196
G. D. Johnson and D. J. Krusienski
9.1.2 Types of BCIs
BCIs can generally be grouped into three categories: Reactive, Active, and Passive;
based on how the brain response is elicited [34].
9.1.2.1 Reactive
The use of sensory stimuli to induce predictable changes in the EEG creates a
“reactive” control paradigm wherein the EEG control signal and its detection are
dependent upon the parameters of the sensory stimulus, which is typically visual,
auditory, or tactile [2]. The P300 oddball paradigm and its extension, the P300
speller, are examples of reactive paradigms [10]. Such paradigms are commonly
designed so that the EEG response differs depending on the characteristics of temporally and/or spatially multiplexed sensory stimuli, which individually represent
various selection options for the control interface. In general, reactive paradigms
are inherently synchronous systems, where timing of the user’s intended device
commands is dependent on the timing of the presented stimuli. Synchronous systems
do not allow for completely spontaneous communication or control.
9.1.2.2 Active
In contrast to reactive paradigms, volitional mental imagery or cognitive state switching creates an “active” control paradigm wherein the changes in the EEG are due
to endogenous cognitive actions and not external stimuli. An example of an active
control paradigm is sensorimotor rhythm cursor control where the user imagines
left/right hand movements to move a computer cursor to the left/right, respectively
[32, 33]. Active paradigms are better suited for achieving asynchronous communication or control, although synchronous designs with cued mental imagery are also
used.
9.1.2.3 Passive
The concept of passive BCIs where the user is not volitionally modulating the EEG
for direct device control will not be presented in this chapter. However, most passive
BCI scenarios and analyses are closely related to active/reactive paradigms and the
techniques discussed in this chapter are directly applicable [35].
G. D. Johnson and D. J. Krusienski
9.1.2 Types of BCIs
BCIs can generally be grouped into three categories: Reactive, Active, and Passive;
based on how the brain response is elicited [34].
9.1.2.1 Reactive
The use of sensory stimuli to induce predictable changes in the EEG creates a
“reactive” control paradigm wherein the EEG control signal and its detection are
dependent upon the parameters of the sensory stimulus, which is typically visual,
auditory, or tactile [2]. The P300 oddball paradigm and its extension, the P300
speller, are examples of reactive paradigms [10]. Such paradigms are commonly
designed so that the EEG response differs depending on the characteristics of temporally and/or spatially multiplexed sensory stimuli, which individually represent
various selection options for the control interface. In general, reactive paradigms
are inherently synchronous systems, where timing of the user’s intended device
commands is dependent on the timing of the presented stimuli. Synchronous systems
do not allow for completely spontaneous communication or control.
9.1.2.2 Active
In contrast to reactive paradigms, volitional mental imagery or cognitive state switching creates an “active” control paradigm wherein the changes in the EEG are due
to endogenous cognitive actions and not external stimuli. An example of an active
control paradigm is sensorimotor rhythm cursor control where the user imagines
left/right hand movements to move a computer cursor to the left/right, respectively
[32, 33]. Active paradigms are better suited for achieving asynchronous communication or control, although synchronous designs with cued mental imagery are also
used.
9.1.2.3 Passive
The concept of passive BCIs where the user is not volitionally modulating the EEG
for direct device control will not be presented in this chapter. However, most passive
BCI scenarios and analyses are closely related to active/reactive paradigms and the
techniques discussed in this chapter are directly applicable [35].