[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Coherent brain frequency measurement (?)



>I have been very interested in a technique described in the following paper
>for determining a nonlinear 'resonance' in the human brain's response to
>auditory stimulation (usually around 40 Hz for awake, unanesthetized human
>subjects):
>
>Sapsford DJ. Pickworth AJ. Jones JG. A method for producing the coherent
>frequency: a steady-state auditory
>evoked response in the electroencephalogram. [Journal Article] Anesthesia &
>Analgesia. 83(6):1273-8, 1996 Dec.
.... The authors claim that
>this technique defines a 'coherent frequency' in the brain's response--i.e.
>a frequency of stimulation to which the brain is already 'tuned' and
>therefore is likely to respond maximally. ....it can be changed in a
>systematic manner
>through the administration of an anesthetic and therefore can be used to
>assess how deeply anesthetized an individual brain has become.....


I haven't read the paper, but I think I heard about this before.  Auditory
evoked potentials are usually divided into three categories:
brainstem potentials, less than 10 msec latency; late cortical potentials,
probably 100 msec and up, something like that, I don't exactly remember;
and the "middle latency potentials", with a latency in the range of about
20-40 msec or so, and a duration of about 25 msec. In other words, the
later the potential, the more it resembles EEG, in several respects,
especially sensitivity to anesthesia and levels of consciousness.

A potential with a duration of 25 msec can be elicited using a 40 hz
stimulus (40 times a second that is; the stimuli can be tone bursts at
various frequencies), thus producing a 40Hz steady-state potential.
In, say,  the Seventies and Eighties, there was a lot of interest in using
middle latency auditory EPs for hearing assessment in difficult-to-test
subjects, particularly with  the 40Hz paradigm.  The late cortical
potentials are too closely linked to state of consciousness for that.  The
early brainstem components, which have in fact been successfully used for
hearing screening, still leave a couple of things to be desired: they lack
much frequency specificity, and they don't occur very close to threshold
intensities, so they aren't great for defining thresholds. For whatever
reason, perhaps because they too are considerably dependent on state of
consciousness, the middle latencies didn't work out that well for hearing
testing.  (There is a newer technique called evoked oto-acoustic emissions
that is working out much better).

Since the middle latency potentials are linked to state of consciousness,
it makes sense to try to use them to assess levels of anesthesia. but I
don't see that it has anything to do with the brain having a "resonance" at
40 Hz.  In fact, since the highest amplitude EEG waves are at frequencies
below 8 Hz (delta, theta), alpha waves (8-13) are in the middle, and beta
(14 and up) are the smallest amplitudes, I would suspect that the resonant
frequencies of cortical tissues are in the 10Hz and under range.  Notice,
too, that the resonant frequencies, if that's indeed what they are, seem to
change with levels of consciousness and activation-- in normal adults, you
don't get frequencies under 7 Hz except in sleep, and you get occipital
alpha activity only in an alert, relaxed state with eyes closed.  Also, if
it really has to do with a resonant frequency characteristic of the brain,
why are there not 40Hz visual and somatosensory potentials?

The one other thing I recall from previous discussions about this technique
of using middle latency responses for assessing level of consciousness is
that there were some paradoxical increases in amplitude, which sounded like
artifacts of the measurement technique, which was automatic.  In recordings
of cortical evoked potentials, one source of artifact (in awake subjects
with eyes closed) is alpha activity, and it didn't sound like that was
being taken into account.





Jerry

Jerry  Larson jerry@neuromon.com
http://www.neuromon.com