1. Reverberations in the Auditory System, (Or: On theoretical models and on various experimental data…)
Petr Marsalek
Prague, CZ, Charles University,
Medical Faculty 1, Institute of Pathological Physiology
Internal Seminar, Biological MPI PKS
March 13, 2017 _

2. outline
1 brief review of hearing, from hair cells to the auditory pathway
2 spikes (discharges, action potentials) and post-synaptic potentials are unitary neuronal events
3 experimental observations
4 sound localization circuits and their models
5 high frequency circuit (LSO) model and perspectives of how the low frequency circuit (MSO) model can be described and constructed
6 some computations with post-synaptic potentials
7 conclusions.
of X

3. outer, middle and inner ear
middle ear
outer ear _
[textbook]

4. outer ear outer ear mechanical model scaling factor? 1:10.. 1:20
[artifact] _

5. middle ear with small ear bones
levering impedance from air -> water
distance indicator
anvil
ear drum
hammer
stirrup
middle ear
mechanical model
scaling factor? 1:10.. 1:20
[artifact] _

6. inner ear (= cochlea) is enclosed in osseous helical and spiral spatial tunnel (like snail shell)
Width at the base
100 mm
Width at the apex
500 mm
Length
33 mm
basilar membrane – from above and unfolded into trapezoid plane, has graded stiffness
[textbook]

7. cochlear cross section, hair cells, auditory nerve
cochlear cross section, hair cells, auditory nerve. organ of Corti is active, mechanical-electrical transducer.
[textbook]
/~X

8. Inner Hair Cell initiates spikes on the auditory nerve.
[textbook/ Bures thesis] _

9. High, Middle and Low Spontaneous Rate auditory nerve fibers
dB SPL,
decibel of sound pressure level
is objective unit of sound intensity
x-axis: stimulus intensity, dB SPL,
y-axis: neuronal spike rate per s
WINTER I.M., PALMER A.R. Intensity coding in low-frequency auditory-nerve fibers of the guinea pig.
J Acoust Soc Am. 1991, 90, pp. 1958–1967
[Winter and Palmer, 1991] _

10. auditory nerve time histogram
x-axis: time; y-axis: neuronal characteristic frequency;
z-axis: spike time histogram
[Kiang, 1965] _

11. auditory nerve interval histogram
x-axis: time, ms;
y-axis: Number of spikes
(A) stimulation frequency 412 Hz;
(B) stimulation frequency 1 kHz.
[Rose JE, 1962] _

12. x-axis: time,
y-axis: position along cochlea
[textbook]

13. filter bank cochlea and auditory nerve
tonotopic organization of human auditory pathway
x-axis: sound frequency, y-axis: gains of individual neurons;
Its granularity is one quarter of octave (multiples of \sqrt[4]{2})
f = 128, 152, 181, 215, 256, 304, 362, 430, 512, ..., 8192 Hz, these steps are critical bands, shown e.g. by masking
filter bank cochlea and auditory nerve _

14. x-axis, sound frequency (Hz), y-axis, sound pressure (Pa)
x-axis, sound frequency (Hz), y-axis, sound pressure (Pa). Red top curve is pain threshold. Hearing threshold is Green bottom curve (=normal audiogram). Blue are curves of subjectively equivalent loudness level. Phon is its unit, like dB SPL. Yellow blob is speech region. Gray line is reference frequency 1 kHz.
[textbook] _

15. Signal discrimination by ideal observer
x-axis: firing rates;
y-axis: response probabilities;
different signals are detected based on variable responses and different mean values
[reviews/ textbook] _

16. sound localization pathways ITD and ILD
auditory pathway
sound localization pathways
ITD and ILD
(interaural time and level differences)
[reviews/ textbook] _

17. top left: candidate mechanisms of sound azimuth in high frequencies.
JND, just noticeable difference in ILD, inter-aural level difference can be equivalently expressed in dB, or in spikes/ s.
right column: shows parameter variation.
[Bures and Marsalek] _

18. candidate mechanisms of calculating sound azimuth in low frequencies with the use of adding excitatory and inhibitory postsynaptic
potentials
[Toth and Marsalek] _

19. [Toth and Marsalek] solution to some questions can be simple:
to calculate a “read-out” curve, it takes to take inversion function of a spike time histogram, pick a proper branch and a proper function normalization
[Toth and Marsalek] _

20. low frequency localization with several different parameter settings/ ideal observer readouts:
left: single neuron, unlimited time
right: reaction time TA, spike timing jitter TJ, window of coincidence detection wCD
[Sanda and Marsalek] _

21. Applications of auditory nerve spike train studies: cochlear implants.
Some of subjects with cochlear implants (CI) have both sides (mostly German speaking countries :-). How is it with binaural hearing and horizontal localization?
In tonal languages (Asian, eg. Chinese) major CI are not useful. How can be pitch impression restored?
In digital sound engineering, in normal hearing, impression can be improved and/or enhanced: viz surround sound…
ETC.
open problems,
interdisciplinary questions,
technology questions _

22. [inner ear and stato-kinetic organ]_

23. TOTH P.G. and MARSALEK P., Analytical description of coincidence detection synaptic mechanisms in the auditory pathway. Biosystems, 136, 90-98, 2015.
BURES Z. and MARSALEK P., On the precision of neural computation with inter-aural level differences in the lateral superior olive. Brain Res., 1536, 16-26, 2013.
SANDA P. and MARSALEK, P., Stochastic interpolation model of the medial superior olive neural circuit, Brain Res., 1434, , 2012.
BURES Z., PhD thesis, 2007.
[References] _