Keynote Speaker

wwwd Nazmul Siddique
Intelligent Systems Research Centre, Ulster University, Northland Road, Londonderry, BT48 7JL, U.K
Title: Brain-Computer Interfacing: Motion Trajectory Prediction from EEG Signal
E-mail: [email protected]

Abstract:
Brain-computer interfacing (BCI) is a popular technology in recent times due to a number of successful applications ranging from orthosis control in the spinal injured, wheelchair control, stroke rehabilitation and gaming applications. A large proportion of BCI applications are aimed at enabling people to gain control of objects in three-dimensional (3D) spaces with motor imagery (MI), where imagined movement enables voluntary modulation of the neural activity in the sensorimotor cortex that can be decoded for control purposes. A motion trajectory prediction (MTP) based BCI involves the reconstruction of the 3D trajectory of upper limb movement using EEG signals. The most common MTP BCI uses potential time-series (PTS model) of band-pass filtered EEG potentials for reconstructing the trajectory of 3D limb movement with a multiple linear regression (mLR). Hitherto, most of the MTP BCI studies report the best accuracy using low delta (0.5-2Hz) band-pass filtered EEG potentials. The traditional sensorimotor rhythm (SMR) based multiclass (MC) BCI uses power values of mu (8-12Hz) and beta (12-30Hz) bands for limb movement classification, e.g., left vs. right hand movement classification as opposed to hand trajectory prediction.

In this research, we show that the band-pass filtered EEG potentials are represented properly by the input time-series of the PTS model if the filter is applied to the 0.5-2Hz (low-delta) band. Based on this evidence we show the PTS model has limited access to the information content which is coded in mu (8-12Hz) and beta (12-30Hz) rhythms. We present an alternative MTP approach (i.e., the bandpower time-series model: BTS model), where time-varying power values estimated over a window within a specified EEG band are the model features. A comprehensive analysis comprising of subjects performing pointing movements with the dominant right arm towards six targets is presented. The results show that the BTS model produces significantly higher MTP accuracy [R~0.4] compared to the standard PTS model [R~0.2]. Additionally, in the case of the BTS model, the highest accuracy was achieved in the theta (4-8Hz), mu (8-12Hz), and low beta (12-18Hz) bands. The finding that mu and beta band are prominent is contrary to other MTP studies but is consistent with the extensive literature on classical MC based SMR BCIs.

Reconstruction of 3D trajectory of imagined limb movements using EEG is very challenging. The solution of this problem can lead to better non-invasive BCIs for the physically impaired people. In this research, we further show the reconstruction of actual and imagined 3D limb movement trajectory with an MTP BCI using a time-series of bandpower values (BTS model). The BTS model provided best results in the mu and beta bands (R~0.5 for actual and R~0.2 for imagined movement reconstruction). Our study shows that mu (8-12Hz) and beta (12-30Hz) activity can be used for decoding imagined 3D hand movement from EEG.

Short Biography:
Nazmul Siddique obtained Dipl.-Ing degree from Dresden University of Technology, Germany in Cybernetics and Automation. He obtained M. Sc. in Computer Science from Bangladesh University of Engineering and Technology (BUET). He received his PhD in intelligent control from the Department of Automatic Control and Systems Engineering, University of Sheffield, England, UK in 2003.
Dr Siddique has published over 150 research papers including four books (published by John Wiley, Springer and Taylor & Francis). His research interests are in the broad area of cybernetics, brain-computer interfacing, artificial emotion, computational intelligence, intelligent control, and robotics. He is on the editorial board of six international journals (Int. J. of Neural Systems, J. of Behavioural Robotics, Int. J. of Machine Learning and Cybernetics, Int. J. of Applied Pattern Recognition, Int. J. of Advances in Robotics Research, and Engineering Letters). He guest edited eight special issues of reputed journals. He has been involved in organizing many national and international conferences and co-edited seven conference proceedings. He is a senior member of the IEEE, and a member of executive committee of IEEE SMC Chapter. He has been with Ulster University since 2001. He was with Khulna University from 1991 to 1998. He was a visiting professor at Prefecture University of Hiroshima in 2012.