Frustration is an emotional response with aversive hedonic value, triggered by an unexpected downshift in reward quality or quantity (Amsel, 1992, Papini and Dudley, 1997). Moreover, the frustrative response is usually transient, as animals recover from the reward downshift within a few sessions. This is illustrated by the consummatory successive negative contrast (cSNC) paradigm measuring consummatory behavior in response to downshifts in sucrose concentrations (Flaherty, 1996, Torres and Papini, 2017). After training rats to lick a high sucrose concentration (e.g., 32 %), an unexpected shift to a low concentration (e.g., 2 %) leads to a transient rejection of the downshifted solution compared to unshifted controls that always received access to the low sucrose concentration. Such consummatory suppression is accompanied by behavioral and physiological outcomes that suggest a negative emotional state traditionally called frustrative nonreward (FNR; Papini et al., 2024). Some of these outcomes include escape responses, release of stress hormones, and changes in aggressive, sexual, and pain-related responses (Papini et al., 2015). However, reward comparison alone is not sufficient to elicit negative emotion. There needs to be an ample disparity in reward value (i.e., difference between obtained and expected sucrose concentration) to see the effect (Papini & Pellegrini, 2006). The present research is centered on the neurobiological substrates of the consummatory suppression that follows an unexpected reward downshift of a relatively large sucrose disparity.

The hypothetical neural circuit explaining consummatory suppression in the cSNC task incorporates brain areas associated with the detection of the reward disparity, negative emotion, and action (e.g., Ortega et al., 2017). The input to this neural circuit comes from taste receptors in the tongue that convey information to brainstem nuclei primarily via the fascial and glossopharyngeal nerves. In rats, this input activates a taste-licking modal action pattern organized at the brainstem level (Flynn & Grill, 1988). Sensory inputs are then conveyed to regions in the diencephalon and telencephalon that process and integrate information to influence behavior. The specific brain areas involved in these circuits were identified largely by lesion studies investigating how individual regions affect consummatory behavior in reward-downshift tasks. Attention has concentrated mainly on areas involved in taste perception, such as the parabrachial nucleus (Grigson et al., 1994), the gustatory thalamus (Reilly & Trifunovic, 2003), the insular cortex (Lin et al., 2009), and areas involved in negative emotion, such as the amygdala (Arjol et al., 2024; Becker and Flaherty, 1982, Guarino et al., 2020, Kawasaki et al., 2015, Kawasaki et al., 2017). However, the role of brain regions involved in the organization of behavior during reward downshifts has not been explored in detail.

The basal ganglia (BG) are potentially important to understand the organization of behavior during episodes of FNR. The BG have been implicated in emotional disorders (e.g., Gray, 1995, Macpherson and Hikida, 2019, Stathis et al., 2007) and shown to be involved in anxiety behavior in rodents (Avila et al., 2020). The BG consists of the striatum (which further includes the caudate nucleus, putamen, and nucleus accumbens), the globus pallidus, the ventral pallidum, the substantia nigra, and the subthalamic nucleus (Lanciego et al., 2012). The BG lie within a larger cortico-basal ganglia-thalamic circuit that regulates a variety of behaviors and thus has multiple inputs and outputs. The striatum specifically receives dopaminergic inputs from the ventral tegmental area (VTA) and substantia nigra pars reticulata (SNr), and glutamatergic inputs from the cortex, hippocampus, amygdala, and thalamus (Britt et al., 2012, Finch, 1996, Phillipson and Griffiths, 1985). These inputs interact with cholinergic and GABAergic interneurons within the striatum (Kita, 1993, Dautan et al., 2014). More than 90 % of neurons within the striatum are GABAergic medium spiny neurons (MSNs) that generate inhibitory signals throughout the circuit (Anderson and Hearing, 2019, Kauer and Malenka, 2007, Kemp and Powell, 1971). Notably, glutamatergic inputs to MSNs directly innervate the head of dendritic spines, whereas dopaminergic inputs innervate the spine neck. The resulting interaction between these two inputs allows for the complex modulation of MSN activity (Freund et al., 1984, Xu et al., 1989).

Based on MSNs’ molecular properties and anatomical projections within the BG, the classical view involves a distinction between the direct and indirect pathways, which modulate downstream thalamic activity in opposite directions. MSNs of the direct pathway project monosynaptically to the output nuclei of the BG: the globus pallidus internus (GPi, also known as entopeduncular nucleus in rodents) and SNr. These MSNs express dopamine D1 receptors along with dynorphin and substance P receptors. MSNs of the indirect pathway first synapse onto neurons of the globus pallidus externus (GPe), which sends axons to the subthalamic nucleus (STN), and from there to output nuclei. These MSNs also express dopamine D2 and enkephalin receptors. Functionally, the direct pathway tends to facilitate thalamic activity via the inhibition of inhibitory signals of the GPi and SNr, whereas the indirect pathway tends to inhibit activity via disinhibition of glutamatergic neurons of the STN, thus resulting in excitation of inhibitory signals of the GPi and SNr (Albin et al., 1989, DeLong, 1990, Yager et al., 2015). The relationship between the direct and indirect pathway is modulated through dopamine release from the substantia nigra acting upon the D1 and D2 receptors to balance inhibitory and excitatory signals from these two pathways (Simonyan, 2019).

In addition to the direct and indirect pathways, the nucleus accumbens (NAc) sends inhibitory GABAergic fibers to the ventral pallidum (VP), a connection that runs parallel and converges with the pathways to the GP. The VP also sends inhibitory fibers to the thalamus and projects onto a variety of other structures including the VTA, STN, hypothalamus, and lateral habenula (LHb), which in turn has reciprocal projections to the VP (Jhou et al., 2009, Root et al., 2015). Both the GP and VP send reciprocal inhibitory fibers back to the NAc (Bevan et al., 1998, Haber et al., 1985).

Several studies suggest that BG neurons play a significant role in situations involving FNR. For example, the NAc shows a decrease in dopamine efflux in a cSNC task (Genn et al., 2004), changes in cellular activity in both the dorsal striatum and NAc regions during reward downshifts using an operant procedure (Webber et al., 2016), and an increase in c-Fos expression (a marker of neural activity) during an episode involving unexpected sucrose downshifts (Pecoraro & Dallman, 2005). Furthermore, experiments using chemogenetic inactivation and excitation of the NAc and GPe indicated that these areas influence consummatory suppression after a sucrose downshift (Guarino et al., 2023). These experiments revealed interactions between these regions and reward downshift. While chemogenic inhibition of the NAc failed to disrupt consummatory suppression after the downshift, excitation significantly enhanced suppression. It was hypothesized that MSN excitation of the NAc and the resulting suppression of behavior was facilitated via activation of the inhibitory neurons in the GPe. This hypothesis was supported by excitation of the GPe which caused a reduction in consummatory suppression following sucrose downshift. These effects were observed in the absence of any evidence of motor effects in the open field (OF) task. Taken together, these experiments gave the first insight into how the BG can influence consummatory downshift.

To investigate the role of specific BG pathways during reward downshift, the current experiments utilized a double-infection chemogenetic procedure (Oguchi et al., 2015). This double-infection procedure utilizes two types of viral vector constructs which are infused intracranially. Excitatory Cre-dependent DREADDs (designer receptors exclusively activated by designer drugs) are delivered bilaterally into the departure area (the NAc) and vector constructs carrying the Cre protein are delivered bilaterally into the destination area (GPe, GPi, or VP). Retroactive transport of the Cre protein from the destination area to the departure area activates the expression of the DREADDs. Thus, although all neurons exposed to each virus are infected, only projection neurons in the NAc containing the Cre protein express the excitatory DREADDs. Therefore, DREADD activation by CNO (clozapine N-oxide) administration only excites neurons projecting to the destination areas. Both vectors also express either red or green fluorescence to allow for an accurate determination of the co-localization of the excitatory DREADDs (red mCherry) and Cre protein (green eGFP), both present in individual cells located in the departure area, the NAc.