Coordinated Ramping of Dorsal Striatal Pathways preceding Food Approach and Consumption

Abstract

The striatum controls food-related actions and consumption and is linked to feeding disorders, including obesity and anorexia nervosa. Two populations of neurons project from the striatum: direct pathway medium spiny neurons and indirect pathway medium spiny neurons. The selective contribution of direct pathway medium spiny neurons and indirect pathway medium spiny neurons to food-related actions and consumption remains unknown. Here, we used in vivo electrophysiology and fiber photometry in mice (of both sexes) to record both spiking activity and pathway-specific calcium activity of dorsal striatal neurons during approach to and consumption of food pellets. While electrophysiology revealed complex task-related dynamics across neurons, population calcium was enhanced during approach and inhibited during consumption in both pathways. We also observed ramping changes in activity that preceded both pellet-directed actions and spontaneous movements. These signals were heterogeneous in the spiking units, with neurons exhibiting either increasing or decreasing ramps. In contrast, the population calcium signals were homogeneous, with both pathways having increasing ramps of activity for several seconds before actions were initiated. An analysis comparing population firing rates to population calcium signals also revealed stronger ramping dynamics in the calcium signals than in the spiking data. In a second experiment, we trained the mice to perform an action sequence to evaluate when the ramping signals terminated. We found that the ramping signals terminated at the beginning of the action sequence, suggesting they may reflect upcoming actions and not preconsumption activity. Plasticity of such mechanisms may underlie disorders that alter action selection, such as drug addiction or obesity.SIGNIFICANCE STATEMENT Alterations in striatal function have been linked to pathological consumption in disorders, such as obesity and drug addiction. We recorded spiking and population calcium activity from the dorsal striatum during ad libitum feeding and an operant task that resulted in mice obtaining food pellets. Dorsal striatal neurons exhibited long ramps in activity that preceded actions by several seconds, and may reflect upcoming actions. Understanding how the striatum controls the preparation and generation of actions may lead to improved therapies for disorders, such as drug addiction or obesity.

Keywords: accumbens; basal ganglia; electrophysiology; feeding; reward; striatum.

Figure 1.

Meta-analysis of neural responses to approach and consumption during electrophysiology recordings. A, Schematic of investigated behaviors. B, Proportion of inhibitory versus excitatory responses during consumption. C, Proportion of inhibitory versus excitatory responses during action and/or approach.

Figure 2.

Dorsal striatal phasic responses to food approach and consumption. A, Schematic of analyzed behavior and histology showing recording array locations. X's indicate histological estimate of the center of recording arrays on a coronal section of a mouse brain. B, Proportion of neural responses to approach, consumption, and overlapping responses. C–F, Top to bottom, Example perievent histograms, its corresponding firing rate, and average firing rate for approach increase, approach decrease, consumption increase, and consumption decrease responses. Average firing rate expressed as z score for the corresponding groups, with shaded error bars indicating SEM. Time 0 (seconds) aligned to pellet retrieval.

Figure 3.

Ramping signals in the dorsal striatum leading up to pellet retrieval. A, Example of a negative ramping neuron. B, Example of a positive ramping neuron. C, Distribution of Pearson correlation coefficients for ramping in all recorded neurons, with significantly ramping neurons in red bars. D, Average firing rate of all negative ramping neurons. E, Average firing rate of all positive ramping neurons. Average firing rate expressed as z score for the corresponding groups, with shaded error bars indicating SEM. Time 0 (seconds) aligned to pellet retrieval.

Figure 4.

Using fiber photometry to record bulk calcium signals from striatal subpopulations. A, Coronal section at (0.5 mm) anterior to bregma for GCaMP6s expression. B, Fiber-photometry system schematic with feeding device. C, Transient rate by genotype. D, Example path plot of behavior in chamber. E, Example calcium traces from iMSN-GCaMP-, dMSN-GCaMP-, and GFP-expressing animals.

Figure 5.

Population calcium activity during approach and consumption. A, Example response around pellet retrieval of a mouse expressing GCaMP6s in iMSNs. Single trials are represented in the heat map, whereas average calcium signal is represented in the trace below. B, Same as in A for a mouse expressing GCaMP6s in dMSNs. C, Average fluorescence around pellet retrieval. D, Average fluorescence power during task periods: baseline, approach, consumption, and postconsumption. E, Velocity around time of pellet retrieval. F, Distribution of Pearson correlation coefficients for ramping in all recorded mice. Circles represent values from individual mice. *Significance from baseline. Red dashed line at time 0 indicates pellet retrieval.

Figure 6.

Striatal neurons also ramp before spontaneous movements that do not result in pellets. A, Schematic of analysis used to identify movement events. B, Average velocity trace of identified movements. C, Distribution of Pearson correlation coefficients for ramping in all recorded neurons, with significantly ramping neurons in red bars. D, Average firing of positive ramping units. E, Average firing of positive ramping units. Average firing rate expressed as z score for the corresponding groups, with shaded error bars indicating SEM. Time 0 (seconds) aligned to movement start.

Figure 7.

Population calcium signals show strong ramping before spontaneous movements that do not result in pellets. A, Example response around start of movement of a mouse expressing GCaMP6s in iMSNs. Single trials are represented in the heat map, whereas average calcium signal is represented in the trace below. B, Same as in A for a mouse expressing GCaMP6s in dMSNs. C, Average fluorescence around start of movement for all genotypes. D, Average velocity around start of movement for all genotypes. E, Distribution of Pearson correlation coefficients for ramping in all recorded mice. Red dashed line at time 0 indicates start of movement.

Figure 8.

Comparison of ramping signals in electrophysiology and population calcium recordings, for pellet retrievals and spontaneous movements. A, B, Schematic of pellet retrieval (left) or spontaneous movement (right) events. C, D, Average firing of all recorded units around pellet retrieval (left) or spontaneous movement (right). E, F, Average population calcium signal (both pathways) around pellet retrieval (left) or spontaneous movement (right). G, H, Distributions of Pearson correlation coefficients for ramping of average firing and population calcium signals in all recorded mice leading up to pellet retrieval (left) or spontaneous movements (right).

Figure 9.

Principal component (PC) analyses of firing around behavioral events. A, First 4 PCs of firing rates of all units leading up to pellet retrieval. B, First 4 PCs of firing rates of all units leading up to spontaneous movement. C, D, Percentage of variance explained by each PC for data in A and B.

Figure 10.

Changes in calcium activity during completed trials. A, Diagram of behavioral task events. B, Example path plot of completed trials. C, Averaged proportion of completed, no movement, and wrong movement trials. D, Example response of successfully completed trials around tone of a mouse expressing GCaMP6s in iMSNs. Single trials are represented in the heat map, whereas average calcium signal is represented in the trace below. E, Same as in D for a mouse expressing GCaMP6s in dMSNs. F, G, Average fluorescence. H, I, Velocity. J, K, Distance to trigger zone during completed trials. Colored bars represent trial periods. Average fluorescence in z score for the corresponding groups with colored error bars (SEM). Time 0 (seconds) indicates tone onset. *Significant difference from baseline.

Figure 11.

Changes in calcium activity during error trials A, Example path plot during no movement trials. B, D, Average fluorescence and velocity trace during no movement trials. C, Average fluorescence power during five trial periods: baseline, pretrial, waiting, tone, and postconsumption. E–H, Same as A–D for wrong movement trials. Colored bars represent trial periods. Average fluorescence in z score for the corresponding groups with colored error bars (SEM). Time 0 (seconds) indicates tone onset. *Significant difference from baseline.