Skip to main content
Fig. 5 | BioMedical Engineering OnLine

Fig. 5

From: A computational paradigm for real-time MEG neurofeedback for dynamic allocation of spatial attention

Fig. 5

State-based neurofeedback (sb-NFB). During training (left), the training window from a block of trials (block i) is used to train the signal transformation, which is tested during the development window. The training component involved unsupervised dimensionality reduction using the data from the training window, extracting the time–frequency representation of these periods, and then learning features that best separate the states. During state decoding (right), each trial from the current block (block i + 1) is collected in real-time and transformed using the trained signal transformation from the previous block to obtain a state signal (bottom right plot). During testing, the features from the training are used to detect attention state over time, from which we extract the timing feature of interest. The timing feature targeted for training was extracted from the state signal for use as the NFB cue

Back to article page