Loading Events

Can a computer extract speech directly from the brain signals?

Every year thousands of people lose their ability to communicate due to neurological disorders caused by damage in speech-related brain areas (such as in aphasia) or paralysis of voluntary muscle movement (such as in locked-in syndrome). The consequences of the disruption of natural speech production on the life quality of such patients are devastating. Therefore, we aim at establishing a bridge between the brain and a computer that would allow participants to interact with the external world using only their brain activity, bypassing the need for speaking aloud. We call such a system a Brain-Computer Interface (BCI). To collect brain signals, we use electroencephalography (EEG)- a non-invasive technology where recording sensors are placed at different locations over the scalp. During the experiment, we ask the participant to imagine pronouncing two different speech units without speaking aloud. Our BCI system then collects their EEG data and tries to distinguish the signal from the two speech units. The participants are then asked to do the same task again as before while the BCI system will give a prediction of the “mind” in their brain. The goal of this research is to better understand the neural mechanisms involved in human speech processing that in the long term might provide rehabilitation solutions for patients with compromised speech functions. At Scientifica 2021, we will demonstrate our ongoing speech-BCI experiment to give visitors a feel of how a scientific experiment of this kind is carried out.