The aim of this study was to determine motor unit firing patterns during ischemic versus non-ischemic sustained submaximal isometric contractions of the tibialis anterior muscle. 10 healthy adults attended two experimental sessions approximately 48 hours apart. Both sessions were identical except that the fatigue task in one was performed with a pressure cuff placed above the knee and inflated to 180 mm Hg. Three 5s maximum voluntary contractions (MVCs) were performed prior to and after the fatigue task. Each participant held a target force of 20% MVC until endurance time (peak-to-peak tremor amplitude exceeded 5% MVC). Single motor unit firing rates (11 non ischemic, 9 ischemic) were recorded with intramuscular fine wire electrodes. Mean interspike intervals over 5s time bins were calculated at every 5% endurance time. The endurance time for the ischemic (3.7 ± 0.58 min) fatigue task was significantly (p<0.001) shorter than the non-ischemic (9.5 ± 0.57 min) task. There was no significant difference in mean motor unit firing rates between the two conditions (p=0.883). Within both tests, there was a significant decline in firing rate (ischemic initial: 12.95 ± 0.71 Hz, minimum: 11.41 ± 0.81 Hz, p=0.023; non-ischemic initial: 13.13 ± 0.87 Hz, minimum: 11.15 ± 0.48 Hz, p=0.012). The time to minimum firing rate was significantly (p<0.001) less in the ischemic (1.29 ± 0.2 min) compared to non-ischemic (3.14 ± 0.23 min) condition. Muscle ischemia significantly reduced endurance time and the time to minimum firing rate. However, there were no differences in average motor unit firing rates between the two conditions across the relative phases of endurance time. / text

Developmental nicotine exposure (DNE) is known to cause deleterious effects in neonatal mammals through nicotine's actions on nicotinic acetylcholine receptors (nAChRs). In this work, we studied how DNE altered the structure and function of the hypoglossal motoneurons (XIIMNs) over the first few days post-parturition. Previous work in XIIMNs demonstrated an increase in cellular excitability (Pilarski et al., 2011), alterations in synaptic transmission among respiratory-related neurons (Wang et al., 2006; Pilarski et al., 2012; Jaiswal et al., 2013), and a reduction in inspiratory drive currents in DNE animals (Pilarski et al., 2011). Here we show that the effects of DNE extend to alterations in the spike-timing precision and reliability of XIIMNs, as well as spike-frequency adaptation. Additionally, simple morphological analysis of XIIMNs following nicotine exposure in utero has revealed a reduction in soma cross-sectional area. We were interested in studying the complete morphology of XIIMNs following DNE to discern its effects on more complex morphological parameters. We advanced this research using a combination of techniques in thin brainstem slices of neonatal rats, including whole cell patch clamp recordings and immunohistochemistry of intracellularly labeled hypoglossal motoneurons. Furthermore, morphological analysis revealed significant differences in the complexity of the dendritic arborization, showing that neurons from DNE animals had shorter dendrites that branched less often. We also used computational analysis to gain insight into mechanisms that may underlie the changes in spike-timing precision and reliability. In a single cell model of XIIMNs, decreases in potassium-dependent conductances such as the calcium-activated potassium current could potentially replicate the alterations seen in vitro. Finally, we also did a systems-level study of the hyoglossus muscle, a tongue retractor, to determine the relation between tongue retraction force and motor unit discharge characteristics. These experiments utilized adult, anesthetized rats to record single motor units, whole muscle electromyography (EMG) activity and tongue retraction force during spontaneous breathing. We determined that during inspiration-related tongue retractions in low and high force conditions, recruitment of motor units plays a crucial role in the control of tongue force output, whereas rate coding of single motor units is present, but appears to play a lesser role. Overall, this study shows that DNE effects the input-output properties of XIIMNs, potentially through changes in intrinsic channel properties; DNE also alters XIIMN morphology, particularly dendritic arborization; and that organization of a tongue retractor muscle depends primarily on recruitment, but also rate coding, to increase force output.

The dexterity of the human hand depends largely on the ability to move the fingers independently, the execution of which requires the coordination of multiple muscles. How these muscle ensembles are recruited by the central nervous system is not clear. Therefore, the objective of this dissertation was to identify some of the neural mechanisms whereby certain hand muscles are recruited into functional groups, or muscle synergies, needed for the generation of specific hand and finger movements.We characterized the organization of synaptic inputs onto the motor neurons supplying different compartments of a multi-tendoned finger flexor, the flexor digitorum superficialis (FDS). We found that the motor neurons controlling different finger compartments of the FDS do not receive entirely segregated inputs, and that the motor neurons supplying adjacent compartments receive substantially more common synaptic input than motor neurons supplying compartments further apart. The FDS and another multi-tendoned finger flexor, the flexor digitorum profundus (FDP), both insert onto each finger and function together to flex the fingers. Surprisingly, we found that the motor neurons controlling the compartments of FDS and FDP to the same finger receive completely independent inputs, despite similar mechanical functions of the two muscles. Thus, there is more neural coupling between motor neurons supplying compartments of the same muscle that move different fingers than there is between motor neurons supplying the compartments of two different muscles that move the same finger.Although the motor neurons supplying the flexors of the tips of the thumb [flexor pollicis longus (FPL)] and index finger [index compartment of the flexor digitorum profundus (FDP2)] receive substantial shared synaptic input during a precision grip task, the removal of the normal tactile feedback from the digit pads did not change the amount of common input to the two motor neuron pools, indicating these last-order divergent neurons do not require tactile afferent inputs for activation. Finally, in contrast to the substantial shared input to motor neurons supplying these two extrinsic muscles (FPL and FDP2), the motor neurons supplying two intrinsic muscles of the thumb [adductor pollicis (AdP)] and index finger [first dorsal interosseous (FDI)] were shown to receive few shared inputs during precision grip.

No description available.

Electromyographic (EMG) signal decomposition is the process of resolving an EMG signal into its component motor unit potential trains (MUPTs). The extracted MUPTs can aid in the diagnosis of neuromuscular disorders and the study of the neural control of movement, but only if they are valid trains. Before using decomposition results and the motor unit potential (MUP) shape and motor unit (MU) firing pattern information related to each active MU for either clinical or research purposes the fact that the extracted MUPTs are valid needs to be confirmed. The existing MUPT validation methods are either time consuming or related to operator experience and skill. More importantly, they cannot be executed during automatic decomposition of EMG signals to assist with improving decomposition results. To overcome these issues, in this thesis the possibility of developing automatic MUPT validation algorithms has been explored. Several methods based on a combination of feature extraction techniques, cluster validation methods, supervised classification algorithms, and multiple classifier fusion techniques were developed. The developed methods, in general, use either the MU firing pattern or MUP-shape consistency of a MUPT, or both, to estimate its overall validity. The performance of the developed systems was evaluated using a variety of MUPTs obtained from the decomposition of several simulated and real intramuscular EMG signals. Based on the results achieved, the methods that use only shape or only firing pattern information had higher generalization error than the systems that use both types of information. For the classifiers that use MU firing pattern information of a MUPT to determine its validity, the accuracy for invalid trains decreases as the number of missed-classification errors in trains increases. Likewise, for the methods that use MUP-shape information of a MUPT to determine its validity, the classification accuracy for invalid trains decreases as the within-train similarity of the invalid trains increase. Of the systems that use both shape and firing pattern information, those that separately estimate MU firing pattern validity and MUP-shape validity and then estimate the overall validity of a train by fusing these two indices using trainable fusion methods performed better than the single classifier scheme that estimates MUPT validity using a single classifier, especially for the real data used. Overall, the multi-classifier constructed using trainable logistic regression to aggregate base classifier outputs had the best performance with overall accuracy of 99.4% and 98.8% for simulated and real data, respectively. The possibility of formulating an algorithm for automated editing MUPTs contaminated with a high number of false-classification errors (FCEs) during decomposition was also investigated. Ultimately, a robust method was developed for this purpose. Using a supervised classifier and MU firing pattern information provided by each MUPT, the developed algorithm first determines whether a given train is contaminated by a high number of FCEs and needs to be edited. For contaminated MUPTs, the method uses both MU firing pattern and MUP shape information to detect MUPs that were erroneously assigned to the train. Evaluation based on simulated and real MU firing patterns, shows that contaminated MUPTs could be detected with 84% and 81% accuracy for simulated and real data, respectively. For a given contaminated MUPT, the algorithm on average correctly classified around 92.1% of the MUPs of the MUPT. The effectiveness of using the developed MUPT validation systems and the MUPT editing methods during EMG signal decomposition was investigated by integrating these algorithms into a certainty-based EMG signal decomposition algorithm. Overall, the decomposition accuracy for 32 simulated and 30 real EMG signals was improved by 7.5% (from 86.7% to 94.2%) and 3.4% (from 95.7% to 99.1%), respectively. A significant improvement was also achieved in correctly estimating the number of MUPTs represented in a set of detected MUPs. The simulated and real EMG signals used were comprised of 3–11 and 3–15 MUPTs, respectively.

Beginning with the introduction of electrically powered prostheses more than 65 years ago surface electromyographic (sEMG) signals recorded from residual muscles in amputated limbs have served as the primary source of upper-limb myoelectric prosthetic control. The majority of these devices use one or more neural interfaces to translate the sEMG signal amplitude into voltage control signals that drive the mechanical components of a prosthesis. In so doing, users are able to directly control the speed and direction of prosthetic actuation by varying the level of muscle activation and the associated sEMG signal amplitude. Consequently, in spite of decades of development, myoelectric prostheses are prone to highly variable functional control, leading to a relatively high-incidence of prosthetic abandonment among 23-35% of upper-limb amputees. Efforts to improve prosthetic control in recent years have led to the development and commercialization of neural interfaces that employ pattern recognition of sEMG signals recorded from multiple locations on a residual limb to map different intended movements. But while these advanced algorithms have made strident gains, there still exists substantial need for further improvement to increase the reliability of pattern recognition control solutions amongst the variability of muscle co-activation intensities. In efforts to enrich the control signals that form the basis for myoelectric control, I have been developing advanced algorithms as part of a next generation neural interface research and development, referred to as Motor Unit Drive (MU Drive), that is able to non-invasively extract the firings of individual motor units (MUs) from sEMG signals in real-time and translate the firings into smooth biomechanically informed control signals. These measurements of motor unit firing rates and recruitment naturally provide high-levels of motor control information from the peripheral nervous system for intact limbs and therefore hold the greater promise for restoring function for amputees. The goal for my doctoral work was to develop advanced algorithms for the MU Drive neural interface system, that leverage MU features to provide intuitive control of multiple degrees-of-freedom. To achieve this goal, I targeted 3 research aims: 1) Derive real-time MU-based control signals from motor unit firings, 2) Evaluate feasibility of motor unit action potential (MUAP) based discrimination of muscle intent 3) Design and evaluate MUAP-based motion Classification of motions of the arm and hand.

<p> The present study was undertaken to investigate the mechanism underlying the observation that the maximal voluntary strength of the two legs acting together or bilaterally in isometric leg extension was less than the summed unilateral (sum of the left and right legs tested separately) strength. Observations were made on this phenomenon under both isometric and concentric contraction conditions by testing young adult males performing unilateral and bilateral leg press contractions on a modified isokinetic dynamometer. </p> <p> Electromyographical evidence indicated that there was a lesser activation of motor units in bilateral contraction as compared to unilateral, under isometric conditions and at a low and high concentric velocity. To determine whether a particular type of motor unit was being activated to a lesser extent in bilateral contractions, two physiological parameters of unilateral and bilateral contractions were compared: the strength-velocity relation and fatigability. This investigative method was based on the known physiological differences between the motor unit types; namely fast-twitch (FT), type two motor units have a faster twitch contraction time, greater force output at high velocities of shortening and lesser resistance to fatigue than the slow-twitch (ST), type one units. </p> <p> Results showed a greater relative decline in the strength of bilateral contractions as the velocity of contraction was increased through a range from 0°/s to 424°/s (0 to 7.40 radians/s). The bilateral to summed unilateral strength ratio (B/U-ratio) decreased from 0.91 under isometric conditions to 0.51 at the highest test velocity. Lesser fatigability was found in the bilateral condition in a 100 consecutive concentric contraction fatigue test. These results provided complementary evidence for the conclusion that FT motor units were active to a lesser degree in bilateral contractions. </p> / Thesis / Master of Science (MSc)

Este trabalho apresenta a modelagem matemática e a simulação de potenciais de ação de unidades motoras de músculos de vertebrados visando a posterior simulação do eletromiograma. Para conseguir isso, inicialmente se fez uma compilação de dados existentes para a distribuição das fibras musculares (FBs) nas unidades motoras (MUs) de vários músculos, e as modelagens matemáticas descritos na literatura para o potencial de ação de uma FB (SFAP) e de uma MU (MUAP). Com base nos dados fisiológicos, primeiro se localizou as FBs em um músculo, por meio de uma aproximação de que as FBs estão rodeadas de outras seis no músculo. Para conseguir isto se construiram hexágonos concêntricos por MU, e posteriormente se localizou as FBs nas MUs, cobrindo uma faixa entre 75 e 2000 FBs, o que corresponde a músculos distais de mamíferos. Depois se fez uma aproximação para a distribuição de 170000 FBs nas 272 MUs da cabeça medial do músculo gastrocnêmio (MG) do gato, conseguindo numa primeira simulação localizar cerca de 70% das FBs para cada MU. Com esta localização das FBs no músculo baseados nos dados da literatura se aproximaram os retardos axonais por uma distribuição gaussiana, com média de 2 ms (gato) ou 10 ms (homem) e com desvio padrão de menos de 0,5 ms, desprezando o atraso axonal nas ramificações axonais, que foi estimado no máximo 29 vezes menor. Para a geração do SFAP trabalhou-se com dois modelos, um analítico, o qual resulta em simulações numéricas demoradas, e, outro numérico baseado na convolução da corrente com uma função peso. Para o modelo numérico dobrou-se imaginariamente o comprimento das FBs, para levar em conta o erro computacional de fim de fibra. O modelo numérico resultou em um tempo de simulação 30 vezes menor que o analítico. Adicionalmente, para simular a captação externa (i.e. na pele), fez-se uma aproximação para a função que modela os eletrodos de superfície de secção circular localizados a uma distância maior que 1,79 mm das FBs, mostrando um espectro similar ao reportado na literatura. Finalmente, os MUAPs obtidos resultavam com formas de onda e espectros similares ao descrito na literatura. Além disto, em certos casos, obtiveram-se MUAPs com indentações, seja localizando as junções neuromusculares em bandas da ordem de 1 mm de espessura, seja quando o tempo de atraso axonal foi considerado junto com a velocidade de condução da FB em função da raiz quadrada do diâmetro da FB. Foram feitas simulações para os MG e bíceps braquial do homem. Neste último caso, foram obtidos MUAPs similares aos captados para pessoas saludáveis, e foi observada a freqüência de disparos dos potenciais de ação do motoneurônio no espectro do MUAP. Quanto às formas dos agrupamentos das FBs em uma MU, não se obtiveram diferenças significativas para as FBs posicionadas homogênea e aleatoriamente, exceto uma ligeira variação nas amplitudes. No entanto, ocurreu uma mudança na faixa espectral, quando as FBs estavam concentradas. / This work presents the mathematical model and simulation of motor unit action potentials of vertebrate muscles aiming at after simulation of the electromyogram. To obtain this, initially, it was made a compilation of several data about the distribution of muscle fibers (FBs) in motor units (MUs) of many muscles, and the mathematical models of the action potential of a single FB (SFAP) and MU (MUAP), reported in previous works. On the basis of this physiological data, first, the FB was located in a muscle, using an approximation in which the FBs are encircled with other six FBs in the muscle. To reach this, concentric hexagons were constructed to build the surface of the MU, and later the FBs were situated in the MU, covering a range between 75 and 2000 FBs, corresponding to mammals extremity muscles. Later, a new approximation were was madein order to distribute the 170000 FBs in the 272 MUs of the medial head of muscle medialis gastrocnemius (MG) of the cat, reaching, in a first simulation, the localization of almost 70% of the FBs at each MU. With the FBs lalready allocated in the muscle, and based in data of previous works, their axonal delay were approximated by a gaussian distribution, with mean of 2 ms (cat) or 10 ms (man) and standard deviation of less than 0,5 ms, discarding the axonal delay in the axonal branching, that were estimated to affectup to 29 times less. To SFAP generation, two models were used, the first analytical, resulting in delayed numerical simulations, and the other based on convolution of the second derivate of the current with a weight function, where the length of the FBs was imaginarily duplicated, in order to consider the end fiber effect. Using this, a simulation time 30 times lesser than the analytical one was obtained. Additionally, so as to simulate the external recording (i.e. in the skin), it was made an approximation to the function that models the circular shape surface electrodes located at distances greater than 1,79 mm of the FBs, showing a similar spectrum reported. Finally, the waves and spectrum of the simulated MUAPs resulted similar to the ones reported in the literature. Beyond this, in certain cases, MUAPs were simulated with some tuned, either locating the neuromuscular junctions with thickness bands of 1 mm, or, when the axonal delay and the FB muscle fiber conduction velocity were considered as a function of the square root fiber diameter. This was simulated for MUAPs of MG and biceps brachii muscles of human beings, in the last case it has reached the waveforms and tuned found in heath subjects, and it was visualized the mean frequency of firing rate at the spectrum. In order to know how much affects grouping for the FBs to waves a MU, they were not found significant differences with FBs located homogeneously and randomly, except a little variation in the amplitude of the MUAP. However, they presented a change in the spectral bandwidth when the FBs are more concentrated.

Este trabalho apresenta a modelagem matemática e a simulação de potenciais de ação de unidades motoras de músculos de vertebrados visando a posterior simulação do eletromiograma. Para conseguir isso, inicialmente se fez uma compilação de dados existentes para a distribuição das fibras musculares (FBs) nas unidades motoras (MUs) de vários músculos, e as modelagens matemáticas descritos na literatura para o potencial de ação de uma FB (SFAP) e de uma MU (MUAP). Com base nos dados fisiológicos, primeiro se localizou as FBs em um músculo, por meio de uma aproximação de que as FBs estão rodeadas de outras seis no músculo. Para conseguir isto se construiram hexágonos concêntricos por MU, e posteriormente se localizou as FBs nas MUs, cobrindo uma faixa entre 75 e 2000 FBs, o que corresponde a músculos distais de mamíferos. Depois se fez uma aproximação para a distribuição de 170000 FBs nas 272 MUs da cabeça medial do músculo gastrocnêmio (MG) do gato, conseguindo numa primeira simulação localizar cerca de 70% das FBs para cada MU. Com esta localização das FBs no músculo baseados nos dados da literatura se aproximaram os retardos axonais por uma distribuição gaussiana, com média de 2 ms (gato) ou 10 ms (homem) e com desvio padrão de menos de 0,5 ms, desprezando o atraso axonal nas ramificações axonais, que foi estimado no máximo 29 vezes menor. Para a geração do SFAP trabalhou-se com dois modelos, um analítico, o qual resulta em simulações numéricas demoradas, e, outro numérico baseado na convolução da corrente com uma função peso. Para o modelo numérico dobrou-se imaginariamente o comprimento das FBs, para levar em conta o erro computacional de fim de fibra. O modelo numérico resultou em um tempo de simulação 30 vezes menor que o analítico. Adicionalmente, para simular a captação externa (i.e. na pele), fez-se uma aproximação para a função que modela os eletrodos de superfície de secção circular localizados a uma distância maior que 1,79 mm das FBs, mostrando um espectro similar ao reportado na literatura. Finalmente, os MUAPs obtidos resultavam com formas de onda e espectros similares ao descrito na literatura. Além disto, em certos casos, obtiveram-se MUAPs com indentações, seja localizando as junções neuromusculares em bandas da ordem de 1 mm de espessura, seja quando o tempo de atraso axonal foi considerado junto com a velocidade de condução da FB em função da raiz quadrada do diâmetro da FB. Foram feitas simulações para os MG e bíceps braquial do homem. Neste último caso, foram obtidos MUAPs similares aos captados para pessoas saludáveis, e foi observada a freqüência de disparos dos potenciais de ação do motoneurônio no espectro do MUAP. Quanto às formas dos agrupamentos das FBs em uma MU, não se obtiveram diferenças significativas para as FBs posicionadas homogênea e aleatoriamente, exceto uma ligeira variação nas amplitudes. No entanto, ocurreu uma mudança na faixa espectral, quando as FBs estavam concentradas. / This work presents the mathematical model and simulation of motor unit action potentials of vertebrate muscles aiming at after simulation of the electromyogram. To obtain this, initially, it was made a compilation of several data about the distribution of muscle fibers (FBs) in motor units (MUs) of many muscles, and the mathematical models of the action potential of a single FB (SFAP) and MU (MUAP), reported in previous works. On the basis of this physiological data, first, the FB was located in a muscle, using an approximation in which the FBs are encircled with other six FBs in the muscle. To reach this, concentric hexagons were constructed to build the surface of the MU, and later the FBs were situated in the MU, covering a range between 75 and 2000 FBs, corresponding to mammals extremity muscles. Later, a new approximation were was madein order to distribute the 170000 FBs in the 272 MUs of the medial head of muscle medialis gastrocnemius (MG) of the cat, reaching, in a first simulation, the localization of almost 70% of the FBs at each MU. With the FBs lalready allocated in the muscle, and based in data of previous works, their axonal delay were approximated by a gaussian distribution, with mean of 2 ms (cat) or 10 ms (man) and standard deviation of less than 0,5 ms, discarding the axonal delay in the axonal branching, that were estimated to affectup to 29 times less. To SFAP generation, two models were used, the first analytical, resulting in delayed numerical simulations, and the other based on convolution of the second derivate of the current with a weight function, where the length of the FBs was imaginarily duplicated, in order to consider the end fiber effect. Using this, a simulation time 30 times lesser than the analytical one was obtained. Additionally, so as to simulate the external recording (i.e. in the skin), it was made an approximation to the function that models the circular shape surface electrodes located at distances greater than 1,79 mm of the FBs, showing a similar spectrum reported. Finally, the waves and spectrum of the simulated MUAPs resulted similar to the ones reported in the literature. Beyond this, in certain cases, MUAPs were simulated with some tuned, either locating the neuromuscular junctions with thickness bands of 1 mm, or, when the axonal delay and the FB muscle fiber conduction velocity were considered as a function of the square root fiber diameter. This was simulated for MUAPs of MG and biceps brachii muscles of human beings, in the last case it has reached the waveforms and tuned found in heath subjects, and it was visualized the mean frequency of firing rate at the spectrum. In order to know how much affects grouping for the FBs to waves a MU, they were not found significant differences with FBs located homogeneously and randomly, except a little variation in the amplitude of the MUAP. However, they presented a change in the spectral bandwidth when the FBs are more concentrated.

EMG signal decomposition is the process of resolving a composite EMG signal into its constituent motor unit potential trains (classes) and it can be configured as a classification problem. An EMG signal detected by the tip of an inserted needle electrode is the superposition of the individual electrical contributions of the different motor units that are active, during a muscle contraction, and background interference. <BR>This thesis addresses the process of EMG signal decomposition by developing an interactive classification system, which uses multiple classifier fusion techniques in order to achieve improved classification performance. The developed system combines heterogeneous sets of base classifier ensembles of different kinds and employs either a one level classifier fusion scheme or a hybrid classifier fusion approach. <BR>The hybrid classifier fusion approach is applied as a two-stage combination process that uses a new aggregator module which consists of two combiners: the first at the abstract level of classifier fusion and the other at the measurement level of classifier fusion such that it uses both combiners in a complementary manner. Both combiners may be either data independent or the first combiner data independent and the second data dependent. For the purpose of experimentation, we used as first combiner the majority voting scheme, while we used as the second combiner one of the fixed combination rules behaving as a data independent combiner or the fuzzy integral with the lambda-fuzzy measure as an implicit data dependent combiner. <BR>Once the set of motor unit potential trains are generated by the classifier fusion system, the firing pattern consistency statistics for each train are calculated to detect classification errors in an adaptive fashion. This firing pattern analysis allows the algorithm to modify the threshold of assertion required for assignment of a motor unit potential classification individually for each train based on an expectation of erroneous assignments. <BR>The classifier ensembles consist of a set of different versions of the Certainty classifier, a set of classifiers based on the nearest neighbour decision rule: the fuzzy <em>k</em>-NN and the adaptive fuzzy <em>k</em>-NN classifiers, and a set of classifiers that use a correlation measure as an estimation of the degree of similarity between a pattern and a class template: the matched template filter classifiers and its adaptive counterpart. The base classifiers, besides being of different kinds, utilize different types of features and their performances were investigated using both real and simulated EMG signals of different complexities. The feature sets extracted include time-domain data, first- and second-order discrete derivative data, and wavelet-domain data. <BR>Following the so-called <em>overproduce and choose</em> strategy to classifier ensemble combination, the developed system allows the construction of a large set of candidate base classifiers and then chooses, from the base classifiers pool, subsets of specified number of classifiers to form candidate classifier ensembles. The system then selects the classifier ensemble having the maximum degree of agreement by exploiting a diversity measure for designing classifier teams. The kappa statistic is used as the diversity measure to estimate the level of agreement between the base classifier outputs, i. e. , to measure the degree of decision similarity between the base classifiers. This mechanism of choosing the team's classifiers based on assessing the classifier agreement throughout all the trains and the unassigned category is applied during the one level classifier fusion scheme and the first combiner in the hybrid classifier fusion approach. For the second combiner in the hybrid classifier fusion approach, we choose team classifiers also based on kappa statistics but by assessing the classifiers agreement only across the unassigned category and choose those base classifiers having the minimum agreement. <BR>Performance of the developed classifier fusion system, in both of its variants, i. e. , the one level scheme and the hybrid approach was evaluated using synthetic simulated signals of known properties and real signals and then compared it with the performance of the constituent base classifiers. Across the EMG signal data sets used, the hybrid approach had better average classification performance overall, specially in terms of reducing the number of classification errors.