Biomedical Engineering Reference
In-Depth Information
nation of speech stimuli than during discrimination of tone stimuli, suggesting that
HGR in auditory cortex depends on the complexity of acoustic stimuli and, by
inference, on the amount of cortical processing necessary for the auditory
discrimination [32]. In contrast, the amplitude of the N100 depends much less on
the type of stimulus.
iEEG studies of auditory cortex have also highlighted the broadband spectral
profile of HGRs. For example, power spectral analyses of HGRs during auditory
discrimination indicated that the greatest energy increases occurred at 80 to 100 Hz
but also extended up to 150 to 200 Hz [32, 37]. In a study by Edwards et al. [72],
HGRs were reported between 60 and 250 Hz, centered at ~100 Hz, and in a study
by Trautner et al. [45], HGRs extended to 200 Hz. Similar frequency responses
have been observed in microelectrode recordings of auditory cortex in monkeys [90,
91]. In addition, studies in both humans and animals have shown that
nonphase-locked responses in the traditional 40-Hz gamma band are more variable
and less sensitive to functional activation of cortex [32, 91]. This may be due to
variability in the upper boundary of frequencies at which event-related power sup-
pression (ERD) occurs. If ERD extends into low-gamma frequencies, it may obscure
any power augmentation in this frequency range. For this reason analysis of
event-related 40-Hz responses that are based on narrowly bandpass-filtered signals
may at times yield misleading results. Time-frequency analyses of iEEG recordings
have shown that the lower boundary of HGR power augmentation may extend into
40-Hz frequencies, but the most consistent responses occur above 60 Hz. The
greater reliability of ECoG power changes in higher-gamma frequencies has
recently been reinforced by a study in which movement classification for different
body parts was best for power at 70 to 150 Hz, which the authors called the “chi”
band. The limited classification accuracy of lower-gamma frequencies (30 to 70 Hz)
was interpreted to result from a superposition in the power spectrum of band-lim-
ited power suppression (ERD) at lower frequencies (ranging up to ~50 Hz) and an
increase in power across all frequencies, obeying a power law.
The broadband spectral profiles of HGRs recorded with iEEG in humans are
difficult to reconcile with earlier conceptualizations of gamma oscillations associ-
ated with functional activation. These ideas of gamma oscillations have been
derived largely from observations of relatively band-limited responses (e.g., in and
around 40 Hz) in microelectrode recordings of animals. Although accumulating
observations in animals of higher-frequency, broader-band oscillations appear to be
gradually extending the frequency range that is quoted for “gamma,” gamma
responses are still largely conceptualized as band-limited network oscillations.
However, the broadband frequency response of HGRs observed with iEEG during
functional activation would presumably require the summation of activity in multi-
ple spatially overlapping neural populations or assemblies, each oscillating at
different, perhaps overlapping or broadly tuned, frequencies [57, 92].
An alternative explanation for the broadband nature of HGRs is that they are
the time-frequency representations of transient responses with a broad range of fre-
quency components. Intertrial jitter in the latency of these transients presumably
renders them invisible when averaged across trials in the time domain. Recent inves-
tigations of microelectrode recordings from macaque SII cortex during tactile stim-
ulation have observed broadband HGRs with spectral profiles identical to those
Search WWH ::
Custom Search