EEG-Based Brain-Computer Interface for Decoding Motor Imagery Tasks within the Same Hand Using Choi-Williams Time-Frequency Distribution

Abstract

This paper presents an EEG-based brain-computer interface system for classifying eleven motor imagery (MI) tasks within the same hand. The proposed system utilizes the Choi-Williams time-frequency distribution (CWD) to construct a time-frequency representation (TFR) of the EEG signals. The constructed TFR is used to extract five categories of time-frequency features (TFFs). The TFFs are processed using a hierarchical classification model to identify the MI task encapsulated within the EEG signals. To evaluate the performance of the proposed approach, EEG data were recorded for eighteen intact subjects and four amputated subjects while imagining to perform each of the eleven hand MI tasks. Two performance evaluation analyses, namely channel- and TFF-based analyses, are conducted to identify the best subset of EEG channels and the TFFs category, respectively, that enable the highest classification accuracy between the MI tasks. In each evaluation analysis, the hierarchical classification model is trained using two training procedures, namely subject-dependent and subject-independent procedures. These two training procedures quantify the capability of the proposed approach to capture both intra- and inter-personal variations in the EEG signals for different MI tasks within the same hand. The results demonstrate the efficacy of the approach for classifying the MI tasks within the same hand. In particular, the classification accuracies obtained for the intact and amputated subjects are as high as 88.8% and 90.2%, respectively, for the subject-dependent training procedure, and 80.8% and 87.8%, respectively, for the subject-independent training procedure. These results suggest the feasibility of applying the proposed approach to control dexterous prosthetic hands, which can be of great benefit for individuals suffering from hand amputations.

Tonic Cold Pain Detection Using Choi–Williams Time-Frequency Distribution Analysis of EEG Signals: A Feasibility Study

“…The EEG modality, which records the electrical potentials produced at different locations in the brain in response to an exogenous pain stimulus, is considered the most commonly used modality for designing objective pain detection systems. This can be attributed to several properties that are associated with the EEG modality, including the high temporal resolution, low-cost, and noninvasive nature [5,6]. Nonetheless, pain detection based on EEG signals analysis is considered challenging due to the nonstationarity nature of the EEG signals, low spatial resolution, and low signal-to-noise (SNR) ratio [7].…”

Functional Connectivity and Classification of Actual and Imaginary Motor Movement

“…Classification results of rough sets method indicated that universal parameters couldn't be decided upon, and each subject had to be considered differently for parameter selection and classification. [1] For subject dependent training procedures, accuracy obtained in [3] is 88.8% for intact participants and 90.2% for amputated participants, and for subject independent training the accuracy was 80.8% and 87.8% respectively. Detecting actual motor activity of both hands with 80% accuracy and imaginary motor activity with just 34% accuracy in [7].…”

“…while EEG signals were recorded in 64 channel BCI2000 systems. [1] Eighteen intact and four amputated participants volunteered to perform and/or imagine to perform eleven motor imagery tasks with a single hand in [3] like finger and wrist movements, grasping tasks, etc. Eleven subjects participated for three hand and wrist movements MI experimenthand opening/ closing, wrist flexion/ extension and forearm pronation/ supination in [4] while EEG data was acquired using G.Nautilus headset with 16 electrodes positioned as per 10/20 system [5].…”

“…[1]. A sliding window approach is used in [3] to decompose each EEG signal into overlapping segments, convert to time-frequency domain and extract five categories of time frequency features-logamplitude, amplitude, statistical, spectral, and spectralentropy based category. Common spatial pattern algorithm was used for feature extraction in [4] from the whole band and then from the five filtered bands.…”

“…It is done by first including elements that may be present in the set, and then elements that are surely present in the set. [1] A hierarchical classification model of four layers is used in [3] that classify each of the overlapping EEG segments into one of the eleven motor imagery tasks considered. First layer classifies rest or movement segments, second layer classifies between functional and basic movements, third layers further classifies several types of wrist and finger related grasps and movements and finally the fourth layer classifies the eleven motor imagery tasks.…”