ERG waveforms were digitally filtered for 60-Hz noise and averaged
across the three stimulus presentations.
8 Spectral
sensitivity functions were calculated for both the a- and b-wave ERG
components when they were apparent. Spectral sensitivities of the 6 to
8 dpf subjects from the LD, DD, and LL conditions were calculated using
the a-wave component from responses to 320- to 440-nm stimuli; after
440 nm, the a-wave was no longer evident in the subjects’
waveforms.
9 The a-wave was measured from baseline
(response before stimulus onset) to the first negative peak. For the
b-wave component, spectral sensitivity was calculated from 320 to 640
nm for all age groups and all conditions. The b-wave was measured from
baseline or the first negative peak to the first positive peak. There
were no apparent differences in the subjects’ ERG waveforms across all
rearing conditions, including normal subjects (LD). The young
subjects’ ERG waveforms had strong a- and b-waves in response to UV
stimuli, as opposed to being dominated by the b- and d-waves as found
in adult responses across wavelengths.
8 9 To obtain
sensitivity to each stimulus wavelength, the reciprocal of the log
stimulus irradiance (quanta s
−1 cm
−2), which produced a criterion response, was
calculated from the log irradiance-response function.
9
To make comparisons across the lighting conditions, the data were first
normalized to values at 640 nm. This is because, based on previous
research, it was hypothesized that the deficits due to light-rearing
conditions would be found in the UV and short-wavelength areas of the
spectrum. Thus, normalizing the data to the peak in the UV area, which
is where larvae zebrafish are normally most sensitive,
9 would be inappropriate because that is where the deficits were
expected. The decision to normalize the curves at the long wavelengths
was supported empirically by attempting to normalize the functions at
other parts of the spectrum. For example, normalizing at the UV
wavelengths (which was the most sensitive portion of all the functions)
gave the impression that the 6 to 8 dpf subjects in the three rearing
conditions were identical in sensitivity to UV wavelengths and that the
LL group was more sensitive to middle and long wavelength stimuli when
compared to normal subjects. This interpretation is inconsistent with
normal zebrafish ERG development.
9 Therefore, the data
were normalized to values at 640 nm. The data were renormalized to the
peak sensitivity of the LD-condition subjects for each age group.
Normalizing the data to the peak sensitivity of the LD-condition
subjects was done so the sensitivity of the DD- and LL-condition
functions could be compared, relative to the normal LD-condition
function. After calculating the relative spectral sensitivities for the
three conditions and the three age groups, a quantitative assessment of
the cone contributions to the spectral sensitivity function was
performed. A multiple mechanism model was used to derive the cone
inputs (λ
max = 362, 415, 480, and 570 nm; U-,
S-, M-, and L-cones, respectively
10 ) to the spectral
sensitivity data (see
Ref. 8 for details).
Figure 1 shows the spectral sensitivity functions of the 6 to 8 dpf subjects
from the LD (squares), DD (circles), and LL (triangles) conditions. The
lines represent the best-fit models and the error bars indicate ±1
SEM. As expected, compared to the LD-condition subjects, the
LL-condition subjects showed the greatest deficit in sensitivity,
especially to UV and short-wavelength stimuli. There were differences
between the LL- and the LD-condition subjects in the middle- and
long-wavelength areas, but they were relatively small. The subjects in
the DD condition also showed a deficit in sensitivity when compared to
the LD-condition subjects, but it was not as large as the deficit of
the LL-condition subjects. Unlike the LL-condition function, there was
a uniform sensitivity deficit of a half log unit in DD-condition
subjects’ relative sensitivity compared to the sensitivity of the
LD-condition subjects.
Figure 2 A shows the spectral sensitivity of the 13 to 15 dpf subjects from the
LD (squares), DD (circles), and LL (triangles) conditions. In general,
all three spectral sensitivities are similar in shape. At this age, the
difference in the spectral sensitivities that was apparent in the 6 to
8 dpf subjects has disappeared. Interestingly, there are no sensitivity
differences for the three groups in the UV area. The LL- and the
DD-condition subjects appear to show only a slight deficit in
sensitivity to the short- and middle-wavelength areas of the spectrum
when compared to the LD-condition subjects.
Figure 2B shows the
spectral sensitivity functions of the 21 to 24 dpf subjects in the LD
(squares), DD (circles), and LL (triangles) conditions. The damage seen
initially in the 6 to 8 dpf subjects raised in the LL and DD conditions
gradually disappeared and by 24 dpf, the sensitivity of the LL- and
DD-condition subjects returned to normal.
To assess whether photoreceptor function was responsible for the
differences seen in the LD, LL, and DD conditions, a-wave spectral
sensitivity functions were calculated.
Figure 3 A shows the spectral sensitivity functions of the a-wave component of
the ERG response from 6 to 8 dpf subjects in the LD (squares), DD
(circles), and LL (triangles) conditions. The data are shown in
absolute sensitivity values (quanta s
−1 cm
−2). The curves from the three groups are very
similar in shape and absolute sensitivity. There do not appear to be
any differences in the a-wave spectral sensitivity functions of the 6
to 8 dpf subjects exposed to the LL and DD conditions when compared to
the LD condition, suggesting no problems with photoreceptor function.
Because the functions across the three conditions were similar, the
data were averaged and modeled. In
Figure 3B , the points represent the
data and the line represents the model. Because the spectrum range was
limited to primarily UV wavelengths, only the U-cone spectra were used
in the model. There appears to be a good fit between the model and the
data, suggesting that this spectral sensitivity function reflects
U-cone activity in the retina. Because the function is derived from the
a-wave component of the ERG response, it suggests normal U-cone
function.