Re: Coherent brain oscillations at 40 Hz

[Gary Goldberg MD <goldberg@VM.TEMPLE.EDU>]

I am very interested in both issues:  the analysis of the EEG in the
gamma-frequency range and the study of EEG signals with nonlinear dynamical
systems analysis techniques.
With regard to the analysis of the EEG in the gamma-frequency range, I have
been very interested in techniques that have been described by D Sapsford
et al (see Appendix below) from Gareth Jones' lab in the department of
Anesthesiology at Cambridge University in which auditory stimulus
frequencies are swept across the gamma range in order to identify a peak of
coherent response typically around 40 Hz in the awake normal human subject.
These researchers from a department of Anesthesiology have found that this
coherent frequency drops systematically in subjects to whom general
anesthesia is administered and that the amount the frequency drops can be
used to monitor the depth of anesthesia and corresponds to changes in
behavioral measures of brain responsiveness (see paper by Andrade et al in
Appendix below).  This technique would have obvious applications in the OR
for monitoring depth of anesthesia and for preventing awareness during
surgery, but I have been very interested in trying to determine whether the
technique could be adapted to the problem of assessing the depth of coma in
patients with traumatic and atraumatic coma, and therefore being able to
project the likelihood of emergence from coma.  We have a Clinical
Neurophysiology Lab set up in a Traumatic brain injury Rehabilitation
Center and a critical issue we are constantly addressing clinically is the
question of prognosis in patients who are slow to emerge from a comatose
condition--i.e. whether they will remain in a persistent vegetative state,
advance to a state of minimal consciousness, or emerge to full
consciousness and communicative ability with a variable state of residual
disability.  We have applied brainstem auditory evoked potentials and
median nerve somatosensory evoked potentials to this problem and have found
them to be useful but somewhat incomplete in their predictions.  We are
always looking to improve our ability to predict emergence from coma using
other approaches, and this seemed to be an interesting technique to
explore.  It is also very interesting when one places this idea of a
coherent brain resonance at around 40 Hz into the context of a large body
of work in animals that shows that oscillations at 40 Hz in cortical field
potentials may serve an important purpose in synchronizing disparate
neuronal populations associated with spatially separate cortical feature
maps (e.g. representing color, shape, motion in the visual cortex) so that
these populations form a synchronized, coherent functional brain network
associated with the emergence of a unitary conscious percept.   This would
suggest that the presence of a coherent resonant response around this
frequency would serve as a necessary condition for the re-emergence of a
capacity for conscious information processing at the cortical level.  I
would be very interested in hearing from anyone else on this list who has
looked at this particular clinical application or who has experience with
the Sapsford et al technique for identifying the coherent response or
nonlinear brain resonance at 40 Hz.  I would also be interested in hearing
from anyone on the list who has used other clinical neurophysiology
techniques to look at the question of predicting recovery from severe brain
injury, either traumatic or atraumatic.
In their paper, Sapsford et al make reference to the possibility of
developing a more efficient algorithm for tracking the coherent frequency
that could be implemented in a monitoring instrument.  I have not seen any
further papers on the subject in my review of the literature and wonder if
anyone knows if they have made further progress with this project?
Thanks for your help with this.
--Gary Goldberg MD

APPENDIX
Unique Identifier
       97098058
Authors
       Sapsford DJ. Pickworth AJ. Jones JG.
Institution
       Department of Anaesthesia, Cambridge University, Addenbrooke's Hospital,
       England.
Title
       A method for producing the coherent frequency: a steady-state
auditory evoked
       response in the electroencephalogram.
Source
       Anesthesia & Analgesia. 83(6):1273-8, 1996 Dec.
Abstract
       Transient and steady-state auditory evoked responses in the
electroencephalogram are used to study the
       effect on the brain of graded changes in the concentration of
general anesthetics. A method is described
       using modern signal processing techniques to improve the analysis of
steady-state auditory evoked responses
       (SSAER). The SSAER was obtained using headphones to give 100-200
auditory click stimuli from 6.5 Hz.to
       50.5Hz in 1-Hz steps. The resulting electroencephalogram signals
were filtered and subject to Fourier
       analysis, after which a series of coherence indexes were derived
based on waves with significant power in
       the fundamental but with minimal harmonic content. These were
plotted against the range of stimuli and
       fitted with a third-order polynomial. The frequency at which the
maximum coherence index was achieved
       (highest possible value = 1) was derived from polynomial
interpolation. The repeatability of the method was
       examined in 10 awake subjects using runs of ascending then
descending stimulating frequencies. The mean
       maximum coherence index was at 38 Hz, with the 95% confidence
interval of 37.3 Hz-38.9 Hz. There was
       no difference between ascending and descending sweeps. The method
provides an automatic analysis of the
       SSAER that obviates the need to make subjective decisions about
which is the dominant wave, a major
       problem in the analysis of the transient auditory evoked responses.

nique Identifier
       97098059
Authors
       Andrade J. Sapsford DJ. Jeevaratnum D. Pickworth AJ. Jones JG.
Institution
       MRC Applied Psychology Unit, Cambridge University, Addenbrooke's
Hospital,
       England.
Title
       The coherent frequency in the electroencephalogram as an objective
measure of
       cognitive function during propofol sedation.
Source
       Anesthesia & Analgesia. 83(6):1279-84, 1996 Dec.
Abstract
       Ten volunteers were studied during six stages of propofol sedation,
namely awake (no propofol), light
       sedation (small dose of propofol), deep sedation (large dose), deep
sedation with stimulation of the ulnar
       nerve, then light sedation again (small dose), and awake (recovery).
Light and deep sedation were
       defined in terms of performance on a test of cognitive function: the
within-list recognition (WLR) test. At
       each stage, the steady-state auditory evoked potential was measured
at different stimulating frequencies
       to derive the frequency needed to achieve the maximal coherence
index. This frequency is called the
       "maximum coherent frequency". WLR performance correlated with
infusion dose (r = -0.71), plasma
       propofol concentration (r = -0.75), and maximum coherent frequency
(r = 0.75). When the correlations
       were examined for the propofol sedation stages only, there remained
a strong correlation between WLR
       performance and maximum coherent frequency (r = 0.47, P < 0.005),
but no significant correlations
       between infusion dose and WLR (r = -0.11) or infusion dose and
plasma concentration (r = 0.13). These
       data suggest that maximum coherent frequency provides a better
measure of depth of sedation than does
       the dose of propofol alone.