Brainbox unit metrics upgrade

[oliche]

A few simple changes of how we output metrics could be beneficial:

1. We could output Tau for the sliding RP. It may have physiological meaning as a proxy for the intrinsic refractory period of the neuron.
2. output the standard deviation of amplitudes. In general any metrics about the stability of units is welcome
3. output the firing rate of units during its active firing only. Here we could either use the presence ratio windows (allowing gaps in firing not taken into account), or more simply just average over the last_spike_time - first_spike_time of each unit.

[bimac]

to discuss at data meeting