Methodology: In the spirit of our many demonstrations of control, our excellent laboratory for investigation of control is our own corporeal experience. The evidence is in our subjective experience, but replicability and actual replication makes the experience objective data within the public constituted by those who carry out the experiment.
Several levels of perception are involved in the phenomena reported here.
Context: I routinely mix a smoothie in the morning with a hand-held blender. I count to 75 seconds.
In my counting I control the auditory perceptions in imagination so that I ‘hear’ the unvoiced number names. Bored with this, I began visualizing the numbers, controlling the visual perceptions in imagination. I controlled ‘counting’ in both sensory modalities concurrently. I counted down from 75 to 1 or up from 1 to 75.
Bored with this, I began counting the visually imagined numbers down from 75 to 1 concurrently as I am counting the auditorily imagined numbers up from 1 to 75.
At first, the visual images lagged the words.
Then I ‘shifted focus’ to counting visually (which is anyway not a usual way of counting). Sometimes the visual image leads the auditory perception, sometimes word and image are concurrent, and with practice that is the regular experience unless my focus of attention weakens. (What are the perceptual inputs to the function(s) for controlling ‘focus of attention’?)
Whatever system controls counting, two distinct instances are in operation at the same time. This suggests that ECSes of all types exist or may be recruited redundantly, at least in pairs. (I say "redundantly’ in common parlance. The statistician who clings to the technical definition of redundancy should feel free to substitute a word like “suppletively”.)
Supplemental observation: With practice, my counting of seconds has become quite accurate, i.e. coincides with a reading of the clock’s second hand when I turn and look at the end of the count. It’s possible I hear the slight tick that the electric clock makes, about ten feet away on the wall, despite the noise of the blender. This could be tested. The pulse of seconds is unrelated to heart rate. What are the inputs to the perceptual input function for this temporal perception? I bet it can be done for arbitrary relatively short time intervals, e.g. 5 seconds, though I haven’t tested that. I’m reminded of workers at the US Government Mint who could accurately count out bundles of 100 newly printed bills by riffling them in the manner of shuffling cards.