[Pd] OT- FFT and human auditory cortex

[Charles Henry]

> > > By phase-locked pulses, do you mean that the hair-cell trigger happens
> > > at the same point in each cycle, and so the resulting pulses correlate
> > > to the same phase for each vibration?  If you mean something else,
> > > it's probably not worth trying to explain.  But please do.
> >
yes, I *think* they are phase-locked, meaning that it happens at the
same phase per oscillation.  I would not know if it is one or two
action potentials per period or at what point they are triggered, but
the auto correlation of the auditory nerve bundles shows strong
correlations at the period of oscillation...implies phase-locked...but
that's all I got.
> >
> > >
> > > > I'm searching my brain for a better term than timing
> > > > encoding, but not sure
> > >
> > > PPM (pulse position modulation/encoding)? Well, that's what I'd call
> > > it with my electronic engineers hat on.
> >
> > >Wow, now there's a term that would get a blank stare from me!  I'd say
> > >temporal, depending on the context.  Or "frequency of impulses".  I'd
> > >also use "localized" as opposed to "tonotopic", but that's just my own
> > >vocab.
> >

I think temporal is as good as any term for it.  Tonotopic means that
there is a map from one domain of parameters to another domain, in
this case, the map proceeds from frequency to place.

> >
> > > so that at the cortical level, timing
> > > encoding only pertains to the low <100-200 Hz frequencies.
> >
> > >The magic number at ten to twenty milliseconds has fascinated me
> > >for ages.
> >
Hey, how about loudness integration?  I think it's in Music,
Cognition, and Computerized sound...a resonator takes time to build up
energy at resonance.  For short time periods, ~25 ms, the energy is
 accumulating in the resonator and beyond that, it evens out.
(of course, impulses are short too, but they don't resonate by
building up energy, they just impart it.  )
So any grains on the order of 10-20 ms don't fully create the steady
state activity of the resonator.  Just an idea...I really don't have a
clue about grains...

I'm technically still in school this summer, until I complete a
literature review for one class.  The topic is neuroscience and
auditory scene analysis.  I've been way too bogged down in speech
readings for awhile, and I never liked studying speech much.  I think
that ASA is about as important of a cognitive psych theory as
any...This is probably old hat, here, but has anyone read David
Huron's "Tone and Voice: a derivation of the rules of voice leading
from perceptual principles" in music perception?  This article got me
thinking about how voice leading rules and ASA go hand in hand.  Or
that ASA even subsumes the category of voice leading rules.

> >
> > >To get somewhat on-topic, Charles, what brings you to Pd?  Have you
> > >used it in research at all?  For auditory stuff or DSP?  I think this
> > >list could benefit from people whose primary interest is other than
> > >programming or composing.
> >
This spring, I was planning to design an experiment for Cg. Psych here
at KU, but I couldn't figure out the analysis protocol...I'm moving
forward, but I chose a difficult topic for the data analysis. oh,
those damn receiver operator characteristics of signal detection...I
just don't have a good grasp on it yet--couldn't complete the proposal
if I couldn't detail the analysis

I talked about it with my advisor, and brought up an interesting idea.
 The exerimental protocol is just so simple, that we could do it with
Pd and use only the most elementary objects.  It can be run simply on
any computer, then, and use Pd to log data to text files.

The big idea is to try to circulate the experiment among Pd users, and
let them email the data back.  Thus, we could collect a fairly large
amount of data for an experiment.  Of course, we run into
institutional issues.  If you wanted the analysis to get published
somewhere, you'd have to get some kind of approval for the work and
faces serious issues of reproducibility and lack of controlled
environment...use headphones (non-standardized) or loudspeakers
(free-field, room dependent)...a whole lot of experimental design
problems...

I started using Pd, because I had a music composition project
(going on 2yrs now and still haven't finished) and Max just couldn't
do what I wanted...I couldn't figure out how to create new externals
 for Max (probably never tried hard enough), so I switched to Pd.
The composition project was to create music via stable feedback
between a loudspeaker and microphone--hence I've tried to build an
adaptive filter bank for fixing the LS/room/mic response.  I'll get it
 done one of these days...

Chuck