Article Figures & Data
Figures
- Figure 1.
Vibration sensing, primary mechanosensory center and DL-Int-1 interneuron in the honeybee. a, Airborne vibration jets produced during the waggle dance are picked up by the flagellum are transduced by sensory neurons of the JO in the pedicel and transmitted to the primary mechanosensory center of the honeybee brain, which consists of the mPPL, DL, and dSEG. Modified with permission from Ai et al. (2007), their Figure 1. b, Projection patterns of sensory afferents (green) and DL-Int-1 (magenta) in the primary mechanosensory center of the honeybee brain. DL-Int-1 has dendrites running close to sensory afferents in the DL. Modified from Ai (2013), their Figure 5. c, Morphology of DL-Int-1 visualized using three 2D projections. We divided DL-Int-1 morphology into four subregions for analysis. Inset, Magnified version of the region around the Main Branch. OL, Optic lobe; PC, protocerebrum; DC, deutocerebrum; AN, antennal nerve; AL, antennal lobe.
- Figure 2.
Changes in radial distribution of dendritic density. Comparison of PDLshell calculated for dendrites contained in concentric spherical shells of thickness 20 μm for the WA, DB, and VB, respectively, in a, b, and c. Solid circles indicate means and error bars indicate SD, both of which were calculated by pooling PDLshell values across registration parameters (Extended data Figure 2-1). Asterisks indicate a significant difference in PDLshell between maturation levels independent of registration parameters (see Extended data Figure 2-1). ART two-way ANOVA was used for factor analysis with a p value cutoff of 5%. These comparisons indicate a redistribution of dendritic length during maturation, with reductions in proximal parts and increases in distal parts of DL-Int-1 morphologies.
- Figure 3.
Region-dependent changes in dendritic density. All 12 DL-Int-1 morphologies visualized together after co-alignment, highlighting regions that show significant differences in PDLvoxel during maturation. a, WA, b, DB, and c, VB. A voxel was highlighted if ART two-way ANOVA indicated that maturation had a significant effect on PDLvoxel independent of registration parameters. The dendrites were colored with normalized change in PDLvoxel (see Materials and Methods, Morphological comparison using 3D voxels; for distributions, see Extended data Figure 3-1). The MBs are colored in black. d, The space containing the morphologies was divided into proximal and distal partitions based on distances along the dendritic tree of a node from the root and terminals in its subtree (for detailed 3D view for all subregions, see Extended data Figure 3-2). e, Changes in median PDLvoxel in proximal and distal partitions of each subregion. “n.s.” indicates that maturation did not have a significant effect on PDLvoxel independent of registration parameters when tested with ART two-way ANOVA (for distributions, see Extended data Figure 3-3).
- Figure 4.
Definition of activity periods and spike timing features of electrophysiological responses. a, An example response of DL-Int-1 to 1 s long vibration stimulus of 265 Hz. Activity before stimulus is colored in red, activity during stimulus in green and activity after stimulus in blue. b, The definitions of the four activity periods used for analyzing electrophysiological properties of DL-Int-1. c, The trace contained in the dotted rectangle in a is magnified and the four spike timing features, T0, T1, T2, and T3 are defined on it.
- Figure 5.
Analysis of electrophysiological properties. a, Comparison of the average firing rate profiles of newly emerged adult and forager DL-Int-1 neurons. Smoothed estimates of time-resolved average firing rates were calculated from responses aligned to stimulus onset using adaptive kernel density estimation (Shimazaki and Shinomoto, 2010). Solid lines indicate average firing rates, whereas shaded regions indicate 95% confidence intervals. Inset, Average firing rate during on-phasic response with expanded time scale. b, Comparison of firing rates during four activity periods. The filled areas represent firing rate distributions estimated using kernel density estimation (see Methods, Analysis of electrophysiology). The distributions were normalized to have equal areas. Horizontal white markers indicate mean values of the distributions. The numbers below the distributions are p values calculated using Mann–Whitney U test. P-values <5% are highlighted in green. Firing rates during spontaneous activity and rebound response showed significant increases. c, Comparison of the strength of inhibition relative to spontaneous activity by plotting the firing rates during the two periods against each other. Lines were fit using linear least-squares regression. The dashed line indicates the line of slope 1. d, Comparison of the timing of the first four spikes of the response using first spike latency, first ISI, second ISI, and third ISI. The distributions were estimated using Kernel Density estimation and normalized to have the same area. The numbers under the distributions are p values calculated using Mann–Whitney U test. Horizontal white markers indicate mean values, which are also shown above the distributions with sample numbers in parentheses. Spiking response was faster in foragers compared with newly emerged adults with significant reductions in first spike latency and first ISI.
Tables
Measure Morphological subregion Newly emerged
(NE, n = 6)Forager
(F, n = 6)Change in median value
[(F-NE)/NE]p Width (along x), μm Main branch 10.4,
22.1,
26.525.2,
34.1,
61.4+54.3% 0.004 Height (along y), μm Dorsal branch 152,
236,
263212,
268,
294+13.6% 0.041 Total dendritic volume, ×103(μm)3 Main branch 0.183,
0.409,
0.8030.301,
0.829,
0.967+103% 0.041 Average partition asymmetry Whole arborization 0.597,
0.661,
0.6530.561,
0.576,
0.636−12.86% 0.041 Maximum centrifugal order Ventral branch 13,
30,
4312,
16,
20−46.7% 0.043 Hausdorff fractal dimension Ventral branch 1.14,
1.24,
1.351.11,
1.14,
1.2−8.07% 0.041 Summary statistics of six scalar morphometric and topological parameters that show significant differences for at least one morphological subregion. The triplets in columns three and four represent minimum, median, and maximum values. P values were calculated using Mann–Whitney U test and a cutoff of 5% was used. Summary statistics for all 19 scalar measures and for all 4 subregions are provided in Extended data Table 1-1, Table 1-2, Table 1-3, and Table 1-4. Morphologies showed significant differences for a few scalar measures, with width, total dendritic volume, and maximum centrifugal order showing large changes.
Figure 2-1
a, Differing Parameters among the nine parameters sets used for registration. Among the nine parameter sets used for registration, the morphologies used as initial references for coregistering separately newly emerged adults and foragers were different and the experimental identifiers of these morphologies are listed here. Three initial references were used each for newly emerged adults and foragers and taking all possible combinations of these resulted in nine sets of parameters. b, Common Parameters among the nine parameters sets used for registration. Parameters other than the initial references were common among the nine parameter sets. For parameter description, see https://web.gin.g-node.org/ajkumaraswamy/regmaxs/src/master/regmaxsn/core/RegMaxSPars.py. Download Figure 2-1, DOC file.
Extended Data 1
This ZIP file contains all the code used to generate the results of this publication. The code is organized into two folders: (1) GJEphys-master, which contains code for analyzing and plotting electrophysiology data and (2) GJMorph-master, which contains code for analyzing morphology data. Download Extended Data 1, ZIP file.
- Table 1-1
Summary statistics of 19 scalar morphometric measures applied to the whole arborization subregion of DL-Int-1 morphologies. The triplets in columns two and three represent minimum, median, and maximum values. Column four contains p values calculated using Mann–Whitney U test for differences between newly emerged adults and foragers. Measures with p values <5% are highlighted in red. Download Table 1-1, DOC file.
- Table 1-2
Summary statistics of 19 scalar morphometric measures applied to the main branch subregion of DL-Int-1 morphologies. The triplets in columns two and three represent minimum, median, and maximum values. Column four contains p values calculated using Mann–Whitney U test for differences between newly emerged adults and foragers. Measures with p values <5% are highlighted in red. It was not possible to calculate statistics for some measures (marked N/A) as one or more morphologies of newly emerged adult or forager DL-Int-1 neurons had no bifurcations in the main branch. Download Table 1-2, DOC file.
- Table 1-3
Summary statistics of 19 scalar morphometric measures applied to the dorsal branch subregion of DL-Int-1 morphologies. The triplets in columns two and three represent minimum, median, and maximum values. Column four contains p values calculated using Mann–Whitney U test for differences between newly emerged adults and foragers. Measures with p values <5% are highlighted in red. Download Table 1-3, DOC file.
- Table 1-4
Summary statistics of 19 scalar morphometric measures applied to the ventral branch subregion of DL-Int-1 morphologies. The triplets in columns two and three represent minimum, median, and maximum values. Column four contains p values calculated using Mann–Whitney U test for differences between newly emerged adults and foragers. Measures with p values <5% are highlighted in red. Download Table 1-4, DOC file.
Figure 3-1
Distributions of normalized change in PDLvoxel for WA, DB, and VB. The distributions were calculated by smoothing histograms using Gaussian filters with standard deviation equal to 0.001 times the SD of the data. The box plots indicate quartile deviations. The distribution for the DB had more positive values, with values as high as 119%, whereas the distribution for the VB had more negative values with values as low as −15%. Both distributions had several points lesser than the first quartile and greater than the third quartile. Download Figure 3-1, EPS file.
Figure 3-2
Proximal and distal partitions. All 12 DL-Int-1 morphologies visualized together after co-alignment, illustrating proximal and distal partitions. a, WA, (b) DB, and (c) VB. The space containing the morphologies was divided into proximal and distal partitions based on the path lengths along the dendritic tree between a morphological node and the root as well as the morphological node and its downstream terminals (see Materials and Methods). Download Figure 3-2, EPS file.
Figure 3-3
Distributions of P DL voxel restricted to proximal and distal partitions. The distributions were calculated by smoothing histograms using Gaussian filters with SD equal to 0.001 times the SD of the data. The distributions were normalized to have equal areas. Horizontal white markers indicate mean values of the distributions. An asterisk indicates a significant difference in p DL voxel between maturation levels independent of registration parameters. ART two-way ANOVA was used for factor analysis with a p value threshold of 5%. Although all subregions showed significant difference for the proximal partitions, only VB showed significant difference for the distal partition. Download Figure 3-3, EPS file.
In this issue
