Multitask Learning for Brain-Computer Interfaces

Morteza Alamgir, Moritz Grosse–Wentrup, Yasemin Altun
Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, PMLR 9:17-24, 2010.

Abstract

Brain-computer interfaces (BCIs) are limited in their applicability in everyday settings by the current necessity to record subject-specific calibration data prior to actual use of the BCI for communication. In this paper, we utilize the framework of multitask learning to construct a BCI that can be used without any subject-specific calibration process. We discuss how this out-of-the-box BCI can be further improved in a computationally efficient manner as subject-specific data becomes available. The feasibility of the approach is demonstrated on two sets of experimental EEG data recorded during a standard two-class motor imagery paradigm from a total of 19 healthy subjects. Specifically, we show that satisfactory classification results can be achieved with zero training data, and combining prior recordings with subject-specific calibration data substantially outperforms using subject-specific data only. Our results further show that transfer between recordings under slightly different experimental setups is feasible.

Cite this Paper

BibTeX
@InProceedings{pmlr-v9-alamgir10a, title = {Multitask Learning for Brain-Computer Interfaces}, author = {Alamgir, Morteza and Grosse–Wentrup, Moritz and Altun, Yasemin}, booktitle = {Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics}, pages = {17--24}, year = {2010}, editor = {Teh, Yee Whye and Titterington, Mike}, volume = {9}, series = {Proceedings of Machine Learning Research}, address = {Chia Laguna Resort, Sardinia, Italy}, month = {13--15 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v9/alamgir10a/alamgir10a.pdf}, url = {https://proceedings.mlr.press/v9/alamgir10a.html}, abstract = {Brain-computer interfaces (BCIs) are limited in their applicability in everyday settings by the current necessity to record subject-specific calibration data prior to actual use of the BCI for communication. In this paper, we utilize the framework of multitask learning to construct a BCI that can be used without any subject-specific calibration process. We discuss how this out-of-the-box BCI can be further improved in a computationally efficient manner as subject-specific data becomes available. The feasibility of the approach is demonstrated on two sets of experimental EEG data recorded during a standard two-class motor imagery paradigm from a total of 19 healthy subjects. Specifically, we show that satisfactory classification results can be achieved with zero training data, and combining prior recordings with subject-specific calibration data substantially outperforms using subject-specific data only. Our results further show that transfer between recordings under slightly different experimental setups is feasible.} }
Endnote
%0 Conference Paper %T Multitask Learning for Brain-Computer Interfaces %A Morteza Alamgir %A Moritz Grosse–Wentrup %A Yasemin Altun %B Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2010 %E Yee Whye Teh %E Mike Titterington %F pmlr-v9-alamgir10a %I PMLR %P 17--24 %U https://proceedings.mlr.press/v9/alamgir10a.html %V 9 %X Brain-computer interfaces (BCIs) are limited in their applicability in everyday settings by the current necessity to record subject-specific calibration data prior to actual use of the BCI for communication. In this paper, we utilize the framework of multitask learning to construct a BCI that can be used without any subject-specific calibration process. We discuss how this out-of-the-box BCI can be further improved in a computationally efficient manner as subject-specific data becomes available. The feasibility of the approach is demonstrated on two sets of experimental EEG data recorded during a standard two-class motor imagery paradigm from a total of 19 healthy subjects. Specifically, we show that satisfactory classification results can be achieved with zero training data, and combining prior recordings with subject-specific calibration data substantially outperforms using subject-specific data only. Our results further show that transfer between recordings under slightly different experimental setups is feasible.
RIS
TY - CPAPER TI - Multitask Learning for Brain-Computer Interfaces AU - Morteza Alamgir AU - Moritz Grosse–Wentrup AU - Yasemin Altun BT - Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics DA - 2010/03/31 ED - Yee Whye Teh ED - Mike Titterington ID - pmlr-v9-alamgir10a PB - PMLR DP - Proceedings of Machine Learning Research VL - 9 SP - 17 EP - 24 L1 - http://proceedings.mlr.press/v9/alamgir10a/alamgir10a.pdf UR - https://proceedings.mlr.press/v9/alamgir10a.html AB - Brain-computer interfaces (BCIs) are limited in their applicability in everyday settings by the current necessity to record subject-specific calibration data prior to actual use of the BCI for communication. In this paper, we utilize the framework of multitask learning to construct a BCI that can be used without any subject-specific calibration process. We discuss how this out-of-the-box BCI can be further improved in a computationally efficient manner as subject-specific data becomes available. The feasibility of the approach is demonstrated on two sets of experimental EEG data recorded during a standard two-class motor imagery paradigm from a total of 19 healthy subjects. Specifically, we show that satisfactory classification results can be achieved with zero training data, and combining prior recordings with subject-specific calibration data substantially outperforms using subject-specific data only. Our results further show that transfer between recordings under slightly different experimental setups is feasible. ER -
APA
Alamgir, M., Grosse–Wentrup, M. & Altun, Y.. (2010). Multitask Learning for Brain-Computer Interfaces. Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 9:17-24 Available from https://proceedings.mlr.press/v9/alamgir10a.html.

Related Material