Search

Recently, the advent of the non-invasive brain-computer interface (BCI) for continuous decoding of upper limb motions opens a new horizon for motor-disabled people. However, the performance of discrete-decoding BCIs based on discriminating different brain states are still more robust. In this study, we aimed to cascade a discrete state decoder with a continuous decoder to enhance the prediction of hand trajectories. EEG data were recorded from nine healthy subjects performing a center-out task with four orthogonal targets on the horizontal plane. The pre-movement data of each trial has been used for training a binary discrete decoder which identifies the axis of the movement based on common spatial pattern (CSP) features. Two non-parametric continuous decoders based on Gaussian process regression (GPR) have been designed for continuous decoding of hand movements along each axis using the envelope features of EEG signals in six frequency bands. In addition to those four principal orthogonal targets, some targets at random directions on the horizontal plane were recorded to evaluate the generalizability of the proposed model. The discrete decoder attained the average binary classification of 97.1% for discriminating movement along the x-axis and y-axis. The proposed state-based method achieved the mean correlation coefficient of 0.54 between actual and predicted trajectories for principal targets over all subjects. The trajectories of random targets were also decoded with a mean correlation of 0.37. The generalizability of the proposed paradigm proved by the findings of this study could open new possibilities in developing novel types of neuroprostheses for rehabilitation purposes.

In this paper, a novel fully-automated state-based decoding method has been proposed for continuous decoding problems in brain-machine interface (BMI) systems, called Gaussian mixture of model (GMM)-assisted PLS (GMMPLS). In contrast to other state-based and hierarchical decoders, the proposed method does not demand any prior information about the desired output structure. Instead, GMMPLS uses the GMM algorithm to divide the desired output into a specific number of states (clusters). Next, a logistic regression model is trained to predict the probability membership of each time sample for each state. Finally, using the concept of the partial least square (PLS) algorithm, GMMPLS constructs a model for decoding the desired output using the input data and the achieved membership probabilities. The performance of the GMMPLS has been evaluated and compared to PLS, the nonlinear quadratic PLS (QPLS), and the bayesian PLS (BPLS) methods through a simulated dataset and two different real-world BMI datasets. The achieved results demonstrated that the GMMPLS significantly outperformed PLS, QPLS, and BPLS overall datasets.