Scoring spines is at best subjective. Here are some strategies.
Score Blind. Score all your spines blind to the experimental condition. This is best done by either having a colleague score spines or to hire an undergrad for scoring.
Over-score your spines. Over-score your spines by being very liberal in creating spines if you are not sure. Later on you can flag spines as ‘bad’, they will NOT be removed from the database but can easily be excluded from your analysis. Basically, false positives are easier to manage than false negatives.
Score timepoints independently. Mark spines in each timepoint independently. A good practice is to score disjoint time points (score them out of order) or score them backwards.
Impose a length threshold. Only score spines that protrude more than a minimum distance from the backbone. You can use the spine line as a guide and the Search panel to browse all spines in a map that are less-than some length.
Impose an intensity threshold. You can use the ‘Show Pixel Intensity’ option to sample a putative spines intensity, compare it to the local backbone intensity and only accept spines that are brighter than their associated backbone by some factor. See Options.