Md. Rabiul Islam

Artifact suppression from EEG signals using data adaptive time domain filtering

This paper presents a data adaptive filtering approach to separate the electrooculograph (EOG) artifact from the recorded electroencephalograph (EEG) signal. Empirical mode decomposition (EMD) technique is used to implement the time domain filter. Fractional Gaussian noise (fGn) is used here as the reference signal to detect the distinguish feature of EOG signal to be used to separate from EEG. EMD is applied to the raw EEG and fGn separately to produce a finite number band limited signals named intrinsic mode functions (IMFs). The energies of individual IMFs of fGn and that of raw EEG are compared to derive the energy based threshold for the suppression of EOG effects. The separation results using EMD based approach is also compared with wavelet thresholding technique. The experimental results show that the data adaptive filtering technique performs better than the wavelet based approach.

Multiband Prediction Model for Financial Time Series with Multivariate Empirical Mode Decomposition

This paper presents a subband approach to financial time series prediction. Multivariate empirical mode decomposition (MEMD) is employed here for multiband representation of multichannel financial time series together. Autoregressive moving average (ARMA) model is used in prediction of individual subband of any time series data. Then all the predicted subband signals are summed up to obtain the overall prediction. The ARMA model works better for stationary signal. With multiband representation, each subband becomes a band-limited (narrow band) signal and hence better prediction is achieved. The performance of the proposed MEMD-ARMA model is compared with classical EMD, discrete wavelet transform (DWT), and with full band ARMA model in terms of signal-to-noise ratio (SNR) and mean square error (MSE) between the original and predicted time series. The simulation results show that the MEMD-ARMA-based method performs better than the other methods.

Data adaptive analysis of ECG signals for cardiovascular disease diagnosis

This paper presents a data adaptive technique of cardiovascular disease diagnosis by analyzing electrocardiogram (ECG) signals. The separation of high-frequency QRS and low frequency signal are performed by employing empirical mode decomposition (EMD). Biomedical signals like heart wave commonly change their statistical properties over time, tending to be nonstationary for which EMD is a powerful tool of decomposition. EMD is used to decompose ECG signal into a finite set of band-limited signals termed as intrinsic mode functions (IMFs). Then the low and high frequency components of ECG signals are obtained partial reconstruction intrinsic mode functions and the residual. The related signal processing tools are applied to extract high and low frequency parts to diagnosis the cardiovascular diseases.

Frequency recognition for SSVEP-based BCI with data adaptive reference signals

Steady-state visual evoked potential (SSVEP) is an effective electrophysiological source to implement a brain-computer interface (BCI). In this paper, a novel frequency recognition method is introduced using two levels of reference signals derived from the training set of real world SSVEP signals with canonical correlation analysis (CCA). The first level reference signals are obtained by averaging the training trials of respective stimulus frequency. Standard CCA with thus obtained reference signals is applied to the training trails to measure the dominance of the stimulus frequency component. Several training trials containing more prominent target (stimulus) frequency component are selected as the second level reference signals. Both the obtained reference signals are used with CCA to derive an effective spatial filter for frequency recognition. The experimental results show that the proposed approach significantly improves the recognition accuracy of SSVEP as well as the information transfer rate (ITR) compared to the state-of-the-art recognition methods.

Artifact suppression and analysis of brain activities with electroencephalography signals

Brain-computer interface is a communication system that connects the brain with computer (or other devices) but is not dependent on the normal output of the brain (i.e., peripheral nerve and muscle). Electro-oculogram is a dominant artifact which has a significant negative influence on further analysis of real electroencephalography data. This paper presented a data adaptive technique for artifact suppression and brain wave extraction from electroencephalography signals to detect regional brain activities. Empirical mode decomposition based adaptive thresholding approach was employed here to suppress the electro-oculogram artifact. Fractional Gaussian noise was used to determine the threshold level derived from the analysis data without any training. The purified electroencephalography signal was composed of the brain waves also called rhythmic components which represent the brain activities. The rhythmic components were extracted from each electroencephalography channel using adaptive wiener filter with the original scale. The regional brain activities were mapped on the basis of the spatial distribution of rhythmic components, and the results showed that different regions of the brain are activated in response to different stimuli. This research analyzed the activities of a single rhythmic component, alpha with respect to different motor imaginations. The experimental results showed that the proposed method is very efficient in artifact suppression and identifying individual motor imagery based on the activities of alpha component.

Subband entropy-based features for clothing invariant human gait recognition

Abstract This paper presents a wavelet-based feature extraction method for human gait recognition. The selection of features with most discriminative information is the key to improve recognition performance. The frequency domain representation of the gait image is obtained by using fast Fourier transforms. Next, a discrete wavelet transform is applied to the obtained spectrum. With single-level wavelet decomposition, four coefficients are generated. The sum of the entropy of these four wavelet coefficients is computed yielding the wavelet Entropy Image (wEnI) which is used here as the potential feature for human gait recognition. A template matching-based approach is used as the classification. The performance of the proposed wEnI feature is evaluated using whole-based and part-based methods. The experimental results show that the wEnI feature performs better compared to state-of-the-art gait features in common use.

Frequency recognition of steady-state visually evoked potentials using binary subband canonical correlation analysis with reduced dimension of reference signals

This paper presents a frequency recognition method of steady-state visual evoked potentials (SSVEPs) using binary subbands with canonical correlation analysis (CCA). The first subband contains all the target frequencies of SSVEPs. The second one includes the SSVEP signal corresponding to a desired number of higher order stimulus frequencies, which is obtained by filtering out of required range of lower order stimuli. The full dimension of artificial reference signals are used for first subband, whereas a reduced dimension of references is employed for second subband to compute canonical correlation. The weighted sum of the obtained correlation values are used to recognize the frequency of an SSVEP. The experimental results show the superiority of the proposed method compared to the state-of-the-art recognition methods.

EEG signal enhancement using multivariate wavelet transform Application to single-trial classification of event-related potentials

Empirical mode decomposition (EMD) has been successfully used in artifact suppression form the recorded electroencephalography (EEG) signals using a data-adaptive subband filtering approach. The higher computation burden of EMD processing is the main obstacle in online implementation of brain-computer interfacing (BCI). To resolve such limitation, multivariate wavelet transform with higher computation speed is introduced in this paper to decompose multichannel EEG signals into a finite set of subbands. The energy based subband filtering is implemented to separate the higher frequency noise components to clean the noisy event-related potential (ERP) signals. An auditory oddball BCI experiment is conducted to test cleaning performance followed by the BCI classification of single trial ERP using linear discriminant analysis (LDA). The experimental results illustrate that the classification performance is increased noticeably with the cleaned single-trial ERP data using proposed algorithm. It requires lower computational cost compared to EMD based cleaning approach.

Frequency recognition for SSVEP-BCI using reference signals with dominant stimulus frequency

Detection of frequency for steady-state visual evoked potentials (SSVEP) is addressed. We propose to use the combination of CCA and training data-based template matching between two level of data adaptive reference signals that can deal with the dominant frequency. On the basis of magnitude of stimulus frequency components, the dominant channels are selected. The recognition accuracy as well as the information transfer rate (ITR) of the proposed method are examined compared to the state-of-the-art recognition method.

Artifact suppression from electroencephalography signals using stationary subspace analysis

Different types of artifacts contaminate the electroencephalography (EEG) signals in brain computer interface (BCI) application. Electrocardiography (ECG) is such potential artifact which negatively affects the BCI performance. This paper presents a novel method for ECG artifact elimination from EEG using stationary subspace analysis (SSA). It is based on the consideration that the ECG components are relatively non-stationary than that of the EEG signals. Applying SSA, the total channels of raw EEG are partitioned into two groups — stationary and non-stationary. The non-stationary channels contain the ECG artifacts. A statistical test is used to measure the degree of non-stationarity. The channel with highest non-stationarity is selected as the source of ECG artifact. The normalized ECG source is used to segregate the target artifact from the measured EEG. The result of the proposed method is compared with that of the well-known statistical method independent component analysis (ICA). The experimental evaluation illustrates that the proposed method is superior to the ICA based approach in terms of ECG artifact suppression from raw EEG.