ORIGINAL ARTICLE Delta band contribution in cue based single trial classification of real and imaginary wrist movements Aleksandra Vuckovic Æ Francisco Sepulveda Received: 17 May 2007 / Accepted: 29 March 2008 / Published online: 17 April 2008 Ó International Federation for Medical and Biological Engineering 2008 Abstract The aim of this study was to classify different movements about the right wrist. Four different movements were performed: extension, flexion, pronation and supina- tion. Two-class single trial classification was performed on six possible combinations of two movements (extension– flexion, extension–supination, extension–pronation, flex- ion–supination, flexion–pronation, pronation–supination). Both real and imaginary movements were analysed. The analysis was done in the joint time–frequency domain using the Gabor transform. Feature selection was based on the Davis-Bouldin Index (DBI) and feature classification was based on Elman’s recurrent neural networks (ENN). The best classification results, near 80% true positive rate, for imag- inary movements were achieved for discrimination between extension and any other type of movement. The experiments were run with 10 able-bodied subjects. For some subjects, real movement classification rates higher than 80% were achieved for any combination of movements, though not simultaneously for all six combinations of movements. For classification of the imaginary movements, the results sug- gest that the type of movement and frequency band play an important role. Unexpectedly, the delta band was found to carry significant class-related information. Keywords Brain computer interface (BCI)  EEG  Motor tasks  Movement imagery  Right hand 1 Introduction Brain computer interface (BCI) systems based on EEG recordings can be utilized for a broad range of applications, from a long term use to assist communication in locked-in patients, driving a wheelchair and giving a command to a neural prosthesis, to short term applications for therapeutic treatment in rehabilitation and in different brain disorders [24]. The main advantages of EEG based BCI systems are that they are non-invasive, easy to use and relatively inexpensive. The main drawback of such BCI over more invasive recording is the low signal to noise ratio and low spatial resolution. Motor imagination is often used for BCIs as it can uti- lize the cortical somatotopic representation of different parts of the body. However, this kind of BCI has typically been limited to four classes corresponding to four different parts of the body, such as left and right hand, both legs and tongue [23, 30]. The main difficulty in detecting different movements of the same limb, especially different move- ments about the same joint, is that they all activate nearly the same area of the motor cortex. Therefore, there have been only a few attempts to detect different real move- ments of a single limb from EEG recordings [6, 7, 19, 28, 33]. In the present study, a method for two class single trial classifications of both imaginary and real wrist movements were proposed. Different combinations of movements were classified in order to find which combination of movements would give the best classification results. Classification of different movements of the same wrist should increase the number of separable classes available for a BCI system (especially if used in combination with movements of different limbs) and should also enable more intuitive prosthesis control. A. Vuckovic (&) Centre for Rehabilitation Engineering, University of Glasgow, Glasgow, UK e-mail: avuckovi@eng.gla.ac.uk F. Sepulveda BCI Group, Computing and Electronic Systems Department, University of Essex, Colchester, UK 123 Med Biol Eng Comput (2008) 46:529–539 DOI 10.1007/s11517-008-0345-8