Alexey Ossadtchi and Alexey Voskoboynikov gave a lecture on modern advances in speech decoding from brain electrical activity using brain-computer interfaces

Alexey Ossadtchi, PhD, director of the Center for Bioelectrical Neurointerfaces at the Higher School of Economics, presented a report on modern approaches in signal decoding in speech prostheses with a small number of implantable electrodes.

Alexey Voskoboynikov, trainee researcher of the Center for Bioelectrical Interfaces of the HSE, researcher of the Laboratory of Medical Neurointerfaces and Artificial Intelligence, FMBA CMNT, discussed speech zone mapping, which is one of the most important stages of neurosurgical intervention in the cerebral cortex.

Report 1: Toward a speech prosthesis: decoding speech from brain electrical activity.
Typical solutions are based on invasive recording of brain activity using a large number of cortically distributed electrodes implanted through an extensive craniotomy. After a brief overview of the state of the art in the field, the possibility of creating a speech prosthesis with a small number of electrodes and based on a compact decoder built according to the principle of explainable artificial intelligence was demonstrated using two patients as examples. 
Report 2: HMM-Simple net for modeling dynamics of speech cortex activity in intraoperative monitoring tasks.The lecture described an alternative approach, which consists in the observation of the electrical activity of the cortex, modeling the complex spatial and temporal dynamics of neural tissue activation, and identifying the key nodes of the neural network that implements the speech function.

The lecture is available at the link.