215
where R 3 is the maximum response at the best stimulus angle (or best axis, BA); R 1
is the response at the adjacent stimulus axis that is −30° from the BA; and R 2 is the
response at the adjacent stimulus axis that is +30° from the BA (illustrated in Fig.
5a, b). This calculation was done separately for the horizontal plane (= azimuth) and
for the vertical plane (or mid-sagittal plane = elevation). For a perfect cosine
response function, the calculated SR is 0.866. The maximum SR possible is 1,
which would be for a cell that has an equivalent response to three adjacent stimulus
axes (and no distinct single best stimulus angle). An SR of 1 was never present in
any of our data sets and would not be expected from saccular afferents unless the
cell’s ability to respond to a stimulus is “saturated” by stimulus levels well above
threshold. An SR near “0” would indicate a very sharpened directional response
wherein the cell responds well to only one axis. Due to the natural variations in
spike counts for stimulus repetitions, particularly for cells with background activity,
a minimum data set at two stimulus levels was required, and the critical value for
considering a cell to be sharpened was an average SR ≤ 0.75 (Fig. 5c). The median
SR value for DON cells was 0.67 in azimuth and 0.62 in elevation (Edds-Walton
and Fay 2005b).
In a subset of DON cells (n = 73), 64 % exhibited sharpening (SR < 0.76) in azimuth and 67 % exhibited sharpening in elevation (Fig. 5c). Moreover, some of the
DON cells were sharpened greatly (SR < 0.56, Fig. 5c) in both planes. An equally
important observation was that there were cells for which sharpening occurred in
one plane only (azimuth or elevation) or for which sharpening was unequal in the
two planes. Taken together, the evidence indicates that sharpening is an important
computation in DON, which occurs by various ways (likely the weighting of inputs)
that result in different degrees of sharpening in different planes. In addition, directional sharpening and frequency tuning appear to be separate computational processes, as one is not predictive of the other (Edds-Walton and Fay 2003, 2008).
The best direction in three-dimensional space was calculated for afferents in
DON and plotted on a flattened globe (northern hemisphere only) to compare the
distribution around the fish with the best directions plotted for saccular afferents
(Fig. 6). The globe’s outer perimeter (equivalent to the equator of the flattened
globe) represents 0° in elevation, and directly above the fish (shown at the center of
the globe) is 90° elevation. Elevation rings (similar to latitude lines on a globe) are
shown for 30 and 60° in elevation around the fish. Azimuth is represented around
the fish with 0° in azimuth at the head of the fish; 30° in azimuth is labeled on the
left side of the fish for the saccular data to identify the angle around which most of
the left saccular data were found. Note that the best direction is shown as the point
on the globe at which the characteristic axis would pierce the northern hemisphere
of a globe.
Comparing the best directions for afferents from the left saccule and cells in the
left DON reveals very different distributions (Fig. 6). The directional plot for saccular afferents reflects the orientation of the saccule in the otic capsule of the fish.
The large number of overlapping data points around 30° left azimuth is consistent
with the orientation of hair cells on the rostral saccule, where much of the physiological recording was done due to the accessibility of the rostral bundle of the saccular
What the Toadfish Ear Tells the Toadfish Brain About Sound
where R 3 is the maximum response at the best stimulus angle (or best axis, BA); R 1
is the response at the adjacent stimulus axis that is −30° from the BA; and R 2 is the
response at the adjacent stimulus axis that is +30° from the BA (illustrated in Fig.
5a, b). This calculation was done separately for the horizontal plane (= azimuth) and
for the vertical plane (or mid-sagittal plane = elevation). For a perfect cosine
response function, the calculated SR is 0.866. The maximum SR possible is 1,
which would be for a cell that has an equivalent response to three adjacent stimulus
axes (and no distinct single best stimulus angle). An SR of 1 was never present in
any of our data sets and would not be expected from saccular afferents unless the
cell’s ability to respond to a stimulus is “saturated” by stimulus levels well above
threshold. An SR near “0” would indicate a very sharpened directional response
wherein the cell responds well to only one axis. Due to the natural variations in
spike counts for stimulus repetitions, particularly for cells with background activity,
a minimum data set at two stimulus levels was required, and the critical value for
considering a cell to be sharpened was an average SR ≤ 0.75 (Fig. 5c). The median
SR value for DON cells was 0.67 in azimuth and 0.62 in elevation (Edds-Walton
and Fay 2005b).
In a subset of DON cells (n = 73), 64 % exhibited sharpening (SR < 0.76) in azimuth and 67 % exhibited sharpening in elevation (Fig. 5c). Moreover, some of the
DON cells were sharpened greatly (SR < 0.56, Fig. 5c) in both planes. An equally
important observation was that there were cells for which sharpening occurred in
one plane only (azimuth or elevation) or for which sharpening was unequal in the
two planes. Taken together, the evidence indicates that sharpening is an important
computation in DON, which occurs by various ways (likely the weighting of inputs)
that result in different degrees of sharpening in different planes. In addition, directional sharpening and frequency tuning appear to be separate computational processes, as one is not predictive of the other (Edds-Walton and Fay 2003, 2008).
The best direction in three-dimensional space was calculated for afferents in
DON and plotted on a flattened globe (northern hemisphere only) to compare the
distribution around the fish with the best directions plotted for saccular afferents
(Fig. 6). The globe’s outer perimeter (equivalent to the equator of the flattened
globe) represents 0° in elevation, and directly above the fish (shown at the center of
the globe) is 90° elevation. Elevation rings (similar to latitude lines on a globe) are
shown for 30 and 60° in elevation around the fish. Azimuth is represented around
the fish with 0° in azimuth at the head of the fish; 30° in azimuth is labeled on the
left side of the fish for the saccular data to identify the angle around which most of
the left saccular data were found. Note that the best direction is shown as the point
on the globe at which the characteristic axis would pierce the northern hemisphere
of a globe.
Comparing the best directions for afferents from the left saccule and cells in the
left DON reveals very different distributions (Fig. 6). The directional plot for saccular afferents reflects the orientation of the saccule in the otic capsule of the fish.
The large number of overlapping data points around 30° left azimuth is consistent
with the orientation of hair cells on the rostral saccule, where much of the physiological recording was done due to the accessibility of the rostral bundle of the saccular
What the Toadfish Ear Tells the Toadfish Brain About Sound