In vitro model networks could provide cellular models of physiological relevance to reproduce and investigate the basic function of neural circuits on a chip in the laboratory. Several tools and methods have been developed since the past decade to build neural networks on a chip; among them, microfluidic circuits appear to be a highly promising approach. One of the numerous advantages of this approach is that it preserves stable somatic and axonal compartments over time due to physical barriers that prevent the soma from exploring undesired areas and guide neurites along defined pathways. As a result, neuron compartments can be identified and isolated, and their interconnectivity can be modulated to build a topological neural network (NN). Here, we have assessed the extent to which the confinement imposed by the microfluidic environment can impact cell development and shape NN activity. Toward that aim, microelectrode arrays have enabled the monitoring of the short- and mid-term evolution of neuron activation over the culture period at specific locations in organized (microfluidic) and random (control) networks. In particular, we have assessed the spike and burst rate, as well as the correlations between the extracted spike trains over the first stages of maturation. This study enabled us to observe intense neurite communications that would have been weaker and more delayed within random networks; the spiking rate, burst and correlations being reinforced over time in terms of number and amplitude, exceeding the electrophysiological features of standard cultures. Beyond the enhanced detection efficiency that was expected from the microfluidic channels, the confinement of cells seems to reinforce neural communications and cell development throughout the network.