This technology improves the accuracy and speed of spelling using EEG signals. Some 7.5 million people in the US have impaired communication. In order for these individuals to communicate, augmented and alternative communication tools are used. These can range from something as simple as a pen to something as complicated as a brain-computer interface. The latter is used in extreme circumstances, such as locked-in syndrome. Currently, brain-computer interfaces are extremely slow, 1 to 10 characters per minute. Selection of letters also lacks accuracy—average accuracy of letter selection is only 70%. This lack of accuracy further slows communication. This technology is an algorithm that improves both the accuracy and the speed of communication with a brain-computer interface.
Greater accuracy in computer-brain interfaces
Brain-computer interfaces track EEG signals while a screen displays certain patterns over a grid of an alphabet. A certain EEG signal called the P300 signal appears when the letter a user wants is selected. This algorithm is, among other things, a better series of patterns for allowing patients to quickly select the letter they desire. This was tested on twenty-one patients and was found to increase accuracy by an average of ~2.5% but as much as ~20% in some cases. Additionally, it includes a dynamic stopping and starting feature which ensures that patient will not accidentally start selecting random letters when not paying attention to the screen. This will assist disabled patients in communicating with doctors and loved ones.
- Reduce typing errors while using the P300 Speller by applying a certainty threshold
- Dynamic stopping based on certainty value of the patient’s letter selection
- Dynamic starting to only spell when the user is paying attention to the screen
- There is currently no similar established method
- Reduces typing errors by withholding erroneous selection
- Allows user to make selections at their own