GFP responses in two-photon imaging of neurons in V1 and ACC.

(A) Average change in apparent GFP fluorescence in L2/3 neurons in V1 on locomotion onset. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(B) As in A, but for grating onsets in L2/3 neurons in V1. Green shading marks the duration of the gratings.

(C) As in A, but for responses to optogenetic stimulation light in L2/3 neurons in V1, with VR on (light red) or in dark (dark red). Pink shading marks the duration of the light stimulus.

(D) As in A, but for visuomotor mismatch in L2/3 neurons in V1. Orange shading marks the duration of visuomotor mismatch. As mismatch events are defined to occur during times of locomotion, and locomotion itself drives GFP responses, we would expect to find an increase in these signals by chance at mismatch. To correct for this, we quantified the distribution of signal amplitude on random triggers during locomotion (95% confidence interval, gray shading).

(E) As in A, but for L5 neurons in V1.

(F) As in B, but for L5 neurons in V1.

(G) As in C, but for L5 neurons in V1.

(H) As in D, but for L5 neurons in V1.

(I) As in A, but for L2/3 neurons in ACC.

(J) As in B, but for L2/3 neurons in ACC.

(K) As in C, but for L2/3 neurons in ACC.

(L) As in D, but for L2/3 neurons in ACC.

Peak responses of GCaMP6f were larger than those observed with GFP.

(A) Example GFP response traces (black) of three neurons with corresponding running speed traces (red).

(B) Peak responses of GCaMP6f (blue) were typically an order of magnitude larger than those observed with GFP (orange). Inset is a zoomed-in version of the same data.

(C) Histogram of trial averaged responses on locomotion onset measured with two-photon imaging of calcium indicator (blue) and GFP (orange). Dotted lines indicate the mean response of the distribution of locomotion onset response in calcium indicator (blue) and GFP (orange).

(D) As in C, but for grating onset.

(E) As in C, but for visuomotor mismatch.

Fraction of neurons with significant GFP responses to different stimuli.

(A) The fraction of GFP expressing V1 L2/3 neurons responsive to locomotion, full-field grating, and visuomotor mismatch onset, quantified for all imaging sites. Each dot is an imaging site, jittered along the x-axis to make them visible. Whiskers mark the SEM. Dashed line indicates chance levels (p = 0.05). n.s.: not significant; *: p < 0.05; **: p < 0.01; ***: p < 0.001; see Table S1 for all statistical information.

(B) As in A, but for V1 L5 neurons.

(C) As in A, but for ACC L2/3 neurons.

Correlation between blood vessel diameter and GFP responses.

(A) Example average fluorescence image from an imaging site in L2/3 of V1 under a two-photon microscope. The fluorescence signal is averaged over 22 trials of optogenetic stimulation light directed at V1 while the mouse was in the dark. Shown on the right is an example blood vessel before (upper panel) and after (lower panel) stimulation.

(A1) Average area of the example blood vessel shown in A (in blue box) on light stimulation. The area of the indicated blood vessel is extracted from the trial averaged two-photon image. Shading in pink indicates the optogenetic light presentation.

(A2) Average change in GFP signal from a ROI positioned within the blood vessel (shown in A in red). Mean (red) and SEM (shading) were calculated over light stimulation trials.

(A3) Average change in GFP signal from all neurons (268 neurons) simultaneously imaged in the same site. Mean (red) and SEM (shading) were calculated over light stimulation trials.

(A4) Variance in average GFP signal in this example L2/3 site in V1 explained by area of the example blood vessel and by the change in GFP signal from a ROI within the blood vessel.

(B) As in A, from an imaging site in L5 of V1.

(C) Percentage of variance in neuronal GFP signal to individual events explained by the change in GFP signal from blood vessel ROIs. Boxes show 25th and 75th percentile, central mark is the median, and the whiskers extend to the most extreme datapoints not considered outliers.

(D) Percentage of variance in neuronal GFP signal over the entire recording explained by blood vessel GFP signal per imaging location. Each datapoint is one imaging site. Boxes show 25th and 75th percentile, central mark is the median, and the whiskers extend to the most extreme datapoints not considered outliers.

GFP responses at locomotion onset depended on visuomotor condition and cortical region.

(A) Average GFP response in L2/3 neurons in V1 on locomotion onset in closed loop (in red), in open loop (in blue), and in darkness (in grey). Here and in subsequent panels, bins with a significant difference (p < 0.01) are indicated by a black line above the plot (lines compared are marked on the left by pairs of colored horizontal bars); those with p > 0.01 are marked gray. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(B) As in A, but for L5 neurons in V1.

(C) As in A, but for L2/3 neurons in ACC.

Locomotion increased pairwise correlation of GFP signals between neurons.

(A) Average pairwise correlation in GFP signal of V1 L2/3 neurons while mice were stationary and while they were locomoting. Each dot is the mean pairwise correlation of one neuron to all other neurons in the same imaging site. Colors indicate data from different mice.

(B) As in A, but for V1 L5 neurons.

(C) As in A, but for ACC L2/3 neurons.

Signals in widefield fluorescence imaging of V1 and ACC in mice expressing GFP.

(A) Average change in signal from V1 on locomotion onset. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(B) As in A, but for grating onset in V1. Grating duration were randomly selected between 4 s and 8 s (green shading).

(C) As in A, but on visuomotor mismatch in V1. Orange shading marks the duration of visuomotor mismatch. Gray shading marks the 95% confidence interval.

(D) As in A, but for change in signal from ACC on locomotion onset.

(E) As in B, but for change in signal from ACC on grating onset.

(F) As in C, but for change in signal from ACC on visuomotor mismatch.

GRAB signals in V1 with two-photon imaging.

(A) Example two-photon image taken in V1 of mouse expressing GRAB-DA1m, with neuropil ROIs (in violet) and blood vessel ROIs (in red).

(B) Average change in GRAB-DA1m signal in V1 on locomotion onset in ROIs over neuropil (in black) and in ROIs in blood vessels (in red). Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(C) As in B, but for response to grating onset. Green shading marks the duration of the gratings.

(D) As in B, but for response to visuomotor mismatch. Orange shading marks the duration of visuomotor mismatch.

(E) Example two-photon image taken in V1 of mouse expressing GRAB-5HT1.0, with neuropil ROIs (in green) and blood vessel ROIs (in red).

(F) Average change in GRAB-5HT1.0 signal in V1 on locomotion onset in ROIs over neuropils (in black) and in ROIs in blood vessels (in red).

(G) As in F, but for response to grating onset.

(H) As in F, but for response to visuomotor mismatch.

(I) Example two-photon image of mouse expressing GRAB-ACh3.0 in V1, with neuropil ROIs (in green) and blood vessel ROIs (in red).

(J) Average change in GRAB-ACh3.0 fluorescence in V1 in response to locomotion onset in ROIs over neuropil (in black) and in ROIs over blood vessels (in red).

(K) As in J, but for response to grating onset.

(L) As in J, but for response to visuomotor mismatch.

Signals of V1 and ACC in widefield imaging in mice expressing GRAB-NE1m.

(A) Average change in signal from V1 on locomotion onset. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin. Here and below, GFP response (in gray) from Figure 7 overlaid.

(B) As in A, but for grating onset in V1. Green shading marks the duration of the gratings.

(C) As in A, but for visuomotor mismatch in V1. Orange shading marks the duration of visuomotor mismatch. Gray shading marks the 95% confidence interval.

(D) As in A, but for change in signal from ACC on locomotion onset.

(E) As in B, but for change in signal from ACC on grating onset.

(F) As in C, but for change in signal from ACC on visuomotor mismatch.

Key Resources Table

Statistical information on all analysis.

Number of mice in different experimental groups.

GFP response in V1 L2/3 on locomotion onset in closed loop was not a linear sum of component responses. Related to Figure 1.

(A) Average change in GFP signal from L2/3 in V1 in response to visual flow onset in open loop during stationary periods. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(B) As in A, but locomotion onset in dark.

(C) As in A, but for locomotion onset in closed loop.

The variance of neuronal GFP signal explained by blood vessel GFP signal as a function of smoothing window. Related to Figure 4D.

Black arrow indicates the smoothing window used in Figure 4D.

Signals in widefield fluorescence imaging of V1 and ACC in mice not expressing a fluorophore. Related to Figure 7.

(A) Average change in signal from V1 on locomotion onset. Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin, in mice not expressing a fluorophore.

(B) As in A, but on grating onset in V1. Grating duration were randomly selected between 4 s and 8 s (green shading).

(C) As in A, but on visuomotor mismatch in V1. Orange shading indicates the duration of visuomotor mismatch.

(D) As in A, but for change in signal from ACC on locomotion onset.

(E) As in B, but for change in signal from ACC on grating onset.

(F) As in C, but for change in signal from ACC on visuomotor mismatch.

GRAB signals in ACC recorded with two-photon imaging. Related to Figure 8.

(A) Average change in GRAB-DA1m signal in ACC on locomotion onset in ROIs over neuropil (in black) and in ROIs over blood vessels (in red). Mean (solid lines) and the bootstrap standard error (shading) are calculated as hierarchical bootstrap estimate for each time bin.

(B) As in A, but for response to full-field grating onset. Green shading marks the duration of the gratings.

(C) As in A, but for response to visuomotor mismatch.

(E) Average change in GRAB-5HT1.0 signal in ACC on locomotion onset in ROIs over neuropil (in black) and in ROIs over blood vessels (in red).

(F) As in G, but for response to grating onset.

(G) As in G, but for response to visuomotor mismatch.