Olfactory modulation of the medial prefrontal cortex circuitry: implications for social cognition

Abstract

Olfactory dysfunction is manifested in a wide range of neurological and psychiatric diseases, and often emerges prior to the onset of more classical symptoms and signs. From a behavioral perspective, olfactory deficits typically arise in conjunction with impairments of cognition, motivation, memory, and emotion. However, a conceptual framework for explaining the impact of olfactory processing on higher brain functions in health and disease remains lacking. Here we aim to provide circuit-level insights into this question by synthesizing recent advances in olfactory network connectivity with other cortical brain regions such as the prefrontal cortex. We will focus on social cognition as a representative model for exploring and critically evaluating the relationship between olfactory cortices and higher-order cortical regions in rodent models. Although rodents do not recapitulate all dimensions of human social cognition, they have experimentally accessible neural circuits and well-established behavioral tests for social motivation, memory/recognition, and hierarchy, which can be extrapolated to other species including humans. In particular, the medial prefrontal cortex (mPFC) has been recognized as a key brain region in mediating social cognition in both rodents and humans. This review will highlight the underappreciated connectivity, both anatomical and functional, between the olfactory system and mPFC circuitry, which together provide a neural substrate for olfactory modulation of social cognition and social behaviors. We will provide future perspectives on the functional investigation of the olfactory-mPFC circuit in rodent models and discuss how to translate such animal research to human studies.

1. Introduction

A large body of clinical literature reports olfactory dysfunction in neurological and psychiatric diseases, including Parkinson’s disease [1], Alzheimer’s disease [2], schizophrenia [3], mood disorders [4], and autism spectrum disorder [5]. Olfactory impairment often precedes other major disease symptoms and predicts disease severity. For example, olfactory function can be considered as a predictive marker of disease progression and prognosis in Parkinson’s disease [6] as well as in schizophrenia [7]. These clinical observations suggest that olfactory deficits may have primary pathogenic roles in these diseases. From a behavioral perspective, olfactory impairments coincide with deficits in cognition, motivation, memory, and emotion, presumably due to close connections between the olfactory system and other brain regions.

How the olfactory system interacts with other brain regions and thus influences higher cognitive functions is the topic of this review. Since deficits in social cognition are manifested in a variety of neurological and neuropsychiatric disorders [8–10], we will use social cognition as a lead dimension to provide in-depth discussion of the relationship between olfactory cortex and higher-order brain areas. Social cognition refers to the complex mental processes through which individuals perceive and respond to social signals from conspecifics. Here we use the term social cognition to include social motivation, social memory, and social hierarchy, common features described in numerous species. We will not cover species specific social behaviors such as those related to reproduction and parental care. It is worth noting that humans show unique higher-level social behaviors (e.g., empathy, mind reading, and moral judgement) that cannot be readily recapitulated in animal models. Humans have the ability to read subtle differences in facial expression and vocal tones as social cues and to infer mental states of self and others (known as theory of mind) [11, 12]. Nonetheless, common themes have emerged from human and animal studies on the neural circuitry involved in social cognition. In particular, the mPFC – spanning the anterior cingulate, prelimbic, and infralimbic cortices in the rodent brain, and homologous to human Brodmann Area 24, 32 and 25, respectively [13] – has been recognized as a key brain region in mediating social cognition across species [14–16] (Table 1). The mPFC is implicated in working memory, motivation, decision making, and adaptive behaviors [17], essential elements for proper social behaviors. With experimentally accessible neural circuits and well-established behavioral tests, rodents provide invaluable preclinical animal models for social cognition studies.

Table 1.

Comparison of human and rodent mPFC subdivisions and brain regions receiving direct inputs from the olfactory bulb. The midcingulate cortex (MCC or BA 24’ in human; posterior to ACC, see Fig. 1) is not included here.

Human (BA: Brodmann Area)	Rodent (Mouse and Rat)
Medial portion of the prefrontal cortex (mPFC)
Anterior cingulate cortex (ACC); BA 24	ACC*
Pregenual ACC (pACC); BA 32	Prelimbic cortex (PL)
Subgenual ACC (sACC); BA 25	Infralimbic cortex (IL)
Brain regions receiving direct inputs from the olfactory bulb (OB)
Anterior olfactory nucleus (AON)	AON#
Taenia tecta (TT)	Taenia tecta, ventral (TTv)
	Taenia Tecta, dorsal (TTd)
	Dorsal peduncular cortex (DP)^
Olfactory tubercle (OT)	OT
Piriform cortex, frontal and temporal portion (Pir-F and Pir-T); BA27	Piriform cotex, anterior and posterior portion (aPir and pPir)
Amygdala (Amg)	Anterior amygdalar area (AAA)
	Piriform-amygdalar area (PAA); also called cortex-amygdala transition zone (CxA) or periamygdalar cortex
	Nucleus of lateral olfactory tract (NLOT)
	Bed nucleus of accessory olfactory tract (BAOT)
	Medial amygdalar nucleus (MEA)
	Cortical nucleus of amygdala (CoA; anterior and posterolateral divisions)
	post-piriform transition area (TR); also called amygdalo-piriform transition area (AmPir)
Rostral Entorhinal Cortex (ventral BA 28, dorsal BA 34)	lateral Entorhihnal Cortex (EC)

^*The rodent anterior cingulate gyrus can also be divided into Cg1 (dorsal) and Cg2 (ventral) spanning both ACC and MCC [13].

^#The rodent AON (also called anterior olfactory cortex [67]) is divided into the external and principal part, and the latter is further divided into the medial, dorsal, lateral, and ventroposterior division. The external part predominantly receives inputs from and projects to the contralateral OB, while the principal part communicates with ipsilateral OB as well as other brain regions.

^{^}The dorsal peduncular (DP) cortex was reported to receive direct OB inputs in earlier tracing studies [57]; however, this projection was potentially due to contamination from the AON, adjacent to the OB. A recent study with more restricted labeling of OB neurons did not confirm this projection [58]. Regardless, the DP is closely related to the olfactory system by receiving direct inputs either from the OB or from the AON.

A social behavior is typically triggered by sensory cues (olfactory, visual, auditory, somatosensory, etc.) that carry social information. In addition to sensing odors, the olfactory system also transmits nasal breathing signals into the brain [18–21], which shapes oscillatory rhythmic brain activity even in the absence of odors [22, 23]. Recent studies highlight under-recognized network connections between the olfactory system and mPFC.

In this review, we will first provide a brief overview of the mPFC, the olfactory pathway, and their roles in social cognition. We will then discuss anatomical and functional connections from the olfactory system to the mPFC. Specifically, we will focus on the anterior olfactory cortical region (including the anterior olfactory nucleus and taenia tecta), which receives direct inputs from the olfactory bulb (the first olfactory synaptic relay) and projects to the mPFC. Lastly, we will discuss future directions on parsing the contribution of olfactory-mPFC connections in social cognition and other cognitive functions using modern neuroscience approaches, while considering the potential translation of rodent circuit-level understanding to human studies in health and disease.

2. The central role of mPFC in social cognition in rodents

Within the vast neural networks involved in social cognition, the mPFC has been recognized as a key regulator from rodents to humans [14–16]. We use the term mPFC to include three distinct subdivisions: anterior cingulate, prelimbic, and infralimbic cortex in rodents, and their human homologous structures (summarized in Table 1). Although rodents lack the granular prefrontal cortex that is present in primates [24], the mPFC under this definition has anatomically and functionally homologous structures across species [13, 14, 16, 24–26]. Since rodents have experimentally accessible neural circuits and well-established behavioral tests, investigation of the neural basis of rodent social behaviors is of high translational value. This review does not cover certain sex-specific social behaviors such as mating and parental care, which are recently reviewed elsewhere [27].

A variety of behavioral paradigms has been developed to assess social cognition in rodents, including social motivation (i.e., sociability or social preference), social recognition (i.e., social memory or social novelty preference), and social dominance (i.e., hierarchy). For example, in a three-chamber arena, a test mouse spends more time investigating a social object (a stranger mouse) than a non-social object, and thus providing a behavioral read-out of social preference or motivation [28]. Typically a juvenile same-sex mouse is used as the social object to avoid potential confounding factors in data interpretation (e.g., the test mouse may be submissive to the stranger mouse). When given the choice between a novel mouse and a familiar one, the mouse prefers interacting with the novel mouse, yielding a metric for assessing social recognition [28]. Social dominance can be assessed by placing two cage mates into a tube to observe that one mouse (dominant) forces the other one out (subordinate) [29]. Additionally, social defeat (i.e., chronic exposure to an aggressor) leads to social avoidance, anxiety, and depression-like behaviors [30]. Since these behaviors depend on internal drive as well as on learning and memory, social cognition contains both innate and acquired components.

Numerous studies suggest that the mPFC and its associated circuitry play a central role in social behaviors [14, 15]. The mPFC contains neural ensembles that dynamically code real-time social behaviroal information [31]. Direct manipulation of the activity of different subtypes of mPFC neurons impacts social approach and social interaction behaviors [32, 33], reflecting altered social motivation. In addition to glutamatergic and GABAergic synatpic transmission, the activity of mPFC neurons is also regulated by neuromodulators including dopamine, acetylcholine and oxytocin [14]. Oxytocin receptors are expressed in glutamatergic neurons in the mPFC and optogenetic activation of axons in the basolateral amygdala (originated from oxytocin receptor positive neurons in the mPFC) impairs social recognition [34]. The mPFC has strong connection with the hippocampus [17], which is essential for learning and memory consolidation to support social recognition [e.g., 35, 36–41]. Similarly, the mPFC has been established as a key regulator in social dominance, likely through its connection with the mediodorsal thalamus [29, 42], as well as in social defeat-induced avoidance and anxiety-like behaviors through its connection with the amygdala, nucleus accumbens, and brainstem [43–45]. The emerging picture suggests that each mPFC subdivision and its specific input and output pathways serve distinct functions in mediating social behaviors [e.g., 46, 47].

3. The olfactory pathway in rodents

The most comprehensive studies of anatomical and functional organization of the mammalian olfactory system have come from rodent models (mice and rats). The rodent nose contains several distinct chemosensory organs including the main olfactory epithelium (MOE) and vomeronasal organ (VNO), which express different types of chemoreceptors and transmit sensory information through the main and accessory olfactory system, respectively [reviewed in 48, 49]. The VNO detects pheromones, which signal social and/or sexual status of conspecifics, and downstream projections to the accessory olfactory pathway mediate species specific social behaviors such as mating and parental care [27, 50]. Because the VNO is non-functional in humans [51], this review is limited to discussion of the main olfactory system only.

The rodent MOE located in the posterior nasal cavity harbors approximately ten million olfactory sensory neurons (OSNs), each of which expresses a single type G protein-coupled odorant receptor from a repertoire of ~1200. OSNs respond to odor molecules as well as respiration-related nasal airflow and transduce sensory information into membrane potential changes and neuronal firing. The axons of OSNs expressing the same OR type typically project to two discrete glomeruli in the ipsilateral main olfactory bulb (OB), where they make synaptic contacts with OB neurons including mitral and tufted cells [52].

The OB efferent projections have been investigated via anterograde tracing from the OB neurons as well as retrograde tracing from the targeted regions [53–58]. The axons of mitral and tufted cells form the lateral olfactory tract, which runs through the lateral part of the olfactory peduncle (the stalk connecting the OB with the basal forebrain) and projects to a number of cortical and subcortical regions (Figs. 1, 2). These regions are mostly three-layered structures including the anterior olfactory nucleus (AON), taenia tecta (TT), olfactory tubercle (OT, the most ventral part of the striatum), piriform cortex (Pir), amygdala, and entorhinal cortex (EC) (Table 1). The olfactory system contains parallel processing pathways. The mitral and tufted cells, the two types of projection neurons in the OB, project to different brain regions [59]. While mitral cells send elaborated axonal collaterals to almost all primary olfactory cortices, tufted cells predominantly target the anterior structures including the AON, TT and OT [60–62] (Fig. 2). A subset of mitral cells in the posteroventral OB have been shown to transmit sex-related attractive social signals to the medial amygdala [63], but more studies are needed to distinguish the sensory information relayed by mitral and tufted cells.

Figure 1.

Comparison of human and mouse mPFC and olfactory areas. A, Mouse whole-brain side view (tilted to view both sides) built from the Scalable Brain Atlas Composer (Allen Mouse Common Coordinate Framework v3) [157, 158]. B, A coronal section across the mPFC and some of the olfactory cortices [157]. C. Human whole-brain viewed from the midline, titled to better view the olfactory and mPFC areas. The mPFC areas are built from the Scalable Brain Atlas Composer (Human Brainnectome Atlas, Human-BN274) [159, 160], while the olfactory parts are added based on [161]. D and E, Coronal sections across the mPFC and some of the olfactory cortices [162]. The taenia tecta (TT) is added based on [159].

Abbreviations: MOB, main olfactory bulb; AON, anterior olfactory nucleus; TTd, taenia tecta dorsal; TTv, taenia tecta ventral; DP, dorsal peduncular cortex; mPFC, medial prefrontal cortex; IL, infra limbic; PL, prelimbic; ACC, anterior cingulate cortex; MCC, midcingulate cortex; OT, olfactory tubercle; aPir and pPir, anterior and posterior piriform cortex; Amg, amygdala; CoA, cortical amygdala; EC, entorhinal cortex; MD mediodorsal nucleus of thalamus; Hip, hippocampus; BA, Brodmann area; OB, olfactory bulb; Pir-F, piriform cortex frontal lobe; Pir-T, piriform cortex temporal lobe.

Figure 2.

Potential connections from the olfactory pathway to the mPFC in rodents. Arrow lines start from neuronal cell bodies and end in axonal targets. Note that most connections are reciprocal. For clarity, not all connections are shown.

Abbreviations: OE, olfactory epithelium; MOB, main olfactory bulb; T, tufted cells; M, mitral cells; AON, anterior olfactory nucleus; TTd, taenia tecta dorsal; TTv, taenia tecta ventral; DP, dorsal peduncular cortex; mPFC, medial prefrontal cortex; IL, infralimbic cortex; PL prelimbic cortex; ACC, anterior cingulate cortex; OT, olfactory tubercle; Pir, piriform cortex; Amg, amygdala; EC, entorhinal cortex; MD, mediodorsal thalamic nucleus; Hip, hippocampus; OFC, orbitofrontal cortex.

There are several noteworthy features of the olfactory pathway. (1) Unlike other sensory modalities, olfactory information is transmitted to the cortices without a thalamic relay. For example, the Pir transmits the sensory information to the orbitofrontal cortex (OFC) directly, although there is a minor pathway via the mediodorsal thalamic nucleus [64]. (2) The fact that the OB projects to multiple olfactory cortices suggests parallel processing of olfactory information. The olfactory regions including the OB, primary cortices, and OFC, are extensively interconnected [65–67]. The recurrent connections provide feedforward and feedback signals to the olfactory network and enable one region (e.g., AON) to impact the activity of the other regions (e.g., Pir) [68, 69]. (3) The OB receives extensive top-down inputs not only from olfactory cortices, but also from neuromodulatory centers (e.g., serotonin from raphe nuclei, acetylcholine from basal forebrain, and norepinephrine from locus coeruleus) [57, 70, 71], underscoring that the peripheral olfactory input is highly regulated by the brain state and experience [72–74]. (4) Only a few synapses separate OSNs from neurons that ultimately drive behavioral outputs and hormone changes [e.g., 75, 76, 77]. Moreover, the olfactory system has strong connections with the limbic system as well as the mPFC, which is discussed in-depth below.

4. Olfactory contributions to social cognition in rodents

Social behaviors in rodents predominantly rely on chemosensory cues released from conspecifics. These chemosensory cues range from small volatile molecules to large non-volatile proteins coming from a variety of sources (e.g., urine, feces, skin, and glands) [78]. Although multiple chemosensory organs in the nose detect these social cues, the main olfactory system is critical for social behaviors by detecting and processing both volatile and non-volatile molecules [79, 80]. Here we focus on potential contributions of primary olfactory cortices to social cognition beyond detection and processing of these social cues.

The OT encodes odor valence and reward information, and promotes motivational behavior [81–83]. Social interaction leads to increased Ca²⁺ signals of the dopaminergic projection from the ventral tegmental area to the OT [84]. The Pir is essential for learned associations between odor and socially significant cues via oxytocin receptor signaling [85]. Notably, the AON plays an important role in social behaviors via several subtypes of neurons. Oxytocin receptor-expressing neurons in the AON receive direct projections from oxytocin neurons in the hypothalamus. Consequently, oxytocin excites these neurons and increases top-down inhibition onto OB granule cells. This signaling ultimately increases the signal-to-noise ratio for social cue-induced OB activity, and deletion of oxytocin receptors in the AON impairs social recognition memory [86]. A recent study on male mice with stable linear dominance hierarchy (alpha, subdominant, and subordinate) shows that the alpha mice had a higher level of oxytocin receptor binding in the AON [87], suggesting potential involvement of this brain region in social hierachy. In addition, exposure to a juvenile rat (but not to non-social objects) activates GABAergic, vasopressin expressing neurons in the AON [88]. Furthermore, the medial AON contributes to odor memory and social behaviors via direct projection from the ventral hippocampus [89, 90]. Taken together, in addition to detecting and processing social cues, the central olfactory circuits also contribute to social cognition through their connections with other brain regions.

5. Olfactory influences on mPFC circuitry in rodents

5.1. Anatomical evidence supporting olfactory-mPFC connections

Whole-brain afferents to the rodent mPFC have been analyzed via retrograde tracing [e.g., 91, 92, 93]. By focally injecting Fluorogold into each subdivision of the rat mPFC, Hoover and Vertes [91] found that the inputs along the dorsoventral axis shift from more sensorimotor inputs to the ACC to more limbic inputs to the prelimbic and infralimbic cortex. Differential inputs into different subregions of the mPFC are recently confirmed in mice [93]. Using rabies virus-based trans-synaptic tracing, whole-brain afferent inputs into different subtypes of neurons in the mPFC have been quantified [92, 93]. The long-range projections to the mPFC come from the entire neocortex, thalamus, striatum and pallidum, hippocampus, hypothalamus, and amygdala, as well as olfactory primary cortices (TT, AON and Pir) [92, 93].

Based on the current understanding of the mPFC connectivity, the primary olfactory cortices can transmit information to the mPFC via the relay of multiple regions, including the OFC, thalamus, hippocampus, and amygdala (Fig. 2). More importantly, the anterior olfactory cortical region (i.e., AON and TT [94]), which have been traditionally more neglected than other olfactory regions, has physical proximity to the mPFC and provides a more direct pathway linking the olfactory system and the mPFC. Retrograde tracing from the mPFC identifies presynaptic cells in the AON and TT [91, 92, 95, 96]. The projection from the AON to the mPFC (PL and IL) is also further confirmed by anterograde tracing via viral infection of glutamatergic neurons in the AON and TT [96]. Interestingly, the tufted cells in the OB primarily project to the anterior olfactory cortical region while the mitral cells project broadly. Since tufted cells receive stronger peripheral inputs [97] and display robust respiration-entrained rhythms and odor-induced responses [98, 99], it is tempting to propose that the OB tufted cells, via the relay of the anterior olfactory cortical region, carry olfactory information to the mPFC and serve distinct functions. In addition, a direct projection from the posterior Pir to mPFC has been reported to play a critical role in social transmission of food safety [100].

5.2. Functional evidence supporting olfactory-mPFC connections

Until now, the functional evidence that supports olfactory modulation of mPFC activity mainly comes from respiration-entrained olfactory signals in the context of odor-independent behaviors. The mPFC displays a predominant respiration-coupled activity during the immobility phase of the tail-suspension test [95, 101] or during auditory-cued, conditioned fear-induced freezing bouts [96, 102, 103]. Bulbectomy or ablation of the olfactory epithelium significantly reduces respiration-related activity in the mPFC [95, 96, 102, 103], highlighting a major contribution from the olfactory pathway even though there are multiple sources of respiratory rhythms in the brain [104]. Furthermore, optogenetic activation of OSNs in the olfactory epithelium is sufficient to entrain mPFC activity at the stimulating frequencies [96]. These findings strongly support that the olfactory input modulates mPFC activity.

Social behaviors depend on chemosensory cues, and rodents change their breathing/sniffing patterns during odor investigation [105, 106]. For instance, in a social hierarchy test, subordinate rats decrease their sniffing frequency when facing dominant rats [107]. Repeated social defeat leads to long-lasting bradypnea, an abnormally slow breathing rate in rats [108]. It is plausible that during social behaviors, both odor-induced and respiration-related olfactory signals may modulate mPFC activity. The different breathing patterns may differentially modulate the activity of the mPFC circuitry and influence the behavioral outcome. Alternatively, social factors may modulate mPFC activity through non-olfactory pathways to influence motor output related to olfaction/respiration. To differentiate the contribution of olfactory (odor-induced or respiration-related) versus non-olfactory input to mPFC activity, experimental manipulation of specific neural circuit(s) mediating one input at a time in social behaviors is required in future studies.

5.3. Olfactory influence on other parts of the mPFC circuitry involved in social cognition

The mPFC has widespread projections to subcortical regions (e.g., hippocampus, amygdala, hypothalamus, nucleus accumbens, ventral tegmental area, and neuromodulatory centers) that are crucial for memory, emotion, motivation, decision making, and adaptive behavior [14–16]. The olfactory pathway also has close connections with these subcortical regions [109, 110]. For example, the OB projects directly to the entorhinal cortex, which provides major inputs to the hippocampus, and the hippocampus also sends strong projections to the olfactory regions [111], including the OB [112]. Recent studies reveal that the ventral hippocampus to the AON projection plays an important role in odor memory and olfactory-guided behaviors [89, 90]. Similarly, through direct projections to subregions of the amygdala [113], the olfactory pathway potentially modulates the neural activity in the basolateral amygdala and mPFC circuitry. Moreover, the activity of most neural networks (including the olfactory system and mPFC circuitry) is also regulated by the neuromodulatory centers in the brainstem in a state- and context-dependent manner [114, 115]. For instance, the reward effect of social interaction activates serotonergic neurons in the dorsal raphe nucleus [116], which may in turn modulate sensory input and mPFC activity to influence behavioral output. During specific behaviors, the neural activity in multiple brain regions often becomes transiently synchronized [117, 118], which has been demonstrated repeatedly in the mPFC and its associated neural network [17, 118]. Respiration-coupled olfactory signals provide a low-frequency rhythmic activity to coordinate the network activity among the mPFC and its associated subcortical brain structures. A recent study reveals that mPFC activity from two socially interacting mice is correlated [119], and it would be interesting to determine whether social partners also have synchronized respiration and/or neural activity in their olfactory pathways.

6. Implications for olfactory contributions to social cognition in humans

In addition to olfactory cues, humans also rely on visual and auditory cues (e.g., facial expression and vocal tones) to interpret the mental states of others in social settings. It is conceivable that visual and/or auditory dysfunction is associated with compromised social cognition in diseased conditions [e.g.,120]. Curiously, olfactory dysfunction is characteristic in a wide range of neurological and psychiatric diseases (see Introduction for references). The close network connection between the olfactory pathway to higher brain regions (including but not limited to mPFC) may provide a clue to the underlying mechanisms.

Compared to the rodent brain, anatomical and functional organization of the human olfactory pathway is less understood. This is partly due to the small sizes of olfactory regions relative to the rest of the brain and their mostly ventral location at the base of the brain. Nevertheless, primary olfactory cortices (including the AON, OT, Pir, and subregions of both the amygdala and entorhinal cortex) as well as the OFC have been investigated anatomically and functionally [121–129] (Table 1 and Fig. 1). A recent study using functional magnetic resonance imaging (fMRI) from healthy human subjects clustered the subdivisions of primary olfactory cortices based on resting-state, whole-brain functional connectivity patterns, providing preliminary support for identifying the presumptive AON, OT, and Pir (frontal and temporal divisions) [130] (Fig. 1). More task-based fMRI studies are needed to dissect out distinct roles of each primary olfactory region and their connections with other brain regions including the mPFC. Interestingly, even though humans breathe at a much slower rate (0.1–0.3 Hz) compared to rodents (2–12 Hz), nasal (but not oral) breathing is effective in entraining the neural activity in the olfactory and limbic system of the human brain and modulating cognitive functions [131, 132]. These findings underscore that the olfactory system can impact brain activity in the absence of odors via involuntary or voluntary changes in respiration.

Studies in healthy individuals have linked alteration of olfactory function to changes in memory, language, and executive function, which include cognitive inflexibility, higher impulsivity, and lower behavioral inhibitory control [133–136]. Other studies also link olfactory dysfunction to impaired perception of olfactory signals, emotions, and reduced social network size [137–142]. It is unclear how much these changes are related to the observations in rodents described above, in particular in association with social cognition. However, an overlap of brain circuits in olfactory function and those in social cognition exists particularly in limbic and paralimbic brain regions [130]. Some studies reported that functional connectivity of amygdala with OFC has been related to both olfactory function (sensitivity) and social network size [142, 143]. In addition, activation of the mPFC and OFC has been observed during olfactory memory tests and subjective perception of pleasantness of odors [144]. Future studies with independent validation cohorts will evaluate these interesting ideas or working hypotheses proposed by the aforementioned pioneer works.

Olfactory dysfunction and social cognition deficits have been intensively studied as separate and independent topics in psychiatric disorders, yet little is known about how these functions interact with each other. As far as we are aware, three reports have been published on these interactions in patients with schizophrenia in a cross-sectional design. One study demonstrates a significant association of odor identification with theory of mind and face recognition in social cognition domains [145], and the other two studies report a significant relationship of odor identification with facial affect recognition [146, 147]. In addition, similar findings are reported in other psychiatric conditions such as bipolar disorder and autism spectrum disorder [148–151]. Although these human studies are informative, the descriptive nature of these studies do not provide mechanistic insights, in particular those translatable with rodents. Again, these clinical studies are expected to be further validated by other independent cohorts or meta-analyses. Brain imaging combined with olfactory functional and cognitive tests may be a route to address mechanisms.

7. Perspectives

Recent technological advances in neuroscience research make it feasible to study the neural basis of complex behaviors at the molecular, cellular, and circuit levels in rodent models. In the context of social cognition and social behavior, we propose that contributions of the olfactory system reach beyond detection and perception of social cues. Through its critical link with the mPFC and its circuitry, the olfactory pathway significantly impacts social cognition even in species that do not heavily rely on olfactory cues. A better understanding of olfactory modulation on the mPFC circuitry will offer novel mechanistic insights into how olfactory dysfunction is frequently observed in early stages of neuropsychiatric disorders and holds potential predictive significance.

To achieve this goal, several questions in basic science remain to be addressed. First, given the almost non-overlapping cortical and subcortical targets of OB mitral and tufted cells, do they relay distinct sensory information in social behaviors? Second, among the multiple pathways through which information flows from primary olfactory cortices to the mPFC, does the anterior olfactory cortical region (i.e., AON and TT) serve a special function? It is important to identify specific molecular marker(s) to gain genetic access to these neurons. One can further dissect out circuit connections of these neurons to each subdivision of the mPFC and test the effects of optogenetic and/or chemogenetic activation and inactivation on mPFC activity and social behaviors. Third, how to separate odor- and respiration-related contributions of the olfactory system to social cognition? Both odor-dependent and odor-independent behaviors need to be tested, and mPFC activity needs to be monitored concurrently with the respiration signal.

Whether olfactory dysfunction contributes to neurological and psychiatric disorders in a causal manner or is one of the symptoms, reflecting central neuropathological abnormalities, cannot be easily answered in human studies, whereas in rodent models, we can manipulate the olfactory circuitry and test the effects on higher-order brain areas and cognitive behaviors. In some mouse models in which disease-associated genes are disrupted, both olfactory and social deficits have been reported [152, 153]. However, their causal relationship also remains elusive. It is important to test this idea by using rodent models in which olfactory deficits are specifically introduced, while examining resultant changes of social behaviors. The bulbectomy model in which the OB is surgically impaired has been considered in this line of efforts [154, 155]. However, the technology of the frontline neuroscience may provide more specific, less invasive models for testing a causal role of olfactory deficits in social behaviors.

Despite our focus on social cognition in this review, neural connectivity between the olfactory system and mPFC also likely contributes to other cognitive functions [e.g., 156]. Olfactory modulation of brain activity and higher cognition warrants future investigations in preclinical animal models and humans.