognition tests, no differences were found. This study indicates that the type of enrollment might influence the memory development of individuals with DS and clearly points the need for future investigations. In the second research, the tests used were the digit span, free word recall and subtests of the WISC-III. The test results showed better performance by adults that received ES before six months of age. The studies showed improvement in STM both in people who attended or were attending regular school, as well as those who benefited from ES before six months of age. However, some issues still need to be better understood. What is the relation between this stimulation with the individual s education? Since ES may reflect a greater family involvement with the individual, what is the role of emotional components derived from this involvement in the cognitive improvement? These and other questions are part of the continuity of this study

Correlato eletroencefalográfico do estado vibracional

O estado vibracional (EV) é descrito como uma sensação de vibração intensa por todo o corpo, em que o sujeito se mantém num estado de relaxamento psicofisiológico que pode ser gerado de forma espontânea ou autoinduzida. Pessoas que aplicam esta técnica relatam alterações do estado mental e emocional, tais como: relaxamento, disposição, limpidez de pensamento, equilíbrio emocional, melhoria do raciocínio, bemestar, entre outros. Estas são, entretanto observações subjetivas, sendo a mensuração deste fenômeno uma lacuna e um desafio para a ciência. O objetivo desta pesquisa é explorar sistematicamente o estado vibracional no âmbito da neurociência. Desta forma, medidas eletroencefalográficas (EEG) foram utilizadas para observar se a sensação subjetiva de EV é acompanhada por mudanças na atividade elétrica cerebral. Além disso, para avaliar se o EV provoca algum efeito positivo em funções cognitivas como atenção e memória, foi utilizado um teste de reconhecimento de palavras antes e depois da aplicação da técnica de EV. Foram também aplicados questionários de dados gerais socioeconômicos e de saúde, do perfil de estados de humor, de qualidade do sono, além de inventários psicológicos. O foco inicial do trabalho foi a análise estatística dos dados de EEG, ficando as outras análises para uma etapa posterior. Dois grupos de voluntários foram analisados, o primeiro formado por 14 sujeitos que praticam a técnica de EV há pelo menos 10 anos (Grupo Experiente - GEXP), e o segundo formado por 11 sujeitos que nunca haviam realizado a técnica (Grupo Controle - GCONT). O GCONT obteve instruções sobre a técnica de EV antes dos experimentos. Foram realizadas análises estatísticas dos registros eletroencefalográficos, para comparar os grupos, em quatro condições: Basal, Relaxamento, Não-EV (período em que o sujeito está engajado na tarefa, mas ainda não percebe o EV) e EV (período em que o sujeito percebe o EV). Uma vez que os sujeitos do GCONT relataram não ter conseguido atingir a condição de EV propriamente, a comparação entre grupos foi feita apenas nas três condições, Basal, Relaxamento e Não-EV. Para isso, foi usado o teste de Mann-Whitney U com um limiar estatístico de p<0,05. De forma geral, o GEXP apresentou maior potência na banda de frequência alfa 2 (9,5-11,0 Hz) em todas as condições. Durante o período Não-EV, o GEXP também apresentou uma maior potência na banda de frequência alfa 3 (11,5-13,0 Hz) na região temporal esquerda, e gama 1 (30,5-55,0 Hz) e gama 2 (65,0-80,0 Hz) em regiões central, parietal e temporal esquerda, mas menor potência na banda de frequência teta 1 (3,5 - 5,0 Hz), em regiões centro-parietais. Para a análise estatística intragrupo, entre as condições, utilizou-se o teste estatístico Wilcoxon pareado. Observaram-se diferenças significativas (p<0,005), principalmente em regiões centrais, em teta 1 (3,5-5,0 Hz), sendo maior no Relaxamento, quando comparado com as condições Basal e Não-EV, no GCONT, e com o Não-EV e EV, no GEXP. No GEXP, a potência de gama 1 (30,5-55,0 Hz) e gama 2 (65,0-80,0 Hz) foi difusamente maior durante o EV, se comparado às outras três condições. Para o GCONT, apenas a condição Basal apresentou maior potência de gama 1 (30,5-55,0 Hz) e gama 2 (65,0- 80,0 Hz), se comparado com o Relaxamento. O aumento de teta 1 no Relaxamento, principalmente no GCONT, pode estar associado a uma maior sonolência deste grupo durante esta condição. Já o aumento de alfa 2 durante o Não-EV e o EV, pode estar associado com processos de atenção e cognição (DOLPPERMAYR et al., 2002; FELL et al., 2010; KLIMESCH et al., 1999; RAY E COLE, 1985). Por outro lado, o aumento da potência de gama em sujeitos experientes na técnica de EV encontrado aqui e em trabalhos anteriores, em meditadores experientes (FELL et al., 2010; LEHMANN et al., 2001; LUTZ et al., 2004), poderia estar associado a alterações nos processos mentais e cognitivos destes praticantes, tais como atenção, memória operacional, aprendizagem e percepção consciente embora, análises adicionais devam ser realizadas para excluir a possibilidade de interferência de artefatos musculares nos dados de EEG. Estes resultados suscitam a hipótese de que no engajamento da tarefa do EV e durante o EV, os sujeitos do GEXP conseguem manter-se em um estado de alerta, porém com maior nível de relaxamento e concentração. Uma inspeção mais detalhada dos dados, além de outros experimentos com diferentes protocolos, um maior número de sujeitos e pesquisas longitudinais são necessários para que testar esta hipótese

Ecóico e nomeação de figuras em crianças com deficiência auditiva pré-lingual com implante coclear

The aim of this work was to verify the effect of teaching the echoic behavior over the pictures naming in four children between eight and nine years old with prelingual hearing impaired, users of cochlear implants. The design adopted was: (a) pre-training that taught the matching-to-sample task; (b) pre-tests that selected three words to teach; (c) teaching of auditory-visual conditional relations; (d) naming pos-test; (e) the teaching of echoic with orofacial clues and, (f) the second naming pos-test. In the pre-test all participants achieved smaller percentage of correct on naming (60%-80%) and echoic (20%-50%) when compared to percentages word recognition (86%-93%). All participants learned the auditory-visual relations. The improvement on naming test occurred after auditory training select based for two participants; for other two participants the improvement on naming test occurred just after the training of echoic. Analysis of data showed that the listening and speaking performances are independent in their establishment and require specific conditions of teaching; in the case of this study, even though the result is not generalized to all participants, the highest correspondence into point to point naming was obtained following the teaching of echoic.

Campos lexicais e desenvolvimento de vocabulário na educação pré-escolar

O presente estudo pretendeu desenvolver o capital lexical de crianças a frequentar a Educação Pré-Escolar a partir da exploração de campos lexicais com recurso a obras da literatura infantil e de jogos. Paralelamente, realizamos com as crianças atividades centradas no reconhecimento de palavras, na divisão silábica e na exploração da correspondência grafema-fonema. Recolhemos dados relativos ao desempenho das crianças, a partir das atividades em que estas participaram. A análise de conteúdo dos dados recolhidos revelou que as crianças tinham efetivamente adquirido novo vocabulário, que procuravam utilizar em novos contextos, e desenvolvido algumas competências no domínio da consciência fonológica.