It is an attempt to reproduce results of the original project. For the original Readme see the original original/readme.txt.

There I started to refactor the existing brian1 version but it is a very time consuming process as the existing code is sufficiently tangled and brian1 can't be run properly on my machine due to pylab/scipy. Old versions of them have to be installed. On my machine those old versions are not installable. That is why I freeze the work there and start clean in Brian2 and py3. Also I am not sure that the original can reproduce the claimed results. Too high probability of an error in this tangled code. For example, I'm highly curios why weights of XeAe are not updated when I was able to run the original without plotting functions.

Basically we just take an idea and try to build the same in the latest Brian2.

• The best one https://github.com/BindsNET/bindsnet/tree/master/examples/mnist
• https://github.com/djsaunde/lm-snn
• https://github.com/whenov/stdp-mnist-brian2
• https://github.com/bamford/Brian2STDPMNIST