Figure 1—figure supplement 1. Minimal stimulation protocol. ; (A) Schematic showing the method used for minimal stimulation. Patch-pipettes were filled with Atto-594 to visualize the dendritic tree of the recorded MLI. To put the stimulating pipettes in close vicinity of MLI dendrite, those pipettes were also filled with Atto-594. Recordings were performed in parasagittal sections which optimized the visualization of MLI morphology. GCL: granule cell layer (B) Minimal configuration is used to stimulate unitary GC contacts. Intensity of stimulation was chosen as the lowest intensity enable to evoke reliable synaptic responses. The plot of the success rate versus stimulation current intensity was systematically performed for each GC-MLI contact. Current values were normalized with respect to the intensity chose for minimal stimulation (star-containing symbols). The plot profile of minimal stimulation was one of the criteria used to include or reject any recording in the dataset for further analysis. The graph displays the results of experiments that were included in the dataset for PCA and k-mean clustering analysis (Figure 2). (C) Representative experiment showing how current intensity was set to ensure the stimulation of a single synaptic contact. The graphs show amplitudes of EPSC#1 and EPSC#2 (cyan and orange points for individual responses and dark blue and red points for mean values) and the success rate (black squares) following stimulation of increasing intensities. In this example, current values of 15 µA, 20 µA and 25 µA were unable to induce reliable stimulation of GC-MLI contact; 15 µA and 20 µA currents were associated with systematic failures at the first and second stimulus and 25 µA currents were associated with a high failure rate at the second stimulus. The current value of 35 µA was chosen for minimal stimulation because the success rate and the mean amplitude of EPSC#1 reached a plateau at this intensity and because of absence of failure at the second stimulus. (D) Corresponding traces at the indicated current intensity. Blue traces correspond to responses associated with failures at the first and red traces correspond to averaged traces. (E) Estimation of current spread from stimulation pipette in cerebellar slices. Currents of increasing intensities were recorded at variable distances from the tip of the stimulation pipette. The graph shows the current drop at variable intensities (inset) plots again the distance from the stimulation pipette. Solid colored lines represent the fit to the data using biexponential decay functions. Current-distance constants were conserved at any current intensity. (F) The percentage of current drop could be fitted by a biexponential decay function with 93.4% ± 1.8 of current drop described with a current-distance constant of 12.1 µm ± 0.6 µm. (G) Current intensity histogram of values used in minimal stimulation experiments. Note that the median of current intensity used for minimal stimulations median was equal to 25 µA and that almost 80% of currents were inferior to 50 µA (inset). In MLI dendrites, intersynaptic distances were estimated to 10 µm (Soler-Llavina and Sabatini, 2006). In our experimental conditions, a current of 50 µA is supposed to drop to 22.16 µA at 10 µm away from the tip of the stimulation pipette.