PROJECTIONS FROM THE FIVE DIVISIONS OF THE ORBITAL CORTEX TO THE THALAMUS IN THE RAT

Abstract

The orbital cortex (ORB) of the rat consists of five divisions: the medial (MO), ventral (VO), ventrolateral (VLO), lateral (LO) and dorsolateral (DLO) orbital cortices. No previous report has comprehensively examined and compared projections from each division of the ORB to the thalamus. Using the anterograde anatomical tracer, Phaseolus vulgaris leucoagglutinin, we describe the efferent projections from the five divisions of the orbital cortex to the thalamus in the rat. We demonstrated that, with some overlap, each division of the ORB distributed in a distinct (and unique) manner to nuclei of the thalamus. Overall, ORB projected to a relatively restricted number of sites in the thalamus, and strikingly distributed entirely to structures of the medial/midline thalamus, while completely avoiding lateral regions or principal nuclei of the thalamus. The main termination sites in the thalamus were the paratenial nucleus (PT) and nucleus reuniens (RE) of the midline thalamus, the medial (MDm) and central (MDc) divisions of the mediodorsal nucleus, the intermediodorsal nucleus, the central lateral, paracentral, and central medial nuclei of the rostral intralaminar complex and the submedial nucleus (SM). With some exceptions, medial divisions of the ORB (MO, VO) mainly targeted “limbic-associated” nuclei such as PT, RE and MDm, whereas lateral division (VLO, LO, DLO) primarily distributed to “sensorimotor-associated” nuclei including MDc, SM and the rostral intralaminar complex. As discussed herein, the medial/midline thalamus may represent an important link (or bridge) between the orbital cortex and the hippocampus and between the ORB and medial prefrontal cortex. In sum, the present results demonstrate that each division of the orbital cortex projects in a distinct manner to nuclei of the thalamus which suggests unique functions for each division of the orbital cortex.

Color coded maps showing the major target sites and relative densities of projections from the five divisions of the orbital cortex (medial, ventral, ventrolateral, lateral and dorsolateral cortices) to nuclei of the thalamus. Of note, the orbital cortex almost exclusively distributes to the nuclei of the “limbic thalamus”; that is, to the midline, intralaminar, submedial, anterior and mediodorsal nuclei.

1. INTRODUCTION

It well recognized that the thalamus and cortex are intimately interconnected and form well defined thalamo-cortico-thalamic loops, involving both non-principal and principal (or relay) nuclei of the thalamus (Sherman, 2016; Halassa and Sherman, 2019; Perry et al., 2021; Shepherd and Yamawaki, 2021). It has further been demonstrated that several non-principal nuclei, mainly of the central thalamus, are strongly reciprocally connected with the medial prefrontal cortex (mPFC), and serve a critical role in various affective and cognitive functions (Vertes, 2015b; Dolleman-van der Weel, 2019, McGinty and Otis, 2020; Cassel et al., 2021; Griffin et al. 2021; Kirouac, 2021). Unlike, however, well documented connections of the mPFC with the thalamus (Vertes, 2002, 2004) considerably less attention has been paid to connections of the orbital cortex (ORB) with the thalamus (Hoover and Vertes, 2011; Barreiros et al., 2021a,b) – despite extensive examinations of the behavioral properties of the ORB.

The ORB has been linked to various functions, with a concentration on its role in reward/appetitive behavior and cognition. With respect to cognition, and based largely on deficits in reversal tasks, ORB dysfunctions were initially characterized as a loss of inhibitory control (Schoenbaum et al., 2002; McAlonan & Brown, 2003; Eagle & Baunez, 2010; Hardung et al., 2017). It is now recognized, however, that failures of inhibition are part of a larger constellation of ORB deficits, characterized by an inability to correctly associate stimuli with actions (Rolls & Grabenhorst, 2008; Schoenbaum et al., 2009; Rudebeck & Murray, 2014; Izquierdo, 2017). For instance, a current dominant theory holds that the ORB is vital for utilizing extant cues to predict the consequences of actions (Wilson et al., 2014; Schuck et al., 2016; Bradfield & Hart, 2020). Accordingly, the ORB is thought to represent “a cognitive map of task space” which entails assessing all relevant task information required for successful goal directed behavior (Wilson et al., 2014; Bradfield & Hart, 2020).

The ORB of the rat has been subdivided into five divisions: the medial orbital (MO), ventral orbital (VO), ventrolateral orbital (VLO), lateral orbital (LO) and the dorsolateral (DLO) orbital cortices (Ray & Price, 1992; Van de Werd & Uylings, 2008). Examinations of ORB functions have, in only a few instances, been restricted to a single division of the ORB -- but rather span two or more divisions. In this regard, comparisons have most commonly been made between medial and lateral regions of the ORB -- often describing opposite functions for these two regions. For instance, Ahmari et al. (2013) showed that optogenetic stimulation of the medial ORB (MO and VO) produced obsessive compulsive-like behavior in mice, whereas Burguiere et al. (2013) reported that optogenetic activation of the lateral ORB suppressed compulsive behavior in mice. Dalton et al. (2016) examined medial vs. lateral orbital effects on a probabilistic (spatial) reversal learning (PRL) task, and demonstrated that inactivation of either medial or lateral orbital regions disrupted PRL -- but differently for the two regions. Disruption of the medial region resulted in the inability of mice to use information on outcomes to guide future choices, while inactivation of the lateral region produced a failure to adjust responses to changing contingencies.

More recently, Robbins and colleagues (Hervig et al., 2020a) described differential effects of inactivation of the medial or lateral ORB on a visual serial reversal task in rats. Specifically, inactivation of the lateral orbital region severely impaired performance on this task, while (paradoxically) inactivation of the medial orbital area “improved” visual reversal learning. The improved performance on the task was attributed to a greater sensitivity to negative feedback, thereby promoting “lose-shift” behavior – supporting reversal learning. Finally, Bradfield and Hart (2020) proposed separate roles for the medial and lateral ORB within the framework of the ORB as “a cognitive map of task space”. In essence, they proposed that the lateral ORB represents a rodent’s initial position within task space, to evaluate potential options, whereas the medial ORB represents a rodent’s “future position” within the task map, to thereby choose the appropriate course of actions to achieve intended goals.

The foregoing suggests that the projections of medial (MO, VO) and lateral (VLO, LO, DLO) regions of the ORB may be distinct, thus supporting very different (or separate) sets of functions for these two regions. Aside from relatively few previous descriptions of ORB projections (Schilman et al., 2008; Kondo & Witter, 2014; Hoover & Vertes, 2011; Barreiros et al., 2021a,b), the overall projections of the ORB have not been systematically examined in the rat. An analysis of the projections of individual divisions of the ORB would greatly enhance our understanding of the functions of each division – as well as the ORB as a whole.

No previous report has described the efferent projections from the five divisions of the ORB to the thalamus. Using the anatomical anterograde tracer, PHA-L, we examined the projections from the five divisions of the ORB to the thalamus in the rat, and demonstrated a unique pattern of projections from each subdivision of the ORB to the thalamus.

2. METHODS

2.1. Animals

Thirty-two male Sprague Dawley rats (Charles River, Wilmington MA) were used in this study. Rats weighing 300 to 375 grams were given a minimum of 7 days of habituation upon arrival before undergoing aseptic surgery to inject the anterograde tracer Phaseolus vulgaris-leucoagglutinin (PHA-L; Vector Laboratories, Cat #: L-1110-5; RRID: AB 2336656) into a single division of the ORB. These experiments were approved by the Florida Atlantic University Institutional Animal Care and Use Committee and conform to all federal regulations and the NIH guidelines for the care and use of laboratory animals.

2.2. Surgical procedure

Rats were anesthetized with 80 mg/kg ketamine (100 mg/ml) and 10 mg/kg xylazine (20 mg/ml). Rats were placed in the stereotaxic apparatus and a protective ophthalmic ointment was applied to the eyes and the scalp was prepped with a betadine scrub solution (Fisher Scientific, Cat #: 19-066452). Next, an incision was made to expose the skull and a burr hole was drilled over the ORB. The following coordinates were used to target the divisions of the ORB (from bregma): MO (AP: 3.5–4.5 mm, L: 0.4–0.6 mm, DV: 4.8–5.0 mm); VO (AP: 3.5–4.5 mm, L: 0.7–0.9 mm, DV: 4.9–5-1 mm); VLO: (3.4–4.4 mm, L: 1.8–2.0 mm, DV: 4.2–5.0 mm); LO (AP: 3.3–4.3 mm; L: 2.5–2.8 mm, DV: 5.0–5.2 mm); and DLO (AP 3.2–4.2 mm, L: 3.0–3.4 mm, DV: 5.2–5.8 mm). The PHA-L solution from powdered lectin was reconstituted to 4–5% in 0.05 M sodium phosphate buffer (PB) at a pH of 7.4], and was iontophoretically deposited in the orbital cortex of rats by means of a glass micropipette with an outside tip diameter of 40–60 μm. Positive direct current (5–10 μA) was applied through a Grass stimulator (Model 88) coupled with a high voltage stimulator (Frederick Haer, Bowdoinham, ME) at 2 seconds “on” / 2 seconds “off” intervals for 40–50 minutes. After a survival time of 7–10 days, rats were deeply anesthetized with 150 mg/kg Euthasol and perfused transcardially with heparinized buffered physiological saline (100 mL/rat) followed by 4% paraformaldehyde, 0.2–0.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4) (300–500 mL/animal). The brains were removed and stored for 2 days at 4°C in 30% sucrose in 0.1 M PB before being processed for analysis.

2.3. Immunohistochemical and anatomical analysis

Six series of 50-μm frozen sections were collected in phosphate-buffered saline (PBS, 0.9% sodium chloride in 0.01 M sodium phosphate buffer, pH 7.4) using a freezing microtome (Leica Biosystems SM2000R, Wetzlar, Germany). A complete series of sections was treated with 1% sodium borohydride (Fisher Scientific, Cat #: S678–10) in 0.1 M PB for 30 minutes to remove excess reactive aldehydes. Sections were then rinsed in 0.1 M PB, followed by 0.1 M Tris-buffered saline (TBS), pH 7.6. Following this, sections were incubated for 60 minutes at room temperature (RT) in 0.5% bovine serum albumin (BSA) in TBS to minimize nonspecific labeling. The sections were then incubated overnight at RT in diluent (0.1% BSA in TBS containing 0.25% Triton X-100) and biotinylated goat anti PHA-L (Vector Labs, Burlingame, CA, Cat #: AS-2224-1; RRID: AB 2315136) at a concentration 1:500. Sections were then washed in 0.1 M PB (4 × 8 minutes) and placed in a 1:500 concentration of biotinylated rabbit antigoat immunoglobulin (IgG; Vector Labs, Burlingame, CA, Cat #: BA-5000-1.5; RRID: AB 2336126) and diluent for 2 hours. Sections were washed and then incubated in a 1:100 concentration of peroxidase-avidin complex from the ABC Elite kit (Vector Labs, Burlingame, CA, Cat #: PK-6100; RRID: AB2336819) and diluent for 1 hour. Following another 0.1 M PB wash the peroxidase reaction product was visualized by incubation in a solution containing 0.022% 3,3′ diaminobenzidine (DAB, Aldrich, Milwaukee, WI, Cat #: D3939) and 0.003% H₂O₂ (Fisher Scientific, Cat #: H325–500) in TBS for 6 minutes. Sections were then rinsed again in PBS before being mounted onto chrome-alum gelatin-coated slides, dehydrated in graded methanols (Fisher Scientific, Cat #: A41220), and placed in xylene (Fisher Scientific, Cat #: A41220), before being cover slipped with Permount (Fisher Scientific, Cat #: SP15500). An adjacent series of sections from each rat was stained with cresyl violet (Thermo Scientific, Cat #: AC229630250) for anatomical reference. Photomicrographs were captured from representative cases from each subdivision of ORB using light and darkfield optics at 100x with a NikonFI-3 camera mounted on a Nikon Eclipse E600 microscope (RRID: SCR018858). Digital images were captured in NIS elements (Nikon Instruments, Melville NY, RRID: SCR014329) and adjusted for brightness and contrast in Adobe Photoshop. Labeled orbital fibers were plotted on representative schematic coronal sections through the thalamus generated in Adobe Illustrator from the Nissl-stained sections. The schematic sections qualitatively depict the patterns and density of labeled fibers in nuclei of the thalamus produced by the orbital injections. This is summarized in Table 1. For Table 1, estimates of the density of anterograde labeling in thalamic nuclei was made, as previously described (Vertes & Hoover, 2008; Hoover & Vertes, 2011; Kondo & Witter, 2014), using a “one to four” + system: with ++++, very dense labeling; +++, dense labeling; ++, moderate labeling; + weak labeling; and – no labeling.

Table 1.

Relative Density of Labeled Fibers in Nuclei of the Thalamus Produced by PHA-L Injections into the Five Divisions of the Orbital Cortex

	MO	VO	VLO	LO	DLO
Midline Thalamus
PV	+	+	−	−	−
PT	++++	++++	−	++	++++
RH	+	−	+	−	++
RE	++++	++++	−	−	+
Anterior Thalamus
AD	−	−	−	−	−
AM	−	−	++	+	−
AV	−	−	+	−	−
IAM	+	+	−	+	+
Mediodorsal Nucleus
MDm	++++	+	+	+	++++
MDc	−	++++	++++	++++
MDl	−	−	+
IMD	++	+	+	+	++
Intralaminar Thalamus
CM	−	−	+++	+	+
PC	−	−	+++	−	−
CL	−	−	+++	−	−
SM	−	−	++++	++	+++

⁺, light labeling;

⁺⁺, moderate labeling;

⁺⁺⁺dense labeling;

⁺⁺⁺⁺, very dense labeling;

^--no labeling. See list for abbreviations.

3. RESULTS

The present report describes the pattern of projections from the five divisions of the orbital cortex (ORB) to the thalamus in the rat. The transverse Nissl-stained section of Fig. 1a, through the prefrontal cortex, shows the locations of the five divisions of the ORB. Figure 1b-f depicts the site of injections in the medial orbital (MO), ventral orbital (VO), ventrolateral orbital (VLO), lateral orbital (LO) and dorsolateral (DLO orbital cortices. The patterns of labeling shown schematically for the illustrated cases are representative of the patterns found with the non-illustrated cases.

Figure 1.

a: Nissl-stained section through the anterior forebrain showing the locations of the five divisions of the orbital cortex in the rat: the medial orbital (MO), ventral orbital (VO), ventrolateral orbital (VLO), lateral orbital (LO) and dorsolateral orbital (DLO) cortices. b-f: Bright field micrographs showing the locations of PHA-L injections in MO (b), VO (c), VLO (d), LO (e) and DLO (f). Scale bar for a, d = 600 µm; for b, c, e, f = 500 µm.

3.1. Medial orbital cortex (MO)

Figure 2a-j schematically illustrates the pattern of distribution of labeled fibers to the thalamus following a PHA-L injection in MO. As depicted, labeled fibers from MO spread rostrocaudally throughout the thalamus, but strikingly were confined to the medial/midline thalamus and absent from lateral regions of thalamus -- or principal nuclei. At the rostral thalamus (Fig. 2a-c), labeled axons were essentially restricted to the paratenial nucleus (PT) and nucleus reuniens (RE), bilaterally – with labeling stronger ipsi- than contra-laterally. The photomicrograph of Figure 3a shows pronounced terminal labeling bilaterally in PT. A few labeled fibers were also present in the paraventricular (PV) and the interanteromedial nucleus (IAM) (Fig. 2a-c). Aside from IAM, the anterior thalamus was essentially devoid of labeled fibers, including the anterodorsal (AD), anteroventral (AV) and anteromedial (AM) nuclei.

Figure 2.

Schematic representation of the pattern of distribution of labeled fibers at select rostral (a) to caudal (j) levels of the thalamus produced by a PHA-L injection in the medial orbital cortex. See list for abbreviations.

Figure 3.

a-c: Brightfield micrographs through the anterior thalamus showing dense collections of labeled fibers within the paratenial nucleus (PT) of the dorsal midline thalamus, stronger ipsilaterally (left side) than contralaterally. Labeling produced by PHA-L injections in the medial orbital (a), ventral orbital (b) and dorsolateral orbital (c) cortices. Abbreviations. PV, paraventricular nucleus of thalamus. Scale bar for a = 600 µm; for b, c = 500 µm.

At mid-levels of the thalamus (Fig. 2d-f), labeled fibers continued to be present in PT and RE and additionally, prominently within the medial division of the mediodorsal nucleus (MDm) (Fig. 2d-f) and the intermediodorsal nucleus (IMD) (Fig 2f). Labeling was sparse in other thalamic nuclei at these levels. The photomicrographs of Figure 4a,b depict dense terminal labeling, rostrally (Fig. 4a) and caudally (Fig. 4b) in MDm as well as pronounced but less dense labeling in the caudal pole of PT (Fig. 4a), the central division of MD (MDc) (Fig. 4b) and RE – with the heaviest concentration of fibers in the lateral wings of RE (or peri-reuniens nucleus, pRE), caudally (Fig. 4b).

Figure 4.

Darkfield micrographs at a rostral (a) and caudal level (b) of the anterior thalamus depicting the pattern of distribution of labeled fibers dorsally and ventrally within the midline thalamus produced by a PHA-L injection in the medial orbital cortex. Note the dense collection of labeled fibers in the paratenial nucleus (PT) (a), the medial division of the mediodorsal nucleus (MDm) (a, b), and the nucleus reuniens (RE) (b). Abbreviations: MDc, central division of the mediodorsal nucleus of thalamus; PV, paraventricular nucleus of thalamus; 3V, third ventricle. Scale bar for a, b = 500 µm.

The main terminal sites at the caudal thalamus (Fig. 2g-j) were MDm and IMD, dorsally, and RE and pRE, ventrally. In addition, MDc, the posterior paraventricular nucleus (PVp), rhomboid nucleus (RH) and central medial nucleus (CM) were lightly to moderately labeled. The photomicrograph of Figure 5a depicts dense labeling in MDm (bilaterally), IMD and RE, strongest in pRE – as well as moderate labeling in PVp and RH.

Figure 5.

Darkfield micrographs at a rostral (a) and caudal level (b) of the posterior thalamus depicting the pattern of distribution of labeled fibers dorsally and ventrally within the midline thalamus produced by a PHA-L injection in the medial orbital cortex. Note the dense collection of labeled fibers in the medial division of the mediodorsal nucleus (MDm) (a), the intermediodorsal nucleus (IMD) (a) and rostrally (a) and caudally (b) in nucleus reuniens (RE) and the peri-reuniens nucleus (pRE) of thalamus. Abbreviations: RH, rhomboid nucleus of thalamus. Scale bar for a = 650 µm; for b = 500 µm.

3.2. Ventral orbital cortex (VO)

Figure 6a-j schematically depicts the pattern of distribution of labeled fibers in the thalamus following a PHA-L injection in VO. As shown, the pattern of thalamic labeling was similar to that of MO, but with some important differences. As with MO, labeling was restricted to medial/midline structures, and absent from principal nuclei of the thalamus. Virtually the sole targets of labeled fibers at the rostral thalamus (Fig. 6a-c) were PT and RE -- both strongly and bilaterally labeled with an ipsilateral predominance. The photomicrograph of Figure 3b shows moderately dense labeling, bilaterally, in the rostral PT.

Figure 6.

Schematic representation of the pattern of distribution of labeled fibers at select rostral (a) to caudal (j) levels of the thalamus produced by a PHA-L injection in the ventral orbital cortex. See list for abbreviations.

At mid-levels of the thalamus (Fig. 6d-f), labeled fibers distributed strongly to the caudal PT (Fig. 6d), to MDc and to RE/pRE (Fig. 6d-f). IAM and RH were sparsely labeled. The photomicrographs of Figure 7a,b depict dense labeling in PT (Fig. 7a) and MDc (Fig. 7b), dorsally, and at two levels of RE, ventrally (Fig. 7a,b). As shown, labeled fibers terminated massively on the dorsal/dorsolateral border of RE, extending to pRE. (Fig. 7b).

Figure 7.

Darkfield micrographs at a rostral (a) and caudal level (b) of the anterior thalamus depicting the pattern of distribution of labeled fibers dorsally and ventrally within the midline thalamus produced by a PHA-L injection in the ventral orbital cortex. Note the dense collection of labeled fibers in the paratenial nucleus (PT) (a), the central division of the mediodorsal nucleus (MDc) (b) and the nucleus reuniens (RE) and peri-reuniens nucleus (pRE) at two levels of the anterior thalamus (a, b). Note also the very intense labeling on dorsal/dorsolateral border of the caudal RE (b). Abbreviations: IAM, interanteromedial nucleus of thalamus; MDm, medial division of the mediodorsal nucleus of thalamus; PV, paraventricular nucleus of thalamus. Scale bar for a = 400 µm; for b = 500 µm.

At the caudal thalamus (Fig. 6g-j), labeling thinned considerably, with fibers distributing moderately to densely to RE and pRE (Fig. 7g-j) and lightly to MDc and the central lateral nucleus (CL) of the intralaminar thalamus. The photomicrograph of Figure 5b depicts dense labeling caudally in RE -- which like seen rostrally, was concentrated dorsal/dorsolaterally in RE, and in pRE.

3.3. Ventrolateral orbital cortex (VLO)

Figure 8a-h schematically depicts the pattern of distribution of labeled fibers in the thalamus produced by a PHA-L injection in VLO. As depicted, the pattern of labeling found with the VLO injection significantly differed from that seen with the MO or VO injections. For instance, unlike MO/VO, there was an absence of labeling of PT and RE (Fig. 8a-d), which differs with the pronounced labeling of the intralaminar nuclei of thalamus (Fig. 8d-h).

Figure 8.

Schematic representation of the pattern of distribution of labeled fibers at select rostral (a) to caudal (h) levels of the thalamus produced by a PHA-L injection in the ventrolateral orbital cortex. See list for abbreviations.

At the rostral thalamus (Fig. 8a,b), labeled fibers coursed dorsolaterally into the thalamus to terminally distribute to the anterior medial nucleus (AM) and also traverse AM in route to caudal regions of the thalamus (Fig. c,d). Further caudally, labeled fibers spread over a wide area terminating moderately to densely in AM, in the adjacent AV, in MDc and in the central medial nucleus (CM) of the intralaminar thalamus (ILt) (Fig. 8c,d). The photomicrograph of Figure 9a shows fibers entering laterally at the rostral thalamus, but a complete lack of terminal labeling at this level -- including that of PT, PV or RE.

Figure 9.

Darkfield micrographs depicting the pattern of distribution of labeled fibers at a rostral (a) and caudal (b) level of the thalamus produced by a PHA-L injection in the ventrolateral orbital cortex. a: Note the presence of labeled fibers coursing laterally into anterior thalamus (left side of a), but a complete lack of terminal labeling at this level in the paraventricular (PV), paratenial (PT) and nucleus reuniens (RE) of the dorsal midline thalamus. b: Note dense collections of labeled fibers in the central lateral (CL), paracentral (PC) and central medial (CM) nuclei of the rostral intralaminar complex of thalamus as well as ventrally in the central division of the mediodorsal nucleus (MDc) and the submedial nucleus (SM) of thalamus. Abbreviations: sm, stria medullaris. Scale bar for a = 650 µm; for b = 500 µm.

At the mid to caudal thalamus (Fig. 8e-h), labeled fibers spread widely throughout the thalamus to terminate moderately to densely in CL, the paracentral nucleus (PC) and CM of the intralaminar thalamus (Fig. 8e-h), in MDc, in IMD (Fig. 8e,f), and in RH (Fig 8g). In addition, the rostro-caudal extent of the submedial nucleus (SM) was densely labeled (Fig. 8f-h). The labeled fibers present ipsilaterally in ventral medial nucleus (VM) of thalamus (Fig. 8e-h) appeared to mainly course through VM, bound for SM. Figure 9b shows strong terminal labeling in CL, PV and CM of the intralaminar complex, in the adjacent ventral part of MDc and in SM, ventrally.

3.4. Lateral orbital cortex (LO)

Figure 10a-h schematically depicts the pattern of distribution of labeled fibers throughout the thalamus produced by a PHA-L injection in LO. At the rostral thalamus (Fig. 10a-c), labeled fibers, traversing AM, distributed lightly to moderately to PT and to the anterior pole of MD. At mid to caudal levels of the thalamus (Fig. 10d-h), labeled fibers predominantly targeted MDc, IMD and SM, terminating strongly (and bilaterally) in MDc and moderately in IMD and SM. In addition, IAM and RH were lightly labeled. Figure 11a,b depicts dense terminal labeling rostrally ( Fig. 11a) and caudally (Fig. 11b) in MDc, and moderate labeling in SM (Fig. 11b).

Figure 10.

Schematic representation of the pattern of distribution of labeled fibers at select rostral (a) to caudal (h) levels of the thalamus produced by a PHA-L injection in the lateral orbital cortex. See list for abbreviations.

Figure 11.

Dark field micrographs depicting the pattern of distribution of labeled fibers at a rostral (a) and caudal (b) level of the thalamus produced by a PHA-L injection in the lateral orbital cortex. Note the dense collection of labeled fibers selectively in the central division of the mediodorsal nucleus (MDc) and the submedial nucleus (SM) at the two levels of the thalamus. Abbreviations: PV, paraventricular nucleus of thalamus. Scale bar for a = 550 µm; for b = 500 µm.

3.5. Dorsolateral orbital cortex (DLO)

Figure 12a-j schematically depicts the pattern of distribution of labeled fibers throughout the thalamus produced by a PHA-L injection in DLO. At the rostral thalamus (Fig. 12a-c), labeled fibers were essentially restricted to PT and the rostral aspect of MD, distributing heavily to PT (Fig. 3c). At mid-levels of the thalamus (Fig. 12d-f), labeled fibers were densely concentrated in MDm (bilaterally), moderately in IMD and SM, and lightly in CM and RH. At the caudal thalamus (Fig. 12g-j), labeling was pronounced in MDm, strong but less dense in IMD, moderate in RH and SM and light in CM. Figure 13 shows dense terminal labeling in MDm and SM and moderate labeling of IMD and RH at the caudal thalamus.

Figure 12.

Schematic representation of the pattern of distribution of labeled fibers at select rostral (a) to caudal (j) levels of the thalamus produced by a PHA-L injection in the dorsolateral orbital cortex. See list for abbreviations.

Figure 13.

Darkfield micrograph depicting dense collections of labeled fibers at a mid-level of the thalamus produced by a PHA-L injection in the dorsolateral orbital cortex. Note dense collection of labeled fibers in the medial division of the mediodorsal nucleus (MDm) and the submedial nucleus (SM) of thalamus as well as moderate labeling of the intermediodorsal nucleus (IMD) of thalamus. Abbreviations: PV, paraventricular nucleus of thalamus, RE, nucleus reuniens of thalamus; RH, rhomboid nucleus of thalamus. Scale bar = 500 µm.

4. DISCUSSION

We described projections from the five divisions of the orbital cortex (ORB) to the thalamus in the rat. As demonstrated, with some overlap, each division of the ORB distributed in a distinct (and unique) manner to nuclei of the thalamus. Overall, ORB projected to a relatively limited number of sites in the thalamus, and strikingly distributed entirely to structures of the medial/midline thalamus (or limbic thalamus), while completely avoiding lateral regions of the thalamus -- or principal (sensorimotor) nuclei. The main ORB termination sites in the thalamus were the paratenial nucleus and nucleus reuniens of the midline thalamus, the medial and central divisions of the mediodorsal nucleus, the intermediodorsal nucleus, the rostral intralaminar nuclei (CL, PC, CM) and the submedial nucleus. All five divisions of the ORB distributed to MD, either to the medial and/or central divisions. Interestingly, PT was a major orbital target which contrasted with minimal projections to the adjacent paraventricular nucleus.

4.1. Highlights of ORB projections -- by thalamic region

The midline thalamus (PT, PV, RE/pRE, RH):

The midline thalamus consists of the PT and PV, dorsally, and RH and RE, ventrally. PT received strong projections from MO, VO and DLO and moderate ones from LO. By contrast, PV received, at best, light projections from any division of the ORB. For the ventral midline thalamus, RE/pRE received pronounced projections from MO and VO, but few (or none) from VLO, LO or DLO. While ORB projections to RH were overall light, all divisions contributed some fibers to RH, heaviest from DLO (Table 1).

The anterior thalamus (AD, AV, AM, IAM):

The anterior thalamus (or complex) was not a major target of ORB; that is, there were no ORB projections to AD and relatively few to IAM. Essentially, the sole source of afferent projections to AV and AM was from VLO. Whereas VLO fibers largely spread throughout (rostral) AM, they were localized to the ventromedial sector of AV, adjacent to AM. Finally, a sizable number of fibers in AM appeared to course through it in route to the intralaminar nuclei (Table 1).

The mediodorsal and intermediodorsal nuclei (MD, IMD):

All divisions of ORB projected to MD and IMD, strongest to MD. Interestingly, the ‘outer divisions’, or medial and lateral walls, of ORB (MO and DLO) projected to MDm, whereas the inner divisions (VO, VLO, LO) distributed to MDc. The lateral sector of MD (MDl) received modest projections, mainly from VLO. While fibers distributed terminally to IMD, predominantly from MO and DLO, a significant percentage appeared to course through IMD to the contralateral thalamus, mainly bound for MD (Table 1).

The central lateral, paracentral and central medial nuclei of the rostral intralaminar thalamic complex (CL, PC, CM).

Comparable to AM and AV, CL and PC received projections virtually solely from VLO, whereas CM received (strong) projections from VLO as well as modest projections from other divisions of the ORB (Table 1).

The submedial nucleus (SM).

SM received moderate to dense projections from VLO, LO and DLO, but none from MO or VO. Notably, SM was a major target of VLO – with VLO fibers distributing heavily throughout the nucleus (Table 1).

4.2. Projections of the five divisions of the ORB to the thalamus: comparisons with previous studies

4.2.1. Medial orbital cortex

We showed that the primary thalamic targets of MO were PT, RE, pRE and MDm, and secondarily IMD (Figure 14). Consistent with present findings, Hoover & Vertes (2011) described pronounced MO projections to these same sites. In contrast with present results, however, Hoover and Vertes (2011) reported somewhat stronger MO projections to MDc, CM and RH than demonstrated here, whereas presently we observed heavier MO projections to IMD than described previously. These (minor) differences likely involve slightly different locations of MO injections in the two studies.

Figure 14.

Schematic representation of patterns and density of medial orbital (red) and ventral orbital (yellow) projections to the thalamus. Color coded chart depicting strength of projections: - none; +, light; ++, moderate; +++, heavy, ++++, dense projections. See list for abbreviations.

PT was shown to be a major MO target. Undoubtedly owing to its small size, only a single older report (Chen & Su, 1990) described afferents to PT. Retrogradely labeled cells following PT injections were found along the medial wall of the PFC but seemingly did not extend ventrally to MO (Chen & Su, 1990). In primates, however, Hsu & Price (2007) showed that retrograde tracer injections in PT produced strong neuronal labeling in area 13a, (the rodent equivalent of MO), and further that anterograde injections in area 13a gave rise to dense terminal labeling in PT. Finally, PT sends substantial return projections to MO (Reep et al., 1996; Berendse & Groenewegen, 1991; Van der Werf et al., 2002; Vertes & Hoover, 2008), indicating pronounced reciprocal connections between these structures.

MO fibers were shown to heavily target RE and pRE. Examining afferents to RE in the rat, McKenna & Vertes (2004) described strong retrograde cell labeling in MO with injections at three (rostro-caudal) levels of RE. In a similar manner, Scheel et al. (2020) recently identified significant numbers of retrogradely labeled cells in MO following RE injections in mice. Regarding return RE projections to MO, Van der Werf et al. (2002) described pronounced terminal labeling in MO with injections in the medial (or central core) of RE, but interestingly even heavier labeling in MO with injections in pRE. More recently, Vertes et al. (2006) confirmed strong RE/pRE projections to MO.

MO strongly targeted MDm of the MD complex. In an early report, Groenewegen (1988), using WGA-HRP as an anterograde and retrograde tracer, described significant interconnections between MO and MDm. Specifically, they stated that “the medial orbital area is reciprocally connected with the medial segment of MD”. Consistent with this, Gabbott et al. (2005) reported that large MD injections, encompassing most of MD, produced dense retrograde cell labeling in MO – and in lateral parts of ORB. Finally, MD to MO projections have been shown to primarily originate from MDm (Groenewegen, 1988; Ray & Price, 1992; Reep et al., 1996).

4.2.2. Ventral orbital cortex

Similar to MO, the principal targets of VO were PT, RE, pRE and MD (Figure 14). In an earlier examination of VO projections in the rat, Hoover & Vertes (2011) similarly reported that VO projects strongly PT, RE/pRE and MDc, but in slight contrast to present findings, described (some) VO projections to the intralaminar thalamus and RH.

VO strongly targeted PT -- but less so than MO. This is consistent with a previous description of quite prominent VO-PT projections (Hoover & Vertes (2011). PT, in turn, has been shown to distribute heavily to VO (Van der Werf et al., 2002; Vertes & Hoover, 2008).

VO distributes massively to RE (and pRE) – particularly to the dorsal/dorsolateral sector of rostral RE (see Fig. 7B). Consistent with this, pronounced numbers of retrogradely labeled cells were observed in the VO of mice (Scheel et al., 2020) and rats (McKenna & Vertes, 2004) following retrograde tracer injections in RE.

VO distributed heavily (and selectively) to MDc of the MD complex. Supporting this, VO was previously shown to strongly target MD, mainly MDc (Hoover & Vertes, 2011), while Groenewegen (1988) described reciprocal VO connections with the central MD. Finally, Reep et al. (1996) confirmed MD to VO projections, showing labeled cells essentially confined to MDc following retrograde tracer injections in VO.

4.2.3. Ventrolateral orbital cortex

The major targets of VLO were AM, CL, PC, CM, MDc, and throughout the submedial nucleus (SM) (Figure 15) The pattern of VLO projections differs from that of other ORB divisions in that VLO was essentially the only ORB division with significant projections to the anterior and rostral intralaminar nuclei and also lacked projections to the midline nuclei (PT, PV and RE) (see Fig. 9A).

Figure 15.

Schematic representation of the patterns and density of ventrolateral orbital (green), lateral orbital (blue) and dorsolateral orbital (purple) projections to the thalamus. Color coded chart depicting strength of projections: -, none; +, light; ++, moderate; +++, heavy; ++++, dense projections. See list for abbreviations.

Unlike the limited reports examining the output from other ORB divisions, the projections VLO have been well studied, owing in part to VLO involvement in nociception (Craig et al., 1982; Tang et al., 2009; Vertes et al., 2015a), and in this regard, VLO-SM connections have received particular attention. For instance, Craig et al. (1982) early on described pronounced, topographically organized, connections between SM and VLO in the cat, with the anterior SM mainly distributing to ventral VLO and the dorsal/ventral SM to the dorsal VLO. In a similar manner, Yoshida et al. (1992), using WGA-HRP as anterograde/retrograde tracer, reported that VLO projects densely to SM, as well as (moderately) to AM, MDc and the rostral intralaminar complex. Finally, SM to VLO projections have been described in several studies (Reep et al., 1996; Alcaraz et al., 2015; Murphy & Deutch, 2018), and Kuramoto et al. (2017), tracing the trajectory of individual SM axons, showed that dorsal SM fibers terminated almost exclusively in VLO, whereas those of the ventral SM in both VLO and LO.

4.2.4. Lateral orbital cortex

The projections of LO were relatively limited, with major projections to MDc, IMD and SM, and minor ones to PT, AM and RH (Figure 15). While no report has systematically examined the output of LO to the thalamus (or brain), Yoshida et al. (1992), investigating VLO (see above), made control injections in LO and described relatively pronounced LO projections to CM and SM. More recently, Alcaraz et al. (2015) similarly demonstrated LO projections to MDc and SM.

With respect to return thalamic projections to LO, Barreiros et al. (2021a) compared inputs to anterior (ALO) and posterior (PLO) regions of LO, and demonstrated inputs to LO from PT, MD and SM – with differences to ALO and PLO. Specifically, (1) lateral regions of MD/MDc distributed to ALO; medial regions of MD (MDm) to PLO; and (2) the posterodorsal SM projected to ALO; the ventral SM to PLO.

4.2.5. Dorsolateral orbital cortex

The main thalamic targets of DLO were PT, MDm, IMD and SM, with moderate projections to RH (Figure 15). Interestingly, DLO, like MO, distributed significantly to PT and MDm. While the output of DLO has not been comprehensively described, some studies have examined projections to the thalamus from regions of the anterior forebrain that included DLO. For instance, Shi & Cassell (1996) described strong terminal labeling in MD following injections in the agranular insular cortex that overlapped with DLO. Jasmin et al. (2004) examined the output from a forebrain region termed, the rostral agranular insular cortex (RAIC), lying “immediately adjacent” to LO, and demonstrated strong RAIC projections to PT, MDm and SM (see their Fig. 3). This pattern of projections is strikingly similar to that seen here with DLO injections. Finally, Groenewegen (1988) described pronounced reciprocal connections between DLO and MDm.

4.3. Brief overview of reciprocal thalamic-orbital projections

Regarding reciprocal thalamo-ORB connections, return thalamic projections (reciprocal) to the ORB range from sparse to abundant. Specifically, some thalamic nuclei are strongly reciprocally connected with the ORB, others not. For instance, VLO projects strongly to SM and SM in turn distributes densely to VLO – as well as moderately to LO and DLO (Yoshida et al. 1992; Alcaraz et al., 2015; Kuramoto et al. 2017; Murphy & Deutch, 2018). All divisions of ORB project to MD, and correspondingly MD distributes throughout ORB, most heavily to LO (Groenewegen, 1988; Ray & Price, 1992; Reep et al., 1996; Wang & Shyu, 2004). MO and VO densely innervate RE/pRE and PT, and both RE/pRE and PT give rise to pronounced return projections to MO and VO (Reep et al., 1996; Berendse & Groenewegen, 1991; Van der Werf et al., 2002; Vertes et al., 2006; Vertes & Hoover, 2008). In contrast, however, to largely reciprocal ORB-thalamic connections, the intralaminar nuclei of thalamus receive pronounced ORB (or VLO) projections, but interestingly send at best modest return projections to the ORB, mainly originating from CM, and (lightly) targeting VLO and LO (Berendse & Groenewegen, 1991; Van der Werf et al., 2002, Wang & Shyu, 2004; Vertes et al., 2012).

4.4. Functional considerations

We showed that, with some overlap, each of the five divisions of the ORB exhibit a unique (and differential) set of projections to the thalamus. This is consistent with the demonstration that various divisions of ORB distribute differentially to non-thalamic sites; namely, to the striatum, piriform cortex and parahippocampal cortices (Burwell & Amaral, 1998; Schilman et al., 2008; Kondo & Witter, 2014; Heilbronner et al., 2016). This suggests that each division of the ORB may participate in a unique (or separate) set of functions – with the challenge to identify the unique function(s) of each division.

There are extensive interconnections between the ORB and the mPFC (Conde et al., 1995; Ongur & Price, 2000; Vertes, 2004; Hoover & Vertes, 2007; Vertes & Hoover, 2008; Bedwell et al., 2017; Murphy & Deutch, 2018). While the ORB projects directly to the mPFC (Hoover & Vertes, 2007, 2011), the present findings indicate that the ORB can indirectly affect the mPFC via the thalamus. Specifically, as shown, several divisions of the ORB distribute to various thalamic nuclei, strongly to PT, RE and MDm, and these thalamic nuclei, in turn, give rise to pronounced projections to the mPFC (Groenewegen, 1988; Berendse & Groenewegen, 1991; Reep et al., 1996; Van der Werf et al., 2002; Hoover & Vertes, 2007; Vertes & Hoover, 2008; Vertes et al., 2006, 2015b; Varela et al., 2014). Accordingly, indirect ORB projections to the mPFC, via the thalamus, would appear to complement direct connections between these structures.

In the virtual absence of direct projections from the mPFC to the hippocampus (HF) (Larouche et al., 2000; Vertes, 2004), mPFC connections with the HF appear to be largely indirectly routed through the nucleus reuniens (RE) -- which is reciprocally connected with the HF and mPFC (Wouterlood et al., 1990; Berendse & Groenewegen, 1991; Van der Werf et al., 2002; Vertes, 2002, 2004; McKenna & Vertes, 2004; Hoover & Vertes, 2007, 2012; Vertes et al., 2006, 2007; Varela et al., 2014; Scheel et al., 2020). This circuitry (mPFC-RE-HF) has been shown to serve a prominent role in various affective and cognitive functions (Vertes et al., 2015b, Dollerman-van der Weel et al., 2019; Cassel et al., 2021; Griffin, 2021). In a similar manner, CA1/SUB projects to the ORB (Jay & Witter, 1991; Cenquizca & Swanson, 2007), but there are no direct return projections from the ORB to CA1/SUB. As such, RE may also serve to critically link the ORB with the hippocampus.

4.5. Medial vs. lateral ORB: anatomy and function

As discussed, very few reports have examined the functional properties of individual (or each) division of the ORB, but rather focus has been on functional differences between medial and lateral parts of the ORB. Specifically, medial versus lateral ORB functional differences have been described for various tasks/behaviors such as reversal learning, cost/benefit decision making, delay discounting, attentional set formation and emotional reactivity (Rolls & Grabenhorst, 2008; Schoenbaum et al., 2009; Rudebeck & Murray, 2014; Izquierdo, 2017; Barreiros et al., 2021b). Further, medial and lateral regions of the ORB have often been shown to exert opposing effects on various behaviors (Gourely et al., 2010; Ahmari et al., 2013; Burguiere et al., 2013; Dalton et al., 2016; Hervig et al., 2020a,b).

As described (see Introduction), Hervig et al. (2020a) initially demonstrated that the inactivation of the medial ORB improved performance on visual reversal learning task, whereas inactivation of the lateral ORB impaired performance on this task. They (Hervig et al., 2020b) subsequently examined the effects of enhancing levels of glutamate or serotonin at the medial and lateral ORB on the same task, and showed that increases in levels of glutamate in the medial ORB, but not in the lateral ORB, improved reversal learning, whereas the blockade of 5-HT_2A receptors (and consequent reduction of glutamate release) at the lateral ORB impaired performance, without affecting the medial ORB. The authors concluded that the results “further support dissociable roles of the rodent mOFC and lOFC in deterministic visual reversal learning.”

Gourely et al. (2010) similarly reported that medial and lateral ORB lesions in mice produced very different effects on an instrumental reversal task; that is, impaired task acquisition (with medial ORB lesions) as compared to marked perseverative behavior (with lateral ORB lesions). Using a delay-discounting task in rats, Mar et al. (2011) demonstrated that medial ORB lesions produced an increase in preference for larger delayed rewards, whereas lateral ORB lesions resulted in a decreased preference for delayed rewards -- indicating that medial ORB lesions suppress impulsive behavior, while lateral ORB lesions accentuate it. Finally, Dalton et al. (2016) demonstrated that inactivation of the medial ORB disrupted performance on the acquisition phase of a probabilistic reversal learning (PRL) task, resulting in perseverative responding, whereas inactivation of the lateral ORB did not affect acquisition but impaired performance during reversal stages of the task.

In summary, the foregoing studies demonstrate dissociable (or opposing) functions for the medial and lateral ORB across a range of behaviors. An emerging view is that the medial ORB is mainly involved in affective and motivational behaviors, while the lateral ORB serves a role in sensory integration, linking cues to actions (Izquierdo, 2017; Barreiros et al., 2021b). For instance, reviewing this issue, Barreiros et al. (2021b) recently commented that “the medial OFC is more commonly associated with value integration and comparison with internally guided choices, whereas the lateral OFC appears more closely linked with sensory processing and updating specific stimulus-reward associations”.

A medial/lateral orbital functional distinction gains support from the present findings showing significant anatomical differences in the projections of the medial and lateral ORB regions to the thalamus. Specifically, the projections of the medial ORB (MO, VO) were primarily directed to ‘limbic-associated’ nuclei of the thalamus such as PT, RE and MDm, whereas those of the lateral ORB (VLO, VO, DLO) mainly targeted sensorimotor thalamic nuclei such as MDc, SM and the rostral intralaminar complex. Further insights into anatomical and functional properties of the ORB will await a complete examination of the projections of each division of the ORB.

Abbreviations

AD

anterodorsal nucleus of thalamus

AM

anteromedial nucleus of thalamus

AV

anteroventral nucleus of thalamus

CL

central lateral nucleus of thalamus

CM

central medial nucleus of thalamus

DLO

dorsolateral orbital cortex

fx

fornix

HF

hippocampus

IAD

interanterodorsal nucleus of thalamus

IAM

interanteromedial nucleus of thalamus

ILt

intralaminar thalamus

IMD

intermediodorsal nucleus of thalamus

LD

laterodorsal nucleus of thalamus

LH

lateral habenula

LO

lateral orbital cortex

LP

lateral posterior nucleus of thalamus

MD

mediodorsal nucleus of thalamus

MDc

mediodorsal nucleus of thalamus, central division

MDl

mediodorsal nucleus of thalamus, lateral division

MDm

mediodorsal nucleus of thalamus, medial division

MH

medial habenula

MO

medial orbital cortex

mPFC

medial prefrontal cortex

mt

mammillothalamic tract

ORB

orbital cortex

PC

paracentral nucleus of thalamus

PHA-L

Phaseolus vulgaris leucoagglutinin

PO

posterior nucleus of thalamus

PT

paratenial nucleus of thalamus

PV

paraventricular nucleus of thalamus

PVp

paraventricular nucleus of thalamus, posterior division

PVh

paraventricular nucleus of hypothalamus

RE

reuniens nucleus of thalamus

pRE

peri-reuniens nucleus of thalamus

RH

rhomboid nucleus of thalamus

RT

reticular thalamus of thalamus

sm

stria medullaris

SM

submedial nucleus of thalamus

SUB

subiculum

VAL

ventral anterolateral nucleus of thalamus

VB

ventrobasal complex of thalamus

VLO

ventrolateral orbital cortex

VM

ventral medial nucleus of thalamus

VO

ventral orbital cortex

ZI

zona incerta

Data availability statement:

All of the basic anatomical data comprising this report is available upon request to the authors.