Biomedical Engineering Reference
In-Depth Information
during multiple trials of a functional task are averaged across trials in the time
domain or in the frequency domain.
Averaging EEG signals in the time domain extracts phase-locked signal compo-
nents as ERPs and discards nonphase-locked components as noise. In contrast,
averaging EEG signals in the frequency domain focuses on event-related changes in
EEG spectral energy that may have both phase-locked and nonphase-locked com-
ponents [2]. The latter approach is still temporally anchored to an event across mul-
tiple trials, but the result is not limited to phase-locked components, as with ERPs.
In either approach, it is necessary to verify the significance of putative functional
responses with respect to some reference, which is usually derived from a baseline
period preceding the event under study.
The distinction between these types of responses, particularly the distinction
between ERPs and nonphase-locked increases in signal energy (often termed
induced responses
), can easily be confused. Averaging in the time domain necessar-
ily yields phase-locked responses. However, significant variability (jitter) in the
latency (or phase) of ERPs can distort their appearance in time-averaged responses.
High-frequency components are more susceptible to latency jitter, and there is usu-
ally more jitter of ERP components (and their corresponding cognitive processes) at
longer latencies. ERPs with high-frequency (e.g., gamma) components are usually
confined to early (
150 ms) latencies. On the other hand, ERPs at longer latencies
(e.g., P300) usually consist of low-frequency components that are more resistant to
jitter [3, 4]. Because averaging in the frequency domain does not require phase lock-
ing, it may be better suited to investigate cortical processes with longer or more vari-
able latencies and to investigate high-frequency EEG responses at longer latencies.
EEG responses may have different combinations of frequencies, latencies, and phase
locking. Nevertheless, the distinction between phase-locked and nonphase-locked
responses is often still a practical one. Furthermore, many studies suggest that these
different classes of EEG responses may have distinct functional response properties
[5-7].
<
14.3.2 Application of Phase-Locked iEEG Responses to Cortical Function
Mapping
A substantial literature has already accumulated from iEEG investigations of
phase-locked responses, that is, ERPs, under a variety of experimental cognitive
paradigms. The most influential of these is the oddball paradigm in which infre-
quent stimuli of any sensory modality (auditory having been studied the most) are
randomly presented in a stream of frequent stimuli. Detection of the infrequent
stimuli may be performed either automatically or intentionally. Infrequent stimuli
produce a variety of phase-locked responses with a positive polarity and a peak
varying from 300 to 600 ms [8], and a vast number of scalp EEG studies have
explored a variety of behavioral and clinical factors affecting the different
components (e.g., P3a and P3b) of this ERP.
Because of the inherent uncertainty regarding the sources of scalp-recorded
ERPs, a number of iEEG studies have investigated the brain structures responsible
for different components of the P300. Insights from iEEG studies have also been
supplemented by studies of the effects of focal lesions on the P3a and P3b [9, 10], as
Search WWH ::
Custom Search