A covert attention P300-based brain-computer interface: Geospell
Related Experiment Video
Updated: May 23, 2026

P300-Based Brain-Computer Interface Speller Performance Estimation with Classifier-Based Latency Estimation
Published on: September 8, 2023
Fabio Aloise1, Pietro Aricò, Francesca Schettini
1Neuroelectrical Imaging and BCI Lab, Fondazione Santa Lucia IRCCS, Rome, Italy. f.aloise@hsantalucia.it
Abstract:
The Farwell and Donchin P300 speller interface is one of the most widely used brain-computer interface (BCI) paradigms for writing text. Recent studies have shown that the recognition accuracy of the P300 speller decreases significantly when eye movement is impaired. This report introduces the GeoSpell interface (Geometric Speller), which implements a stimulation framework for a P300-based BCI that has been optimised for operation in covert visual attention. We compared the Geospell with the P300 speller interface under overt attention conditions with regard to effectiveness, efficiency and user satisfaction. Ten healthy subjects participated in the study. The performance of the GeoSpell interface in covert attention was comparable with that of the P300 speller in overt attention. As expected, the effectiveness of the spelling decreased with the new interface in covert attention. The NASA task load index (TLX) for workload assessment did not differ significantly between the two modalities.
Practitioner Summary:
This study introduces and evaluates a gaze-independent, P300-based brain-computer interface, the efficacy and user satisfaction of which were comparable with those off the classical P300 speller. Despite a decrease in effectiveness due to the use of covert attention, the performance of the GeoSpell far exceeded the threshold of accuracy with regard to effective spelling.

