Christopher D. Harvey, Philip Coen & David W. Tank
Long time no posting about article.
This is a study for decision making with two-photon calcium imaging.
- Recording area was 2/3 layers of Posterior Pariental Cortex (PPC) of mice.
- Calcium indicator was AAV-GCaMP3.
- Sampling frequency was about 16Hz (256 x 64 pixel).
- Behavior tasks were performed by their own system (Dombeck DA et al, 2007, Neuron).
Points
Q. Whether brain encode decision as recurrent attractor network with few neurons or high-dimentional dynamics with many neurons?
- High ratio of PPC neuron (~73%) participates in population activity for decision making
- Different neural classes response to different choices
- The responses for the decision (working memory?) task is sequentially in PPC.
- Spatial distribution of Trial preference (left or right) and Time preference (when did it burst) is not clustered.
Comments
I think this study is well controlled.
For example, they performed tracer experiment to identify PPC, because PPC were not identified in mice,
additional behavior experiments to eliminate the possibility that visual stimulus or motor activity affect PPC activity difference between the choices.
They performed many control experiments.
However, I feel scientific achieve seems to be low.
(Of course impact of technical aspect is very high)
I wonder whether there are any statistical analysis method without PCA (principal component analysis) and correlation.
There should be more powerful mathematical methods for population analysis.
But I have no idea now.
Population imaging study like this work will be increase in near future, I guess.
No comments:
Post a Comment