EEG-based brain–computer interfaces exploiting steady-state somatosensory-evoked potentials: a literature review
Jimmy Petit,José Rouillard,François Cabestaing
DOI: https://doi.org/10.1088/1741-2552/ac2fc4
IF: 5.043
2021-10-01
Journal of Neural Engineering
Abstract:Abstract A brain–computer interface (BCI) aims to derive commands from the user’s brain activity in order to relay them to an external device. To do so, it can either detect a spontaneous change in the mental state, in the so-called ‘active’ BCIs, or a transient or sustained change in the brain response to an external stimulation, in ‘reactive’ BCIs. In the latter, external stimuli are perceived by the user through a sensory channel, usually sight or hearing. When the stimulation is sustained and periodical, the brain response reaches an oscillatory steady-state that can be detected rather easily. We focus our attention on electroencephalography-based BCIs (EEG-based BCI) in which a periodical signal, either mechanical or electrical, stimulates the user skin. This type of stimulus elicits a steady-state response of the somatosensory system that can be detected in the recorded EEG. The oscillatory and phase-locked voltage component characterising this response is called a steady-state somatosensory-evoked potential (SSSEP). It has been shown that the amplitude of the SSSEP is modulated by specific mental tasks, for instance when the user focuses their attention or not to the somatosensory stimulation, allowing the translation of this variation into a command. Actually, SSSEP-based BCIs may benefit from straightforward analysis techniques of EEG signals, like reactive BCIs, while allowing self-paced interaction, like active BCIs. In this paper, we present a survey of scientific literature related to EEG-based BCI exploiting SSSEP. Firstly, we endeavour to describe the main characteristics of SSSEPs and the calibration techniques that allow the tuning of stimulation in order to maximise their amplitude. Secondly, we present the signal processing and data classification algorithms implemented by authors in order to elaborate commands in their SSSEP-based BCIs, as well as the classification performance that they evaluated on user experiments.
engineering, biomedical,neurosciences