LOGLIKELIHOOD_SPIKETRAIN - Compute the log likelihood of seeing spike counts
LL = vlt.neuro.spiketrains.loglikelihood_spiketrain(BIN_RATES, BIN_DURATIONS, SPIKECOUNTS)
Computes the log likelihood of observing SPIKECOUNTS(i) in
bins of length BIN_DURATION(i) with firing rates BIN_RATES(i).
The following equation is used to calculate the log likelihood:
logP = sum over i of
bin_rates(i)*bin_durations(i) + spikecounts(i)*log(bin_rates(i)*bin_durations)-log(k!)
See "A Novel Circuit Model of Contextual Modulation and Normalization in Primary Visual Cortex"
Columbia University 2012