245
it is possible that variations in tank dimensions and position of the saccule relative
to the water surface has important consequences for threshold determination.
However, tank dimensions and fi sh position were essentially identical between
saccular potential recordings and single neuron recordings in the brain, suggesting
that the differences in threshold between these techniques are due to biological
rather than methodological variations. Nevertheless, future studies should carefully
consider and characterize particle motion and sound pressure levels throughout
their experimental tank, as well as any other subtle procedural variations.
One important auditory sensitivity measurement missing from our data set in the
Hawaiian sergeant fi sh is a behavioral audiogram determined by classical conditioning or psychophysical methods. Behavioral auditory thresholds are often, but not
always, lower than any electrophysiologically determined thresholds and may be
the best indicator of true hearing abilities in a species. However, they are extremely
time-consuming and diffi cult to generate in some fi sh species, especially those that
do not respond to the training paradigms. Physiologically determined audiograms
are valuable because they provide a good estimate of the frequency hearing abilities
of a species (i.e., spectral range), including a measure of best frequency, in a comparatively shorter amount of time, even though they may underestimate hearing
sensitivity at certain frequencies in some species. However, this underestimation is
not a universal relationship among all fi shes. For example, behavioral thresholds are
lower than (Fay 1974 ; Kojima et al. 2005 ), greater than (Kenyon et al. 1998 ), or
similar to (Fay 1978a , b ; Kenyon et al. 1998 ; Ladich 1999 , 2000 ) physiologically
determined thresholds in different species [see also Ladich and Fay 2013 for a
review], suggesting that differences may be species-specifi c and dependent on
experimental factors that vary among labs. Based on their extensive comparison of
AEP and behavioral tuning curves in many fi shes, Ladich and Fay ( 2013 ) note that
AEPs tend to produce higher thresholds at low frequencies (<1000 Hz) and lower
thresholds at high frequencies (>1000 Hz) compared to behaviorally generated
audiograms, suggesting there is also a frequency dependent effect between these
two methods.
The type of auditory recording method employed in a study will depend largely
on the research question addressed, species used, and the available resources. For
example, AEPs have become popular in recent years because they are relatively
quick to perform, easy to learn, inexpensive to setup, applicable to almost any species, and are minimally invasive allowing repeated measurements in the same individuals. AEPs are therefore valuable for obtaining rapid information on the frequency
range and threshold tuning for a particular species, as well as doing before and after
comparisons following manipulation or “intervention” to test some aspect of hearing
(e.g., exploring temporary hearing changes that result from noise exposure). Single
neuron recordings, on the other hand, require more expensive equipment, invasive
surgical approaches, complex analysis tools, and expertise to perform and interpret.
Neural recordings that examine auditory responses at different points along the
ascending pathway, however, are quite valuable for providing important information
on specifi c auditory processing and fi ltering mechanisms that occur at different
levels within the central auditory system. This type of information cannot be obtained
Comparison of Electrophysiological Auditory Measures in Fishes
it is possible that variations in tank dimensions and position of the saccule relative
to the water surface has important consequences for threshold determination.
However, tank dimensions and fi sh position were essentially identical between
saccular potential recordings and single neuron recordings in the brain, suggesting
that the differences in threshold between these techniques are due to biological
rather than methodological variations. Nevertheless, future studies should carefully
consider and characterize particle motion and sound pressure levels throughout
their experimental tank, as well as any other subtle procedural variations.
One important auditory sensitivity measurement missing from our data set in the
Hawaiian sergeant fi sh is a behavioral audiogram determined by classical conditioning or psychophysical methods. Behavioral auditory thresholds are often, but not
always, lower than any electrophysiologically determined thresholds and may be
the best indicator of true hearing abilities in a species. However, they are extremely
time-consuming and diffi cult to generate in some fi sh species, especially those that
do not respond to the training paradigms. Physiologically determined audiograms
are valuable because they provide a good estimate of the frequency hearing abilities
of a species (i.e., spectral range), including a measure of best frequency, in a comparatively shorter amount of time, even though they may underestimate hearing
sensitivity at certain frequencies in some species. However, this underestimation is
not a universal relationship among all fi shes. For example, behavioral thresholds are
lower than (Fay 1974 ; Kojima et al. 2005 ), greater than (Kenyon et al. 1998 ), or
similar to (Fay 1978a , b ; Kenyon et al. 1998 ; Ladich 1999 , 2000 ) physiologically
determined thresholds in different species [see also Ladich and Fay 2013 for a
review], suggesting that differences may be species-specifi c and dependent on
experimental factors that vary among labs. Based on their extensive comparison of
AEP and behavioral tuning curves in many fi shes, Ladich and Fay ( 2013 ) note that
AEPs tend to produce higher thresholds at low frequencies (<1000 Hz) and lower
thresholds at high frequencies (>1000 Hz) compared to behaviorally generated
audiograms, suggesting there is also a frequency dependent effect between these
two methods.
The type of auditory recording method employed in a study will depend largely
on the research question addressed, species used, and the available resources. For
example, AEPs have become popular in recent years because they are relatively
quick to perform, easy to learn, inexpensive to setup, applicable to almost any species, and are minimally invasive allowing repeated measurements in the same individuals. AEPs are therefore valuable for obtaining rapid information on the frequency
range and threshold tuning for a particular species, as well as doing before and after
comparisons following manipulation or “intervention” to test some aspect of hearing
(e.g., exploring temporary hearing changes that result from noise exposure). Single
neuron recordings, on the other hand, require more expensive equipment, invasive
surgical approaches, complex analysis tools, and expertise to perform and interpret.
Neural recordings that examine auditory responses at different points along the
ascending pathway, however, are quite valuable for providing important information
on specifi c auditory processing and fi ltering mechanisms that occur at different
levels within the central auditory system. This type of information cannot be obtained
Comparison of Electrophysiological Auditory Measures in Fishes