During my Master Thesis in Computer Science, I studied how unknown Brain-Computer Interface (BCI) data could be explored, for the purpose of finding patterns within brain activity. Data exploration is the basis to which any labeling and classified problem can be initiated, and can be used as a sanity check to determine whether a more expensive labeling experiment is worth investigating.
To aid this, I created BCI Explorer, which enables a visual and hierarchical investigation of clustered BCI data. As covered in greater detail by the Theory section, BCI Explorer uses Quick Shift to hierarchically cluster data points of windowed BCI data. From these clusters, similar windows of brain activity can be retrieved and intuitively visualized by BCI Explorer, making pattern exploration easy.
To be written..
To be written..