Armando Malanda

Independent component analysis as a tool to eliminate artifacts in EEG: A quantitative study

Summary Independent component analysis (ICA) is a novel technique that calculates independent components from mixed signals. A hypothetical clinical application is to remove artifacts in EEG. The goal of this study was to apply ICA to standard EEG recordings to eliminate well-known artifacts, thus quantifying its efficacy in an objective way. Eighty samples of recordings with spikes and evident artifacts of electrocardiogram (EKG), eye movements, 50-Hz interference, muscle, or electrode artifact were studied. ICA components were calculated using the Joint Approximate Diagonalization of Eigen-matrices (JADE) algorithm. The signal was reconstructed excluding those components related to the artifacts. A normalized correlation coefficient was used as a measure of the changes caused by the suppression of these components. ICA produced an evident clearing-up of signals in all the samples. The morphology and the topography of the spike were very similar before and after the removal of the artifacts. The correlation coefficient showed that the rest of the signal did not change significantly. Two examiners independently looked at the samples to identify the changes in the morphology and location of the discharge and the artifacts. In conclusion, ICA proved to be a useful tool to clean artifacts in short EEG samples, without having the disadvantages associated with the digital filters. The distortion of the interictal activity measured by correlation analysis was minimal.

Electromyographic models to assess muscle fatigue

Muscle fatigue is a common experience in daily life. Many authors have defined it as the incapacity to maintain the required or expected force, and therefore, force, power and torque recordings have been used as direct measurements of muscle fatigue. In addition, the measurement of these variables combined with the measurement of surface electromyography (sEMG) recordings (which can be measured during all types of movements) during exercise may be useful to assess and understand muscle fatigue. Therefore, there is a need to develop muscle fatigue models that relate changes in sEMG variables with muscle fatigue. However, the main issue when using conventional sEMG variables to quantify fatigue is their poor association with direct measures of fatigue. Therefore, using different techniques, several authors have combined sets of sEMG parameters to assess muscle fatigue. The aim of this paper is to serve as a state-of-the-art summary of different sEMG models used to assess muscle fatigue. This paper provides an overview of linear and non-linear sEMG models for estimating muscle fatigue, their ability to assess power loss and their limitations due to neuromuscular changes after a training period.

Beta electroencephalograph changes during passive movements: sensory afferences contribute to beta event-related desynchronization in humans

Non-phase-locked beta oscillatory changes during passive movements were studied in six healthy volunteers, and compared with those observed in a similar group during ballistic movements. Passive movements consisted of brisk wrist extensions done with the help of a pulley system. Changes in the beta band were determined by means of wavelet and Gabor transforms, and compared statistically with a pre-movement period. In this paradigm, a marked beta energy loss (event-related desynchronization, ERD) was present after the beginning of the movement, followed by a beta energy increase (event-related synchronization, ERS). The ERD/ERS was similar to that observed during ballistic movements, but without pre-movement components. Although both changes were maximal in the contralateral central electrode, the beta ERD showed a more bilateral topography. These findings suggest that afferent proprioceptive inputs may play a role in the final part of the beta ERD observed during voluntary movements.

EMG spectral indices and muscle power fatigue during dynamic contractions.

The purpose of this study was to examine acute exercise-induced changes on muscle power output and surface electromyography (sEMG) parameters (amplitude and spectral indices of muscle fatigue) during a dynamic fatiguing protocol. Fifteen trained subjects performed five sets consisting of 10 leg presses (10RM), with 2min rest between sets. Surface electromyography was recorded from vastus medialis (VM) and lateralis (VL) and biceps femoris (BF) muscles. A number of EMG-based parameters were compared for estimation accuracy and sensitivity to detect peripheral muscle fatigue. These were: Mean Average Voltage, median spectral frequency, Dimitrov spectral index of muscle fatigue (FI(nsm5)), as well as other parameters obtained from a time-frequency analysis (Choi-Williams distributions) such as mean and variance of the instantaneous frequency and frequency variance. The log FI(nsm5) as a single parameter predictor accounted for 37% of the performance variance of changes in muscle power and the log FI(nsm5) and MFM as a two factor combination predictor accounted for 44%. Peripheral impairments assessed by sEMG spectral index FI(nsm5) may be a relevant factor involved in the loss of power output after dynamic high-loading fatiguing task.

Gamma band activity in an auditory oddball paradigm studied with the wavelet transform.

OBJECTIVES To examine the characteristics of evoked and induced gamma band oscillatory responses occurring during P300 development in an auditory oddball paradigm. METHODS A time-frequency analysis method was applied to an auditory oddball paradigm in 7 healthy subjects. This method combines a multiresolution wavelet algorithm for signal extraction and the Gabor transform to represent the temporal evolution of the selected frequency components. Phase-locked or evoked activity and also non-phase-locked activity were computed for both standard and target stimuli. RESULTS The gamma band frequency components differed between target and non-target stimuli processing. The study showed an early and mainly phase-locked oscillatory response appearing around 26--28 ms after both standard and target stimuli onset. This response showed a spectral peak around 44 Hz for both stimuli. A late oscillatory activity peaking at 37 Hz with a latency around 360 ms was observed appearing only for target stimuli. The latency of this late oscillatory activity had a high correlation (P=0.002) to the latency of the P300 wave. CONCLUSIONS EEG signal analysis with wavelet transform allows the identification of an early oscillatory cortical response in the gamma frequency range, as well as a late P300-related response.

Neuromuscular fatigue after resistance training.

This study examined the effects of heavy resistance training on dynamic exercise-induced fatigue task (5 x 10RM leg-press) after two loading protocols with the same relative intensity (%) (5 x 10RM(Rel)) and the same absolute load (kg) (5 x 10RM(Abs)) as in pretraining in men (n=12). Maximal strength and muscle power, surface EMG changes [amplitude and spectral indices of muscle fatigue], and metabolic responses (i.e.blood lactate and ammonia concentrations) were measured before and after exercise. After training, when the relative intensity of the fatiguing dynamic protocol was kept the same, the magnitude of exercise-induced loss in maximal strength was greater than that observed before training. The peak power lost after 5 x 10RM(Rel) (58-62%, pre-post training) was greater than the corresponding exercise-induced decline observed in isometric strength (12-17%). Similar neural adjustments, but higher accumulated fatigue and metabolic demand were observed after 5 x 10RM(Rel). This study therefore supports the notion that similar changes are observable in the EMG signal pre- and post-training at fatigue when exercising with the same relative load. However, after training the muscle is relatively able to work more and accumulate more metabolites before task failure. This result may indicate that rate of fatigue development (i.e. power and MVC) was faster and more profound after training despite using the same relative intensity.

Alpha and beta oscillatory activity during a sequence of two movements.

OBJECTIVE We studied movement-related electroencephalographic oscillatory changes in the alpha and beta range during a sequence of two movements in 7 healthy volunteers, in order to investigate the relationship between these changes and each component in the sequence. METHODS The sequence consisted of a wrist active extension-passive flexion followed by a first and second finger pincer. A total of 10.5 s sweeps were recorded using the level of surface electromyographic (EMG) activity in wrist extensors as trigger, including a 7.5 s pre-stimulus. The sweeps were also realigned manually offline using as trigger the end of the first EMG burst, or the beginning of the second movement. An index of the changes in non-phase-locked energy in the 7-37 Hz range was obtained by averaging single-sweep time-frequency transforms. RESULTS The duration of each of the movements in the sequence and the relationship between them were compatible with the use of two different motor programmes in the sequence. In the beta band, a decrease in energy (event-related desynchronisation, ERD) began 1.5 s before the onset of the first movement, and was sustained until the end of the second movement. No energy increases were observed until the end of the second movement. In the alpha band, the ERD began 0.5 seconds before the first movement and was sustained throughout the recording. CONCLUSION These findings suggest that the beta-event-related synchronisation is related to the end of the whole motor process, and not to the end of each motor programme.

Effects of Strength Training on Muscle Fatigue Mapping from Surface EMG and Blood Metabolites

PURPOSE this study examined the effects of heavy resistance training on the relationships between power loss and surface EMG (sEMG) indices and blood metabolite concentrations on dynamic exercise-induced fatigue with the same relative load as in pretraining. METHODS twelve trained subjects performed five sets consisting of 10 repetitions in the leg press, with 2 min of rest between sets before and after a strength training period. sEMG variables (the mean average voltage, the median spectral frequency, and the Dimitrov spectral index of muscle fatigue) from vastus medialis and lateralis muscles and metabolic responses (i.e., blood lactate, uric acid, and ammonia concentrations) were measured. RESULTS the peak power loss after the posttraining protocol was greater (61%) than the decline observed in the pretraining protocol (46%). Similar sEMG changes were found for both protocols, whereas higher metabolic demand was observed during the posttraining exercise. The linear models on the basis of the relations found between power loss and changes in sEMG variables were significantly different between pretraining and posttraining, whereas the linear models on the basis of the relations between power loss and changes in blood metabolite concentrations were similar. CONCLUSIONS linear models that use blood metabolites to map acute exercise-induced peak power changes were more accurate in detecting these changes before and after a short-term training period, whereas an attempt to track peak power loss using sEMG variables may fail after a strength training period.

sEMG wavelet-based indices predicts muscle power loss during dynamic contractions.

The purpose of this study was to investigate the sensitivity of new surface electromyography (sEMG) indices based on the discrete wavelet transform to estimate acute exercise-induced changes on muscle power output during a dynamic fatiguing protocol. Fifteen trained subjects performed five sets consisting of 10 leg press, with 2 min rest between sets. sEMG was recorded from vastus medialis (VM) muscle. Several surface electromyographic parameters were computed. These were: mean rectified voltage (MRV), median spectral frequency (F(med)), Dimitrov spectral index of muscle fatigue (FI(nsm5)), as well as five other parameters obtained from the stationary wavelet transform (SWT) as ratios between different scales. The new wavelet indices showed better accuracy to map changes in muscle power output during the fatiguing protocol. Moreover, the new wavelet indices as a single parameter predictor accounted for 46.6% of the performance variance of changes in muscle power and the log-FI(nsm5) and MRV as a two-factor combination predictor accounted for 49.8%. On the other hand, the new wavelet indices proposed, showed the highest robustness in presence of additive white Gaussian noise for different signal to noise ratios (SNRs). The sEMG wavelet indices proposed may be a useful tool to map changes in muscle power output during dynamic high-loading fatiguing task.

Motor unit action potential duration, I: variability of manual and automatic measurements.

The aim of this work is to analyze the variability in manual measurements of motor unit action potential (MUAP) duration and to evaluate the effectiveness of well-known algorithms for automatic measurement. Two electromyographists carried out three independent duration measurements of a set of 240 MUAPs. The intraexaminer and interexaminer variabilities were analyzed by means of the Gage Reproducibility and Repeatability method. The mean of the three closest manually marked positions was considered the gold standard of the duration markers positions (GSP). The results of four well-known automatic methods for estimating MUAP duration were compared to the GSP. Manual measurements of duration showed a lot of variability, with the combined intraoperator and interoperator variability greater than 30%. The greatest difference between manual positions was 11.2 ms. The mean differences between the GSP and those obtained with the four automatic methods ranged between 0.6 and 8.5 ms. Both manual and automatic measurements of MUAP duration show a high degree of variability. More precise methods are needed to improve the accuracy and reliability of the estimates of this parameter.