Re: Rationale for Critical Bands

[Martin Braun <nombraun@xxxxxxxxx>]

Dick and Alain wondered if the following two bodies of evidence were likely 
to be related.

1) Spectral band filters, as observed in psychoacoustics and in behavioral 
experiments, are sliding spectral windows. A catalogue of fixed borders 
between filter bands along the frequency scale have never been found.

2) The central nucleus of the inferior colliculus (ICC) in the mammalian 
auditory midbrain is organized in stacked laminae that represent frequency 
bands approximately corresponding to the behavioral band filters. The lamina 
edges constitute an catalogue of fixed frequencies.

The answer is that the two bodies of evidence appear to be related when one 
considers the neural connectivity within and across the laminae.

Neural connectivity for band filtering:
The general mechanisms here are neuron-centered bands of lateral inhibition 
and lateral facilitation. These mechanisms can be anatomically realized 
independent from the shape of neuron arrays. Linear arrays, as in the 
cochlea, or arrays in stacked laminae, as in midbrain and thalamus, are both 
possible solution for neuron-centered bands of lateral inhibition and 
lateral facilitation.

Band filtering in stacked laminae:
Neuron arrays in stacked laminae provide the advantages of space economy and 
connectivity economy. A typical observation in ICC physiology is lateral 
inhibition within a frequency range of a lamina (reviewed in Braun, 1999). 
The underlying neural connectivity can be organized most economically by 
spreading out a neuron-centered inhibitory network (a) within the neuron's 
own lamina and (b) toward the near-side halves (frequency-wise) of the two 
adjacent laminae. The necessary within-lamina and oblique-across-laminae 
fiber networks of inhibitory GABA neurons has indeed been established in 
extensive anatomical studies of this nucleus (same ref.). The principles of 
connectivity economy for beyond-lamina-space lateral facilitation have also 
been shown in anatomical studies (same ref.).

In summary, if one wants to understand how band filtering is realized in the 
mammalian ICC, it is not enough to read the physiological reports on neuron 
recordings. It is also necessary to read the anatomical reports on neuron 
connectivity.

Braun (1999):
http://www.ncbi.nlm.nih.gov/pubmed/10190753


Martin


---------------------------------------------------------------------
Martin Braun
Neuroscience of Music
S-671 95 Klässbol
Sweden




----- Original Message ----- 
From: "Alain de Cheveigne'" <Alain.de.Cheveigne@xxxxxx>
To: <AUDITORY@xxxxxxxxxxxxxxx>
Sent: Monday, June 28, 2010 6:13 PM
Subject: Re: Rationale for Critical Bands


Malmierca et al's paper shows a discontinuous