72 × σMN were set to 0 (3.72 is the approximate Z score value of p = 0.9999, i.e., only 1 in 10,000 values click here will be above this threshold by chance). The median σMN value was 0.078, and 95% of σMN values were below 0.2. Cells were considered active if they crossed the 3.72 × σMN threshold for at least 500 ms

(nine frames). Cells were considered predominantly activated by a stimulus condition if average response during one condition was at least twice as high as average response during any of the other conditions. We thank Tara Keck and Tony Movshon for comments on the manuscript. This work was supported by the Max Planck Society and fellowships from the Swiss National Science Foundation and the Human Frontier Science Ku-0059436 mw Program to G.B.K. “
“Despite the widespread belief that neural

circuit formation is the central theme of vertebrate neural development, there is ample evidence of the opposite: postsynaptic target cells in various parts of the central and peripheral nervous system appear to be innervated by more axons early in postnatal life than later on (Purves and Lichtman, 1980). The reduction in the number of converging axons, known as synapse elimination, may play a role in establishing permanent synaptic circuits based on experience (Lichtman and Colman, 2000). In the neuromuscular system, this phenomenon has been studied by us and others, especially during the second postnatal week in rodents when muscle fibers Endonuclease make the transition from double and occasionally triple innervation to their adult state of single innervation (Sanes and Lichtman, 1999 and Tapia and Lichtman, 2013). For technical reasons, it has remained unclear whether much more extensive circuit alterations occur in the first postnatal week or even prenatally. Knowing the extent of the early

developmental reorganization would be helpful in resolving several outstanding questions. For example, in mature muscles, motor neurons tend to innervate muscle fibers of a single type. The origin of this so-called motor unit homogeneity remains incompletely understood, with a number of different factors putatively playing a role including the following: specific targeting of axons to certain muscle fibers and not others, conversion of axons by retrograde signals from the muscle fibers, conversion of muscle fibers by activity or other signals from nerves, and synapse elimination of mismatched nerve-muscle connections. Knowing which axons initially contact each muscle fiber would be helpful in understanding the importance of several of these possibilities. Moreover, study of the developing neuromuscular system can reveal detailed circuit information, such as the number of postsynaptic cells innervated by an axon or the contact areas of all the different axons innervating the same postsynaptic cell, data that would be difficult to obtain in less accessible parts of the nervous system.