Supplementary Figure 10: Synchrony of glutamate transients measured by raster scanning.

We performed high speed raster scanning of yGluSnFR activity to compare signals with SLAP imaging (Fig. 6). a) Emx1-IRES-Cre transgenic mice were injected with 5x10¹¹ gc/ml AAV2/1.Syn.FLEX-yGluSnFR.A184S into visual cortex (40nL per site, 300 um deep), resulting in dense labeling of excitatory neurons. We performed strip-shaped two photon raster scan recordings (2.33 x 100 um, 8x300 pixels, 222.2Hz, 4 sec duration, 35-38 mW power) at various angles within the focal plane, 85-110um below dura, while presenting visual motion stimuli as in our other experiments. Recordings were motion-registered, and any with displacements greater than 2 pixel along the short axis of the strip were discarded. 10µm x 1.33µm regions of interest were defined along the strip ((b, left); 9 ROIs across the strip; edges were discarded). (b,right) Example trial, demonstrating both shared and local ΔF/F₀ transients across ROIs. Transients were not correlated to sample motion (bottom). c) Normalized shuffle-subtracted correlations (errorbars denote S.E.M.) binned by pairwise distance between ROIs, for SLAP and strip scanning. Normalization consisted of subtracting the mean pairwise correlation among shuffled pairs over the distance range supported by the strip scans (0-80um) for each session, and division by the standard deviation across bins. Mean correlations prior to normalization were 0.34 (raster), 0.38 (SLAP). Pairwise instantaneous correlations measured by both methods show distance-dependence over tens of microns. Raster: N = 3 mice, 154 strips. SLAP: N = 10 sessions.