A team at Chiba University in Japan has proposed Diffusion Inverse Filtering (DIF), a signal-processing method that undoes the spatial smearing volume conduction introduces into EEG, sharpening the functional-connectivity features that brain-computer interfaces (BCIs) rely on. Tested on an emotion-recognition task, DIF generally improved performance as electrodes were thinned out, and it is compatible with existing BCI pipelines. The work appeared in Brain Sciences on August 27, 2026.
