combining Self Organizing Maps and Hidden Markov Models

Hi all

I have used SOM and i can see all the neurons' weight vectors by using 'net.iw'

Then , i have separated the x and y coordinates of these neurons and plotted them by using 'scatter(x,y)'. I have attached the screenshot of scatter also.

Now, i want to convert these points into states (like states in State Transition diagram) and then, i want to calculate the state transtion and emission probability using these states and 'hmmestimate'. Please help me with the same.

Main focus of my project is to combine SOM and HMM. To be very honest, i donot even know if this is possible. So, please be tolerable if this sound stupid.

Also, if someone have any another approach, please do share. Thanks in advance.