Dynamic Cholinergic Tone in the Basal Forebrain Reflects Reward-Seeking and Reinforcement During Olfactory Behavior

Abstract

Sensory perception underlies how we internalize and interact with the external world. In order to adapt to changing circumstances and interpret signals in a variety of contexts, sensation needs to be reliable, but perception of sensory input needs to be flexible. An important mediator of this flexibility is top-down regulation from the cholinergic basal forebrain. Basal forebrain projection neurons serve as pacemakers and gatekeepers for downstream neural networks, modulating circuit activity across diverse neuronal populations. This top-down control is necessary for sensory cue detection, learning, and memory, and is disproportionately disrupted in neurodegenerative diseases associated with cognitive decline. Intriguingly, cholinergic signaling acts locally within the basal forebrain to sculpt the activity of basal forebrain output neurons. To determine how local cholinergic signaling impacts basal forebrain output pathways that participate in top-down regulation, we sought to define the dynamics of cholinergic signaling within the basal forebrain during motivated behavior and learning. Toward this, we utilized fiber photometry and the genetically encoded acetylcholine indicator GAChR2.0 to define temporal patterns of cholinergic signaling in the basal forebrain during olfactory-guided, motivated behaviors and learning. We show that cholinergic signaling reliably increased during reward seeking behaviors, but was strongly suppressed by reward delivery in a go/no-go olfactory-cued discrimination task. The observed transient reduction in cholinergic tone was mirrored by a suppression in basal forebrain GABAergic neuronal activity. Together, these findings suggest that cholinergic tone in the basal forebrain changes rapidly to reflect reward-seeking behavior and positive reinforcement and may impact downstream circuitry that modulates olfaction.

Keywords: GABA; acetylcholine; basal forebrain; discrimination; go/no-go; olfaction; reward; top-down.

Figure 1

Fiber photometry of a genetically encoded acetylcholine sensor in the HDB reveals real-time cholinergic signaling during freely moving behavior. (A) Coronal section schematic showing AAV injection and implant targeting the HDB. (B) IHC of a coronal section showing GACh2.0 expression and implant targeting in the HDB. Scale bar = 1 mm. (C) Schematic of photometry system showing light paths, LED control systems, filtering, and photodetection. Timeline of surgery, recovery, and photometry recording during open field exploration, followed by behavioral training and testing. (D) Still frame from video of open field exploration with photometric recording. (E) (Top panel) Isosbestic-subtracted GACh dF/F trace during open-field arena exploration. Y scale bar = 1 dF/F and X scale bar = 60 s. (Bottom panel) zoom of blue shaded portion of trace in top panel with excitation events marked with red asterisks and suppression events marked with blue asterisks. Y Scale bar = 1 dF/F, X scale bar = 10 s. (F) Track of mouse location over 20 min of open field exploration with locations corresponding to increases in HDB cholinergic signaling (excitation events) marked with red dots and decreases (suppression events) marked with blue dots.

Figure 2

Basal forebrain cholinergic signaling reflects reward-seeking and positive reinforcement in an olfactory-cued go/no-go discrimination task. (A) (Top panel) Schematic of olfactory-cued go/no-go discrimination task showing odor presentation, decisions, and possible trial outcomes (Hit, Miss, False Alarm, and Correct Reject). (Bottom panel) Timeline of surgery, recovery, and behavioral shaping and testing with photometry recording. (B) Picture of mouse performing go/no-go task during photometric recording. (C) Accuracy in blocks of 20 trials for a go/no-go testing session with novel odors highlighting chance (50%) and criteria (85%) levels. Accuracy is from the same session as the trials shown in (D,E). (D) Heatmap showing isosbestic-subtracted GACh dF/F from individual trials in a single go/no-go testing session. Trials are aligned by trial initiation time and divided by trial outcome. (E) Average GACh dF/F traces for each trial type in the session shown in (D). Shaded areas represent 95% confidence intervals. Black line marks trial initiation time. Green line marks the average reward port entry time in Hit and False Alarm trials. (F) Area under the curve of suppression below baseline across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. *p < 0.05, ***p < 0.001 two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (G) Slopes of GACh dF/F after trial initiation across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. *p < 0.05, **p < 0.01, ***p < 0.001 two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (H) Average of mean GACh dF/F traces for across all sessions, separated by trial type, and baseline subtracted at the time of trial initiation. Shaded areas represent 95% confidence intervals. (I) Mean GACh dF/F traces for Hit and False alarm trials, averaged across all sessions, aligned to reward port entry time. Shaded areas represent 95% confidence intervals.

Figure 3

Temporal profile of cholinergic signaling in the HDB does not change with within-session discrimination learning. (A) Accuracy in blocks of 20 trials for a go/no-go testing session highlighting blocks analyzed as “pre-learning” (light blue shading) and “learned” (dark blue shading). (B) Average GACh dF/F traces for Hit and False Alarm trials separated into pre-learning and learned trials. Shaded areas represent 95% confidence intervals. Black line marks trial initiation time. Green line marks the average reward port entry time. (C) Area under the curve of suppression below baseline across trial types in pre-learning and learned blocks. Transparent circles and diamonds represent individual testing sessions. Hollow opaque circles and diamonds represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. *p < 0.05, two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (D) Slopes of GACh dF/F after trial initiation across trial types in pre-learning and learned blocks. Transparent circles and diamonds represent individual testing sessions. Hollow opaque circles and diamonds represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. Lines and error bars show mean ± SEM of means from each animal. Two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons.

Figure 4

Pseudo-learning reveals task-dependence of dynamic basal forebrain cholinergic tone. (A) Schematic of olfactory-cued go/no-go pseudo-learning task showing odor presentation, decisions, and possible trial outcomes (Pseudo-Hit, Pseudo-False Alarm, Reject). (B) Accuracy in blocks of 20 trials for a pseudo-learning testing session showing performance near chance. (C) Average GACh dF/F traces for Pseudo-Hit, Pseudo-False Alarm and Reject trials. Shaded areas represent 95% confidence intervals. Black line marks trial initiation time. Green line marks the average reward port entry time in reward-seeking trials. (D) Slopes of GACh dF/F after trial initiation across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. **p < 0.01, ***p < 0.001, two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (E) Area under the curve of suppression below baseline across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. Two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (F) Average of mean GACh dF/F traces for across all sessions, separated by trial type, and baseline subtracted at the time of trial initiation. Shaded areas represent 95% confidence intervals. (G) Mean GACh dF/F traces for Hit and False alarm trials, averaged across all sessions, aligned to reward port entry time. Shaded areas represent 95% confidence intervals.

Figure 5

HDB GABAergic neuronal activity mirrors cholinergic tone in response to positive reinforcement. (A) (Top panel) Coronal section schematic showing AAV injection and implant targeting the HDB. (Bottom panel) IHC of a coronal section showing GCaMP6M expression and implant targeting in the HDB. Scale bar = 1 mm. (B) Accuracy in blocks of 20 trials for a go/no-go testing session with novel odors highlighting chance (50%) and criteria (85%) levels. Accuracy is from the same session as the trials shown in (D,E). (D,C) Heatmap showing isosbestic-subtracted GCaMP dF/F from individual trials in a single go/no-go testing session. Trials are aligned by trial initiation time and divided by trial outcome. (D) Average GCaMP dF/F traces for each trial type in the session shown in (C). Shaded areas represent 95% confidence intervals. Black line marks trial initiation time. Green line marks the average reward port entry time in Hit and False Alarm trials. (E) Area under the curve of suppression below baseline across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. Lines and error bars show mean ± SEM of means from each animal. ***p < 0.001 two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons. (F) Slopes of GCaMP dF/F after trial initiation across trial types and testing sessions. Transparent circles represent individual testing sessions. Hollow circles represent mean values from all sessions completed by individual mice. #p < 0.05, nested one sample t-test comparing trial type values to 0. Two-way nested repeated measures ANOVA with Tukey correction for multiple comparisons shows p = 0.36 for differences between trial types. (G) Average of mean GCaMP dF/F traces for across all sessions, separated by trial type, and baseline subtracted at the time of trial initiation. Shaded areas represent 95% confidence intervals. (H) Mean GCaMP dF/F traces for Hit and False alarm trials, averaged across all sessions, aligned to reward port entry time. Shaded areas represent 95% confidence intervals.