Frog Auditory System: Dorsal Medullary Nucleus (DMN)
by David D. Olmsted (Copyright - 1998, 2000, 2006. Free to use for personal and
Last Revised November 4, 2006
medullary nucleus is analogous to the cochlear nucleus in mammals because
the first region of the frog brain where inputs from each ear converge. These inputs
arrive from each ear via the eighth cranial nerve. The tuning curves for representative
neurons found in the dorsal medullary nucleus are shown below in figures 1 and 2. In figure 1 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 to the right are the sound levels required at those
frequencies to totally inhibit the neuron activated to a level indicated by its
dB number next to the line. The thin solid lines give
the percent of maximum firing rate. In "A" the excitatory tuning curve
probably extends to even lower frequencies than could be measured for it never
stopped firing unless it was inhibited at frequencies marked "ambient".
Compared to the eighth nerve tuning curves these tend to be more complex and are
spread over the whole frequency range where the eighth nerve tuning curves cover
only half of the frequency range. This clearly indicates that some sort of signal
combination is taking place. The two sets of tuning curves are the same in that
the inhibitory tuning curves (dashed lines) rise in step with the sound intensity
applied to the excitatory curves. The inhibitory tuning curves rise approximately
10 dB for every increase of 10 dB in the excitatory sound level indicating
that they work via inhibition.
Tuning Curves Showing Two Tone Inhibition from Single Neurons in the Dorsal Medullary Nucleus of Rana pipiens for Tones Presented to One Ear. (Fuzessery and Feng - 1983).
Tuning Curves NOT Showing Two Tone Inhibition from Single Neurons in the Dorsal Medullary Nucleus of Rana pipiens for Tones Presented to One Ear. (Fuzessery and Feng - 1983).
The auditory thresholds for 142 neural cells in Rana catesbeiana at their
best excitatory frequencies ranged from
22 to 155 dB as found by Albert Feng and
Robert Capranica (1976). The thresholds for a smaller sampling of Dorsal Medullary
Nucleus in Rana
pipiens are shown below in figure 3. Open circles represent neurons inhibited
by higher frequencies. Filled circles represent neurons exhibiting no inhibition
and stars represent neurons for which inhibition was not tested.
Distribution of the Best Excitatory Frequencies (BEF's) and Their Thresholds in the Dorsal Medullary Nucleus of Rana pipiens. (Fuzessery and Feng - 1983).
Figure 4 gives the sound response latencies for each of the frog's ears. The dashed
line gives the response latencies for sounds through the ear from the same
side of the head (ipsilateral) while the solid line gives the response latencies
for sound through the ear on the opposite side of the head (contralateral).
Latencies were measured at 10 dB above sound threshold with single tone burst having
1 millisecond rise times. In general the sound latencies from the
opposite ear are 2 milliseconds longer than those from the ear on the same side.
Also the distribution is quit large (5 milliseconds) for a median response
time of 7 milliseconds.
Response Time Latencies for 30 Monaural Cells in the Dorsal Medullary Nucleus of Rana catesbeiana. (Feng and Capranica - 1976)
In a major study of the Dorsal Medullary Nucleus in the frog Rana catesbeiana,
Albert Feng and Robert Capranica (1976) managed
to isolate a total of 142 neural cells. Most were spontaneously active averaging
5 spikes per second.
Nearly half (75 cells) responded only to stimulation of one
ear with 80% of those
responding to stimulation from the ear on the same side
(ipsilateral) as the Dorsal
Medullary Nucleus while the remaining 20% responded
to inputs from the ear on the opposite side (contralateral). The other 67
neural cells responded to inputs from both ears with 51 of these binaural cells
excited by the contralateral ear and inhibited by the ipsilateral ear (EI cells)
while only 2 binaural cells were excited by the ipsilateral ear. The remaining 14
could be excited by either ear (EE cells).
Figures 5 and 6 show that the tuning
curves for both the excitatory-inhibitory (EI) and excitatory-excitatory (EE)
neurons match up in terms of frequency response.
This means that the axons from similar sensory cells from
opposite ears converge upon the same neuron. The dark circles in each figure represent the excitatory
tuning curve when the contralateral (opposite side) ear was stimulated in Rana catesbeiana.
The open circles represent the tuning curve in which stimulation of the ipsilateral
(same side) ear inhibited 75% of a contralateral signal at 510 Hz and 10dB above
its threshold. These tuning curves are similar to "A" in figure 2
Tuning Curves Match up for an EI Neuron in the Dorsal Medullary Nucleus Having Excitatory Inputs from an Ear on One Side and Inhibitory Inputs from the Other Ear. (Feng and Capranica - 1976)
Tuning Curves also Match up for an EE Neuron in the Dorsal Medullary Nucleus Having Excitatory Inputs from Each Ear. (Feng and Capranica - 1976)
Significantly, the endings of the eighth nerve axons in the Dorsal Medullary Nucleus
from the ears are organized approximately tonotopically
with low frequencies located ventrally (stomach side) and high frequencies located
dorsally (spinal cord side). Consequently, fibers from the basilar papilla responsible
for high frequency inputs are found in the extreme dorsomedial region while
those from the amphibian papilla responsible for the other frequencies are found
throughout most of the nucleus (Fuzessery and Feng - 1981). Yet a single axon from
the amphibian papilla terminates over a smaller area than a single axons from
the basilar papilla (Lewis, et al - 1980). The neurons do not seem to have any positional
organization with regards to threshold amount as Feng and Capranica observed
(1976 - page 876):
"In a single electrode pass two neighboring units with
similar best frequencies often were encountered with thresholds differing by 40
dB or more."
Unfortunately the Feng and Capranica (1976) did not mention what
the responses of the figure 5 EE neurons were when both excitatory inputs
are simultaneously activated. Consequently one cannot tell whether this is an INCLUSIVE
OR neuron or a simple summation neuron. Yet in a paper two years later in
which they investigate the Superior Olivary Nucleus they do describe the properties
of EE neurons there and those neurons do show the classic response of multivalued
(fuzzy) logic INCLUSIVE OR operations within 20% of the ideal definition of passing
the greatest valued input (Feng and Capranica - 1978b)
Figure 7 shows the firing
rate response of two representative excitatory - inhibitory (EI) neurons. The " IIL - ICL " represents the difference in sound intensity between
the ipsilateral (same side) ear and the contralateral (opposite side) ear.
The dashed line represents the firing rate when the contralateral ear alone
is stimulated at 10 dB above threshold. Unit cell 26-12 has a best excitatory
frequency (BEF) as well as a best inhibitory frequency from the ipsilateral
ear at 435 Hz with an excitatory threshold of 25 dB. Unit cell 30-7 has a
best excitatory frequency (BEF) at 575 Hz with a threshold of 23 dB.
Firing Rate Characteristics of Two Excitatory-Inhibitory (EI) Neurons in Rana catesbeiana. (Feng and Capranica - 1976)
Notice the difference in output sensitivity between the two neurons. The top cell's
(unit 26-12) output is reduced by 30 spikes per second with a 10 dB difference
while the bottom cell's (unit 30-7) output is reduced by only 15 spikes per
second over a greater 25 dB difference. The bottom cell type is the most common.
The size of the recording electrodes used by Feng and Capranica (1976) was recording
neural cell bodies. They say this on page 878:
"According to Guinan, et al (1972),
extracellular activity picked up with large microelectrodes most probably
originates from cell bodies, while extracellular activity recorded with small microelectrodes
can be from fibers or cell bodies. Since the tips of our recording electrodes were
quite large (4-10 um), the neural activity that we detected likely represents
the response properties of cell bodies and not of afferent fibers or passing fibers
originating in the contralateral ear. We have tried to use similar indium-filled
micropipettes in the eighth nerve but seldom could we isolate spike activity."
distributions of the best excitatory frequencies for the monaural neurons shown
in figure 8. The top histogram (A) shows best excitatory frequencies
from 75 monaural (single input) neurons. The bottom histogram (B) shows best
excitatory frequencies from 67 binaural (two input) neurons. The frequency
sample size bin width is 100 Hz. The neural responses shown at the top are similar to the distribution of the best
excitatory frequencies in the eighth nerve for Rana catesbeiana. Both show three
primary groupings with most neurons responsive to the lower frequency ranges. This is in contrast to the distribution of best excitatory frequencies
of binaural cells at the bottom of figure which shows only one large cluster.
This low frequency binaural clustering certainly
suggests that most of the binaural cells in the Dorsal Medullary Nucleus have a purpose
different from call recognition which would require the neuron to
combine a variety of frequencies from all frequency
bands. Yet neither would the major purpose be sound localization as suggested
by the researchers. If this was the case then most of the neuron responses
would be at the high frequency range since that provides a better sound intensity
differential. 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. The most likely purpose is some sort of dynamic response filtering
based upon the detected sound clutter in the environment and controlled by sound
centers further downstream. This is not to say sound localization does not occur here, just that
it is not the function of the majority of the neurons.
Distributions of Monaural (top) and Binaural (bottom) Cells in the Dorsal Medullary Nucleus of Rana catesbeiana. (Feng and Capranica - 1976).
The general response characteristics are shown in figure 9 for a different species
of frog, Rana pipiens. “A” shows the best excitatory frequencies for all
neurons, “B” shows the threshold distribution, “C” shows the width of tuning curves
at 10 dB also known as the Q factor, and “D” gives the spontaneous firing rates.
Distribution of Various Neuron Characteristics in the Dorsal Medullary Nucleus of Rana pipiens. (Feng and Capranica - 1976).
Temporal Responses of the Neurons
Feng and Capranica (1976)
examined the temporal firing pattern of 79 binaural and monaural neural cells in
the bullfrog Rana catesbeiana at their best excitatory frequencies. Of these:
"59 cells responded tonically throughout the duration of the tone burst. The rest
of the cells responded phasically: 19 of these cells fired only at the onset
of a tone burst with 1-2 spikes, independent of stimulus intensity, while 1 unit
was found which responded only to the offset of a tone burst.(page 876)"
of the sampled cells in Rana catesbeiana are tonic. This compares to the 89% figure
found in Rana pipiens in which 62 out of 70 neurons had tonic responses to single
frequency tones 100 milliseconds long (Fuzessery and Feng - 1983). In this study
Fuzessery and Feng observed that most of the phasic-on neurons displayed a tonic
firing pattern in response to noise which they suggested was due to the ever changing
intensity of differing frequencies (page 109). This suggests that these phasic neurons
may be used to characterize either the frequency sweeps or the fast pulsating sounds
often found in frog calls. The best excitatory frequencies for the phasic
neurons were in the lower frequencies and no phasic-off neurons were found.
example of filter type neurons which pass their input signals in the absence of
certain inhbitory characteristics is shown in figure 10. Sections “A” to “D”
are fast-rise pass neurons while sections “E” and “F” are all-pass neurons. Responses
are based upon mean spike count. The all-pass neurons pass
the signal on (although one suspects that they test for some as yet undetermined signal
feature) while the fast pass neurons only pass those signals having fast rise times.
Filter Type Neurons Selecting for the Rise Times of Pure Tones. (Hall and Feng - 1991)
The types of Dorsal Medullary Nucleus neuron responses to differing frequencies of sound are shown below in
figure 11. These peristimulus time histograms show the
responses to various pure 200 millisecond long tones
at 10 dB above threshold. Most phase lock to the peak pressures of the sound waves up to a certain
Temporal Responses from Five Types of Dorsal Medullary Neurons. (Feng and Wen-Yu Lin - 1994)
Connections of the Dorsal Medullary Nucleus
Figure 12 below shows the connections
of the dorsal medullary nucleus (Feng - 1986). First notice that it sends
(indicated by crosses) and receives (indicated by filled triangles) information
from the dorsal medullary nucleus on the opposite side. It also sends and receives
information from the superior olivary nucleus on both sides.
HRP (horseradish peroxidase) was injected into the dorsal medullary nucleus
(DN of illustration B) 6 days prior to the frog's termination thus giving
HRP a chance to be absorbed and transported by the various active processes
of the cell. HRP absorbed into the axon side of the synapses is transported
back to the cell body (retrograde transport). HRP absorbed from the dendrite
side of the synapses is transported to the axon terminals (anterograde transport).
Axon terminals are represented by crosses, axons by lines, and cell bodies
by filled triangles. Illustrations "a" to "g" represent a caudal (tailward)
to rostal (headward) direction.
The Output Connections of the Dorsal Medullary Nucleus in Male Rana pipiens. (Feng - 1986).
A - Aquaduct, CER - Cerebellum, DN - Dorsal Medullary Nucleus, LLN - Lateral Lemniscal Nucleus, MT - Midbrain Tegmental Nucleus, Ni Nucleus Isthmi, OT - Optic Tectum, 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
Interestingly the Dorsal Medullary Nucleus receives some information from two regions of the nearby reticular
formation with one region to the side and one region below. These projections may
be part of the circuit projecting back to the sensory ear neurons or perhaps
they block call identification during the frogs own call in order to prevent it
from reacting to its own call. Suggestive evidence that call blocking occurs is
provided by Peter Narins (1982) during his investigation of call training
in which the calls of certain frogs tend to occur during the quiet times between
the calls of other frogs. He found that male Puerto Rican treefrogs (Eleutherodactylus
coqui) are not able to respond to any call for 1.13 seconds after its own
call. Another treefrog (Hyla ebraccata) found in Panama could not react for
Notice that the projecting axons (indicated by the lines) exit
the Dorsal Medullary
Nucleus in two directions on either side of the ventral nucleus
(VN). One projection exits to the right forming the dorsal arcuate tract while the
other projection exits to the left and forms the ventral arcuate tract. Some
of the axons in the dorsal accuate tract send a branch (collateral) the ipsilateral Superior Olive and another branch to the contralateral Superior Olive before
coursing headwards mostly on the opposite side of the brain via the lateral
bulbotectal tract (lateral lemniscus in mammals) to the Torus Semicircularis. Some
if not all of these axons send a branch to the Lateral Lemniscal Nucleus. Some of the axons in the ventral arcuate tract
also send a branch into the ipsilateral Superior Olive before coursing on to terminate
in the Dorsal Medullary Nucleus on the opposite
side. These axons did not appear to project any further headward in the lateral
bulbotectal tract. It was also unclear whether or not these axons sent a branch
into the contralateral Superior Olive.
The Lateral Lemniscal Nucleus but not the Torus Semicircularis sends information
to the Dorsal Medullary Nucleus. The terminations in the Torus Semicircularis are mostly in its principal nucleus with
some in the Lateral Laminar Nucleus and some in the midbrain Tegmental Nucleus. The Laminar Nucleus of the Torus Semicircularis
is also involved in tactually triggered
The projections to the Dorsal Medullary Nucleus are organized tonotopically
with low frequency areas projecting to low frequency areas and high frequency areas
projecting to high frequency areas. This also seems to be the case for projections
to the Superior
Olive and Torus Semicircularis with low frequency projections terminating
in the lateral and dorsal regions in the Superior Olive and the central regions of the
Principle Nucleus of the Torus (Tp) while the high frequency projections
terminated in the ventral and medial regions in the Superior Olive and the
outer margins of the Principle Nucleus of the Torus (Tp).
Neuron Shapes (Morphology)
The dominant neurons in the Dorsal Medullary Nucleus are the bushy cells shown in
figure 13 to the left (Feng and Lin - 1996). These neurons have one thick central
dendritic trunk from which project small short dendritic branches. The bushy cells
can be different sizes but the larger ones completely cross the nucleus as shown
in figure 14. The scale of these bushy neurons relative to the brain stem is shown
in figure 15.when the frog is still in a semi-tadpole (larval) stage
having small fore legs and well developed hing legs. Transverse section is
at a magnification of 65.
Representative Bushy Neurons in the Dorsal Medullary Nucleus. (Feng and Lin - 1996)
Illustration B - dendrite is longer than shown. Scale bar is 10 micrometers. ax - axon.
Bushy Cells in the Right Side Dorsal Medullary Nucleus of Rana pipiens. (Feng and Lin - 1996)
Scale bar is 50 micrometers. D = dorsal (towards the back), V = ventral (towards the stomach), M = medial (towards the middle), L = lateral (towards the side)
The Scale of the Bushy Neurons Compared to the Size of the Brain Stem in Hyla regilla. (from figure 16 of Larsell - 1933
nu. VIIId is the Dorsal Medullary Nucleus
Significantly, the tonotopic organization of the nucleus is
perpendicular to the direction of the bushy cells. The terminations of the eighth
nerve axons from the ears are organized approximately tonotopically with low
frequencies located ventrally (stomach side) and high frequencies located dorsally
(spinal cord side). This suggests that these are the cells responsible for the tuning
curves discussed above.
Bushy neurons are not the only neurons in the Dorsal Medullary Nucleus. Feng and Lin (1996) have identified the following:
cells having dendrites originating on opposite sides of the cell body and
projecting obliquely across the bushy cell dendrites.
- Stellate cells having three
or more dendrite trunks originating on a triangular or polygonal shaped body.
They tend to be located in the center of the Superior Olivary nucleus. Their
dendrites radiate in all directions although those of the larger ones (the
Giant cells) tended to direct their dendrites dorsally and ventrally across the
bushy cell dendrites. Some of the dendrites appeared smooth while others had
- Radiate cells which are similar to the stellate cells but have round bodies.
Often one of these dendrites tends to be extra long so it curves around the superior
- Octopus neurons having two or three dendrites originating
on the same side of the cell projecting ventrally or dorsally.
- Small neurons
which were difficult to classify into any of the above categories.
For the exact
shape of these other neurons see the paper by Feng and Lin (1996).
Their distribution is shown in figure 16.
Distribution of the Variously Shaped Neurons in the Left Side Dorsal Medullary Nucleus of Rana pipiens. (Feng and Lin - 1996)
Scale bar is 100 micrometers.
eighth nerve fibers terminate in the dorsal medullary nucleus is shown in figure
The terminations are sparse and fairly limited in the dorsal-ventral (spinal-stomach)
direction (the tonotropic direction) yet quite extensive in the rostal-caudal (head-tail)
direction. The left of the illustration shows the cross
sectional view of the same fiber shown on the right.
Eighth Nerve Fiber Termination in the Dorsal Medullary Nucleus. (Lewis, Leverenz, and Koyama
Feng, A.S. and Capranica, R.R. (1976) Sound Localization in
Anurans: I. Evidence of Binaural Interaction in Dorsal Medullary Nucleus of Bullfrogs (Rana catesbeiana).
Journal of Neurophysiology 39:871-881
Feng, A.S. and Capranica,
R.R. (1978) Sound Localization in Anurans: II. Binaural Interaction in the
Superior Olivary Nucleus of the Green Tree Frog (Hyla cinerea).
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 Cochlear Nucleus (Dorsal Medullary Nucleus) of
the Leopard Frog. Brain Research 367:183-191
Feng, A.S. and Lin, W.-Y (1996)
Neuronal Architecture of the Dorsal Nucleus (cochlear Nucleus) of the Frog, Rana pipiens pipiens. J. Comp. Neurol. 366:320-334
Feng, A.S. and Lin, W.-Y
(1994) Phase-Locked Response Characteristics of Single Neurons in the Frog “Cochlear
Nucleus” to Steady-State and Sinusoidal-Amplitude-Modulated Tones. Journal of Neurophysiology
Fuzessery, Z.M. and Feng, A.S. (1981) Frequency Representation in
the Dorsal Medullary Nucleus of the Leopard Frog, Rana p. pipiens. J. Comp Physiol
Fuzessery, Z.M. and Feng, A.S. (1983) Frequency Selectivity in
the Anuran Medulla: Excitatory and Inhibitory Tuning Properties of Single Neurons
in the Dorsal Medullary and Superior Olivary Nuclei. J. Comp Pysiol. 150:107-119
Guinan, J.J., Guinan, S.S., and Norris, B.E. (1972) Single Auditory Units
in the Superior Olivary Complex. I. Responses to Sounds and Classifications
Based on Physiological Properties. Intern. J. Neurosci. 4:101-120
Hall, J.C., and
Feng A.S. (1990). Classification of the Temporal Discharge Patterns of Single Auditory
Neurons in the Dorsal Medullary Nucleus of the Northern Leopard Frog. Journal of
Neurophysiology 64 (5): 1460-1473
Hall, J.C., and Feng A.S. (1990). Temporal Processing
in the Dorsal Medullary Nucleus of the Northern Leopard Frog (Rana pipiens pipiens).
Journal of Neurophysiology 66(3): 955-973
Larsell, O. (1934) The Differentiation
of the Peripheral and Central Acoustic Apparatus in the Frog. J. Comp Neurol.
Lewis, E.R. Leverenz, E.L. and Koyama, H. (1980) Mapping Functionally
Identified Auditory Afferents from the Peripheral Origins to their Central
Terminations. Brain Res. 197:223-229
Narins, P.M. (1982) Behavioral Refractory Period
in Neotropical Frogs. J. Comp. Physiol. 148:337-344