The two main objectives of the proposal are: - The study of man-machine interaction for the design of an appropriate online adaptative BCI - The development of an automated adaptive multiclass BCI using the results of the aforementioned topic for a high in formation transfer rate.
They will be achieved:
- Designing different automated adaptive BCI systems with different types of adaptation according to the analysis and assessment of the man-machine interaction (long-term changes, such as learning) with the aim to design a BCI which adapts where and when it is required.
- Applying different feature extraction methods, which describe the classes of the system, focusing adaptive feature extraction to deal with short-term non-stationarities.
- Using different adaptive classifiers to deal with long-term non-stationarities.
The increase of classes will increase the dimensionality because more features will be necessary to describe them, but also the information transfer rate. This expertise is important in BCI research and the same techniques can be applied in related research areas (EEG analysis or fMRI analysis). The applicant is expert in online adaptation of statistical classifiers; she has much experience in univariate adaptive autoregressive parameters and band power estimates.
The project will extend her knowledge to SVMs, which are also very powerful methods used in BCI research, with the aim of online adapt them. The application of the signal processing methods used in FIRST IDA will diversify her training in the feature extraction part. The knowledge acquired will be of great impact in her future career as well as in the development of BCI in Spain.
Acquired training can be applied to analyse fMRI and EEG. Furthermore the machine learning and signal processing techniques that the applicant is trained in have a multitude of applications beyond biomedical data analysis; they will become more and more salient as key technologies for the European industry.
Call for proposal
See other projects for this call