1886 Rutherford
Hz
0
20
• Spike frequency codes tone frequency
• Refractory period limits firing rate
Timing Code
1949 Wever proposed volley principle
• nerve fibers work together
• 1970 Phase Locking
500
Von Bekesy won Nobel Prize in 1961 for
studies of cochlear Traveling Wave
4000
20,000
Place Code
• Pressure vibrates Basilar membrane
• peak of vibration depends upon frequency
• maximal transduction occurs at peak
Phase-locking
•
In response to low-frequency (< 5 kHz) pure tones, spikes tend
to occur at a particular phase within the stimulus cycle.
•
Phase locking can be assessed with period histograms, which
display the distribution of spikes within a stimulus cycle.
•
In interspike interval histograms, phase locking creates modes
at integer multiples of the stimulus period.
Phase-locking:
response synchrony to acoustic waveform
Russell and Sellick
Volley Principle (Wever)
Although a single fiber may not discharge on
every cycle of a tonal stimulus, frequency can
still be accurately coded in the synchronized
discharges of a population of fibers
e.g., those innervating the same hair cell.
Vector Strength
Synchronization index for circular distributions
Neural sensitivity to the modulation frequency is evident
as synchronization to the modulation period (T)
Mean Phase
"
$ # sin '
"i )
& i
= arctan&
)
cos
#
&
" i )(
% i
Vector Strength
!
where
" = 2#
mod ti , T
i
!
2
2
$
'
$
'
r = 1 &" cos# i) + &" sin# i )
N % i
( % i
(
1st Fourier coefficient of period histogram
spectral magnitude of the response at !m
!
normalized by mean spike-rate
F1
F0
(
T
)
Each spike time ti is treated
as a vector of unit length
and with phase !i between 0
and 2" measured as the
spike time modulo the
stimulus period of interest.
The N spikes in a response
are combined by vector
addition, and the resultant
vector is normalized to N
significance determined by
the Rayleigh statistic:
2r 2 N
Cochlear Frequency Map
There is a precise mapping between the
CF of an AN fiber and cochlear place
nearly
logarithmic
for CFs > 2 kHz
Type I
95%
Type II
5%
Relative Sound Level (dB)
Auditory Nerve: Frequency selectivity
Frequency (kHz)
Brainstem modulation of cochlear
function
Olivocochlear pathway modulates dynamic range
Auditory Nerve Classifications
Type I
Type II
~95%
~5% (of all AN)
myelinated
unmyelinated
large
smaller
synapse on IHCs
synapse on OHCs
(20:1)
OHCs
IHCs
(1:many)
OHCs
IHC
Spiral ganglion
Type II
Type I
~30,000 nerve fibers with afferent projections
from cell bodies in the spiral ganglion and
efferent axons from cells in the olivary
complex.
95% Type I
5% Type
II
All existing physiological data are from Type I; nothing is
known about responses of Type II neurons.
Spontaneous Activity and Threshold
low/med-SR
high-SR
% in population
~40
~60
Syn term on IHCs
small
large
Bimodal distribution of
spontaneous discharge rates
spontaneous rate
Temporal coding in
Auditory nerve
Cariani & Delgutte (1996ab)
Dial-anesthetized cats.
100 presentations/fiber
60 dB SPL
Population-interval distributions are
compiled by summing together
intervals from all auditory nerve fibers.
The most common intervals present in
the auditory nerve are invariably
related to the pitches heard at the
fundamentals of harmonic complexes.
AN Population interval distribution
Morphology
and function
• CN has 3 major subdivisions: AVCN, PVCN and DCN
• Each subdivision contains distinct cell types differing in
morphology, cytochemistry, patterns of inputs & outputs,
and responses to sound
Superior Olivary Complex (cat)
MSO LSO
MNTB
Binaural Cues for Sound Localization
Interaural Time Difference (ITD):
Time-of-arrival between the two ears indicates
source location primarily for low-frequency
sounds.
For sinusoids, this can be expressed as a phase
difference. As the maximum delay approaches
a lag of 180°, the phase difference becomes
ambiguous. This occurs for f > 1.5 kHz.
Max natural ITD in human: ~ 680!s
(c " 340 m/s; max. path difference " 0.23 m)
Binaural Cues for Sound Localization
Interaural Level Difference (ILD):
Sound pressure level differences between the two
ears indicate source location primarily for highfrequency sounds. The head starts to create an
acoustic shadow at frequencies above about 500
Hz. The effect is most apparent above 2 kHz.
Human data:
Medial Superior Olive
processes ITD
ITD circuit:
MSO receives excitatory
inputs from both sides.
MSO cells are coincidence
detectors
Delay line - Jeffress
Delay line
Lateral Superior Olive
processes ILD
ILD circuit:
LSO receives excitatory &
inhibitory inputs from
opposite sides
Sound localization in the barn owl
Barn owls can detect a 1-msec ITD and catch
mice in complete darkness. Their external
ears are highly asymmetric, allowing them to
utilize ILD for localizing in the median
vertical plane. Their localization along the
horizontal plane is largely based on ITD even
for frequencies as high as 8 kHz.
Unique specialization of binaural cues in the barn owl
The barn owl's asymmetric external ears
mean that ILD provides a cue for elevation,
whereas ITD is a better cue for azimuth.
Spaced-tuned neurons in barn-owl midbrain
• Neurons in the barn owl’s external nucleus
of the inferior colliculus (ICx) are tuned to a
specific azimuth and elevation. Unlike
neurons in the central nucleus (ICc), they
have broad frequency tuning.
• Azimuth tuning of ICx neurons depends
primarily on ITD, whereas elevation tuning
depends primarily on ILD (not shown).
Knudsen and Konishi (1978)
Midbrain map of auditory space
• The barn owl’s ICx (the region which
contains space-tuned neurons) forms a
neural map of acoustic space.
• Contralateral azimuths are represented
caudally within ICx, and medial
azimuths rostrally (C). Low elevations
are represented ventrally, and high
elevations dorsally.
• There are also maps of auditory space
in both mammalian and avian superior
colliculi. On the other hand, no
convincing space map has been found
in the mammalian IC.
Barn owls are specialized for phase locking at high frequencies
• Barn owl auditory-nerve fibers
show a much higher upper
frequency limit of phase locking
(as measured by the vector
strength) than any other species
of birds or mammals.
Köppl (1997)
Two separate auditory pathways in the barn owl
Forebrain
Thalamus
Midbrain
• The barn owl’s auditory system processes
ITD and ILD in separate parallel pathways —
via n. angularis & n. magnocellularis.
• Both pathways and different frequency
bands converge in the auditory midbrain,
creating a map of auditory space based on
selectivity to combinations of ITD and ILD.
• This map projects to the optic tectum (OT)
to form a bimodal map of space.
Hindbrain
Summary of specializations for sound localization
in the barn owl
Barn owls show specializations for sound localizations that
distinguish them from mammals and other birds:
• Asymmetric arrangement of the external ears allows ILD to
code elevation
• Neural phase locking at high frequencies
• Separate pathways for ITD and ILD processing
• Spaced-tuned neurons with broad frequency selectivity
• Neural map of auditory space in the IC
ITD processing may be different in birds and mammals
ITD
• Birds and mammals independently evolved the
ability to hear airborne sounds
• The Jeffress model is consistent with data from
studies in birds, but less consistent for
mammals.
Jeffress model features
• phase-locked excitatory inputs
• coincidence detection
• delay line
for different tone frequencies:
• ITD curves are narrower for higher frequencies (green)
• curves coincide at one “characteristic delay”
• for pure excitation, characteristic delay is at a common peak
Different strategies for encoding ITDs
bird laminaris
mammalian MSO
Avian ITD detection
Evidence for delay lines in mammals
meandering route of a AVCN spherical
bushy cell projection to ipsilateral MSO
Reconstruction of the course of an intracellularly-filled
spherical bushy cell axon and its primary collaterals in the
ipsilateral brainstem, illustrating the "round-about'' route
taken by the collateral innervating the LSO (*) and MSO.
Arrows indicate the direction of action potential conduction along the main axon and
its collaterals. Axon collaterals going to MSO and LSO are not drawn to their
termination sites for the sake of simplicity.
Implication: the indirect path accomplishes the necessary
delay in time to allow for coincidence at MSO with activity
conducted from the more distant contralateral AVCN.
400!m
Smith PH, Joris PX and Yin TCT. Projections of physiologically characterized
spherical bushy cell axons from the cochlear nucleus of the cat: Evidence for
delay lines to the medial superior olive. J. Comp. Neurol. 1993, 331:245-260.
Evidence for delay lines in mammals
CONTRALATERAL
Smith, Joris & Yin (1993)
projection to contralateral MSO looks
like a delay line for cells # 1, 2, 3.
Cells 4-9 show different organization.
IPSILATERAL
projection to ipsilateral MSO is
inconsistent with the Jeffress Model
in all cases (cells 10-16)
One interpretation:
the contralateral path creates the delay line (although perhaps not
quite as Jeffress speculated), and the indirect path accomplishes the
necessary time delay to allow for coincidence at MSO.
Another interpretation:
the best evidence for delay lines in mammals is very unconvincing.
MSO receives bilateral
excitation and inhibition
• MSO principal cells receive excitatory
inputs from spherical bushy cells in
AVCN bilaterally.
• MSO cells also receive some
inhibitory inputs from the ipsilateral
MNTB and LNTB. The role of these
inhibitory inputs is unknown.
• MSO principal neurons are bipolar,
with both dendrites largely covered by
excitatory synaptic terminals from
AVCN. Inputs from the two AVCNs are
segregated, each one ending on one of
the two dendrites of MSO cells.
• MSO neurons project to the ipsilateral
central nucleus of the inferior colliculus
(ICC) and dorsal nucleus of the lateral
lemniscus (DNLL).
Stotler (1953)
Evidence for
inhibition in MSO
responses
• The coincidence detection model of
MSO neurons does not account for all
aspects of the responses: When the ILD
of a binaural tone stimulus is varied at a
constant ITD, responses of some MSO
neurons vary non-monotonically.
• Since the excitatory inputs to the MSO
(AVCN spherical bushy cells) have
largely monotonic responses, the
nonmonotonicities in MSO responses
suggest an effect of inhibitory inputs,
which are known to exist anatomically.
• The function of inhibitory inputs to the
MSO in binaural processing is an active
area of research.
Yin and Chan (1990)
Different strategies for encoding ITDs
bird laminaris
mammalian MSO
Mammalian ITD detection
Inhibitory inputs to MSO are refined during development
Juvenile
Adult
effect of blocking inhibition:
Role of inhibition in MSO
slope shifts outside the physiological range
Possible mechanism for this effect:
well-timed, phase-locked inhibition, bilaterally
End bulb of Held: Medial Nucleus of Trapezoid Body
Large axons synapse on principal
cells of the contralateral MNTB at
the calyx of Held, a glutamatergic
synaptic terminal.
MNTB cells are glycinergic and provide
contralateral inhibition to LSO, which
also receives excitatory input from AVCN.
The calyx is an exceptionally reliable and
temporally precise synapse, allowing
MNTB to as a fast, sign-inverting relay
station that maintains phase locking.
LSO neurons are “EI”
ipsi-excited, contra-inhibited
Tuning curves for ipsilateral excitation and
contralateral inhibition closely match.
the response to a fixed ipsi stimulus is inhibited
when the contra stimulus is loud enough
Latencies of ipsi-excitation and contra-inhibition
typically match, despite the added distance and
intervening synapse for contra-inhibition.
Cellular specializations for precise timing in auditory pathway
AVCN bushy cells have two specializations:
1.inputs via end bulbs of Held which ensure
secure synaptic transmission
2.a low-threshold K+ conductance that rapidly
resets the membrane voltage after each
spike or EPSP.
MNTB has similar specializations:
1.input via calyx Held.
2.a low-threshold K+ conductance that rapidly resets
the membrane voltage after each spike or EPSP.
Calyx of Held
preserves timing
EPSCs are much shorter in VCN bushy
cells than in hippocampal neurons