Human primary muscle cells cultured aneurally in monolayer rarely contract spontaneously because, in the absence of a nerve component, cell differentiation is limited and motor neuron stimulation is missing(1). These limitations hamper the in vitro study of many neuromuscular diseases in cultured muscle cells. Importantly, the experimental constraints of monolayered, cultured muscle cells can be overcome by functional innervation of myofibers with spinal cord explants in co-cultures. Here, we show the different steps required to achieve an efficient, proper innervation of human primary muscle cells, leading to complete differentiation and fiber contraction according to the method developed by Askanas(2). To do so, muscle cells are co-cultured with spinal cord explants of rat embryos at ED 13.5, with the dorsal root ganglia still attached to the spinal cord slices. After a few days, the muscle fibers start to contract and eventually become cross-striated through innervation by functional neurites projecting from the spinal cord explants that connecting to the muscle cells. This structure can be maintained for many months, simply by regular exchange of the culture medium. The applications of this invaluable tool are numerous, as it represents a functional model for multidisciplinary analyses of human muscle development and innervation. In fact, a complete de novo neuromuscular junction installation occurs in a culture dish, allowing an easy measurement of many parameters at each step, in a fundamental and physiological context. Just to cite a few examples, genomic and/or proteomic studies can be performed directly on the co-cultures. Furthermore, pre- and post-synaptic effects can be specifically and separately assessed at the neuromuscular junction, because both components come from different species, rat and human, respectively. The nerve-muscle co-culture can also be performed with human muscle cells isolated from patients suffering from muscle or neuromuscular diseases(3), and thus can be used as a screening tool for candidate drugs. Finally, no special equipment but a regular BSL2 facility is needed to reproduce a functional motor unit in a culture dish. This method thus is valuable for both the muscle as well as the neuromuscular research communities for physiological and mechanistic studies of neuromuscular function, in a normal and disease context.