It seems that predicted cells, predicted active cells, and next predicted cells arays are empty in ApicalTMSequenceRegion. Do you know why this is so?
Can you give more details of the circumstances in which you observed this?
I am experimenting with different temporal memory implementations. At the moment I have a sensor connected to a spatial pooler which is connected to an apical TM region which projects further to another spatial pooler and receives feedback from a higher-up apical TM. I could email the actual python code. Let me know. Perhaps I should be using apicalTieBreakSequenceMemory?