Which statistical test?

Is there a clear definable baseline for how often the neuron would be expected to fire by chance? If so you could use a Wilcoxon signed-rank test to decide whether the participants had significantly more response spikes than the normal average number.

Or even better, you could determine your own baseline by recording how many times the neuron fired in a ten second period before the stimulus. Then, again, you could use a Wilcoxon paired signed-rank test to compare how many times the neuron fired in the pre-stimulus period to the during/post-stimulus period.

To be clear I know very little about brain science or the design of this kind of experiment - this is just the statistical test that seems most suitable to me. You're trying to compare a numerical variable to a given value, or to the same numerical variable in what is essentially another group, for which you would normally use a T-test. But your sample size is 8 so you can't assume normality, so you use the non-parametric equivalent which is the Wilcoxon signed-rank test.