The Journal of Neuroscience, February 15, 2003, 23(4):1555
CORRECTION
In the article "Firing Properties of Rat Lateral
Mammillary Single Units: Head Direction, Head Pitch, and Angular Head
Velocity," by Robert W. Stackman and Jeffrey S. Taube, which appeared
on pages 9020-9037 of the November 1, 1998 issue, the authors reported the presence of head direction (HD) cells in the lateral mammillary nuclei (LMN). In the course of conducting data analyses for another project, the authors realized that the timing clocks for two pieces of
equipment (a timestamper and video tracker) were not synchronized in
time with one another in one of the recording rooms. Each piece of
equipment had its own clock, and the two clocks ran at slightly different speeds. Consequently, the disparity between the two clocks
increased linearly the longer a recording session was conducted. It was
determined that the difference in clock speed between the two clocks
was 3.0055 µsec per video field (1/60th sec). Thus, over the 8 min
recording sessions conducted, an error was introduced between when the
cell fired (determined by the timestamper) and the rat's current
directional heading (determined by the video tracker, which tracked the
locations of the two light-emitting diodes attached along the midline
of the rat's body axis). This error was negligible at the beginning of
the recording session, but grew to ~86 msec during the last minute of
recording, which is a little more than five video samples. This error
effected 15 of the 20 LMN HD cells reported; five HD cells were not
affected, because they were recorded in a different room that used
different recording equipment that was not susceptible to this timing error.
The authors have now corrected for this error in their data analysis
programs. The error caused little change in the mean values for the HD
cell firing properties reported in Table 1, including observed peak
firing rate, directional firing range, information content, asymmetry
ratio, background firing rate, signal-to-noise ratio, and correlations
to linear and angular velocities. However, this error significantly
changed the values for the time shift analyses and altered
the anticipatory time interval (ATI) for LMN cells. The
table below shows the corrected values
and replaces Table 2 from the November 1, 1998 article. This table
indicates that the estimated ATI for LMN HD cells is ~66.7 msec,
compared with the incorrect value of 95.8 msec reported previously.
This corrected value, although still somewhat larger than the value of
38.5 msec reported by Blair et al. [Blair HT, Cho J, Sharp PE
(1998) Role of the lateral mammillary nucleus in the rat head
direction circuit: a combined single unit recording and lesion study.
Neuron 21:1387-1397] in their study on LMN HD cells, is less than the
originally reported value and closer to the value reported by
Blair et al. (1998). By calculating the ATI using the
procedures of Blair et al. (1998), a mean ATI of 78.3 msec was obtained for the LMN HD cells, again, a value that is lower
than the value reported previously but higher than the value reported
by Blair et al. (1998). In examining the table below, it
is noteworthy that although all of the optimal time shift values for
LMN HD cells are more anticipatory than the values for anterodorsal
thalamic nucleus (ADN) HD cells, there were no significant differences
for any of the parameters (Scheffe's F test; all
p values >0.05), which is different from the original results reported on page 9030. Also shown below are the corrected graphs for Figures 5D-F and
6D-F of this article. The plots from the original
Figure 5A-C do not change significantly and are not shown
as revised. The plots from the original Figure 6A-C
(optimal time shift curves for LMN HD cells) are altered and can be
viewed in the revised graphs for Figure 6D-F shown
below. These revised plots show that LMN HD cells are a little more
anticipatory than ADN HD cells, because each of the LMN curves is
shifted to the right of the ADN curves around the peaks (see
inset of each plot for an expanded view around the curve
peaks for the ADN and LMN functions; the dashed vertical
line in all plots is time shift = 0).
