Decoding imagined syllables for speech neuroprosthesis
Prof. Anne-Lise Giraud
Aphasia is an impairment in the comprehension and formulation of language caused most commonly by stroke, traumatic brain-injuries, and other pathological conditions affecting the motor production of speech. Current neuro-rehabilitation techniques are effective only partially and in a fraction of patients, leaving the remaining in the impossibility to communicate with the
external world.
The proposed internship is part of Braincom, a multi-disciplinary European project, whose main goal is the development of a new generation of speech neuroprosthesis to restore communication in aphasic patients. The approach is based on the real-time decoding of inner speech from sensory and speech motor cortex through the use of non-invasive (EEG/MEG) and intracranial recordings (ECoG).
The first part of this internship will consist in characterizing the neural signatures of mental imagery of syllables - the “building blocks” of words – from electroencephalography (EEG) recordings in healthy participants. More specifically, we aim at identifying (1) the best time window around the onset of the syllable-imagery and (2) the most relevant brain regions able
to optimally discriminate between two different imagined syllables.
In the second part of the project, these spatio-temporal features will be integrated into a closed-loop brain-machine interface, capable of decoding in real-time participants’ imagined
syllable and, according to the classifier output, provide an auditory feedback.
Tasks:
Application:
Please provide the following information in the e-mail: