Frog Auditory System: Superior Olivary Nucleus (SON)
by David D. Olmsted (Copyright - 1998, 2000, 2006. Free to use for personal and
educational purposes)
Last Revised November 6, 2006
Differences Between the Superior Olive and the Dorsal Medullary Nucleus
The Superior Olivary Nucleus
does not exist in either fish or salamanders (Campbell and Boord - 1974).
Salamanders make little or no sounds and like early mammals rely heavily on olfactory
cues for mating and other behaviors. Consequently, the Superior
Olivary Nucleus
seems to have evolved to accomplish the additional auditory discrimination needed
for call recognition and localization in frogs, toads (anurans) and later species.
Compared to the Dorsal Medullary Nucleus, the Superior Olive has twice the percentage
of neurons exhibiting a transient phasic response for signaling the beginning of
a sound burst. In the leopard frog Rana pipiens, 23 % of the Superior Olive
neurons were phasic compared to 11% for the Dorsal Medullary Nucleus
(Fuzessery and Feng - 1983, page 113). Nearly the same percentage holds for
the green tree frog Hyla cinerea (Feng and Capranica - 1978).
The Superior Olivary Nucleus in the frog receives all its auditory information from
the Dorsal Medullary Nucleus (DMN) since does not receive any projections from the
eighth nerve. This fact is further supported by its longer and more varied neuron
response times than the 3 to 10 milliseconds typical of neurons in the Dorsal Medullary Nucleus
as reported below by Fuzessery and Feng (1983, page 113):
"The response latencies
of SON neurons showed a greater distribution than DMN neurons, ranging from
10-50 ms. The majority of SON neurons (60%) had response latencies clustered
at the low end of this range (10-15 ms)."
Monaural Interactions
In general, the tuning curves of the Superior Olivary Nucleus shown in figures 1 and 2 are even more complex then those seen in the
Dorsal Medullary Nucleus.
The dark solid line with filled circles marks the minimum sound level required
to activate that neuron and thus represents its excitatory tuning curve. The
dashed lines represent the sound levels required at those frequencies to totally
inhibit the neuron activated to a dB level indicated by the number next to the line.
Full auditory
templates are shown in illustrations "A" and "B" at the top of figure 1 for
they have inhibitory regions (dashed lines) on BOTH sides of the facilatory
region (dark solid line). In contrast most of the templates in the Dorsal Medullary
Nucleus
are partial templates with an inhibitory region on only one side. Consequently,
these responses are most likely formed by the inhibition of a partial template
response by a plain tuning curve response. In illustration "D" the small dark
filled circles in the middle of the tuning curve indicate the shift in frequency
of the excitatory tone required to produce the maximum response at 10, 20, and
30 dB.
The "W" shaped tuning curve of illustration "D" of figure 2 clearly
shows the combination of two simpler tuning curves using either a summation or a
multivalued logic INCLUSIVE OR operation.
Figure 1
Tuning Curves in the Superior Olivary Nucleus of Rana pipiens> Showing Inhibition. (Fuzessery and Feng - 1983).
 |
Figure 2
Tuning Curves in the Superior Olivary Nucleus of Rana pipiens> NOT Showing Inhibition. (Fuzessery and Feng - 1983).
Same as above but the narrow solid lines represent the percent of maximum response level
 |
Out of 94 neurons isolated in Rana pipiens
23% of the neurons in the Superior Olivary nucleus showed phasic-on responses
while only 4% showed phasic-off responses. Phasic-on neurons produce a transient
response of action potentials at the beginning of a sound signal while phasic-off
neurons produce transient responses after the sound ends.
Yet, significantly three neurons defined as transition neurons by the researchers
changed their firing characteristics in response to intensity changes. Two of these
are shown in figure 3. Illustration "A" to the left shows the tuning curve of a
neuron which had a tonic response at one intensity level and a phasic response
at an intensity level only 10 dB higher. Illustration "B" to the right shows a neuron
with a phasic-on response at the lower intensity which changes to a phasic-off
response at higher intensities. This neuron had a narrow intensity transition zone in which both phasic-on
and phasic-off responses were seen.
The
phasic-on neurons generally produced a pulse lasting 10 milliseconds (Feng and
Capranica - 1978).
Figure 3
Tuning Curves for Two Kinds of Transition Neurons in the Superior Olivary Nucleus of Rana pipiens. (Fuzessery and Feng - 1983).
 |
The thresholds and frequency distributions for
the neurons in the Superior Olivary
Nucleus are shown in figure 4. Like the distributions in the Dorsal Medullary Nucleus
it shows clusterings at
both the high and low frequencies. Unlike the
threshold distribution
in the Dorsal Medullary Nucleus it shows a greater percentage of neurons clustered at the higher frequencies.
In addition this higher frequency cluster now contains neurons having inhibition
where none were found in the Dorsal Medullary nucleus.
Figure 4
Distribution of the Best Excitatory Frequencies (BEF's) and Their Thresholds in the Superior Olivary Nucleus of Rana pipiens . (Fuzessery and Feng - 1983).
Open circles represent neurons inhibited by higher frequencies. Filled circles represent neurons exhibiting no or little inhibition. Stars represent neurons for which inhibition was not tested.
 |
Binaural Interactions
Like the binaural neurons in the Dorsal Medullary Nucleus the tuning curves for
binaural neurons in the Superior Olivary Nucleus match up for both the EE neurons
having excitatory inputs from both ears and the EI neurons having an excitatory
input from one ear and an inhibitory input from the other (Feng and Capranica -1978).
They report this on page 46 regarding the EI neurons:
"In all but two EI cells,
the best excitatory frequency and the best inhibitory frequency (BIF) were
within 0.25 octave of each other. for the other two cells, the BEF and the
BIF differed by more than 1 octave."
They tested 146 neurons in the Green Tree frog (Hyla cinerea). Of these, 85 were monaural with 67 excited by the ear on the
opposite side of the head and 61 were binaural with only 11 being EE cells, 6 of those being
complex (explained below).
The rest of the binaural neurons showed EI responses. Of the EI neurons 3 were
excited from the ear on the same side and 41 were excited from the ear on
the opposite side.
One of the six binaural EE neurons shown in figure 5 seems to show a response closer to the definition of the multivalued (fuzzy) logic
INCLUSIVE OR operation than two a summation operation.
The multivalued logic INCLUSIVE OR operation passes the greatest value among its
inputs. This figure shows the average responses
of a single neuron (20 trials at 1 per second) to pure 500 Hz tones presented
via headphones for 200 milliseconds. The IL tone was presented to ipsilateral ear
(same side) while the CL tone was presented to the contralateral ear (opposite side).
The CL + IL tone was presented to both ears. The auditory threshold for the
contralateral ear was 44 dB and for the ipsilateral ear 49 dB which is the
reason for the lower spike rates of the ipsilateral ear at the same sound intensity.
This neuron was maximally responsive to 500 Hz tones. Spontaneous firing rate for
this neuron was about 1 spike per stimulus duration.
Figure 5
A Multivalued Logic INCLUSIVE OR Neuron in the Superior Olivary Nucleus of Hyla cinerea. (figure 6 of Feng and Capranica - 1978) .
 |
Out of the 6 binaural neurons only one showed the classic summation of its inputs and the response
of this neuron is shown in figure 6. This result occurred
with the same test setup as figure 5. Both ears had similar tuning curves
with a best excitatory frequency at 225 Hz at a threshold of 45 dB. The spontaneous
neural firing rate was less than one spike per stimulus duration.
Figure 6
Summation Neuron in the Superior Olivary Nucleus of Hyla cinerea. (Feng and Capranica - 1978) .
 |
The remaining four binaural neurons showed intermediate
responses between the INCLUSIVE
OR and summation responses like the neuron shown in figure 7. Interestingly the response shown in figure 7 is
more like an INCLUSIVE OR at low intensities and more like a sum at high intensities.
This is shown below by listing the percent deviations for the three sample
points: The difference away from the ideal INCLUSIVE OR response
could be due to the hidden effects of other active biases or due to the past tuning
of the neuron's membrane charge parameters allowing it to fit the typical timing
constraints of its neural circuit.
-
Left: 14 + 22 = 36 (actual combined response is 25) Percent sum difference
from actual = 44%, Percent INCLUSIVE OR differencefrom actual = 12%
- Middle:
18 + 26 = 44 (actual combined response is 33) Percent sum difference from actual
= 33%, Percent INCLUSIVE OR difference from actual = 21%
- Right: 28 + 29 = 57 (actual
combined response is 46) Percent sum difference from actual = 20%, Percent
INCLUSIVE OR difference from actual = 37%
Figure 7
Partial Multivalued INCLUSIVE OR Neuron in the Superior Olivary Nucleus of Hyla cinerea. (Feng and Capranica - 1978) .
 |
All six complex cells found by Feng and Capranica (1978)
had best excitatory frequencies below 2 kHz. An example of a complex cell
response is given in figure 8. This neuron
was excited by the contralateral (opposite side) ear and had a best excitatory frequency
at 575 Hz as shown by the peak in the curve. It had sound response threshold of
40 dB shown in the figure by the dotted line. The dark solid line gives the average
response when both ears are stimulated (20 repetitions with a stimulus duration
of 300 milliseconds repeated once a second). Notice that this is not a tuning curve but is a response
curve giving the average neuron firing rate versus frequency.
The response of this particular neuron has a narrow
band of inhibition just to the
right of its best excitatory frequency. What is especially significant
here is that the inhibition is so narrow, much narrower than is possible from
any single inhibitory
neuron's tuning curve. Why it was inhibited is not known.
Figure 8
Complex Binaural Neuron in the Superior Olivary Nucleus of Hyla cinerea Showing Response to Stimulation of Both Ears. (Feng and Capranica - 1978) .
 |
Responses of EI binaural neurons in the Superior Olivary nucleus to auditory clicks
of two different intensity differences between the ears are shown in figure 9. The vertical axis gives the probability of generating
an action potential (spike). It has a best excitatory frequency at 3,000 Hz. Clicks
were 0.2 milliseconds wide and presented at 5 per second. Data points are the average
response for 100 clicks. The contralateral (CL) ear was facilatory and by itself
it had a firing probability of 0.72 with a click intensity of -35dB and 0.38 with a
click intensity of -45dB relative to the sound level produced by an 8 volt pulse to the earphone. In
figure 9 the top black dot curve gives its probability of response when the average
sound level was -35dB. The bottom open dot curve gives its probability of
response when the average sound level was -45dB. In each case the probability of
firing off an impulse was reduced as the sound of the opposite ear became relatively
more intense. But notice that in the 4 to 6 dB difference range at the tested 3,000
Hz which head turning
produces the curves are flat indicating that this neuron is not used for sound localization
behaviors. Later Albert Feng (1980) measured the difference in
sound intensity between the sides of a the head of Rana pipiens to be 4 dB at
1900 Hz and 1 to 2 dB at 170 Hz.
Since this click only produces one action potential several neurons must converge
onto this one EI neuron forming a spatial summation to produce the indicated effects
(as opposed to a temporal summation). The more intense the click the greater is the probability
that these feeder neurons will produce an action potential for summation by the
EI neuron which in turn responds in a probabilistic fashion proportional
to the number of its action potential inputs.
Figure 9
Response of an Excitatory-Inhibitory (EI) Neuron to Clicks of Varying Intensity Differences in the Superior Olivary Nucleus of Hyla cinerea (Feng and Capranica - 1978).
 |
The neuron shown in figure 10 shows an amplification effect when the sound reaching
the ear on the same side as the neuron both leads and is louder than the sound reaching
the other ear. In this case the probability of firing is greater than when the sound
is less on the same side relative to the other ear. This neuron has a best excitatory
frequency at 675 Hz. The clicks were 0.2 milliseconds wide and presented at 5 per
second. The data
points are the average response for 100 clicks. The contralateral (CL) ear
was facilatory and by itself it had a firing probability of 0.70 to a click
intensity of -68 dB relative to the sound level produced by an 8 volt
pulse to the earphone.
At the tested frequency of 675 Hz the time difference between the sound wave peaks
is 0.15 milliseconds well within range of the amplification effect shown by this neuron.
Yet if the frog's ears are spaced 2 cm apart the time difference for one
peak to pass through the air from one ear to the other is 60 microseconds (sound travels at 332 m/sec at 0 degrees C). The Hyla cinerea frogs used in this experiment ranged in size
from 3 to 5 cm in length resulting in a head width from 1 to 1.7 cm
in width. So the time difference between ears would seem to be too small
for this neuron to be involved in sound localization. Also arguing against sound
localization is the fact that the sound intensity difference from head turning will
only be around 2 to 3 dB
and not the 5 dB used in the experiment. Later Albert Feng
(1980) measured the difference in sound intensity between the sides of a the head
of Rana pipiens to be 4 dB at
1900 Hz and 1 to 2 dB at 170 Hz. So again this neuron is probably not used for sound
localization but instead is used for something else.
Interestingly, only those neurons having low (below 2,600 Hz) best excitatory
frequencies (BEF's) showed any variation in response to time differences.
Figure 10
Sound Intensity Amplification by Click Timing Differences in an Excitatory-Inhibitory (EI) Neuron in the Superior Olivary Nucleus of Hyla cinerea (Feng and Capranica - 1978) .
 |
Connections
The sizes of the neurons in and near the superior olive are shown in figure 11.
Figure 11
Neuron Sizes Near the Superior Olivary Nucleus of an Adult Rana pipiens. (figure 26 of Larsell - 1934).
Neurons stained with the golgi method and shown at a magnification of 30. nu. VIIId - dorsal medullary nucleus, r. VII - root of the eighth nerve, su. ol. - superior olive
 |
For many years a mystery surrounded the reported connections of the Superior Olive
with the Tectum (responsible for interesting object localization).
The first HRP
study involving the tectum of Rana pipiens by Wilczynsky and Northcutt (1977)
reported that the tectum sends some axons to the dendritic field (neuropil)
of the superior olive. They say this on page 226:
"A large number of ipsilateral
fibers terminated in the Superior Olive staining the neuropil areas a pale
brown."
Later a study with HRP and cobalt-filling in Rana esculanta and Rana
ridibunda
by Lazar, Toth, Csank, and Kicliter (1983) reported the same thing. They
say this on page 116:
"Many fibers terminated in the Superior Olive and the Medial Reticular Nucleus."
Yet an even later HRP study by Masino and Grobstein (1990) reported
the opposite conclusion. They say this:
"Some of these collaterals (axon branches
from the Tectum) cross the midline to terminate in the vicinity of the Olive on
the opposite side of the brain. Previous reports have suggested that this
collateralization reflects a bilateral input to the Olive itself. Our material
shows the fibers to extend well beyond the olivary cell group and terminate as well
near cells of the Medial Reticular Nucleus. In fact, most of the observed
bouton-like swelling were closer to Medial Reticular
neurons than to olivary
neurons."
The findings that the tectum sends projections to the Superior Olive is
probably the result of killing the frogs too early so that the HRP did not
have a chance to be transported to the end of the axons. Wilczynski and Northcutt
killed their frogs between 2 to 9 days while Lazar, Toth, Csank, and Kicliter did
so between 2 and 6 days. Neither gives the survival times of the frogs from which
they drew their conclusions. In contrast the frogs of Masino and Grobstein
lived for 5 to 9 days before being killed. Wilczynski and Northcutt report
that the optimal survival time was between 5 and 6 days for by 9 days the HRP in
the neurons had begun to fade (page 222).
The lack of a direct projection
to the Superior Olive from the Tectum is confirmed by an HRP study of the Superior Olive itself by Albert Feng (1986). Yet, the association of visual
cues with auditory cues should occur somewhere with the Torus Semicircularis
being
the best candidate. Such an association would be expected to enhance the identification
and localization of a buzzing insect or a moving mate.
The connections of the Superior Olivary Nucleus as determined by HRP injection into that nucleus are shown below
in figure 12 (Feng- 1986). Its major projections are back to the Dorsal Medullary Nucleus on each side, sideways to the Superior Olivary
Nucleus on the opposite side, forward to the Lateral Lemniscal Nucleus on both sides, forward to the various Torus Semicircularis nuclei: { the Pinciple Nucleus (both
sides), the Laminar Nucleus (same side) and the Magnocellular Nucleus (same side)}, forward to the midbrain Tegmentum (same side), and finally forward to the Posterior
and Central Thalamic Nuclei (same side).
Of these, those sending return projections are: the Dorsal Medullary Nucleus (both
sides), the Torus Semicircularis - Principle, Laminar, and Magnocellular nuclei
(same side), the Midbrain Tegmentum (same side), and the Posterior Thalamic Nucleus.
Significantly, neither the Torus Semicircularis on the opposite
side nor the central Thalamic Nucleus sends return projections.
Centers which
send projections to the the Superior Olivary Nucleus but which do not receive
any in return are the Reticular Formation and the CCaudalis Nucleus
in the Medulla located
just tailward (caudal) to the Dorsal Medullary Nucleus.
The projections from the Superior Olivary Nucleus to the Dorsal Medullary Nucleus
are organized tonotopically
meaning that low frequency regions send projections to low frequency
regions and high frequency regions send projections to high frequency regions.
Figure 12 below shows the projections from the low frequency region in the Superior Olivary Nucleus. Notice that the resulting projection site in the Dorsal
Medullary Nucleus on both sides is a line parallel to the axis of the bushy neurons. A similar
line is seen in the projection site in the Superior Olivary nucleus on
the opposite side of the brain. Such line projection sites are also seen in the
Laminar and Principle Nuclei of the Torus Semicircularis (same side), the midbrain
Tegmentum (same side) and the two Thalamic nuclei.
Significantly, the bilateral
projection site to the Principle Nucleus of the Torus
Semicircularis is a circle
with the low frequency sites exhibiting a small circle while the high frequency
sites exhibit a larger circle. As pointed out by Albert Feng this means that
the frequency projections overlap the line projections in the Principle Nucleus of the Torus. This is the first time such convergence of differing frequencies
occurs in the auditory system.
Finally, the forward projection of the Superior Olivary Nucleus to the laminar nucleus
of the Torus Semicircularis terminates on
the cell bodies (suggesting inhibition) while the projections to the other Torus nuclei terminate on the dendrites. Likewise the projections back to
the Dorsal Medullary nuclei on both sides also terminated on the cell bodies
again indicating inhibition.
Figure 12
The Connections of the Superior Olivary Nucleus of the Leopard Frog Rana pipiens. (Feng - 1986)
Axon terminals are represented by crosses, axons by lines, and cell bodies by filled triangles. Illustrations (a) to (j) represent a caudal (tailward) to rostal (headward) direction.
(A - Aquaduct, C - Central Thalamic nucleus, CER - Cerebellum, DN - Dorsal Medullary Nucleus, H - Hypothalamus, LLN - Lateral Lemniscal Nucleus, MT - Midbrain Tegmental Nucleus, Nc - caudalis nucleus in the medualla, Ni Nucleus Isthmi, OT - Optic Tectum, P - Posterior Thalamic nucleus, PG - Pretectal Gray, R - Medullary Reticular Formation, SO Superior Olivary Nucleus, Tl - Laminar Nucleus of the Torus Semicircularis, Tmc - Magnocellular Nucleus of the Torus Semicircularis, Tp - Principle Nucleus of the Torus Semicircularis, TV - Tectal Ventricle, VIIIth - Eighth Cranial Nerve, VN - Ventral Nucleus)
 |
References
Campbell, C.B.G. and Boord, R.L. (1974)
The Central Auditory Pathways of Non-mammalian Vertebrates. In: Handbook of
Sensory Physiology, Volume V, edited by W.D. Keidel and W.D. Neff. Berlin:
Springer, pages 337-362
Feng, A.S. and Capranica, R.R. (1978) Sound Localization
in Anurans. II. binaural Interaction in Superior Olivary Nucleus of the Green
Tree Frog (Hyla cinerea). J. Comp. Physiol. 41:43-54
Feng, A.S.(1980) Directional
Characteristics of the Acoustic Receiver of the Leopard frog (Rana pipiens):
a Study of Eighth Nerve Auditory Responses J. Acoust. Soc. Am. 68:1107-1114
Feng, A.S. (1986) Afferent and Efferent Innervation Patterns of the Superior
Olivary Nucleus of the Leopard Frog. Brain Res. 364:167-171
Fuzessery, Z.M. and
Feng, A.S. (1983) Frequency Selectrivity in the Anuran Medulla: Excitatory
and Inhibitory Tuning Properties of Single Neurons. J. Comp. Physiol. 150:107-119
Larsell, O. (1934) The Differentiation of the Peripheral and Central Acoustic
Apparatus in the Frog. J. Comp Neurol. 60: 473-527
Lazar, GY., Toth, P., Csank,
GY. and Kicliter, E. (1983) Morphology and Location of Tectal Projection neurons
in Frogs: A Study with HRP and Cobalt-Filling. J. Comp. Neurol. 215:108-120
Masino, T., and Grobstein, P. (1990) Tectal Connectivity in the Frog Rana pipiens:
Tectotegmental Projections and a General Analysis of Topographic
Organization. J. Comp. Neurol. 291:103-127
Shepherd, Gordon M. (ed.) 1998. The Synaptic
Organization of the Brain (fourth edition), Oxford University Press.
Wilczynski,
W. and Northcutt, R. G. (1977) Afferents to the Optic Tectum of the Leopard
Frog: An HRP Study. J. Comp. Neurol. 173:219-229