What coding strategy brain uses? Are the timings of spikes important (temporal coding), or simply the number of spikes in a given time window is sufficient information carrier (rate coding)? For sure the former is faster and more efficient and the later is more reliable and robust. There are many papers showing evidence for one or the other.
Most likely brain is not all that picky and simply uses whatever is more appropriate for the task, so I would hypothesise that different brain areas employ different strategy and possibly can even switch between the two depending on the task (this is a bit of a wild statement).
I will use this place to gather the references to the studies making claims towards one or the other (or possibly yet other) coding strategy in the brain.
Update (16.2.2011)
Tiago Branco et al. from the lab of Michael Hausser show indications what distal dendrites use inputs in the rate-coding manner, while the proximal dendrites can use time-coding manner. Not sure this was published yet. See Haussers’Â report of his talk.