EEG-based BCI Systems in Neuropsychiatric Diseases

177

Therefore, there was an increasing focus on research exploring Event-Related

Potentials (ERPs) as a neurological reaction to both external and internal

stimuli until the end of the 1980s (see the review by [13]). In the 1990s, EEG

was started to be considered as an important signal of the brain with the

different oscillatory systems rather than a background noise [14, 15].

Over the past few decades, EEG has been employed for various pur-

poses, including BCI applications, neuro-marketing, gaming, emotion, and

face recognition as well as clinical diagnosis (see Section 7.4).

7.2.3

Categories of EEG-based BCI

In the literature, the current BCI systems are classified based on various fac-

tors, including the way of recording brain signals, the elicitation way of brain

signals, synchronization, and dependability and timing (Please see the com-

prehensive reviews, [16–18]).

a) The way of recording brain signals: EEG-based BCI systems can be

divided into three groups including invasive, semi-invasive, and non-invasive

based on the recording style of brain signals. In BCI terminology, non-invasive

methods, like EEG, are called procedures without penetrating the brain to

record the electrical activity, which means that no surgical operation is needed.

Therefore, they are accepted as secure methods despite poor signal quality and

poor spatial resolution and are preferred by most scientists and researchers.

In contrast, the methods with neurosurgical operation, such as iEEG, implant

electrodes to obtain brain signals for BCI systems. As a type of iEEG, Elec-

trocorticography (ECoG) is called a semi-invasive method that uses electrodes

placed on the exposed surface of the cortex [19] whereas stereotactic electroen-

cephalogaphy (sEEG), called invasive method, uses depth electrodes that are

penetrated in deeper structures of the brain [20]. iEEG could not be broadly

used since they have a risk for brain damage as well as they are expensive

systems.

b) The elicitation way of brain signals: EEG-based BCI systems could also

be classified into four categories according to the elicitation way of brain sig-

nals. One of them is active, also called endogenous, BCIs, which transform

the users’ self-induced brain activity to commands to use machines, indepen-

dently of external stimulations. The neural signals used in these systems are

generated through a mental task, such as motor imagery (MI) by which the

user imagines a movement of a limb to use a BCI system. In contrast, passive

BCIs utilize the brain activities that arise without voluntary control upon

application of various types of stimuli (e.g., visual, auditory, or emotional),

especially cognitive tasks. Although the initial motivation was to develop ac-

tive BCI systems in this field, passive systems become an emerging trend in

recent years. Besides these two types, there is also one more BCI type that

uses the brain signals generated as a reaction to stimuli, often tactile, pain,

audio, or video. Such systems are called reactive BCI systems. Several studies