Limits...
Translation of EEG spatial filters from resting to motor imagery using independent component analysis.

Wang Y, Wang YT, Jung TP - PLoS ONE (2012)

Bottom Line: To our knowledge, a state-to-state translation, which applies spatial filters derived from one state to another, has never been reported.This study proposes a state-to-state, zero-training method to construct spatial filters for extracting EEG changes induced by motor imagery.Independent component analysis (ICA) was separately applied to the multi-channel EEG in the resting and the motor imagery states to obtain motor-related spatial filters.

View Article: PubMed Central - PubMed

Affiliation: Swartz Center for Computational Neuroscience, Institute for Neural Computation, University of California San Diego, San Diego, California, United States of America. yijun@sccn.ucsd.edu

ABSTRACT
Electroencephalogram (EEG)-based brain-computer interfaces (BCIs) often use spatial filters to improve signal-to-noise ratio of task-related EEG activities. To obtain robust spatial filters, large amounts of labeled data, which are often expensive and labor-intensive to obtain, need to be collected in a training procedure before online BCI control. Several studies have recently developed zero-training methods using a session-to-session scenario in order to alleviate this problem. To our knowledge, a state-to-state translation, which applies spatial filters derived from one state to another, has never been reported. This study proposes a state-to-state, zero-training method to construct spatial filters for extracting EEG changes induced by motor imagery. Independent component analysis (ICA) was separately applied to the multi-channel EEG in the resting and the motor imagery states to obtain motor-related spatial filters. The resultant spatial filters were then applied to single-trial EEG to differentiate left- and right-hand imagery movements. On a motor imagery dataset collected from nine subjects, comparable classification accuracies were obtained by using ICA-based spatial filters derived from the two states (motor imagery: 87.0%, resting: 85.9%), which were both significantly higher than the accuracy achieved by using monopolar scalp EEG data (80.4%). The proposed method considerably increases the practicality of BCI systems in real-world environments because it is less sensitive to electrode misalignment across different sessions or days and does not require annotated pilot data to derive spatial filters.

Show MeSH

Related in: MedlinePlus

Spatial patterns and spatial filters of the motor components for all nine subjects.(A) spatial patterns of the resting state; (B) spatial patterns of the motor imagery state; (C) spatial filters of the resting state; (D) spatial filters of the motor imagery state. Black dots in each scalp map indicate positions of C3 and C4 electrodes. In each subfigure, the left and right motor ICs for all subjects were grouped on the left and the right panel respectively.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3362620&req=5

pone-0037665-g005: Spatial patterns and spatial filters of the motor components for all nine subjects.(A) spatial patterns of the resting state; (B) spatial patterns of the motor imagery state; (C) spatial filters of the resting state; (D) spatial filters of the motor imagery state. Black dots in each scalp map indicate positions of C3 and C4 electrodes. In each subfigure, the left and right motor ICs for all subjects were grouped on the left and the right panel respectively.

Mentions: To quantitatively investigate to what extent one can translate the motor-related spatial filters derived from resting to motor-imagery BCI practice, this study first compares spatial patterns and spatial filters of motor components in resting and motor-imagery experiments. Figure 5 shows spatial patterns and spatial filters of the motor components in the resting state and the motor imagery state for all subjects. All the components show a typical dipolar-like topography, which is widespread over the sensorimotor cortex on left or right hemisphere of the brain, and shows the highest amplitudes at C3 and C4 electrodes. These findings are consistent with previous motor-related EEG studies [3]. Generally, the motor-related spatial filters show both positive and negative weights around the sensorimotor area, functioning through eliminating the motor irrelevant background activities while keeping the motor related activities. To quantitatively evaluate the topographical similarity, this study calculated the correlations of spatial patterns and spatial filters of the motor components between the two states for each subject. For simplicity, the correlations were obtained by directly computing correlation coefficients of the vectors (shown in Table 2). As can be seen, spatial patterns (i.e. projections of the components to the scalp) between the resting and the motor imagery states were very comparable (mean correlation coefficients of 0.95±0.05 and 0.94±0.06 for left and right ICs) for all subjects. The spatial filter, the unmixing vector, was more variable. For example, spatial patterns are highly correlated for Subject 5 with correlation coefficients of 0.92 and 0.96 for the left and right motor IC respectively, however the correlation of spatial filters is very weak (0.09 and 0.34 for left and right ICs). Although the spatial filters might be different, their effectiveness for extracting the motor-related EEG components should be as effective, judging from the similarity of the corresponding spatial patterns. Therefore, in practice, the selection of motor-related components was based on the spatial patterns instead of spatial filters.


Translation of EEG spatial filters from resting to motor imagery using independent component analysis.

Wang Y, Wang YT, Jung TP - PLoS ONE (2012)

Spatial patterns and spatial filters of the motor components for all nine subjects.(A) spatial patterns of the resting state; (B) spatial patterns of the motor imagery state; (C) spatial filters of the resting state; (D) spatial filters of the motor imagery state. Black dots in each scalp map indicate positions of C3 and C4 electrodes. In each subfigure, the left and right motor ICs for all subjects were grouped on the left and the right panel respectively.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3362620&req=5

pone-0037665-g005: Spatial patterns and spatial filters of the motor components for all nine subjects.(A) spatial patterns of the resting state; (B) spatial patterns of the motor imagery state; (C) spatial filters of the resting state; (D) spatial filters of the motor imagery state. Black dots in each scalp map indicate positions of C3 and C4 electrodes. In each subfigure, the left and right motor ICs for all subjects were grouped on the left and the right panel respectively.
Mentions: To quantitatively investigate to what extent one can translate the motor-related spatial filters derived from resting to motor-imagery BCI practice, this study first compares spatial patterns and spatial filters of motor components in resting and motor-imagery experiments. Figure 5 shows spatial patterns and spatial filters of the motor components in the resting state and the motor imagery state for all subjects. All the components show a typical dipolar-like topography, which is widespread over the sensorimotor cortex on left or right hemisphere of the brain, and shows the highest amplitudes at C3 and C4 electrodes. These findings are consistent with previous motor-related EEG studies [3]. Generally, the motor-related spatial filters show both positive and negative weights around the sensorimotor area, functioning through eliminating the motor irrelevant background activities while keeping the motor related activities. To quantitatively evaluate the topographical similarity, this study calculated the correlations of spatial patterns and spatial filters of the motor components between the two states for each subject. For simplicity, the correlations were obtained by directly computing correlation coefficients of the vectors (shown in Table 2). As can be seen, spatial patterns (i.e. projections of the components to the scalp) between the resting and the motor imagery states were very comparable (mean correlation coefficients of 0.95±0.05 and 0.94±0.06 for left and right ICs) for all subjects. The spatial filter, the unmixing vector, was more variable. For example, spatial patterns are highly correlated for Subject 5 with correlation coefficients of 0.92 and 0.96 for the left and right motor IC respectively, however the correlation of spatial filters is very weak (0.09 and 0.34 for left and right ICs). Although the spatial filters might be different, their effectiveness for extracting the motor-related EEG components should be as effective, judging from the similarity of the corresponding spatial patterns. Therefore, in practice, the selection of motor-related components was based on the spatial patterns instead of spatial filters.

Bottom Line: To our knowledge, a state-to-state translation, which applies spatial filters derived from one state to another, has never been reported.This study proposes a state-to-state, zero-training method to construct spatial filters for extracting EEG changes induced by motor imagery.Independent component analysis (ICA) was separately applied to the multi-channel EEG in the resting and the motor imagery states to obtain motor-related spatial filters.

View Article: PubMed Central - PubMed

Affiliation: Swartz Center for Computational Neuroscience, Institute for Neural Computation, University of California San Diego, San Diego, California, United States of America. yijun@sccn.ucsd.edu

ABSTRACT
Electroencephalogram (EEG)-based brain-computer interfaces (BCIs) often use spatial filters to improve signal-to-noise ratio of task-related EEG activities. To obtain robust spatial filters, large amounts of labeled data, which are often expensive and labor-intensive to obtain, need to be collected in a training procedure before online BCI control. Several studies have recently developed zero-training methods using a session-to-session scenario in order to alleviate this problem. To our knowledge, a state-to-state translation, which applies spatial filters derived from one state to another, has never been reported. This study proposes a state-to-state, zero-training method to construct spatial filters for extracting EEG changes induced by motor imagery. Independent component analysis (ICA) was separately applied to the multi-channel EEG in the resting and the motor imagery states to obtain motor-related spatial filters. The resultant spatial filters were then applied to single-trial EEG to differentiate left- and right-hand imagery movements. On a motor imagery dataset collected from nine subjects, comparable classification accuracies were obtained by using ICA-based spatial filters derived from the two states (motor imagery: 87.0%, resting: 85.9%), which were both significantly higher than the accuracy achieved by using monopolar scalp EEG data (80.4%). The proposed method considerably increases the practicality of BCI systems in real-world environments because it is less sensitive to electrode misalignment across different sessions or days and does not require annotated pilot data to derive spatial filters.

Show MeSH
Related in: MedlinePlus