Figure 1. Distinct activity dynamics in ventral tegmental area (VTA) and locus coeruleus (LC) axons during navigation of familiar environments.

(a) Experimental setup (top), created with BioRender.com. Example virtual reality environment. (b) Schematic representation of injection procedure (left). Representative coronal brain sections immunostained for tyrosine hydroxolase (TH) from a DAT-Cre mouse showing overlapping expression of axon-GCaMP (green) and TH (red) in VTA neurons (top) and from a NET-Cre mouse showing overlapping expression of axon-GCaMP (green) and TH (red) in LC neurons (bottom). (c) Example CA1 field of view of VTA axons (top) and LC axons (bottom). Extracted regions of interest (ROIs) used for example VTA and LC activity throughout the figure. (d) Example DAT-Cre mouse (left) and NET-Cre mouse (right) with aligned reward delivery (top, green), mouse track position (black), $\mathrm{\Delta }F/F$ from example ROI (VTA – orange, LC – blue), and mouse velocity (bottom, gray). (e, i) Position binned $\mathrm{\Delta }F/F\pm$ s.e.m in example VTA (orange) and LC (blue) ROIs during navigation of the familiar rewarded environment. (ii) Population average position binned $\mathrm{\Delta }F/F\pm$ s.e.m. in VTA ROIs (orange, n = 9 ROIs from 8 mice) and LC ROIs (blue, n = 87 ROIs from 27 sessions in 17 mice) . Linear regression analysis (on all data points, not means) shows that the population of VTA ROIs increase activity during approach of the end of the track while the population of LC ROIs have consistent activity throughout all positions. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}<1e-21}$, LC, ${\displaystyle \mathrm{p}<0.01}$. (iii) The LC dataset was resampled 1000× using n = 9 ROIs to match the number of VTA ROIs and the slope and intercept of the regression line were measured each time (blue dots). The VTA slope is steeper than all LC slopes indicating a stronger positive relationship between position and activity for VTA axons. (iv) Linear regression of position binned activity of individual VTA (orange diamonds) and LC (blue, circles) ROIs. The majority (8/9) of VTA ROIs show a significant positive relationship with position while LC ROIs show both a positive (25/87) and negative (37/87) relationship. (F, i) Same example ROIs as (d) binned by velocity. (ii) Same data as (d, ii,) binned by velocity. Linear regression shows that the population of VTA and LC ROIs have a significant relationship with velocity. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}<0.05}$, LC, ${\displaystyle \mathrm{p}<1e-68}$. (iii) Resampling shows the VTA slope and intercept is within the resampled LC slopes and intercepts indicating similar relationships with velocity. (iv) Linear regression of individual VTA and LC axons shows the majority (63/87) of LC ROIs have a significant positive relationship with velocity while only two VTA ROIs show this relationship. (g, i) Same example ROIs as (d) aligned to motion onset. (ii) Same data as (d, ii) aligned to motion onset. Linear regression shows that the population of VTA axons have a negative slope prior to motion onset while LC axons have positive slope. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}<0.01}$, LC, ${\displaystyle \mathrm{p}<1e-65}$. (iii) Resampling shows the VTA slope is negative while all resampled LC slopes are positive. (iv) Linear regression of individual VTA and LC ROIs shows the majority (56/87) of LC ROIs have a significant positive slope prior to motion onset while the majority (6/9) of VTA ROIs have a negative slope.

Figure 1—figure supplement 1. LC and VTA axon activity as a function of time and distance to reward.

(a, i) Population activity $\mathrm{\Delta }F/F\pm$ s.e.m. binned by the virtual distance to reward for ventral tegmental area (VTA) regions of interest (ROIs) (orange, 200 m track n = 9 ROIs form 8 sessions in 8 mice) and locus coeruleus (LC) ROIs (blue, 300 m track, n = 87 ROIs from 27 sessions in 17 mice) in the familiar environment. Linear regression, F test, VTA, ${\displaystyle p=1.83e-28}$, LC, ${\displaystyle p=8.77e-06}$. (ii) The LC dataset was resampled 1000× using n = 9 axons to match the number of VTA ROIs and the slope and intercept of the regression line were measured each time (blue dots). The VTA slope is steeper than all LC slopes indicating a stronger positive relationship between position and activity for VTA inputs. (iii) Linear regression of position binned activity of individual VTA (orange diamonds) and LC (blue, circles) axons. The majority (8/9) of VTA axons show a significant positive relationship with position while LC axons show both a positive (21/87 axons from 9 sessions in 8 mice) and negative (32/87 axons from 15 sessions in 9 mice) relationship. (b, i) Same data as (a, i) averaged by time to reward. Linear regression shows that the population of VTA axons has a significant positive relationship with time to reward. Linear regression, F test, VTA, ${\displaystyle p=4.44e-25}$, LC, ${\displaystyle p=0.119}$. (ii) Resampling shows the VTA slope is above the resampled LC slopes indicating VTA ROIs have a stronger positive relationship with time to reward. (iii) Linear regression of individual VTA and LC axons shows the majority (8/9) of VTA axons have a significant positive relationship with time to reward while LC axons show both a significant positive (31/87 axons from 14 sessions in 11 mice) and negative (17/87 axons from 7 sessions in 4 mice) relationship.

Figure 1—figure supplement 2. VTA DA axons expressing axon-GCaMP6s or axon-GCaMP7b show the same trends as a function of behavioral variables.

(a) Population average position binned ${\displaystyle \mathrm{\Delta }F/F\pm \mathrm{s}.\mathrm{e}.\mathrm{m}.}$ of ventral tegmental area (VTA) GCaMP6s regions of interest (ROIs) (orange, n = 5 ROIs in 4 mice) and VTA GCaMP7b ROIs (dark orange, n = 4 ROIs, in 4 mice) in the familiar environment. Linear regression, F test, GCaMP6s, ${\displaystyle \mathrm{p}=2.4264e-32}$, GCaMP7b, ${\displaystyle \mathrm{p}=1.9576e-14}$. (b) Same data as (a) binned by velocity. Linear regression, F test, GCaMP6s, ${\displaystyle \mathrm{p}=0.21076}$, GCaMP7b, ${\displaystyle \mathrm{p}=0.14113}$. (c) Same data as (a) aligned to motion onset. Linear regression, F test, GCaMP6s, ${\displaystyle \mathrm{p}=2.0974e-8}$, GCaMP7b, ${\displaystyle \mathrm{p}=0.026893}$. The slopes for the two GCaMP variants were compared using a one-way ANCOVA with Tukey HSD post hoc test * ${\displaystyle \mathrm{p}<0.05}$, ***${\displaystyle \mathrm{p}<1e-4}$.

Figure 1—figure supplement 3. Distinct activity dynamics in VTA DA axons expressing axon-GCaMP6s and LC axons expressing axon-GCaMP6 during navigation of familiar environments.

(a, i) Population average position binned ${\displaystyle \mathrm{\Delta }F/F\pm \mathrm{s}.\mathrm{e}.\mathrm{m}.}$ s.e.m. in axon-GCaMP6s expressing ventral tegmental area (VTA) regions of interest (ROIs) (orange, n = 5 ROIs in 4 mice) and locus coeruleus (LC) ROIs (blue, n = 87 ROIs from 27 sessions in 17 mice) . Linear regression analysis (on all data points, not means) shows that the population of VTA ROIs increase activity during approach of the end of the track while the population of LC ROIs have consistent activity throughout all positions. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}=2.42\mathrm{e}-32}$, LC, ${\displaystyle \mathrm{p}=0.00396}$. (ii) The LC dataset was resampled 1000× using n = 5 axons to match the number of VTA ROIs and the slope and intercept of the regression line were measured each time (blue dots). The VTA slope is steeper than all LC slopes indicating a stronger positive relationship between position and activity for VTA inputs. (iii) Linear regression of position binned activity of individual VTA (orange diamonds) and LC (blue, circles) axons. The majority (4/5) of VTA axons show a significant positive relationship with position while LC axons show both a positive (25/87) and negative (37/87) relationship. (b, i) Same data as (a, i) binned by velocity. Linear regression shows that the population of LC ROIs have a significant relationship with velocity. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}=0.211}$, LC,${\displaystyle \mathrm{p}<1\mathrm{e}-68}$. (ii) Resampling shows the VTA slope is within the resampled LC slopes indicating similar relationships with velocity. (iii) Linear regression of individual VTA and LC axons shows the majority (63/87) of LC axons have a significant positive relationship with velocity while only one VTA axon shows this relationship. (c, i) Same data as (a, i) aligned to motion onset. Linear regression shows that the population of VTA axons have a negative slope prior to motion onset while LC axons have positive slope. Linear regression, F test, VTA, ${\displaystyle \mathrm{p}=2.10e-081}$, LC, ${\displaystyle \mathrm{p}=5.51e-66}$. (ii) Resampling shows the VTA slope is negative while all resampled LC slopes are positive. (iii) Linear regression of individual VTA and LC axons shows the majority (56/87) of LC axons have a significant positive slope prior to motion onset while the majority (3/5) of VTA axons have a negative slope.