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

Re: pitch discrimination (fwd)

Dear List,

I would agree that 10-30% is very high.  We have tested an admittedly
nonrandom sample of 50 adult subjects (oops I meant participants :-)) on
up-down discrimination using two 250ms tones with 0ms ISI (and matched
phase at the juncture) estimating a threshold Delta-f in the neighborhood
of 1 kHz. There was an increased representation of poor readers in our
sample, including persons with a history of dyslexia (this was the focus
of our investigation).  Of the 50 there were two or three with
sufficiently high thresholds that would be considered floor performance
given the parameters of the adaptive estimation procedure (500 Hz max).
We do not know if those subjects would be able to perform the task with a
wider frequency separation.  The subjects who so failed at this task were
among those with extremely elevated frequency discrimination thresholds
(for 250 ms tones separated by 750 ms ISI), that is, of several hundred Hz
(for a same/different task).  So I would guess that the up/down judgment
is secondary to a fundamental deficit in frequency representation.
Needless to say, these profound impairments in auditory processing were
correlated with reading ability (and somewhat less with word memory and
with nonverbal IQ, BTW).

So with respect to the reviewer's claim, I would doubt both the "10-30%"
and the "normal" part of the subject characterization.

However, given our data mentioned above, my problem is that they do
not square with Dr. Watson's assertion that "virtually  everyone can
discriminate pitch differences of 2-3% or less", which is based on 1000
subjects, no less.  I suppose that either his same/different measurement
was somehow different from ours (0 ISI between tones? using harmonic
complexes instead of pure tones) or they happened to not have any poor
readers/achievers in their population (which I find unlikely given the
focus of some of their studies I know about).  If someone has more on this
issue I would like to hear about it.


Thanassi Protopapas


On Thu, 27 May 1999, Robert J. Zatorre wrote:

>Dear List
>A reviewer of a recent paper of ours has written the following:
>"...I thought it was well established that 10 to 30% of normal subjects
>cannot judge which frequency [in a pair of pure tones] is higher or lower
>(above chance), while all subjects are generally able to judge if two
>frequencies are identical or not."
>My question: is this, in fact, well established, as the reviewer claims,
>and if so, what might be the reference for this?
>Any leads you all can give us would be appreciated
>Robert J. Zatorre, Ph.D.
>Montreal Neurological Institute
>3801 University St.
>Montreal, QC
>phone: 514-398-8903
>fax 514-398-1338