Questions about Allen Mouse Brain Connectivity Atlas projection and injection data

Science
1

Hi, I am a PhD student working on stochastic connectome-based modeling of pathological particle propagation in the mouse brain. I am using the Allen Mouse Brain Connectivity Atlas, especially the projection dataset from Oh et al., “A mesoscale connectome of the mouse brain.”

I would be grateful if you could clarify a few points about the data and methods:

  1. What is the difference between the segmented injection volume and the regional voxel count?
  2. In the image-data-processing section, the paper mentions that total signal is computed for each voxel by summing signal-positive pixels. Is this voxel-level value equivalent to the amount of projection signal in that voxel, or should it be interpreted differently?
  3. For modeling purposes, should the anatomical volume of a given brain region be treated as the same whether the region is used as a source/injection region or as a target region?
  4. Does the AAV tracer used in the study replicate after injection, or is it non-replicating? Also, is there any known effect of microglial phagocytosis on the measured viral/tracer signal over the 21-day survival period?

I also have a few data-availability questions:

  1. Is voxel-level total signal data available for each experiment, including the total signal within the injection site after the 21-day survival period?
  2. Is the injected viral concentration available, for example viral particles per unit volume of injected solution, or ideally an estimate of viral particles per unit brain volume for each source region?
  3. Is there any dataset that reports estimated viral/tracer concentration or signal concentration in each brain region after 21 days?
  4. The supplementary information and connectivity platform modeling description mention segmented injection volumes for each experiment. Is the analysis table and/or code used to compute these segmented injection volumes available, especially for the region-level experiments used in the connectome matrix?

My goal is to use the biological and imaging information properly in a mathematical model, without misinterpreting projection signal, voxel counts, injection volume, or viral-particle concentration.

Thank you very much for your time and guidance.

2

Hi,

Please see responses to your questions below:

Answers to the data and methods:

1, The segmented injection volume is defined by manually drawn polygons for each injection site across multiple sections. The regional voxel count is a calculation of voxel for each brain region.

2, The voxel-level value is equivalent to the amount of projection signal segmented in the image-processing section.

3, Anatomical volume of a given brain region cannot be treated as the same whether the region is used as a source/injection region or as a target region because there are different sizes of injection sites or contamination.

4, There are injection replicates for most of the brain regions, particularly large brain structures. However, some brain regions only have one injection. We did not observe the effect of microglial phagocytosis on the measured viral/tracer signal.

Answers to the data-availability questions:

1, Total axonal projection signal data at voxel-level is available for each experiment. However, total signal within the injection site is not available because axons and dendrites there were not separated.

2, Since rAAV was delivered by iontophoresis we don’t know how many rAAV particles were injected into each source region.

3, No, we did not report the estimated viral/tracer concentration and signal concentration in each brain region (Oh et al., 2014).

4, Injection volumes were reported in Oh et al. 2014. The segmented injection volumes were calculated from manually drawn polygons across multiple sections. No code was used for injection volume calculation.