"Chemistry is, well technically, chemistry is the study of matter. But I prefer to see it the study of change"- Walter White
I want to use my mammalian neurone simulations to test various theories about biology and get some data on these points. The first point is the chemical concentrations in a cell are treated as set and do not move in the Hodgkin and Huxley models for the mammalian cell.
I started looking at first using data mining to identify changes that would work and used the error propagation to add or remove chemicals from the cell. This turned out to have worked and the cell showed lower error.
What I get is that with using the checks I found version one that uses the chemical change was 2 standard errors better and that the chance that new data could change this distribution is 2%.
T-statistic: 2.280637455367785,P-value:0.026615525098580763
Standard Deviation Group 1
-0.04207299175270728
0.13171618873423263
Standard Deviation Group 2
Max Output Group 1
-4.4958892767059865e-116
Max Output Group 2
-5.880274626363901e-67
Population Size 1
36
Population Size 2
41
This approach requires much testing but I think the brain changes the way the differential equations work within the brain as a learning mechanism. It is obvious this could be done using neuro transmitters changing ion gate openings.
This is just a simulation but it gives a sense of one way the brain might learn is by changing the chemicals in its cells.
This test might feel intuitively right but it has to be done because most of the cells chemical's serve multiple uses and so in changing it I worried it would damage the function by second and third order affects.
Add comment
Comments