I am wondering what is the expected behavior of a sequence where the same input is repeated. With my understanding of the system, I would expect to see bursting followed by a single set of cells that activate each time after some connections reach the permanence threshold (due to sparse connections between cells in one column to cells in other columns that are repeating). However, in my test implementation I am seeing a slightly different behavior (bursting is followed by a series of different cells in the columns activating for each iteration of the same input). I am looking at this as a likely indication of a bug in my code, so tracing through to determine the culprit. Just thought I’d confirm whether I am correctly understanding what the expected behavior should be in the case of one input repeating.
Answering my own question again, actually the expected behavior should not be the same set of cells activating each time, because a cell cannot be in both predictive state and in active state simultaneously (there is also probably a “cooldown” period in the biology, but since I haven’t implemented that yet, I wouldn’t expect it to be a factor in my case). Also since the cells that fired during one cycle would be weakened when another set of cells fired during the next cycle, the output could potentially be more complex than simply two alternating patterns. Any obvious mistakes I am making in my logic? 
You’re right that in biology it might be difficult for a cell to become active and then enter depolarized state shortly afterward - really depends on the timing of everything. In the temporal memory code, I think it is possible for a cell to be both predicted and active. My guess is that after bursting, you will probably end up with multiple cells in each column becoming predictive and then multiple cells per column becoming active when the next input is presented.