This is a nice paper, and addresses an important problem. Three quick thoughts:

1. There is a mutant line of zebrafish, "relaxed", that does not twitch, and is reported under certain conditions to have normal electrophysiological properties (1,2)
2. The authors note that "Additionally, the subset of CNS neurons that use the AChR alpha 7 subunit could also be affected." I am curious to know if, when using the fluorescent protein-conjugated bungarotoxin, if the authors ever observed central nervous system fluorescence? This would seem to definitively answer the question as to whether or not one should worry about central effects of alpha bungarotoxin. For what it is worth, we have not seen any such binding.
3. There are a wide variety of NMJ blockers reported to work in fish (e.g. pancuronium bromide, tubocurarine, etc.). Many are more cost effective than alpha-bungarotoxin, and, as they use a soaking protocol, less labor intensive than microinjection. While the space is vast, had the authors tried/rejected any others for purposes of paralysis?

(1) Zhou et. al. http://www.ncbi.nlm.nih.g...
(2) Koyama et. al. http://www.ncbi.nlm.nih.g...