Chemogenetic manipulation of GABAergic neurons in the ZI bidirectionally modulates motivation.

(A) Schematic of experiment: CRE-dependent inhibitory (hM4DGi) or excitatory (hM3DGq) DREADDs or GFP control were bilaterally injected into the ZI of vGAT-CRE mice. (B) Representative histological image showing DREADD expression in the ZI. (C) Schematic of the operant conditioning box. (D) Experimental timeline and Progressive Ratio test results for the inhibitory DREADD (hM4DGi) cohort. Chemogenetic inhibition of GABAergic neurons in the ZI significantly reduced breakpoints following CNO administration relative to vehicle and relative to GFP controls. GFP controls showed no treatment-dependent change. (E) Experimental timeline and Progressive Ratio test results for the excitatory DREADD (hM3DGq) cohort. Chemogenetic activation of GABAergic neurons in the ZI significantly increased breakpoints following CNO administration compared to vehicle, with no effect in GFP controls. Whisker plots show median and interquartile range; individual data points are colored by sex. **p < 0.01, *p < 0.05, ns = not significant.

Chemogenetic activation of GABAergic ZI neurons rescues stress-induced deficits in motivation.

(A) Schematic of viral strategy: excitatory DREADDs (hM3DGq) or GFP control were bilaterally injected into the ZI of vGAT-Cre mice. (B) Schematic of the chronic unpredictable stress protocol. (C) Experimental groups: no stress/GFP control, stress/GFP, and stress/hM3DGq. (D) Experimental timeline showing concurrent stress exposure and operant training on FR schedules, followed by PR testing under vehicle and CNO conditions. (E) Breakpoint data under vehicle (left) and CNO (right) conditions. Under vehicle, both groups exposed to stress showed significantly reduced breakpoints compared to non-stressed controls. Following CNO administration, hM3DGq mice exposed to stress showed breakpoints no longer significantly different from controls. Whisker plots show median and interquartile range; individual data points are colored by sex. **p < 0.01, *p < 0.05, ns = not significant.

GABAergic neurons in the ZI respond to sensory stimuli across multiple modalities.

(A) Experimental timeline: vGAT-CRE mice expressing GCaMP in GABAergic neurons in the ZI underwent FR1 operant training followed by a modified FR1 task with delayed reward delivery, during which fiber photometry recordings were performed. (B) Mean z-scored calcium signal (dF/F) from GABAergic neurons in the ZI time-locked to lever press and reward delivery during the delayed-reward FR1 task. Shaded regions represent SEM. (C) Schematic of multimodal sensory stimulation paradigm in naive mice. (D) Mean z-scored calcium transients in ZI-located GABAergic neurons in response to auditory (left) and visual (right) stimuli, showing significant time-locked responses to both sensory modalities.

GABAergic ZI neurons encode the learned motivational valence of sensory cues.

(A) Schematic of Pavlovian discrimination paradigm: one visual cue (CS+) was paired with reward delivery and a second visual cue (CS-) was presented without reward, while calcium transients were recorded via fiber photometry. (B) Representative histological image showing DIO-GCaMP expression and fiber placement in the ZI. (C) Mean z-scored calcium traces in response to CS+ and CS-before training (top) and after training (bottom). Data represented as mean ± SEM (shaded area). Dashed red line indicates light onset. (D) Quantification of calcium responses (area under the curve of z-scored dF/F) to rewarded and non-rewarded cues before and after training (n=9, 5 males, 4 females). After training, responses to the CS+ were significantly greater than to the CS-. Data represented as boxplots (median, interquartile range, whiskers) with individual data points connected across conditions by lines. **p < 0.01, ns = not significant.

Inhibition of GABAergic ZI neurons does not suppress reward-associated conditioned responding.

(A) Experimental timeline: vGAT-CRE mice expressing inhibitory DIO-DREADDs (Gi) or DIO-GFP control underwent Pavlovian conditioning, FR1 operant training, and an operant probe test with CS presentations under CNO. (B) Pavlovian conditioning design: an auditory cue (CS+) was paired with reward and a visual cue (CS-) was unreinforced. (C) Pressing rate during the probe test across alternating CS+, CS-, and acclimation time blocks for GFP controls (left) and Gi DREADD mice (right). GFP controls showed a progressive decline in pressing with no differential response to CS+ versus CS-. Gi DREADD mice showed selective increases in pressing rate during CS+ blocks relative to both acclimation and CS-blocks. (Gi: n=18, 7 males, 11 females; GFP: n=11, 4 males, 7 females). Comparisons tested: acclimation: CS+ (first presentation); acclimation: CS-(first presentation); CS+ (first presentation): CS-(first presentation); CS+ (first presentation): CS-(second presentation); CS+ (second presentation): CS-(second presentation). **p < 0.01, *p < 0.05, Only significant comparisons are shown. Data represented as boxplots (median, interquartile range, whiskers) with individual data points colored by sex and group mean connected by lines.

Temporally precise optogenetic activation of GABAergic ZI neurons during reward-predictive cues enhances motivation.

(A) Representative histological image showing DIO-ChR2 expression and fiber placement in the ZI. (B) Experimental timeline: vGAT-Cre mice expressing ChR2 in the ZI underwent FR1 training, contingent FR1 training with a visual cue signaling reward availability, and PR testing under three counterbalanced stimulation conditions. (C) Pressing rates during light on (lever active) and light off (lever not active) periods across contingent FR1 training days for females (left) and males (right), showing acquisition of cue-contingent responding. (D) Total lever presses on the PR schedule under no stimulation, random stimulation, and stimulation time-locked to the light cue (CS+) for females (left) and males (right). In females, contingent optogenetic stimulation significantly increased lever pressing compared to both random and no stimulation conditions. This effect was not observed in males. Data represented as boxplots (median, interquartile range, whiskers) with individual data points, *p < 0.05, ns = not significant.

Chemogenetic inhibition of GABAergic ZI neurons does not affect locomotion.

Total distance traveled (cm) in the open field for GFP controls and inhibitory DREADD (Gi) mice under vehicle and CNO conditions. No significant differences were observed between any group-treatment combinations, indicating that the reduced pressing after inhibition of GABAergic neurons in the ZI is not attributable to changes in locomotor activity. (Gi: n=18, 10 males, 8 females; GFP: n=17, 9 males, 8 females). ns = not significant.

Chemogenetic inhibition of GABAergic ZI neurons does not affect memory.

Correct lever index (proportion of presses on the correct (rewarding) lever for GFP controls and inhibitory DREADD (Gi) mice under vehicle and CNO conditions. During training, animals were presented with two levers, only one of which was paired with a food reward; the second non-rewarding lever was present but inactive. No significant differences were observed between any group-treatment combinations, indicating that inhibition of GABAergic ZI neurons does not impair memory. ns = not significant. (GFP: n=7, 4 males, 3 females; Gi: n=10, 5 males, 5 females)

Chemogenetic inhibition of GABAergic ZI neurons does not affect food consumption.

Number of pellets consumed by GFP controls and inhibitory DREADD (Gi) mice following CNO administration. No significant difference was observed between groups, indicating that the reduction in breakpoint in PR test is not because the mice become less hungry. (Gi: n=7, 4 males, 3 females, GFP: n=8, 5 males, 3 females).

Chronic stress does not impair acquisition during operant conditioning.

Number of reinforcements earned across 14 training days for no stress/GFP control, stress/GFP, and stress/hM3DGq groups, representing learning. Separated by sex (females, left; males, right). All three groups acquired the operant response at comparable rates, with no significant differences in reinforcements earned across training days, indicating that stress exposure did not impair task learning.

Calcium responses of GABAergic neurons in the ZI of individual animals to auditory and visual stimuli.

Z-scored calcium traces (dF/F) from three individual animals in response to auditory (top row) and visual stimuli (bottom row). Individual trial traces (colored lines) and mean response (black line) are shown for each animal, confirming consistent sensory-evoked calcium transients in GABAergic ZI neurons across subjects.

Behavioral evidence of Pavlovian discrimination during training.

Inter-response time (IRT) between light onset and detector entry across training days for rewarded (CS+) and non-rewarded (CS-) cues. IRT decreased selectively for the CS+ across training days (significant light x day interaction, p < 0.001), confirming successful acquisition of cue-reward discrimination. (n=9, 5 males, 4 females). Data represented as mean ± SEM.

No significant difference in operant training acquisition between Gi DREADD and GFP groups.

FR1 training progress showing number of retrieved rewards across 10 training days for GFP and inhibitory DREADD (Gi) groups, separated by sex. (Gi: 7 males, 11 females; GFP: 4 males, 7 females). All groups acquired the operant response at comparable rates. Data represented as mean ± SEM.

No effect of testing day on PR performance in optogenetics experiment.

Total lever presses across three testing days (d1, d2, d3) for females (left) and males (right). No significant differences were observed across days for either sex, ruling out order or practice effects on the optogenetic stimulation results. ns = not significant.

Inter-response time during contingent FR1 training.

Inter-response time (IRT, seconds) between light onset and lever press across contingent FR1 training days (d5-d11) for females and males. Both sexes showed comparable decreases in IRT over training, indicating similar learning trajectories.(6 males, 7 females). Data represented as mean ± SEM.

Success rate during contingent FR1 training.

Success rate (proportion of retrieved rewards) across contingent FR1 training days (d5-d11) for females and males. Both sexes showed increasing success rates over training, with a trend toward lower success rates in females. (6 males, 7 females). Data represented as mean ± SEM.