Supplementary Figure 6: Estimation of the number of boutons from different neurons.

(a,b) Matrices of pair-wise correlation coefficients from all boutons of an example imaged region containing dLGN axons (a) or LP axons (b). Correlation coefficients were obtained by correlating calcium traces from receptive field mapping data only. (c) Cumulative distribution functions of correlation coefficients of all pairs of visually responsive thalamocortical boutons (dLGN and LP data combined). The average correlation coefficient is close to zero (gray line ('All pairs'), median = 0.037, n = 187590 pairs). Some pairs of boutons belonging to the same axon could be identified by visual inspection of averaged images. These exhibited, on average, much larger correlation coefficients (median = 0.476, n = 539). (d) In order to estimate how many boutons from different neurons were present in an imaged region we built clusters of correlated boutons that putatively belonged to the same neuron. We seeded the first cluster with the bouton exhibiting the strongest calcium response and assigned all boutons it was correlated with to the same cluster, then seeded the next cluster with the most strongly responding unassigned bouton, and so on. The graph shows the mean estimated number of visually-responsive boutons from distinct neurons per imaging region for different correlation coefficient cut-off values used for assigning boutons to the same cluster (0.18 is equal to P = 0.05 on the same-axon pair-wise correlation coefficient distribution in a). The column without cut-off shows the mean total number of responsive boutons per imaged area. Error bars represent s.e.m.