search

open-connectome-classes / StatConn-Spring-2015-Info

introductory material
18 stars 4 forks source link

Question about Hypothesis for Bock #130

Open akim1 opened 9 years ago

akim1 commented 9 years ago

I can't recall if this was mentioned in class. But would it be an easier problem if we tried to show that neuron orientation is (in)dependent of its connectivity?

Could this be demonstrated by constructing two independent distributions (one consisting of connectivity/cluster and the other consisting of orientation) and then showing that the likelihood of sampling the graph from this particular distribution is unlikely?

whock commented 9 years ago

Do you mean it's connectivity to down-stream neurons like the inhibitory neurons studied in Bock? Or that orientation is not dependent on any connectivity? Because I think it's pretty likely that connectivity plays some role - the neurons derive their orientation preference from how they are wired to the retinal output cells (RGCs).

edunnwe1 commented 9 years ago

Whock brings up a good point, but assuming that you meant the down-stream inhibitory neurons like those studied in Bock: I'm a little bit confused by the construction you have here. In this case, we're trying to see if the inhibitory neurons receive convergent input from 1) neurons with a particular selective orientation or 2) neurons with no particular selective orientation. So I think that to frame it as whether the neuron's orientation is independent of its connectivity might be confusing because this connectivity is directed, and it's the neurons with preferred orientations that are converging onto inhibitory interneurons. But, perhaps what you meant was that the convergent input is independent of the preferred orientation of the neurons, rather than does the preferred orientation of neuron A converging onto neuron C predict that neuron B with also converge onto neuron C. I think that that's an interesting idea but I can't quite visualize how your method would work to demonstrate this, in particular how you construct the distribution for excitatory neurons converging onto inhibitory neurons, and then once you establish that the observation is unlikely how you could then say that this implies dependence or independence.