The mesopontine tegmentum in reward and aversion: From cellular heterogeneity to behaviour

The mesopontine tegmentum, comprising the pedunculopontine tegmentum (PPN) and the laterodorsal tegmentum (LDT), is intricately connected to various regions of the basal ganglia, motor systems, and limbic systems. The PPN and LDT can regulate the activity of different brain regions of these target systems, and in this way are in a privileged position to modulate motivated behaviours. Despite recent findings, the PPN and LDT have been largely overlooked in discussions about the neural circuits associated with reward and aversion. This review aims to provide a timely and comprehensive resource on past and current research, highlighting the PPN and LDT ’ s connectivity and influence on basal ganglia and limbic, and motor systems. Seminal studies, including lesion, pharmacological, and optogenetic/chemogenetic approaches, demonstrate their critical roles in modulating reward/aversive behaviours. The review emphasizes the need for further investigation into the associated cellular mechanisms, in order to clarify their role in behaviour and contribution for different neuropsychiatric disorders.


Introduction
Every organism is innately bound to seek and respond to rewards for their survival.The ability to recognize, interpret, and respond to rewards is critical to motivate animals to seek water and food, to mate and nurture progeny.Similarly, the recognition of an aversive stimulus triggers avoidance behaviours, ensuring that the individual will avoid a potentially dangerous stimulus or situation.
Many neuropsychiatric disorders present alterations in the response to positive and/or negative stimuli, for example in addiction and depression, and are associated with marked changes in neuronal circuits involved in reward/aversion (Bressan and Crippa, 2005;Cooper et al., 2017;Heshmati and Russo, 2015;Russo and Nestler, 2013;Scheres et al., 2007;Sternat and Katzman, 2016;Wang et al., 2021).
For decades, research has focused on the brain reward circuit comprising dopaminergic neurons projecting from the ventral tegmental area (VTA) to the nucleus accumbens (NAc), prefrontal cortex, amygdala, hippocampus and the bed nucleus of the stria terminalis (BNST) (Russo and Nestler, 2013).With the development of technologies such as viral monosynaptic tracing and optogenetics, other components have been identified in the so-called reward circuit, namely the lateral habenula, but also nuclei from the mesopontine tegmentumthe pedunculopontine tegmentum (PPN) and the laterodorsal tegmentum (LDT) (Floresco et al., 2003;Lammel et al., 2012;Lavoie and Parent, 1994;Mena-Segovia and Bolam, 2017;Proulx et al., 2014).
The PPN and LDT are characterized as being the major cholinergic source of the thalamus and midbrain, containing the cholinergic cell groups Ch5 and Ch6 (Mesulam et al., 1983;Semba and Fibiger, 1992).The mesopontine tegmentum has been traditionally associated with the functions of the reticular activating system and modulation of the sleep-wake cycle and arousal, as well as with functions of the mesencephalic locomotor region and modulation of goal-directed locomotion (Brudzynski, 2014;Kayama et al., 1992;Martinez-Gonzalez et al., 2011;Ruan et al., 2022;Xiao et al., 2016).Considering that the LDT and PPN are the primary sources of acetylcholine to the VTA/Substantia nigra (SN), and via the activation/desensitization of acetylcholine receptors, it constitutes a key modulator of these regions, which emphasizes the unique role of LDT and PPN in shaping basal ganglia, limbic, and motor systems.Interestingly, recent studies suggest a differential role of these nuclei in reward/aversion and reinforcement learning.For example, due to its privileged position in controlling VTA dopaminergic activity and consequent dopamine release in the NAc, the LDT has started to be considered a major locus in the neurobiology of drug abuse (Ishibashi et al., 2009;Kaneda, 2019;Kohlmeier, 2013;Kohlmeier and Polli, 2020).Others have shown that optogenetic modulation of the LDT strongly regulates not only rewarding behaviours (Coimbra et al., 2019(Coimbra et al., , 2021;;Lammel et al., 2012) but also aversion-related behaviours (Broussot et al., 2022;Liu et al., 2022;Wolfman et al., 2018).On the other hand, the PPN appears to be important for the control of gait/posture and goal-directed voluntary movement (Caggiano et al., 2018;Inagaki et al., 2022;Lau et al., 2015;Lin et al., 2023;Tl et al., 2014), and also for stimulus-outcome associations learning (Mena-Segovia and Bolam, 2017) which reflects an important function within basal ganglia nuclei (Mena-Segovia et al., 2004;Mori et al., 2016).
Considering the excellent work developed over the years by many teams using different methodological approaches, it is now timely to discuss seminal findings related to the role of PPN and LDT in reward and aversion.In this review, we provide a brief anatomical and electrophysiological description of the mesopontine nuclei, followed by a summary of it's inputs and outputs, with a focus on those important for rewarding/aversive behaviours.We then discuss lesion, pharmacological and optogenetic/chemogenetic-based studies that demonstrate that the PPN and LDT are involved in rewarding and aversive responses, and finish with a critical review of studies about human LDT and PPN function.

Neuroanatomy of mesopontine nuclei: PPN and LDT
Anatomical studies reveal a high similarity of the LDT and PPN regions in vertebrate species.PPN and LDT neurons that express the neurotransmitters acetylcholine (ACh) (Hirsch et al., 1987;Manaye et al., 1999;Pienaar et al., 2013), gamma-aminobutyric acid (GABA), glutamate (Barroso-Chinea and Bezard, 2010;Martinez-Gonzalez et al., 2012;Mena-Segovia et al., 2009;Wang and Morales, 2009), glycine (Pienaar et al., 2013) and corticotrophin-releasing factor (CRF) (Benarroch and Schmeichel, 2001;Tang et al., 2021) are found in humans, non-human primates, and rodents.Classically defined by the borders of the cholinergic neuron population, comprising the cholinergic groups Ch5 and Ch6, the PPN and LDT have similar characteristics with respect to cellular composition and connectivity, forming the mesopontine tegmentum (Armstrong et al., 1983;Mesulam et al., 1983;Satoh et al., 1983).This region contains a rostrocaudal continuum of cholinergic (but not only) neurons that extends from the caudal border of the substantia nigra (SN) pars reticulata (SNr) to the lateral part of the central gray matter, in the periventricular area near the border of the fourth ventricle (Armstrong et al., 1983;Paxinos and Watson, 2014).No clear borders divide the PPN from the LDT, except the density of cholinergic neurons along the rostrocaudal axis (Honda and Semba, 1995;Rye et al., 1987;Satoh et al., 1983).Both areas, in different proportions, contain cholinergic, GABAergic and glutamatergic neurons (Fig. 1) (Coulombe et al., 2024;Ford et al., 1995;Lavoie and Parent, 1994;Luquin et al., 2018;Mena-Segovia et al., 2009;H.-L.Wang and Morales, 2009).The distribution of neuronal subtypes is not homogeneous in either region, but it is estimated that cholinergic neurons account for approximately 27% of all neurons in the PPN and 22% in the LDT (Coulombe et al., 2024;Wang and Morales, 2009).
The PPN is located in the upper brainstem and is often described as closely interacting with the basal ganglia through direct or indirect connections (Garcia-Rill, 1991;Mena-Segovia et al., 2004;Pahapill and Lozano, 2000).The PPN can be divided in two subterritories with distinct properties: a rostral region (pars rostralis, pr) with a diffuse non-cholinergic part (pars dissipata) and a caudal region (pars caudalis, pc) with a compact cholinergic part (pars compacta).The caudal subdivision (PPNpc) is delimited by the dorsolateral border of the superior cerebellar peduncle and the rostral region (PPNpr) is surrounded by the trochlear nucleus (Lavoie and Parent, 1994;Rye et al., 1996).Anatomical segregation of the PPN from the center of the SNr has been described, and is still used to describe PPNpr and PPNpc (Mena-Segovia et al., 2009).The distinct distribution of neurochemical cell composition in the rostral and caudal regions of the PPN seems to suggest a GABAergic/cholinergic output from the PPNpc and a glutamatergic/cholinergic output from the PPNpr region (Mena-Segovia et al., 2009;Wang and Morales, 2009).Interestingly, a spatial distribution of subgroups of cholinergic neurons was reported with rostral and caudal cholinergic cells diverging in electrophysiological properties (Baksa et al., 2019).
The LDT is a small nucleus bordered dorsally by the fourth ventricle (V4) and ventrally by the cerebropontine formation and is positioned caudally to the PPN, in the pontine central gray (Cornwall et al., 1990;Maley et al., 1988;Mesulam et al., 1986;Woolf, 1991).At the caudal end, the LDT is bordered medially by the dorsal tegmental nucleus (DTg) and laterally by Barrington's nucleus (Bar) and the LC (Paxinos and Franklin, 2001).Initial studies claimed that GABA or glutamate could co-express in a subset of the population of cholinergic LDT neurons (Clements et al., 1991;Jia et al., 2003;Lavoie and Parent, 1994).However, more recent data has questioned this and it is now believed that the GABAergic, glutamatergic and cholinergic neurons represent distinct populations, as only around 2-5% of ChAT+ LDT cells co-expressed either vGluT2 or GAD (Luquin et al., 2018;Steinkellner et al., 2019;H.-L. Wang and Morales, 2009).Interestingly, using a genetic fate-mapping strategy, it was found that around 90% of cholinergic LDT cells expressed VGluT2 at some point during development (Steinkellner et al., 2019).The PPN is classically divided in two subterritories the pars rostralis (PPNpr) and the pars caudalis (PPNpc).The PPNpr is characterized by a large concentration of GABAergic neurons, similar in density to the surrounding areas, and a low density of glutamatergic cells (Mena-Segovia et al., 2009;Takakusaki et al., 1996;Wang and Morales, 2009).Caudally, in the PPNpc, the density of cholinergic cells increases, and the orientation of their dendrites becomes random and less structured.The population of glutamatergic neurons is larger

Electrophysiological properties of PPN and LDT Neurons
In the PPN, recordings in vivo identified two types of cholinergic firing patterns: 1) neurons with a phasic synchronous activity with the cortex, and 2) neurons with bursting activity during the active state of cortical slow wave activity (Mena-Segovia et al., 2008).Interestingly, recordings of PPN firing activity in rodents, cats and monkeys revealed a type of neurons that display tonic firing together with locomotion behaviour that decrease activity when there is cessation of movement (Garcia-Rill, 1991;Petzold et al., 2015;Skinner and Garcia-Rill, 1984) and another type of neurons that present a higher basal firing rate that increases with movement initiation (Matsumura et al., 1997;Mori et al., 2016), the spike activity increasing or decreasing linearly with speed (Carvalho et al., 2020).In the PPN, putative GABAergic neurons can either be interneurons or neurons that project to other brain structures and present a tonic and irregular firing (Martinez-Gonzalez et al., 2014;Mena-Segovia et al., 2009;Ros et al., 2010).An increase in the firing rate of these neurons was shown to reduce REM sleep (Kroeger et al., 2017) and cease locomotion in mice (Collazo et al., 2019;Roseberry et al., 2016), whereas a decrease in the activity of these neurons is able to control muscle tone (Benarroch, 2013).
Physiological characteristics obtained in vivo demonstrated that glutamate-containing neurons can be sub-divided into a population with fast firing activity correlated with cortical slow oscillations and another with slow firing activity not associated with cortical activity (Mena--Segovia et al., 2008;Ros et al., 2010).A recent report classified three subgroups of glutamatergic PPN neurons according to spike frequency adaptation with increasing depolarization (Dautan et al., 2021).Stimulation of glutamatergic neurons in the PPN was shown to excite dopaminergic neurons in the VTA and SN, inducing behavioural reinforcement and locomotion (Galtieri et al., 2017;Roseberry et al., 2016;Yoo et al., 2017).If activation of glutamatergic neurons was followed by sensory stimulation, the increase in the firing rate was maintained beyond the stimulation period (Petzold et al., 2015).
PPN neurons respond to visual and auditory stimuli earlier than activation of midbrain dopaminergic cells (Pan and Hyland, 2005) and have exhibited responses to the sensory and motor task events rather than the task reward, with no distinction on the nature of such neurons (Matsumura et al., 1997;K. Okada et al., 2009;Pan and Hyland, 2005).However, some studies also point out that PPN encodes reward-related information as activity of neurons in this nucleus is related to reward prediction cues, reward acquisition and changes in the reward predictive value, i.e. reward prediction error (Hong and Hikosaka, 2014;Kobayashi and Okada, 2007;Norton et al., 2011;K. Okada and Kobayashi, 2016;K.-I. Okada and Kobayashi, 2013;Ruan et al., 2022;Thompson and Felsen, 2013).Importantly, there is recent evidence obtained in humans that the PPN is implicated in reward processing, as PPN local-field potentials recordings in Parkinson's Disease patients, during a probabilistic instrumental learning task, showed an increase in activity when reward was obtained (Skvortsova et al., 2021).
Similarly to PPN, functionally distinct types of cholinergic neurons are found in the LDT (Dort et al., 2015;Maloney et al., 1999).In vivo recordings identified that putative cholinergic neurons have a short spike duration (<1 ms) and a large range of firing rates (0.08-5 Hz), while non-cholinergic neurons have a long spike duration (±2 ms) and a regular firing frequency (2 Hz) (El Mansari et al., 1989, 1990;Kayama et al., 1992;Koyama et al., 1999).In fact, many studies explore the firing activity of LDT neurons and their role in controlling states of arousal, wakefulness and rapid eye movement (REM) sleep, especially during cortical activation.Glutamatergic and GABAergic neurons of the LDT display a wider array of electrophysiological properties than cholinergic neurons, with different patterns of activity changing according to the state of the animal and during brain state transitions (Boucetta et al., 2014;Petzold et al., 2015;Ros et al., 2010).These neurons present higher firing rates during states of cortical activation (Boucetta et al., 2014).
LDT neurons respond in the presence of rewards, according to reward magnitude and expectation (Redila et al., 2015).LDT neurons also been shown to encode information about movement behaviour leading up to expected reward locations (Redila et al., 2015).Furthermore, when performing electrical stimulation of the LDT, an increase in ACh in the VTA is detected, activating dopaminergic neurons and stimulating the mesocorticolimbic pathway via nAChRs and mAChRs (Forster et al., 2002;Forster and Blaha, 2000;Lodge and Grace, 2006).Additionally, LDT neurons responded to a variety of sensory stimuli of different modalities, including auditory, visual and somatosensory (Koyama et al., 1994).
These results suggest that the mesopontine nuclei, capable of integrating reward and context information, may act as a limbic-motor and sensory-motor interface involved in initiating movement, evaluating expected outcomes and responding to rewards, likely contributing for action selection.

Functional connectivity of PPN and LDT: inputs and outputs
With the use of conventional tracers and viral tracing strategies, the connectivity of the mesopontine nucleus has been identified in recent years, revealing novel circuits.For example, it was long thought that the only source of striatal ACh was from cholinergic interneurons, but now it is known that both the LDT and the PPN also directly innervate the striatum (Coimbra et al., 2019;Dautan et al., 2014).
Importantly, a recent report using resting-state functional magnetic resonance imaging (fMRI) in humans, confirmed the PPN and LDT connectivity previously identified in other species.Specifically, mesopontine connections to frontal cortex, basal ganglia (VTA, dorsal and ventral striatum and SN) and the limbic system (thalamus, hippocampus and amygdala) have been identified (Cauzzo et al., 2022;García-Gomar et al., 2022;Singh et al., 2022).
Together, the pattern of connectivity suggests the PPNpr plays a more prominent role in circuits associated with motor function, while the LDT appears to be more associated with limbic functions.Interestingly, PPNpc appears to influence both circuits, suggesting this region may act as a bridge between motor and limbic systems.

A role For PPN and LDT In reward and aversion
The PPN and the LDT contain neurochemically diverse neuronal populations and different afferent and efferent connections, so it is not surprising that these two brain regions can be involved in reward-related behaviours.In this section, we will discuss the role of PPN and LDT in different dimensions of rewarding/aversive behaviours, and the extent to which these can be associated with particular neuronal populations of these brain regions.A summary of lesion, pharmacological, opto-and chemogenetic studies are provided in Tables 1 and 2.

PPN lesion studies
It has long been established that the PPN is important for attention, learning and the formation of action-outcome (Dellu et al., 1991;Keating and Winn, 2002;Kozak et al., 2005;Leri and Franklin, 1998).PPN lesions altered several aspects of sexual behaviour of sexually naive male rats.Specifically, the lesion disrupted the copulatory behaviour and the conditioned place preference with a receptive female as reward (Kippin and Van Der Kooy, 2003).In a reward association operant task, excitotoxic lesions of the PPN impaired discrimination of a conditioned stimulus to delivery of a food reward, disrupting lever pressing (Alderson et al., 2004;Diederich and Koch, 2005;Florio et al., 1999;Wilson et al., 2009), or even between reinforced and non-reinforced levers (Inglis et al., 2000), suggesting a role for the PPN in conditioned stimulus-reward associations and for selecting the appropriate behavioural output.In spatial discrimination using a probabilistic reinforcement learning procedure, bilateral PPN lesion did not affect acquisition of spatial discrimination but impaired behaviour flexibility, as rats failed to update their behaviour when contingency (probability of reward) was reversed (Syed et al., 2016).In food deprivation conditions, non-specific PPN lesions did not affect CPP acquisition with food (Bechara and van der Kooy, 1992a) though lesioned animals displayed increased sucrose intake (Ainge et al., 2006;Alderson et al., 2001;Keating et al., 2002).More recently, analysis of lick patterns led to the conclusion that PPN lesions disrupt the processing of the salience of the solution, irrespective of whether it is perceived as rewarding or aversive (D. A. MacLaren et al., 2015).
Despite the proposed role of PPN in conditioned stimulus-reward associations, this region does not seem to be required for a previously learned association.Contrary to naïve rats, rats with previous sexual experience did not fail to express sexually rewarded conditioned place preference after PPN lesion (Kippin and Van Der Kooy, 2003).Animals previously conditioned for morphine in a CPP task still expressed this preference after a bilateral PPN lesion (Bechara and Kooy, 1989;Bechara and van der Kooy, 1992a).Furthermore, intravenous self-administration of heroin is impaired by PPN lesions before operant training occurs, but not by lesions made afterwards (Olmstead et al., 1998).Despite being important in addressing the behavioural role of the PPN, these non-specific lesion studies do not provide insights into the cell-specific contribution for the behavioural output.So, targeted approaches were also used, specific cholinergic neuron lesion by diphtheria toxin conjugated to urotensin II (D. A. MacLaren et al., 2015;Steidl et al., 2014).Interestingly, when cholinergic neurons were targeted in the PPNpc specifically, rats did not display deficits in the learning or performance of lever pressing for food reward.Therefore, deficits in learning may be mediated by non-cholinergic neurons within the PPNpc (D. A. A. MacLaren et al., 2016).In agreement, PPT cholinergic neurons were not necessary for cocaine or heroin self-administration or the development of cocaine or heroin CPP (Steidl et al., 2014).
PPN-lesioned animals showed fewer infusions and a lower breakpoint in a progressive ratio schedule of reinforcement to self-administer amphetamine (Alderson et al., 2004).When the PPNpc is lesioned, animals increase intravenous nicotine self-administration (Alderson et al., 2006), but display a lower lever press rate and initial phase learning impairments in an operant response task for food reward (Wilson et al Decreased sexual behaviours and prevented the acquisition of place preference using a female in naïve rats (Kippin and Van Der Kooy, 2003) ICSS Disrupted lever pressing for selfstimulation (Lepore and Franklin, 1996) Heroin self-administration on FR and PR reinforcement Disrupted the acquisition of selfadministration in naïve rats and reduced the reinforcing properties (motivational drive) (Olmstead et al., 1998) Morphine and amphetamine CPP Blocked the development of place preference in drug-naïve rats (Olmstead and Franklin, 1993, 1994, 1997) Intra-VTA nicotine CPP and locomotion Blocked the development of place preference in drug-naïve rats and attenuated drug-induced locomotion (Laviolette et al., 2002) Spatial discrimination testing with probabilistic reinforcement Impairment in updating behaviour when reward probability was reversed (Syed et al., 2016)

Ibotenic acid lesion
Sucrose intake (and sucrose CPP) Lesioned rats over-consumed sucrose solution at high concentrations (no effect on sucrose solution-induced place preference) (Ainge et al., 2006;Alderson et al., 2001;Keating et al., 2002;D. A. MacLaren et al., 2015;Olmstead et al., 1999) Food/chow CPP and selfadministration on PR schedule Blocked acquisition of food-conditioned place preference and increased inappropriate responses towards the food hopper and the control lever as PR increment increased (leading to a reduction in breaking point) (Alderson et al., 2002;Bechara and van der Kooy, 1992a;Dunbar et al., 1992) Amphetamine,morphine and heroin CPP and locomotion Blocked the development of place preference in drug-naïve rats and druginduced locomotion (Bechara and Kooy, 1989;Bechara andvan der Kooy, 1992a, 1992b;Nader et al., 1994;Nader and van der Kooy, 1997;Olmstead and Franklin, 1994;Steiniger and Kretschmer, 2004) Amphetamine selfadministration Disrupted the acquisition of selfadministration in naïve rats (Alderson et al., 2004) Intra-NAc amphetamine self-administration Previously trained rats were not able to discriminate between the active and control levers (Inglis et al., 1994 Increased anxiety-like behaviours only when lesions had also damaged the cuneiform nucleus.When the cuneiform was spared, PPN lesions disinhibited performance in the tests (Walker and Winn, 2007) Electrolytic or quinolinic acid lesions ASR test Reduced prepulse inhibition (Koch et al., 1993;Swerdlow and Geyer, 1993) EPM and OF Anxiety-like behaviours and locomotion were not increased (Homs-Ormo et al., 2003) EPM and social interaction Lesions concentrated to the caudal portion increased anxiety-like behaviours (Podhorna andFranklin, 1998, 2000) Taste aversion learning Impaired the acquisition of taste aversion (Mediavilla et

2009
). PPNpr-lesioned animals showed no differences in learning to press a lever for food, but presented higher perseverative and anticipatory behaviours, as well as learning impairments in the reward extinction test, suggesting disorganized response control.Contrastingly, PPNpc NMDA lesions rather than ibotenic acid lesions (Wilson et al., 2009) did not impair lever-press learning for food reward, suggesting that the type of lesion may result in different outcomes (Leblond et al., 2014).Lesions with urotensin II (D. A. MacLaren et al., 2015;Steidl et al., 2014) also did not induce differences in sensitization to repeated administration of nicotine when compared to sham controls.PPN lesion studies have also shown that this region is important for aversive stimuli.In a shuttle box paradigm, PPN-lesioned animals presented a shorter latency to enter the compartment previously associated with a foot shock (Fujimoto et al., 1992) and display a lower avoidance rate in an active avoidance paradigm (Fujimoto et al., 1989(Fujimoto et al., , 1992;;Satorra-Marıń et al., 2001).Nevertheless, if training in the active avoidance task occurred prior to PPN lesion, animals displayed an avoidance rate comparable to control animals (Fujimoto et al., 1992).
These studies indicate that the PPN is important for the formation of positively-and negatively-valenced action-outcome or stimulusoutcome associations and in selecting the appropriate behavioural output, even though the specific contribution of each neurochemical connection is still not fully understood.
Regarding the LDT, apart from a role in sleep and wake functions, it is also implicated in reward-related functions (Coimbra et al., 2019;Kamii et al., 2015;Lammel et al., 2012;Lester et al., 2008;Nelson et al., 2007).Lesions of the LDT were shown to modulate drug-induced locomotion (Dobbs and Cunningham, 2014;Laviolette et al., 2000).Stereotypical behaviours were increased in LDT-lesioned animals after exposure to nicotine (Forster and Blaha, 2000;Ishibashi et al., 2009) or amphetamine (Forster et al., 2002).In contrast, bilateral lesions of the LDT were shown to decrease stereotypy in response to morphine (Forster

Cocaine and nicotine selfadministration
Reduced self-administration (Corrigall et al., 1999(Corrigall et al., , 2002) ) AChR antagonist ICSS and nicotine or cocaine self-administration Facilitated lever-pressing for rewarding self-stimulation and reduced nicotine but enhanced cocaine self-administration (Corrigall et al., 1999(Corrigall et al., , 2001(Corrigall et al., , 2002;;Lança et al., 2000;Yeomans et al., 1993) AChR agonist ICSS and nicotine or cocaine self-administration Blocked lever-pressing for rewarding selfstimulation and reduced nicotine and cocaine self-administration on different schedules of reinforcement (Corrigall et al., 2002;Yeomans et al., 1993) Cholinesterase inhibitor Nicotine or cocaine selfadministration Reduced nicotine and cocaine selfadministration (Corrigall et al., 2002) Nicotine CPP Induced place preference (Iwamoto, 1990  Alcohol locomotion and 2-bottle (alcohol/water) free choice Attenuated alcohol-induced locomotion and reduced alcohol intake (Vallöf et al., 2019) Sexual interaction test and food consumption Decreased sexual interactions and food intake (antagonism increased food intake) (Reiner et al., 2018;Vestlund and Jerlhag, 2020) Amylin and calcitonin receptor agonist (+ GABA receptors antagonists) Food consumption and sucrose PR reinforcement Reduced food intake and motivated feeding behaviours (antagonists reversed the food intake suppressive effects) (Reiner et (Dobbs and Cunningham, 2014).Specific cholinergic lesioning of the LDT, after operant self-administration training, increased the latencies for rats to initiate cocaine operant self-administration resulting in reduced cocaine intake, but once the response was initiated or with priming, rats responded similarly to pre-lesion conditions (Steidl et al., 2015).This suggests that the LDT is necessary for the responsiveness to cocaine-predictive stimuli/cues, which trigger drug-seeking behaviour.

LDT pharmacological studies
Apart from data obtained in lesion studies, the functions of the PPN and LDT have also been inferred from data obtained following the local application of various pharmacological agents.Manipulation of the PPN by local drug microinjections has produced mixed results.Intra-PPNpc injection of muscimol decreased sensitivity to degradation of the contingency between actions (lever pressing) and outcomes (food reward).Here, treated animals maintained lever pressing, showing that the PPNpc is critically required for updating associations between actions and outcomes, but not in the continued performance of previously learned associations (D. A. A. MacLaren et al., 2013).Administration of the muscarinic agonist carbachol reduced nicotine self-administration on a FR or PR schedule (Corrigall et al., 2002).The same outcome was achieved when GABA agonists muscimol and baclofen were administered into the PPN, on FR schedule but not on PR (Corrigall et al., 2001).Moreover, intra-PPN muscimol administration decreased ethanol self-administration (Samson and Chappell, 2001) and reduced the breaking point of instrumental responding to obtain food pellets (Diederich and Koch, 2005).Intra-PPN microinjection of the nicotinic receptor agonist DHBE also increased cocaine self-administration (Corrigall, 1999).However, stimulation of the PPN, via injection of the muscarinic antagonist scopolamine, increased dopaminergic burst firing and efflux in the dorsal striatum and behavioural activation by increasing lever pressing in a self-stimulation paradigm (Chapman et al., 1997;Lokwan et al., 1999;Yeomans et al., 1993).These data are difficult to interpret and may be the result of the complex local microcircuitry and inputs/outputs of this brain region.
Pharmacological blockade of neurons of the LDT results in motor (Dobbs and Cunningham, 2014), learning (Shinohara et al., 2014) and behavioural state deficits (Kohlmeier and Kristiansen, 2010).Inactivation of the LDT significantly impaired choice accuracy in a spatial-reward discrimination task (Redila et al., 2015).The inactivation of LDT (baclofen/muscimol) reduces the phasic firing of VTA dopaminergic neurons (Lodge and Grace, 2006), which is thought to be sufficient for behavioural conditioning (Tsai et al., 2009).Intra-VTA cholinergic antagonists blocked the three-component LDT electrically evoked dopamine response in the NAc (Forster and Blaha, 2000).Intra-LDT administration of a muscarinic agonist decreased intravenous cocaine self-administration in a fixed ratio (FR) schedule and breakpoint for cocaine responding (Shabani et al., 2010).Additionally, this agonist was able to attenuate food self-administration (Shabani et al., 2010).In a self-administration paradigm, systemic cocaine was shown to induce LDT-stimulated levels of dopamine in the NAc that were blocked by microinjection of scopolamine into the VTA (Lester et al., 2010).Additionally, inactivation of the LDT (blockade of AMPA and NMDA receptors) or intra-VTA injection of scopolamine or mecamylamine inhibited acquisition of cocaine CPP paradigm (Shinohara et al., 2014).In accordance, LDT microinjection of an AMPA receptor antagonist attenuated the reinstatement of drug seeking induced by a priming injection of cocaine, and similar results were obtained when injecting the AMPA receptor antagonist or cholinergic receptor antagonists into the VTA (Schmidt et al., 2009).Furthermore, VTA injection of carbachol produced CPP (Ikemoto and Wise, 2002) and VTA administration of nicotine enhances drug-seeking behaviour during self-administration paradigms and is able to produce changes in plasticity in dopamine neurons that may underlie addiction (Volkow and Morales, 2015).This suggests that glutamatergic activity in the LDT and LDT-to-VTA cholinergic projections are important in modulating the rewarding effects of drugs of abuse and natural rewards, and NAc dopamine release.Indeed, the inactivation of LDT with baclofen/muscimol reduced the phasic firing of VTA dopaminergic neurons (Lodge and Grace, 2006).Intra-VTA cholinergic antagonists blocked the three-component LDT electrically evoked dopamine response in the NAc (Forster and Blaha, 2000).
Acute application of nicotine activated pre and postsynaptic nAChRs in the LDT, which would be expected to alter output to the VTA (Ishibashi et al., 2009), and ex vivo cocaine exposure elicited rises in intracellular calcium within LDT neurons (Lambert et al., 2017).Increased acetylcholine levels within the VTA of mice after intraperitoneal administration of methamphetamine are attenuated by intra-LDT microinjections of the muscarinic 2 agonist, oxotremorine (Dobbs and Mark, 2012;Laviolette et al., 2000).Greater excitability was found in the LDT-VTA-NAc pathway after chronic exposure to amphetamine, suggesting that the LDT might contribute to long-term potentiation of DA VTA neurons relevant for behavioural sensitization (Nelson et al.,

2007
).Thus, these findings suggest an important role of the LDT within the circuitry underlying the neurobiology of addiction and highlight that drugs of abuse can alter the function of neurons in this nucleus.Interestingly, in recent years several classical energy balanceregulating peptides have been implicated in reward-related behaviours.LDT injections of glucagon-like peptide 1 (GLP-1) agonist decreased sexual interaction behaviours in sexually naïve male mice and decreased food intake, whereas antagonism increased food intake (Reiner et al., 2018;Vestlund and Jerlhag, 2020).Likewise, ghrelin signaling in the LDT regulated sexual reward and sexual behaviour in mice (Prieto-Garcia et al., 2015).Furthermore, this peptide administered into the LDT modulated accumbal dopamine through nAChRs in the VTA (Jerlhag et al., 2012).An agonist of amylin and calcitonin, locally administered into the LDT reduced hedonic feeding behaviours in male rats, likely through the LDT GABAergic neuron population (Reiner et al., 2017).Nevertheless, these peptides also regulate drug-related behaviours for example, GLP-1 signaling in the LDT, from the NTS, recruits GABAergic cells and attenuates the reinstatement of cocaine seeking self-administration (Hernandez et al., 2021).These peptides also modulate, via LDT, alcohol intake, alcohol-induced locomotion, alcohol-elicited NAc DA release and cocaine locomotor stimulation (Jerlhag et al., 2009;Kalafateli, Aranäs, et al., 2021;Kalafateli, Satir, et al., 2021;Reiner et al., 2018;Vallöf et al., 2019).Altogether, these data suggests that endogenous appetite-regulatory peptides in the LDT might modulate sexual and food reward and drug-associated behaviours, indicating a possible neurochemical overlap between the hedonic reward system and those regulating energy balance.

PPN and LDT optogenetic/chemogenetic studies
Optogenetics provide an elegant way of modulating neuronal activity with cell specificity and temporal control, which has dramatically facilitated behavioural studies, and allowing researchers to infer causality of activating/inhibiting specific neuronal populations with a particular behavioural outcomes.
Regarding effects on locomotor behaviour, recent studies revealed that optogenetic stimulation of the PPN (unspecific) or glutamatergic PPN neurons elicited locomotion (Caggiano et al., 2018;Masini and Kiehn, 2022;Roseberry et al., 2016).While others report a reduction in motor activity (Dautan et al., 2021).Additionally, optogenetic stimulation of cholinergic neurons induced and maintained locomotion only in moving mice, however the effect is less pronounced than that produced by the optogenetic stimulation of glutamatergic neurons (Josset et al., 2018).Stimulation of PPN GABAergic neurons, during running, caused deceleration (Roseberry et al., 2016).For more details see (Lin et al., 2023).
Regarding reinforcement, photoactivation of PPN glutamatergic cell bodies or PPN-VTA glutamatergic terminals is sufficient to drive intracranial self-stimulation reinforcement (Yoo et al., 2017).Using an appetitive Pavlovian conditioning paradigm, photoinhibition during cue presentation of PPN-VTA excitatory (cholinergic and glutamatergic), but not GABAergic projections, disrupted cue-reward associative learning, possibly through an effect mediated by non-dopaminergic VTA neurons (Yau et al., 2016).Moreover, in an attentional set-shifting task, photometry recordings demonstrated that PPN cholinergic neurons respond to changing stimulus-outcome contingencies (reward location switching).While chemogenetic inhibition of these cells suppressed reversal learning, general or PPN terminals activation in specific downstream regions (SN pars compacta, mediodorsal thalamus (MDT) and parafascicular nucleus (PN)) improved the performance (Ruan et al., 2022).These findings suggest that cholinergic PPNs neurons are necessary to inhibit previous strategies to facilitate a switch/adjustment in behaviour, further supporting a role of PPN in updating behaviour.Noteworthy, inhibition of PPN cholinergic cells did not result in deficits in the initial reward discrimination learning (Ruan et al., 2022), which indicated that other populations of PPN neurons may account to reinforcement learning deficits observed after lesion.
Optical activation within the PPN of GABAergic projections from the SNr (Hormigo et al., 2019), zona incerta (Hormigo et al., 2020) or NAc (Hormigo, Zhou, et al., 2021) block avoidance responses to a cue signaling a footshock, while sparing escape responses to footshock.Optogenetic inhibition of PPN neurons or excitation of local GABAergic neurons also block signaled active avoidance, but not escape (Hormigo et al., 2019).Additionally, excitation of PPN serves as an effective CS to drive avoidance in the absence of a natural CS and drived innate escape responses (Hormigo et al., 2019).Employing fiber photometry, the PPN was found to be active during the auditory CS; inhibition of PPN neurons during CS presentation impaired naïve animals to avoid the shock (Hormigo, Shanmugasundaram, et al., 2021).Together, the results point that the PPN is essential for the acquisition and expression of signaled active avoidance.
General optogenetic activation of PPN projections to the central amygdala (CeA) induced strong place avoidance in the real-time place preference (RTPP) and CPP tasks, decreased time spent in the center of an open field (OF) arena and decreased the time exploring the open arms in the elevated plus maze (EPM) (A.Liu et al., 2023).In a contingent place preference-like task, mice significantly avoided the side paired with the PPN-CeA cholinergic projections activation.This avoidance was also observed in a subsequent laser-extinction test phase (Aitta-aho et al., 2018).Additionally, mice displayed, while being in the stimulated-side, reduced activity, which could be associated with defensive behaviour (Aitta-aho et al., 2018).Photoinhibition of PPN cholinergic neurons or PPN-VTA cholinergic terminals have also been shown to induce place aversion (Xiao et al., 2016).
Optical stimulation in the VTA of either PPN or LDT axons increased motor activity, either during PPN axon stimulation or during increasing cumulative LDT axon stimulation (Dautan, Souza, et al., 2016).Intriguingly, the two types of cholinergic inputs exert distinct modulations on particular neurons: NAc-projecting dopaminergic neurons are preferentially excited, whereas the NAc-projecting non-dopaminergic neurons are selectively inhibited by LDT cholinergic activation, with PPN stimulation having no effect (Dautan, Souza, et al., 2016).The LDT appears to be uniquely positioned to exercise control of burst firing and increase in dopamine release via their direct excitatory projections to mesoaccumbens-projecting VTA dopamine neurons.In the scope of this review, both VTA and NAc regions are treated broadly as singular entities.However, it is important to highlight the heterogeneity within these regions, with specific subregions exerting distinct influences on reward and aversive processing.For instance, studies have demonstrated that the VTA's subregions display divergent patterns of connectivity and neurotransmitter release, which in turn distinctly modulate reward-related behaviors and aversive responses (for a review see (Lu et al., 2024;Morales and Margolis, 2017).Similarly, NAc core and shell contribute differentially for valence and motivated behaviors (see (G. Chen et al., 2023;R. Chen et al., 2021;Domingues et al., 2023;Soares-Cunha et al., 2016).
Non-selective optogenetic activation of all LDT-VTA terminals is reinforcing since it can enhance operant behaviour for a lever associated with laser stimulation (Steidl and Veverka, 2015).Recently, our laboratory demonstrated that LDT-VTA activation/inhibition increased/decreased value of a stimulation-paired reward in a two-choice lever task, enhanced/decreased motivational drive in a progressive ratio task and induced place preference or aversion (Coimbra et al., 2021).Steidl and colleagues showed that specific optogenetic stimulation of LDT cholinergic, but not glutamatergic, terminals in the VTA is able to induce place preference in a non-contingent paradigm (Steidl, Wang, et al., 2017).Also, optogenetic stimulation of both glutamatergic and cholinergic LDT-VTA projections is sufficient to promote reward-related behaviours in rodents, specifically, inducing place preference and delaying extinction in a lever-pressing task, where sugar pellets were replaced by laser stimulation (Coimbra et al., 2021;Dautan, Souza, et al., 2016;Lammel et al., 2012;Steidl and Veverka, 2015;Xiao et al., 2016).
Interestingly PPN neurons, on the other hand, appear to preferentially influence motor control via cholinergic projections to the SN, but not VTA (Xiao et al., 2016), whereas glutamatergic PPN neurons target non-dopamine VTA neurons, which have been demonstrated to be necessary for the acquisition of stimulus-reward associations (Yau et al., 2016).Chemogenetic activation of the nucleus tractus solitarius (NTS)-to-LDT circuit or LDT-VTA GABA projections was sufficient to attenuate the reinstatement of cocaine seeking after a cocaine priming injection (Hernandez et al., 2021).Specific activation of LDT SOM+ GABAergic neurons induced release of DA in the lateral NAc and elicited place preference, an effect presumably involving the disinhibition of excitatory neurons in the hypothalamus (Du et al., 2023).Chronic social defeat stress causes hyperactivity of LDT neurons that in turn triggers cellular adaptations in the VTA.Interestingly, chemogenetic-mediated inhibition of cholinergic LDT neurons prevented stress-induced VTA dopaminergic neurons dysregulation and depressive-like behaviours (Fernandez et al., 2018).
We revealed that optogenetic activation of LDT-NAc projections enhanced motivational drive, shifted preference to an otherwise equal reward in a two choice task, whereas inhibition induced the opposite outcome, and induced place preference (Coimbra et al., 2019).Targeting the main populations of the LDT, we demonstrated that activation of cholinergic inputs to NAc was sufficient to recapitulate all the effects of general stimulation.Nevertheless, glutamatergic projections also enhanced preference for the laser-associated reward and increased motivation while GABAergic inputs decreased the value of the laser-paired reward and decreased motivation (Coimbra et al., 2019).
Conversely, general stimulation of IPN terminals in the LDT elicited avoidance behaviour in a RTPP paradigm, suggesting that these projections contribute to aversion (Wolfman et al., 2018).Furthermore, in the same study, these projections were implicated in the aversive effects of high doses of nicotine, as photoinhibition blocked the conditioned place aversion elicited by high concentrations of the stimulant (Wolfman et al., 2018).Recently, inhibition of LDT GABAergic terminals in the VTA prevented nicotine aversion (C.Liu et al., 2022).Activation of LHb glutamatergic inputs to the LDT interneurons generated fear-like responses (Yang et al., 2016).Interestingly, both activation and inhibition of LDT GABAergic neurons decreased DA signaling in NAc and induced various aversive behavioural phenotypes (Du et al., 2023).Selectively stimulating LDT PV+ GABAergic cells induced fear-and anxiety-like responses including freezing, accelerated heart rate and increased serum corticosterone (Yang et al., 2016).Note that activation of SOM+ GABAergic neurons had a positive valence effect (Du et al., 2023), which highlights the complexity of the LDT microcircuitry.
Optical inhibition of LDT to VTA projections or LDT-NAc glutamatergic inputs activation induced avoidance in a RTPP task (Coimbra et al., 2019(Coimbra et al., , 2021)).Chemogenetically silencing LDT neurons or LDT-VTA projections prior to administration of electrical footshocks reduced the freezing responses to the aversive stimulant, whereas LDT-VTA chemogenetic activation induced strong freezing responses (Broussot et al., 2022).Selective activation of LDT-VTA GABAergic terminals, but not other populations, dampened freezing responses and optogenetic activation induced freezing and bradycardia in the absence of shock (Broussot et al., 2022).In an OF, activation of LDT glutamatergic terminals in the lateral hypothalamus induced freezing reactions (Du et al., 2023).Furthermore, optical activation of LDT cholinergic projections to the ventral posterior complex of the thalamus during the EPM led to a significant decrease in time spent in the open arms, indicating that the circuit induced more anxiety in animals, indicative of an internal negative state (P.Zhao et al., 2022).Photostimulation of corticotropin-releasing hormone (CRH)-positive LDT neurons during an OF test resulted in less time and shorter distances exploring the center of the arena, and during EPM, resulted in fewer entries and less time and distance exploring in the open arm, suggesting that transient activation of LDT CRH neurons induces an internal aversive state (Tang et al., 2021).
Overall, the available data suggests that different neuronal populations in the PPN and LDT can mediate reward and aversion processes.Additional studies are urgently needed to identify the genetical, anatomical (connectivity and/or spatial organization) and/or functionality/activity markers that can be used to further (or better) distinguish subsets of mesopontine neurons involved in different valence behaviours.

PPN and LDT in humans: clinical considerations
Considering evidence for the involvement of PPN and LDT in different behaviours, it is important to understand how these regions are altered in neuropsychiatric disorders.
Post-mortem data has shown that PPN undergoes alterations that may potentially contribute to the development of several neurological disorders, including Parkinson's disease (PD), Alzheimer's disease and schizophrenia due to the connections between this region and the basal ganglia and thalamus.Initially, an increase in cholinergic neurons in the mesopontine region of schizophrenia patients was reported (Garcia-Rill et al., 1995;Karson et al., 1991), however, using a different labelling method, a decrease of cholinergic cells was later observed (Karson et al., 1996).A subsequent work revealed no changes in this neuronal population (German et al., 1999).These discrepancies likely reflect small sample sizes, different ages between the schizophrenic and control subjects or different labelling of cholinergic cells (ChAT-vs.nicotinamide adenosine dinucleotide phosphate (NADPH)-staining) used in the studies.Additionally, there is loss of cholinergic neurons in the PPN in PD, and in progressive supranuclear palsy and multiple system atrophy (Benarroch, 2013;Benarroch et al., 2002;Hepp et al., 2013;Hirsch et al., 1987;Jellinger, 1988;Schmeichel et al., 2008;Zweig et al., 1987).These disorders are characterized by prominent postural and gait abnormalities, which may in part reflect PPN involvement (Bakker et al., 2008;Rochester et al., 2012).In fact, PD patients with damages to cholinergic neurons in the PPN are prone to falls due to postural instability and gait dysfunction (Bohnen et al., 2009(Bohnen et al., , 2013;;Rinne et al., 2008).PPN neuronal degeneration caused DA non-responsive gait and balance impairment (Grabli et al., 2013;Karachi et al., 2010).PPN degeneration might also be related to non-motor functions that are impaired in PD, like mood, cognition, sleep and sensory functions (Jankovic, 2008).For instance, the role of the PPN in controlling REM sleep may explain why PD patients frequently present abnormal REM muscle tone and concomitant REM sleep behaviour disorder (Chambers et al., 2020;Dugger et al., 2012;Gagnon et al., 2009;Iranzo et al., 2006;Kotagal et al., 2012;Marion et al., 2008;Romigi et al., 2008).
Deep brain stimulation (DBS) of the PPN has been considered as a treatment in the late stages of PD, in order to activate the remaining neurons in this region to compensate for degeneration and improve postural instability, freezing of gate (FOG) and falling (Plaha and Gill, 2005;Romigi et al., 2008;Stefani et al., 2007;J.-W. Wang et al., 2017;Wilcox et al., 2011).Several studies demonstrate that PPN DBS can ameliorate FOG or general gait and frequency of falls in PD patients (Khan et al., 2012;Mestre et al., 2016;Pereira et al., 2008;Welter et al., 2015).Imaging studies in PD patients indicate that unilateral PPN DBS increased cerebral blood flow into the central thalamus and cerebellum (Ballanger et al., 2009) and improved movement and/or decreases falls (Mazzone et al., 2014;Moro et al., 2010).Others supported bilateral stimulation since it appeared to be more successful in improving PD-related symptoms (Thevathasan et al., 2010(Thevathasan et al., , 2011(Thevathasan et al., , 2012)), particularly with electrode placement in the caudal PPN (Goetz et al., 2019;Khan et al., 2012).However, further studies indicated variable outcomes with some mild alleviation (Ferraye et al., 2010) or no benefit at all R. Bastos-Gonçalves et al. (Thevathasan et al., 2018;Yu et al., 2020).This may be due to patient choice (small sample size and differences in disease progression), target option heterogeneity, surgical procedure differences and stimulation protocols (Chambers et al., 2020;Lin et al., 2023;Mestre et al., 2016;J.-W. Wang et al., 2019).In addition to effects on motor symptoms, PPN DBS has produced alterations in presumed attentional aspects in PD patients.PPN DBS at therapeutic frequencies (20-35 Hz) improved speed of reaction/reaction time in a simple reaction time task, where the subjects had to press a button upon presentation of a visual stimulus (Thevathasan et al., 2010).However, DBS did not improve the reaction time in more complex and longer tasks, that required inhibitory control and different responses according to the visual stimuli.Another report (with 8 Hz and 20 Hz stimulation frequencies) recapitulated the finding with similar simple alertness tasks and likewise, in a more complex go/no-go task, there were no variations in performance (Fischer et al., 2015).Importantly, the authors argued that the improvement in reaction time could have been due to higher motor control rather than augmentation of attention and, therefore, the conclusion that PPN DBS modulates attentional processing should be accepted with caution.
Few human studies have identified the function of the mesopontine in higher cognitive functions like decision-making and learning.However, Khalighinejad et al. (2020) used ultrahigh-field functional magnetic resonance imaging (fMRI) in healthy subjects to understand how contextual factors and internal state, shaped by present and past environment, influence when an action is made and to elucidate the neural circuits mediating these processes.Subjects had to track stimuli on a screen, that potentially contained a monetary reward and changed accordingly to magnitude and probability of reward, and had to choose when to make a response at a time of their choice (Khalighinejad et al., 2020).The authors described that PPN activity encoded time to act before the initiation of action, earlier in the trial compared to regions like NAc and dorsal striatum.PPN activity was also related to past reward outcome.Additionally, Skvortsova et al. (2021) reanalyzed a previous fMRI dataset collected in healthy volunteers performing a task that involved choosing between two visual cues with either a left-hand or right-hand button press and learning the probabilistic associations with monetary reward.The results suggested that PPN activity discriminates between reward and no reward outcomes (Skvortsova et al., 2021).Then, in PD patients on a clinical trial for DBS (only 3 subjects), they recorded PPN local field potentials (LFPs) while subjects performed the same instrumental learning task.The analysis revealed an increased activity around the time of choice in a low-(alpha or beta) frequency band and identified increased activity in reward compared to no reward outcomes.This does not mean that PPN activity is causally involved in learning however, as PPN DBS stimulation, in a second learning task involving learning to maximize monetary reward and minimize monetary loss (punishment), improved reward learning (increased choice rate for the most rewarded cue) in all 3 patients, while the effect on punishment learning was inconsistent.Even though these results should be considered with caution, the data suggests that the PPN is implicated in reward processing and that PPN stimulation enhances reinforcement in humans, which is in line with considerable evidence from animal studies (Tables 1 and 2).
From a more clinical perspective, these findings suggest that the impact of PPN DBS could go beyond the expected motor effects on gait and posture and that this strategy could be used to treat common nonmotor symptoms in PD patients.

Conclusions and future perspectives
Accumulating evidence supports a key role for PPN and LDT in reward and aversion and supports a differential contribution of distinct neuronal populations.However, it is still do not fully understood how these brain regions integrate/process reward/aversive signals to shape valenced behaviours.Functional specialization of the brainstem in three subterritories, namely PPNpr, PPNpc and LDT, is supported by neurochemical, anatomical and behavioural evidence.These regions interact forming the basis of a more complex mechanism of motor-toassociative-to-limbic function regulation along the rostrocaudal axis.Rather than considering them as separate structures, these regions may operate in coordination as a single and integrated functional complex to provide context-dependent signals in response to behavioural demands.
It is also important to recognize that within each of these brain regions, there is also anatomical, molecular, and functional heterogeneity.A recent study using single nucleus RNA sequencing and spatial transcriptomics, has identified 27 neuronal clusters within the LDT (55% GABAergic, 32% glutamatergic and 13% cholinergic) (Nardone et al., 2024).This emphasizes the need for additional studies focusing on the contribution of specific neuronal subtypes in neuronal modulation and behavior.Considering the molecular signature of mesopontine nuclei, it is tempting to speculate that some of these clusters may also release non-classical neurotransmitters.In fact, there is some evidence showing that mesopontine cholinergic neurons can also produce substance P (Kohlmeier et al., 2002;Vincent et al., 1983) or CRF (Romero-Leguizamón and Kohlmeier, 2020) This provides another layer of complexity to how mesopontine nuclei can influence neuronal activity in downstream regions and consequently, behavior.Future studies leveraging advanced techniques such as spatial transcriptomics and in vivo neurotransmitter biosensors could unveil the complex interplay of classical and non-classical neurotransmitters in these regions.This would not only deepen our understanding of the molecular diversity within the mesopontine tegmentum, but also pave the way for eventually translating some of these findings to the clinical context.

Fig. 1 .
Fig. 1. -Main neuronal populations of the PPN and LDT.The PPN is classically divided in two subterritories the pars rostralis (PPNpr) and the pars caudalis (PPNpc).The PPNpr is characterized by a large concentration of GABAergic neurons, similar in density to the surrounding areas, and a low density of glutamatergic cells(Mena-Segovia et al., 2009; Takakusaki et al.,  1996;Wang and Morales, 2009).Caudally, in the PPNpc, the density of cholinergic cells increases, and the orientation of their dendrites becomes random and less structured.The population of glutamatergic neurons is larger than the cholinergic and GABAergic populations(Martinez-Gonzalez et al., 2012;Wang and Morales, 2009).The LDT contains a ventral part with acetylcholine esterase (AChE) -positive cells.The rostral portion contains a high concentration of glutamatergic cells, the medial portion presents the highest concentration of cholinergic cells of the three subdivisions, but it has a similar proportion of the three main populations, and the caudal end presents more GABAergic cells(Wang and Morales, 2009).Cholinergic neurons (ACh, green), Glutamatergic neurons (Glu, blue) and GABAergic neurons (GABA, red).Representation based on(Pienaar et al., 2017).

projections to Fig. 2. Mesopontine connections functionally involved in rewarding and/ or aversive responses. Aq
have been identified.More specifically, the PPNpr cholinergic neurons present more robust monosynaptic

Table 1
., Summary of findings implicating the PPN in reward/aversion-associated behaviours. )

Table 2
Summary of findings implicating the LDT in reward/aversion-associated behaviours.