Cross-correlation is often used as the primary technique to compare two biological signals. The cross-correlation technique is an effective means to measure the synchronization of two signals if the relative phases at all frequencies are distributed linearly, that is, there is a group delay. The group delay assumption of cross-correlation analysis imposes an unfavourable restriction on signals with relative phase correlation which varies at different frequencies. Traditional Fourier analysis applied to a short data segments, namely the Short Time Fourier Transform (STFT), provides phase information for each frequency component, but it is not suitable for biological signals with non-stationary statistics for which the ideal segment length is unknown. The application of a wavelet based phase analysis technique is discussed in this study. The frequency decomposition and temporally localized nature of the wavelet transform provides localized phase-frequency information for two signals. A wavelet frequency temporal relative phase pattern (WFT-RPP) technique to extract relative phase information at specific frequencies over the time course of a time-varying signal was developed. The technique was tested on simulated data and surface electromyographic (sEMG) data recorded from upper limb muscles in human subjects as they performed a series of dynamic push and pull tasks. Selected sEMG channel pairs are compared against each other using the WFT-RPP technique to extract the relative phase information and repetitive relative phase patterns for certain muscle pairs were observed. The properties of the WFT-RPP and the merits and weaknesses of using the technique for determining intermuscular sEMG synchronization is discussed. / Thesis (Master, Electrical & Computer Engineering) -- Queen's University, 2011-09-13 11:58:16.014

Phagocytosis is the process by which cells engulf foreign bodies. It is the hallmark behavior of white blood cells, being the process through which those cells ingest and degrade pathogens and debris. To date a large amount of research has focused on documenting the existence and role of biochemical components involved with phagocytosis. Scores of signaling molecules have been implicated in the complex signal cascade which drives the process. These molecules are small (typically no larger than 5 nanometers) and operate in a crowded, chemically “noisy,” environment, yet they coordinate the cell's activity over comparatively expansive distances (as large as 20 micrometers). How these molecular processes scale-up to coordinate the activities of the cell over such massive distances is largely unknown. Using a planar analog of phagocytosis termed “frustrated phagocytosis,” we experimentally demonstrate that phagocytosis occurs in three distinct phases: initial cell-antigen binding, symmetric spreading, and late-stage contraction. Initial binding and symmetric spreading appears to be both mechanically and chemically similar to the quasi-universal cellular behaviors of adhesion and migration. Adhesion and migration have received extensive attention from the biophysics community in recent years. Leveraging these similarities, we adapt the biomechanical frameworks used in models of migration to phagocytosis. We show that macroscopic properties such as a cell's effective viscosity and membrane cortical tension can be used to model cell behavior during phagocytosis. Our experiments reveal that late-stage contraction distinguishes frustrated phagocytosis from other spreading behaviors. This contraction is myosin dependent. Additionally we demonstrate, for the first time, that late-stage contraction corresponds with formation of a contractile F-actin belt. Based on the dynamic contraction model (DC) developed to explain actin structure during cell migration we propose a DC model of phagocytosis which posits that contractile belt formation is the result of a late-stage myosin activity coupled with F-actin.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / The muscle fatigue can be caused by multiple factors, and the most common one is bodywork. As a result, the muscle stress signal becomes part of atlets life. However, this phenom may show injuries incident, neuromuscular diseases, and it is related to the general human being health, as well as with its nutrition. To determine the fatigue level from a muscle or from a person is not that simple, because multiple subjective factors are envolved, including psychological and hormonal matters, thus maybe is not possible to determine an universal method for quantification of muscle fatigue. The electromyographic signal (EMG) is well known and studied for reflecting the musculature condition from which it was generated. The electromyography is an important tool for the health muscle assessment, and counts on various studies and advances in its formation and interpretation understanding.Thus, it is expected that the muscle fatigue that affects the natural muscle behavior, affects also the EMG signal. This work aims to understand how the fatigue action appears in the signal, through the study of different EMG signal characteristics. From literature, several studies analyzed isometric contractions, thus it was decided to make a dynamic contractions evaluation, which are more natural in the daily life. For the sake of simplicity, the biceps braquii was chosen. This muscle was estimulated by a scott biceps curl exercise, an exercise known to well isolate the working muscle, so that the weight lifting is almost all done by the biceps action. Pilot trial was done, collecting EMG signals from both biceps braquii, and also measuring the force applied to the bar. For the EMG signal analysis, three software packages were developed. One of them was a programm for the electromyographer control, and for the signals record- ing in text files without header. For this development were used C Sharp and .NET. One library for signals processing was developed using Matlab, including fil- ter functions, muscle activity detection and features extraction, such as amplitude, frequency, entropy, and stationarity. Finally, was developed a programm for feature analysis that uses the previous mentioned library, and that also applies the Kohonen algorithm of self-organizing maps.This programm was also developed using Matlab. All created programms are open source, and they are available for download on GitHub platform. A temporal analysis of the features was performed in order to cluster the results of the features extracted from the signals of 21 volunteers. This analysis showed that signal s amplitude increases as the fatigue occurs while there is a spectral shift for the left. This shift indicates that the main frequencies have decreased. The trends observed for amplitude and frequency are the same reported in the literature. The results also show decreasing in the entropy as effect of the fatigue progres- sion. Two stationarity features indicate decreasing in the stationarity, these were influenced by the amplitude raise, though. A third stationarity feature, which is not dependent on amplitude, show that there is not significant modification on the stationarity. The data clustering attempt using the Kohonen algorithm was frustrated, gener- ating inconclusive results. It can be concluded that the features related to amplitude, frequency and entropy are somehow related to the muscular fatigue. So that it is possible, during future work, the development of a fatigue classifier based on these features. / A fadiga muscular pode ser causada por diversos fatores, e o mais comum deles e o exercício físico. Isso faz com que esse sinal de estresse muscular faça parte da vida de atletas. No entanto, esse fenômeno pode indicar a ocorrência de lesões, doenças neuro-musculares e está ligado à saúde geral do indivíduo, bem como com a alimentação. Determinar o nível de fadiga de um músculo ou de um indivíduo em geral não é simples, pois vários aspectos subjetivos estão envolvidos, incluindo questões psicológicas e hormonais, e talvez não seja possível a determinação de um método universal de quanticação da fadiga muscular. O sinal eletromiográfico (EMG) é conhecido e estudado por refletir o estado da musculatura que o gerou. A eletromiografia é uma ferramenta importante para a avaliação da saúde muscular e conta com diversos estudos e avanços tanto no entendimento de sua formação quanto na sua interpretação. Assim, de antemão, espera-se que a fadiga muscular, que afeta o comportamento natural dos músculos, afete também o sinal eletromiográfico. Nesse trabalho, procurou-se entender, por meio do estudo de diferentes características do sinal EMG, como a ação da fadiga se manifesta no sinal. Na literatura, vários estudos analisam as contrações isométricas, assim decidiu-se por fazer uma avaliação de contrações dinâmicas, as quais são mais naturais no cotidiano. Por uma questão de simplicidade, o músculo escolhido foi o bíceps braquial. Esse músculo foi estimulado por um exercício de rosca em banco scott, um exercício conhecido por isolar bem o músculo trabalhado, de forma que o levantamento do peso é quase todo feito por ação do bíceps. Coletas piloto foram realizadas, nas quais o sinal EMG dos dois bíceps foi registrado em conjunto com a medida de força aplicada na barra. Para a análise dos sinais EMG, três pacotes de software foram desenvolvidos. Um deles foi um programa para controle do eletromiógrafo e registro dos sinais em arquivos texto com cabecalho. Para esse desenvolvimento, utilizou-se C Sharp e .NET. Uma biblioteca para processamento de sinais biológicos foi desenvolvida em Matlab, na qual encontram-se funções de filtragem, detecção de atividade muscular e extração de características tais como amplitude, frequência, entropia e estacionaridade. Por fim, desenvolveu-se um programa para análise de características que usa a biblioteca mencionada e também aplica o algortimo de mapas auto-organizáveis de Kohonnen. Esse programa também foi desenvolvido em Matlab. Todos os programas criados sâo de código aberto e estão disponíveis para download na plataforma GitHub. Uma analise temporal das características foi realizada de forma a agrupar os resultados das características extraídas dos sinais dos 21 voluntários. Essa análise mostrou que a amplitude do sinal aumentou com o avanço da fadiga muscular enquanto a frequência dos sinais se deslocou para esquerda no espectro. Isso indica que as frequências principais diminuiram. Essas tendências para amplitude e frequência são as mesmas registradas na literatura. O estudo mostrou ainda que a entropia diminui com a progressão da fadiga. Duas características de estacionaridade indicaram diminuição, no entanto foram influenciadas pela amplitude. Uma terceira característica, indepentende da amplitude, mostrou que não há alteração signicativa na estacionaridade. A tentativa de agrupamento dos dados com o algortimo de Kohonnen foi frustrada, ja que gerou resultados inconclusivos. Concluiu-se que as características de amplitude, frequência e entropia estão relacionadas com a fadiga muscular. Assim acredita-se ser possível desenvolver, em estudos futuros, um classificador de sinais EMG que faca inferência do nível de fadiga baseado nessas características. / Mestre em Ciências