A detailed investigation of classification methods for vowel speech imagery recognition

Accurate and fast decoding of speech imagery from electroencephalographic (EEG) data could serve as a basis for a new generation of brain computer interfaces (BCIs), more portable and easier to use. However, decoding of speech imagery from EEG is a hard problem due to many factors. In this paper we focus on the analysis of the classification step of speech imagery decoding for a three-class vowel speech imagery recognition problem. We empirically show that different classification subtasks may require different classifiers for accurately decoding and obtain a classification accuracy that improves the best results previously published. We further investigate the relationship between the classifiers and different sets of features selected by the common spatial patterns method. Our results indicate that further improvement on BCIs based on speech imagery could be achieved by carefully selecting an appropriate combination of classifiers for the subtasks involved.

Avaliação da suscetibilidade antimicrobiana e do perfil de macrorrestrição do DNA de isolados humanos de Salmonella serovar Typhimurium fagotipo 193, no período de 1970 a 2008

Para analisar cepas de Salmonella ser. Typhimurium isoladas de processos entéricos e extraintestinais humanos ocorridos no período de 1970 a 2008 de diferentes regiões do país foram selecionadas, com base nos registros contidos no banco de dados do Laboratório de Enterobactérias do IOC/FIOCRUZ, RJ, amostras do fagotipo prevalente 193, visando precipuamente o reconhecimento de clones epidêmicos. Foram selecionadas 553 cepas de Salmonella ser. Typhimurium fagotipo 193 representadas por 91, 65, 70 e 327 amostras referentes as décadas de 70, 80, 90 e ao período de 2000 a 2008, respectivamente. Na análise global da sensibilidade destas cepas, 52% apresentaram um ou mais marcadores de resistência a antibióticos incluídos no perfil ACSSuT. Este perfil de resistência completo foi verificado em 20,9% dos isolados, sendo os 21,9% restantes, sensíveis a todas as drogas testadas, especialmente no período de 2000 a 2008, representadas por 121 amostras (37,0%) em relação as 327 culturas dessa época. O maior percentual de resistência foi observado nas amostras da década de 70 (99%) sendo o perfil ACSSuT detectado em 35,2% dos isolados, ressaltando-se que todas as amostras foram isoladas de processos gastroentéricos ocorridos na cidade de São Paulo. Ao longo das quatro décadas de estudo, descreve-se um ponto de ruptura entre a prevalência de resistência e a suscetibilidade na transição entre as décadas de 80 e 90. Embora o número de isolados de Salmonella ser. Typhimurium fagotipo 193 tenha aumentado no último período considerado, o percentual de mono e multirresistência aos antimicrobianos se situou em nível elevado (63,0%). A análise do polimorfismo obtido após macrorrestrição com a enzima XbaI revelou que cepas isoladas na década de 90 apresentaram elevado percentual de similaridade (≥85%) com cepas isoladas recentemente (período de 2000-2008), sendo agrupadas nos mesmos subclusters. Por outro lado, as cepas da década de 70 inserem-se em subclusters independentes, embora o percentual de similaridade entre tais subclusters e os demais seja ≥70%; o mesmo sendo observado para as cepas isoladas durante a década de 80. Em conclusão, este estudo mostrou que a prevalência de isolados humanos de Salmonella ser. Typhimurium fagotipo 193 no Brasil vem progredindo desde a década de 1990, enquanto a detecção do modelo R (ACSSuT) está diminuindo e a avaliação através da PFGE indicou a presença de multiplicidade de perfis de macrorrestrição no fagotipo 193, entretanto com elevados percentuais de similaridade entre si, sugerindo alguma clonalidade, tendo em vista o período entre o isolamento e a análise

Feature Selection for Speech Emotion Recognition in Spanish and Basque: On the Use of Machine Learning to Improve Human-Computer Interaction

Study of emotions in human-computer interaction is a growing research area. This paper shows an attempt to select the most significant features for emotion recognition in spoken Basque and Spanish Languages using different methods for feature selection. RekEmozio database was used as the experimental data set. Several Machine Learning paradigms were used for the emotion classification task. Experiments were executed in three phases, using different sets of features as classification variables in each phase. Moreover, feature subset selection was applied at each phase in order to seek for the most relevant feature subset. The three phases approach was selected to check the validity of the proposed approach. Achieved results show that an instance-based learning algorithm using feature subset selection techniques based on evolutionary algorithms is the best Machine Learning paradigm in automatic emotion recognition, with all different feature sets, obtaining a mean of 80,05% emotion recognition rate in Basque and a 74,82% in Spanish. In order to check the goodness of the proposed process, a greedy searching approach (FSS-Forward) has been applied and a comparison between them is provided. Based on achieved results, a set of most relevant non-speaker dependent features is proposed for both languages and new perspectives are suggested.

Validation of a method for estimating realized annual fecundity in a multiple spawner, the long-snouted seahorse (Hippocampus guttulatus), using underwater visual census

The long-snouted seahorse (Hippocampus guttulatus) (Cuvier, 1829), was used to validate the pre-dictive accuracy of three progressively realistic models for estimating the realized annual fecundity of asyn-chronous, indeterminate, multiple spawners. Underwater surveys and catch data were used to estimate the duration of the reproductive season, female spawning frequency, male brooding frequency, and batch fecun-dity. The most realistic model, a generalization of the spawning fraction method, produced unbiased estimates of male brooding frequency (mean ±standard deviation [SD]=4.2 ±1.6 broods/year). Mean batch fecundity and realized annual fecundity were 213.9 (±110.9) and 903.6 (±522.4), respectively. However, females prepared significantly more clutches than the number of broods produced by males. Thus, methods that infer spawning frequency from patterns in female egg production may lead to significant overestimates of realized annual fecundity. The spawning fraction method is broadly applicable to many taxa that exhibit parental care and can be applied nondestructively to species for which conservation is a concern.