Botulinum neurotoxin A (BoNT/A) cleaves SNAP-25, a key protein for exocytosis, and thus blocks the release of the neurotransmitter acetylcholine from the motor nerve terminal at the neuromuscular junction (NMJ) (Schiavo et al. 2000). Despite its high toxicity, BoNT/A (Botox®, Allergan Inc.) in small doses has clinical applications in inhibiting undesirable muscular spasticity and in reducing facial wrinkling (Osborne et al. 2007). In addition to muscle paralysis, BoNT/A induces profuse sprouting of motor nerve terminals (Brown et al. 1981). Whether and how these nerve sprouts may affect the treatment is not clear. Previous studies suggest that the nerve sprouts play a major role in the initial functional recovery from the BoNT/A treatment (Angaut-Petit et al. 1990; de Paiva et al. 1999). However, whether synaptic transmission indeed occurs at the sprouts has not been directly demonstrated. In this issue of The Journal of Physiology, Rogozhin et al. (2008) combined electrophysiological recording with imaging approaches, and found that the original synaptic sites play the predominant role in functional restoration from BoNT/A, rather than the nerve sprouts as previously thought. Taking advantage of a thin muscle (epitrochleoanconeus) and a vital nerve terminal dye, 4-Di-2-ASP (Magrassi et al. 1987), Rogozhin et al. (2008) were able to position extracellular electrodes precisely over individual synaptic boutons at the mouse NMJ for focal recordings of synaptic events (del Castillo & Katz, 1956). Because of its fine spatial resolution of 5–10 μm, the focal recording allows comparisons of synaptic efficacy between the original NMJ and the new synaptic site induced by a single BoNT/A injection. By combining the focal recording with intracellular recording, Rogozhin et al. (2008) estimated that the mean quantal content at the new synaptic site was about 17% of that from the original junctional site during the period of 2–12 weeks following a single injection of BoNT/A. The finding by Rogozhin et al. (2008) of the relatively minor role of new sprouts in functional restoration from BoNT/A is in sharp contrast to the prevailing thinking based primarily on a study by de Paiva et al. (1999). To assess nerve terminal function, de Paiva et al. (1999) labelled motor nerve terminals with FM1-43, a fluorescent dye that binds endocytic synaptic vesicles in an activity-dependent manner (Cochilla et al. 1999). They found much brighter FM1-43 staining in the sprouts than in the original synaptic sites during the initial phase of functional recovery (e.g. 28 days) from BoNT/A, but the opposite was the case at the later phase (e.g. 91 days). Although the brighter FM1-43 staining indicates a stronger release efficacy at the nerve sprouts, it may not effectively translate into a larger postsynaptic potential since only a small fraction of the total length of the nerve sprouts is associated with clusters of acetylcholine receptors (AChRs), which are in general smaller and poorly defined. Indeed, Rogozhin et al. (2008) found that the total area of AChRs associated with the nerve sprouts was only about 30% of that associated with the original area during most of the recovery phase. Thus, while the sprouts are highly capable of vesicle recycling, their role in functional recovery is limited. The findings of Rogozhin et al. (2008) raise several important questions regarding synaptic plasticity at the NMJ. For example, why are only a small fraction of the nerve terminal sprouts associated with AChR clusters? What is the role, if any, of nerve sprouts in synaptic repair? Could the vesicle recycling seen at nerve sprouts result in increased uptakes of the neurotrophic factors essential for motoneuron survival? What mechanisms allow nerve terminals to sprout when vesicle exocytosis and subsequent membrane fusions are blocked by BoNT/A? Why do the nerve sprouts and other structure abnormalities at the NMJ persist after functional recovery from multiple injections of BoNT/A as shown in Rogozhin et al. (2008)? Given that sprouting of nerve terminals and the associated Schwann cells is a common form of synaptic plasticity following the blockade of pre- or postsynaptic activity, as well as by partial denervation (Brown et al. 1981; Kang et al. 2003), the study by Rogozhin et al. (2008) is likely to help address these questions to better understand the mechanisms of synaptic plasticity.