Limits...
A Generalizable Brain-Computer Interface (BCI) Using Machine Learning for Feature Discovery.

Nurse ES, Karoly PJ, Grayden DB, Freestone DR - PLoS ONE (2015)

Bottom Line: The classifier does not use extensive a-priori information, resulting in reduced reliance on highly specific domain knowledge.Instead of pre-defining features, the time-domain signal is input to a population of multi-layer perceptrons (MLPs) in order to perform a stochastic search for the best structure.Our new approach has been shown to give accurate results across different motor tasks and signal types as well as between subjects.

View Article: PubMed Central - PubMed

Affiliation: NeuroEngineering Laboratory, Department of Electrical & Electronic Engineering, The University of Melbourne, Melbourne, VIC, Australia, 3010; Centre for Neural Engineering, The University of Melbourne, Melbourne, VIC, Australia, 3010.

ABSTRACT
This work describes a generalized method for classifying motor-related neural signals for a brain-computer interface (BCI), based on a stochastic machine learning method. The method differs from the various feature extraction and selection techniques employed in many other BCI systems. The classifier does not use extensive a-priori information, resulting in reduced reliance on highly specific domain knowledge. Instead of pre-defining features, the time-domain signal is input to a population of multi-layer perceptrons (MLPs) in order to perform a stochastic search for the best structure. The results showed that the average performance of the new algorithm outperformed other published methods using the Berlin BCI IV (2008) competition dataset and was comparable to the best results in the Berlin BCI II (2002-3) competition dataset. The new method was also applied to electroencephalography (EEG) data recorded from five subjects undertaking a hand squeeze task and demonstrated high levels of accuracy with a mean classification accuracy of 78.9% after five-fold cross-validation. Our new approach has been shown to give accurate results across different motor tasks and signal types as well as between subjects.

No MeSH data available.


Related in: MedlinePlus

Comparison of accuracy between hand-squeeze detection and left vs. right detection of three-class hand squeeze classifier.The first bar for each participant is the accuracy for correct classification of a hand squeeze (left or right), given that a hand squeeze had occurred. The second bar for each participant is the accuracy for detection of a hand squeeze regardless of laterality. The bars show the mean accuracy across the resulting confusion matrices after five-fold cross validation for each participant. The error bars are the standard deviation over the five confusion matrices. The asterisks represent the confidence level to reject the  hypothesis that the classification accuracy once a hand squeeze is detected is significantly different from the accuracy of detecting if a hand squeeze has occurred using a two-tailed t-test (** for p < 0.01). White dashed lines indicate minor gridlines. For the exact p-values see S2 Appendix. For the full confusion matrices, see S1 Appendix.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4482677&req=5

pone.0131328.g007: Comparison of accuracy between hand-squeeze detection and left vs. right detection of three-class hand squeeze classifier.The first bar for each participant is the accuracy for correct classification of a hand squeeze (left or right), given that a hand squeeze had occurred. The second bar for each participant is the accuracy for detection of a hand squeeze regardless of laterality. The bars show the mean accuracy across the resulting confusion matrices after five-fold cross validation for each participant. The error bars are the standard deviation over the five confusion matrices. The asterisks represent the confidence level to reject the hypothesis that the classification accuracy once a hand squeeze is detected is significantly different from the accuracy of detecting if a hand squeeze has occurred using a two-tailed t-test (** for p < 0.01). White dashed lines indicate minor gridlines. For the exact p-values see S2 Appendix. For the full confusion matrices, see S1 Appendix.

Mentions: Fig 7 shows that, for all participants, the three-class classifier was significantly better at distinguishing between left and right hand squeezes (once a hand squeeze had been correctly identified) than it was at detecting whether a hand-squeeze had occurred or not. A two-tailed t-test was used to determine the confidence level to reject the hypothesis that the classification accuracies were not significantly different (using the five-fold cross validation results). It should be noted that although both the two- and three-class classifiers used the same datasets, there are well known difficulties in comparing the results of classifiers with different numbers of classes, such as uneven probability distributions for different classes [18, 22].


A Generalizable Brain-Computer Interface (BCI) Using Machine Learning for Feature Discovery.

Nurse ES, Karoly PJ, Grayden DB, Freestone DR - PLoS ONE (2015)

Comparison of accuracy between hand-squeeze detection and left vs. right detection of three-class hand squeeze classifier.The first bar for each participant is the accuracy for correct classification of a hand squeeze (left or right), given that a hand squeeze had occurred. The second bar for each participant is the accuracy for detection of a hand squeeze regardless of laterality. The bars show the mean accuracy across the resulting confusion matrices after five-fold cross validation for each participant. The error bars are the standard deviation over the five confusion matrices. The asterisks represent the confidence level to reject the  hypothesis that the classification accuracy once a hand squeeze is detected is significantly different from the accuracy of detecting if a hand squeeze has occurred using a two-tailed t-test (** for p < 0.01). White dashed lines indicate minor gridlines. For the exact p-values see S2 Appendix. For the full confusion matrices, see S1 Appendix.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4482677&req=5

pone.0131328.g007: Comparison of accuracy between hand-squeeze detection and left vs. right detection of three-class hand squeeze classifier.The first bar for each participant is the accuracy for correct classification of a hand squeeze (left or right), given that a hand squeeze had occurred. The second bar for each participant is the accuracy for detection of a hand squeeze regardless of laterality. The bars show the mean accuracy across the resulting confusion matrices after five-fold cross validation for each participant. The error bars are the standard deviation over the five confusion matrices. The asterisks represent the confidence level to reject the hypothesis that the classification accuracy once a hand squeeze is detected is significantly different from the accuracy of detecting if a hand squeeze has occurred using a two-tailed t-test (** for p < 0.01). White dashed lines indicate minor gridlines. For the exact p-values see S2 Appendix. For the full confusion matrices, see S1 Appendix.
Mentions: Fig 7 shows that, for all participants, the three-class classifier was significantly better at distinguishing between left and right hand squeezes (once a hand squeeze had been correctly identified) than it was at detecting whether a hand-squeeze had occurred or not. A two-tailed t-test was used to determine the confidence level to reject the hypothesis that the classification accuracies were not significantly different (using the five-fold cross validation results). It should be noted that although both the two- and three-class classifiers used the same datasets, there are well known difficulties in comparing the results of classifiers with different numbers of classes, such as uneven probability distributions for different classes [18, 22].

Bottom Line: The classifier does not use extensive a-priori information, resulting in reduced reliance on highly specific domain knowledge.Instead of pre-defining features, the time-domain signal is input to a population of multi-layer perceptrons (MLPs) in order to perform a stochastic search for the best structure.Our new approach has been shown to give accurate results across different motor tasks and signal types as well as between subjects.

View Article: PubMed Central - PubMed

Affiliation: NeuroEngineering Laboratory, Department of Electrical & Electronic Engineering, The University of Melbourne, Melbourne, VIC, Australia, 3010; Centre for Neural Engineering, The University of Melbourne, Melbourne, VIC, Australia, 3010.

ABSTRACT
This work describes a generalized method for classifying motor-related neural signals for a brain-computer interface (BCI), based on a stochastic machine learning method. The method differs from the various feature extraction and selection techniques employed in many other BCI systems. The classifier does not use extensive a-priori information, resulting in reduced reliance on highly specific domain knowledge. Instead of pre-defining features, the time-domain signal is input to a population of multi-layer perceptrons (MLPs) in order to perform a stochastic search for the best structure. The results showed that the average performance of the new algorithm outperformed other published methods using the Berlin BCI IV (2008) competition dataset and was comparable to the best results in the Berlin BCI II (2002-3) competition dataset. The new method was also applied to electroencephalography (EEG) data recorded from five subjects undertaking a hand squeeze task and demonstrated high levels of accuracy with a mean classification accuracy of 78.9% after five-fold cross-validation. Our new approach has been shown to give accurate results across different motor tasks and signal types as well as between subjects.

No MeSH data available.


Related in: MedlinePlus