The postnatal development of retinal projections in strepsirrhine galagos (Otolemur garnettii)

Abstract

Here we describe the postnatal development of retinal projections in galagos. Galagos are of special interest as they represent the understudied strepsirrhine branch (galagos, pottos, lorises, and lemurs) of the primate radiations. The projections of both eyes were revealed in each galago by injecting red or green cholera toxin subunit B (CTB) tracers into different eyes of galagos ranging from postnatal day 5 to adult. In the dorsal lateral geniculate nucleus, the magnocellular, parvocellular, and koniocellular layers were clearly labeled, and identified by having inputs from the ipsilateral or contralateral eye at all ages. In the superficial layers of the superior colliculus, the terminations from the ipsilateral eye were just ventral to those from the contralateral eye at all ages. Other terminations at postnatal day 5 and later were in the pregeniculate nucleus, the accessory optic system, and the pretectum. As in other primates, a small retinal projection terminated in the posterior part of the pulvinar, which is known to project to the temporal visual cortex. This small projection from both eyes was most apparent on day 5 and absent in mature galagos. A similar reduction over postnatal maturation has been reported in marmosets, leading to the speculation that early retinal inputs to the pulvinar are responsible for the activation and early maturation of the middle temporal visual area, MT.

Graphical Abstract

The terminal projections of retinal ganglionic neurons of the eye were labeled and studied in newborn, young, and mature galagos. Those to the pulvinars were sparse at birth and were lost as galago matured. Possibly, this pathway promotes the prenatal development of visual cortex, and then is not needed.

Introduction

Galagos are members of the strepsirrhine (prosimian) radiation of primates that include galagos and pottos of Africa, the lorises of India and southeast Asia, and the lemurs of Madagascar. Their brains, and especially their visual systems, have been of great interest to researchers as these primates have long been recognized as most closely resembling early primates (Simpson, 1940; Le Gros Clark, 1959; Fleagle and Seiffert, 2017). Like present-day galagos, the first primates were small, nocturnal and well adapted to the fine branch niche of the tropical rainforest (e.g. Martin, 1990; Geho, 2004; Kemp, 2005). Early primates were characterized by large eyes to promote useful vision in dim light, and presumably they had a well-developed visual system overall to productively use their visual information. Studies of the organization of the visual systems of galagos and other strepsirrhine primates, in comparison with those of haplorrhine (anthropoid) primates, allow inferences about the visual systems of early primates and suggest that the visual systems of early primates were very much like those of extant galagos (Kaas et al., 2022).

Here, we studied the postnatal development of projections from both eyes to central brain structures in galagos ranging in age from a few days to adult. Retinal projections to central structures in galagos are now largely known from early studies of adults based on tracing degenerating axons after damage to the retina (Campos-Ortega and Glees, 1967; Campos-Ortega and Cluver, 1968; Tigges and Tigges, 1970), and studies limited to the lateral geniculate nucleus or superior colliculus. Axons from the retinal ganglion neurons travel to the optic chiasm at the base of the diencephalon, where axons from the nasal hemiretinal cells decussate in the chiasm to the contralateral hemisphere, and those from the temporal hemiretinal cell axons continue to the ipsilateral hemisphere. Most of the retinal projections terminate in the dorsal lateral geniculate nucleus (dLGN, see Table 1 for abbreviations) of thalamus, which is known to mediate conscious visual perception (Bear et al., 2020). These studies revealed an arrangement of six layers in the dLGN of galagos that receive inputs from either the contralateral eye or the ipsilateral eye. In galagos, dLGN contains six complete layers (Ionescu and Hassler, 1968; Kaas et al, 1978; Turner et al., 2020), which consists of two magnocellular layers (M; layers 1 and 2) with large-sized neurons, two parvocellular layers (P; layers 3 and 6) with smaller-sized neurons, and two koniocellular layers (K; layers 4 and 5) with very small-sized neurons (Norton and Casagrande, 1982; Turner et al., 2020). One additional incomplete and variable layer next to the optic tract was reported, which was more notable in the sagittal plane (Tigges and Tigges, 1970). To label layers, we used an alternative method (Kaas et al., 1972; Kaas et al., 1978) to designate layers based on their external (E) or internal (I) location in the nucleus and P, M, and K cell types, which facilitates comparisons between galagos and other species that have different dLGN organizations. ME and MI correspond to the layers 1 and 2, PI and PE correspond to the layers 3 and 6, and KI and KE correspond to the layers 4 and 5. Previous tracing studies revealed that geniculate relay neurons of different size-classes in each layer are innervated by a corresponding size class of retinal ganglion neurons (Itoh et al., 1982). As previously reported, ME, KE, and PE receive inputs from the contralateral eye, and MI, PI, and KI receive inputs from the ipsilateral eye (Conley et al., 1987; Kaas et al., 1978).

TABLE 1:

Abbreviations

AOS	Accessory optic system
DTN	Dorsal terminal nucleus
LTN	Lateral terminal nucleus
MTN	Medial terminal nucleus
OPT	Olivary pretectal nucleus
nOT	Nucleus of the optic tract
APT	Anterior pretectal nucleus
PPT	Posterior pretectal nucleus
MPT	Medial pretectal nucleus
CTB	Cholera toxin subunit B
MT	Middle temporal visual area
CO	Cytochrome oxidase
VGLUT2	Anti-vesicular glutamate transporter, type 2
PV	Parvalbumin
dLGN	Dorsal lateral geniculate nucleus
M,P,K	Magnocellular, Parvocellular, and Koniocellular layers of LGN
SC	Superior colliculus
uSGS	Upper stratum griseum superficiale of the SC
LP	Lateral posterior nucleus of the thalamus
MPul	Medial pulvinar
bSC	Brachuim of the SC

Projections to the superior colliculus (SC) were to the superficial grey layer, or stratum griseum superficiale, with those from the contralateral eye superficial to those from the ipsilateral eye. The SC is a laminated structure lying on the roof of the midbrain that integrates multisensory information through extensive subcortical and cortical inputs and intrinsic connections (Kaas and Huerta, 1988), and transforms the sensory inputs into eye movement outputs for visual attention. In galagos, the superficial layers of SC, especially the upper stratum griseum superficiale (uSGS), receive direct retinal ganglion inputs from both eyes (Laemle and Noback, 1970; Tigges and Tigges, 1970). Axons from ganglion cells in the nasal hemiretina primarily terminate in the deeper tier of uSGS, and axons from the ganglion cells in the temporal hemiretina mostly project to the superficial tier of uSGS (Kaas and Huerta, 1988). Parts of the superficial layers of SC receive cortical inputs from primary and higher visual areas (Baldwin and Kaas, 2012) and project to subdivisions of pulvinar (Baldwin et al., 2013; Baldwin et al., 2017) and dLGN (Harting et al., 1986; Lachica and Casagrande, 1993). The SC is also innervated by direct projections from visuomotor areas including the posterior parietal cortex, premotor cortex, and frontal eye fields (Baldwin and Kaas, 2012).

In addition to the LGN and SC, the visual pulvinar, sometimes included in the lateral posterior nucleus (LP) of non-primates, is a thalamic complex with nuclei that are considered as higher-order thalamic relays for visual processing (Guillery and Sherman, 2002). In primates, the pulvinar is a complex structure, and its proposed nuclei and nomenclatures have varied from the three earliest identified subdivisions, the medial, lateral, and inferior divisions of the pulvinar (MPul, LPul, and IPul; e. g., Walker, 1938). Six or more nuclei are now identified in architectonic and anatomical connection studies (see reviews in Baldwin and Bourne, 2020; Baldwin et al., 2017; Kaas and Baldwin, 2020). The lateral pulvinar, LPul, includes most of the ventrolateral pulvinar, while the IPul is further divided into posterior, medial, centromedial, and centrolateral subdivisions. While most of the pulvinar is reciprocally connected to the visual areas for higher-order processing of visual information (Baldwin and Bourne, 2020), some anterograde tracing studies in Old World and New World monkeys indicate that the medial part of the IPul is directly innervated by retinal ganglion projections (Cowey et al., 1994; Kwan et al., 2019; O’Brien et al., 2001; Warner et al., 2010). In mature macaques (O’Brien et al., 2001), it appears that the retinal projections, although sparse, terminate in the medial nucleus of the inferior pulvinar that projects to visual area MT and to the central medial nucleus that projects to MT surround (Kaas and Lyon, 2007). A developmental study in New World marmoset monkeys showed that this pathway is most pronounced around P18, and decreases during the maturation into adulthood (Warner et al., 2012). However, whether this retinopulvinar pathway also exists in strepsirrhine galagos, and whether it appears at particular postnatal periods or remains over the entire lifespan is unclear.

Other retinal projections were to the pregeniculate and pretectal nuclei, and parts of the accessory optic system (Tigges and Tigges, 1969). The pregeniculate nucleus (PrG) of primates, or the ventral lateral geniculate nucleus in other mammals, has been reported as a thalamic site that has retinotopic organization for visual and oculomotor interactions. Architectonic studies revealed that PrG is a bilaminar thalamic structure that consists of a bigger, external “pars grisea”, packed with neurons, and a smaller “pars fibrosa” internal portion of mostly mylineated fibers (Babb, 1980; Jones, 2012). As an extension of the ventral thalamus, the PrG receives collateral axons of thalamocortical and corticothalamic projections with the adjacent dLGN (Mitrofanis, 1994; Vaingankar et al., 2012), and is directly innervated by retinal projections predominately from the contralateral eye (Kaas et al., 1978; Theoret et al., 2000; Tigges and Tigges, 1970). In galagos, the pregeniculate nucleus is located at the anteroventral pole of the dLGN as a triangular nucleus in parasaggital sections, and lies on top of the most rostral dLGN in the coronal plane (Ionescu and Hassler, 1968; Symonds and Kaas, 1978; Tigges and Tigges, 1970).

As a conserved feature of the mammalian visual system, the AOS consists of the dorsal (DTN), lateral (LTN), and medial (MTN) terminal nuclei that participate in visuomotor reflexes and stablizing retinal images (see reviews in Brecha et al., 1980; Simpson, 1984). In nearly all primates, the three terminal nuclei receive retinal projections that are primarily from the contralateral eye (Cooper and Magnin, 1987; Itaya and Van Hoesen, 1983; Kaas et al., 1978; Weber, 1985). However, the retinal innervations of the AOS in prosiman primates remained uncertain in earlier studies (Campos-Ortega and Clüver, 1968; Hassler, 1966; Laemle and Noback, 1970; Tigges and Tigges, 1970), mostly due to the criticality of the survival time in degenerative methods. A later study by Cooper (1986), based on the intraocular injections of 3H-proline and horseradish peroxidase (HRP) in gray mouse lemurs (Microcebus murinus), successfully identified the labeled axonal boutons in all three terminal nuclei, suggesting the existence of a complete set of AOS nuclei in strepsirrihine primates.

In primates, the pretectal complex is comprised of five nuclei including the pretectal olivary nucleus (OPT), nucleus of the optic tract (nOT), anterior pretectal nucleus (APT), posterior pretectal nucleus (PPT), and medial pretectal nucleus (MPT) (Gamlin, 2006; Hutchins and Weber, 1985; Simpson et al., 1988; Weber, 1985). While parts of the pretectal complex are innervated by direct retinal ganglion projections, the density and extent of the inputs from the ipsilateral eye vary with the proportion of the binocular visual field (e.g., see comparisons between tree shrews and squirrel monkeys in Weber, 1985). The nOT and OPT are consistently reported to receive extensive projections from both eyes with a slightly denser contralateral component. With extensive connections with visual cortex, thalamus and midbrain, the nOT and OPT are related primarily to the oculomotor control and involved in relaying oculomotor signals to visual relay nuclei (see review in Gamlin, 2006). The PPT receives sparse projections mostly from the contralateral eye, which are almost totally absent in the MPT and APT (Weber, 1985). The functions of these three pretectal nuclei are less clear, and they may be not fully oculomotor in function. For example, the anterior pole of APT was reported to play a role in processing noxious stimuli due to the widespread inputs originating from cortical and subcortical somatosensory regions involved in nociception in rats (Foster et al., 1989) and in cats (Berkley and Mash, 1978).

In the present experiments, different tracers [cholera toxin subunit B (CTB)- Alexa 594 (red) or CTB-Alexa 488 (green)] were injected into each eye, allowing the labeled axonal terminations from each eye to be directly compared in the same structures, with the possibility of revealing developmental changes, such as the segregation of overlapping projections during postnatal development. A special interest was in the possibility of revealing retinal projections to a nucleus or nuclei in the pulvinar complex in postnatal galagos that may decrease in number or be totally lost in adults. In early studies, labeled axons from the retina were sometimes detected in the pulvinar and dismissed as “fibers of passage” on the way to the pretectum or superior colliculus (e.g., Campos-Ortega and Cluver, 1968; Tigges and Tigges, 1970), but it is now known that a significant projection of the retina is to part of the pulvinar in various monkeys (Cowey et al., 1994; O’Brien et al., 2001; Warner et al., 2010; Warner et al., 2012; Warner et al., 2015; see Baldwin and Bourne, 2017 for review). In a study of the postnatal development of projections to the pulvinar of marmosets (Warner et al., 2015), the projections from the retina appeared to be most dense in early postnatal development, and then quite sparse in adult marmosets. As neurons in the pulvinar with direct inputs from the retina project to cortical visual area MT, Warner et al. (2015) proposed that the early retinal projections to the pulvinar activated MT and promoted an early development of MT, which resulted in MT having a major role in the further development of the dorsal stream of visual processing. Our study of the postnatal development of retinal projections in galagos allowed us to determine if inputs to the pulvinar also exist in galagos and if they also appear to have a short-term role in promoting the development of the visual cortex.

Materials and Methods

Nine strepsirrhine galagos (Otolemur garnetti) were used in this study to reveal the pattern of retinal projections in the contralateral and ipsilateral hemispheres. Animals were divided into infant (5–20 days, n = 5), juvenile (78–79 days, n = 2), and adult (over 1 year, n = 2) groups based on the age when tissue was collected (Table 2). All surgical and animal care procedures were conducted in accordance with Guides for the Care and Use of Laboratory Animals by the National Institutes of Health and were in compliance with protocols approved by the Vanderbilt University Animal Care and Use Committee.

Table 2.

Case summary

Group	Case#	Gender	Age	Survival (day)	Figures injections	CTB
Infant	19–14	--	P5*	4	2–5, 9–11	CTB-red CTB-green#
	19–07	male	P19	4	6	CTB-red CTB-green
	19–15	--	P19	2		CTB-red CTB-green
	18–13	--	P20	7		CTB-red CTB-green
	18–14	--	P20	7	2, 3, 9, 10	CTB-red CTB-green
Juvenile	18–15	--	P78	2		CTB-red CTB-green
	18–16	female	P79	7	2–3 ,7, 9–10	CTB-red CTB-green
Adult	19–03	female	4.3y	7	2, 8–10, 12	CTB-red CTB-green
	18–11	female	9.6y	8	3	CTB-red CTB-green

^*postnatal days

^#Not transported

Eye injection procedure

Adult and pre-weaned juvenile galagos were initially tranquilized with an intramuscular injection of ketamine hydrochloride (10–25 mg/kg) and the anesthesia was maintained by isoflurane (1–2% mixed in O₂) during the procedures. Infant galagos were anesthetized with isoflurane (1–2% mixed in O₂) as an alternative to the injected agent. The pupils were dilated with a tropical application of atropine drops into the eyes. Proparacaine was applied to the eye for local anesthetic and analgesic purposes, and Betadine drops were applied to prevent eye infection. The eyeball was slightly rotated to expose the sclera. Different fluorescence-tagged tracers, cholera toxin subunit B (CTB) conjugated with Alexa Fluor 488 (or CTB-green; Invitrogen, Carlsbad, CA, Catalog# C34775) and Alexa Fluor 594 (or CTB-red; Invitrogen, Catalog^# C34777), were injected into each eye. In each galago, a small amount (3–5 μl) of CTB-green was injected into the posterior chamber of the left eye through the sclera, and CTB-red was injected into the posterior chamber of the right eye (Fig. 1). The tracer was loaded into a Hamilton microsyringe (802RN, 25 μl, 22s-gauge, 2 inches, point style 2; Catalog# 84855) that was attached to a small, removable needle (33-gauge, 0.75 inch, point style 2; Catalog# 7803–05). CTB tracers have been extensively used to trace retinal projections in primates (de Sousa et al., 2013; Hannibal et al., 2014; Nakagawa et al., 1998) and in other animals (Santana et al., 2018; Su et al., 2013).

Figure 1.

Experimental design of eye injections in nine galagos at different ages. A small volume (3–5 μl) of cholera toxin subunit B (CTB) conjugated with Alexa Fluor 488 (CTB-green) was injected into the posterior chamber of the left eye, and CTB conjugated with Alexa Fluor 594 (CTB-red) was injected into the posterior chamber of the right eye.

After eye injections, the pre-weaned galagos, juvenile galagos, and adult galagos, were allowed to recover from anesthesia and returned to their home cage. To reduce and eliminate the chance of maternal rejection of pre-weaned infant and juvenile galagos upon return to the dams after the eye injections, their mothers were also sedated with ketamine hydrochloride before the procedure on the experimental animals. No extra interventions were performed to these mother galagos. The offspring and mothers recovered together, and offspring were all accepted by the mothers.

Perfusion and histology

After 2–8 days of survival time to allow tracers to be transported to the destinations, the experimental galagos were euthanized with a high dose of sodium pentobarbital (120 mg/kg, intraperitoneal injection) and perfused with 0.1M phosphate buffered saline (PBS; pH 7.4) through the ascending aorta, followed by 4% paraformaldehyde in 0.1M phosphate buffer (PB) and 10% sucrose-containing fixative. The brain was removed and saved in 30% sucrose PB for cryoprotection. The brain was cut with a freezing microtome in the coronal plane at a thickness of 40 μm. The sections were divided into series for various purposes. One series was directly mounted, air dried and coverslipped for fluorescent observation of CTB-green and CTB-red labeling. Other series were processed stains for Nissl, cytochrome oxidase (CO; Wong-Riley, 1979), vesicular glutamate transporter 2 (VGLUT2), or parvalbumin (PV) that revealed the architectonic features of brain structures. See details in previous publications (Liao et al., 2021; Turner et al., 2020).

Antibody characterization and immunohistochemistry

Anti-vesicular glutamate transporter 2 (VGLUT2) primary antibody (RRID: AB_2187552): Mouse anti-VGLUT2 recombinant protein (monoclonal), Millipore Catalog no. MAB5504, Burlington MA. The immunogen is a KLH-conjugated linear peptide. In Western Blots of primate neocortex, the antibody recognizes a 56 kDa band, the molecular weight of VGLUT2 (Baldwin et al., 2013). This primary antibody was used in a concentration of 1:5000. Also see Turner et al. (2020).

Anti-parvalbumin primary antibody (RRID: AB_477329): Mouse anti-calcium binding spot of PV (IgG1 isotype; monoclonal), Catalog no. P3088, Sigma-Aldrich, St. Louis. The immunogen is derived from the PARV-19 hybridoma produced by the fusion of mouse myeloma cells and splenocytes from an immunized mouse. Purified frog muscle parvalbumin was used as the immunogen. This primary antibody was used in a concentration of 1:2000. See Turner et al. (2020).

Immunohistochemistry steps included use of secondary antibody horse anti-mouse IgG (Vector Laboratories, Inc.) in a concentration of 1:500, and normal horse serum (5% in 0.01 M PBS) was used in blocking solutions to reduce nonspecific binding. An avidin-biotin conjugation kit (Vector ABC Elite kit, Vector) was used to amplify labeling. Details are described in Turner et al. (2020).

Data analysis

A Zeiss Axio Imager 2 microscope (Carl Zeiss AG, Oberkochen, Germany) coupled with a Neurolucida system (MBF Bioscience, Williston, VT) was used to analyze the results. The locations of retinal projections from the left and right eyes that were respectively labeled by CTB-green and CTB-red tracers were plotted using Stereo Investigator version 2019 (MBF Bioscience). Special care was taken to mark blood vessels and landmarks for the alignment of plotted profiles to the adjacent sections stained for CO, Nissl, VGLUT2, or PV in Adobe Illustrator 2021 (Adobe Inc., San Jose, CA). Photomicrographs of labeled axons and boutons were taken with 20X magnification, and brightness and contrast were adjusted in Adobe Photoshop 2021 (Adobe). While the ganglion neurons in retina project to many brain areas (Beier et al., 2021; Hutchins and Weber, 1985; Itaya and Van Hoesen, 1983; Weber, 1985), here we focused on the LGN, SC, pulvinar, pregeniculate nucleus (PrG), accessory optic system (AOS), and pretectum.

Results

The primary goal of this study was to investigate the difference and/or similarities in the organization of retinal projection pathways in strepsirrhine (prosimian) galagos at the infant (5–20 days), juvenile (78–79 days), and adult (over 1 year) developmental periods. In nine galagos, CTB-green was injected into the posterior chamber of the left eye, and CTB-red was injected into the posterior chamber of the right eye (see Table 2). Our microscopic examination of brain sections revealed that the injections of tracers in the left and right eyes were successfully transported in 8 galagos. In galago 19–14, the CTB-green injection in the left eye was not transported. The patterns of labeled profiles in brain recipient areas of two hemispheres are described first in infant galagos, then in juveniles, and last in adult galagos.

Dorsal lateral geniculate nucleus (dLGN)

Our results from three age groups are largely consistent with the established organization of retinogeniculate connections in galagos. In one infant galago (P5, Case 19–14), CTB-red injection in the right eye labeled a great number of axon boutons in layers MI, PI, and KI of the ipsilateral dLGN, as identified when aligned with the adjacent Nissl sections that revealed the lamination structure (Fig. 2a–d). Labeled axons were also seen in the optic tract that lies along the ventral fringe of the LGN. CTB-green injection in the left eye was not transported in the infant case from day P5 (Fig. 2b,d). The overall patterns of the retinogeniculate projection to the ipsilateral and contralateral dLGN in infant galagos were better shown in another case that had effective CTB-green and CTB-red injections in the left and right eyes, respectively (P20, Case 18–14). After comparing to the adjacent CO sections, we found CTB-green labeled axonal boutons in layers ME, KE, and PE, and CTB-red labeled boutons in layers KI, PI, and MI in the right dLGN (Fig. 2e–h). The labeling intensity differed across layers, possibly due to the difference in tracer absorptive rate of classes of retinal ganglion neurons. A similar labeling pattern was also observed in juvenile (Fig. 2i–l; P79, Case 18–16) and adult (Fig. 2m–p; 4.3y, Case 19–03) galagos. However, we noticed that parts of the lateral portion of ME and PE in the adult galago Case 19–03 were innervated with both CTB-green and CTB-red labeled terminals (Fig. 2p).

Figure 2.

The organization of retinal ganglion projections to the dorsal lateral geniculate nucleus (dLGN) in galagos at infant (P5, a-d; P20, e-h), juvenile (P79, i-l), and adult (4.3yr, m-p) stages. The laminar organization of dLGN was characterized with several architectonic staining procedures including those for Nissl, cytochrome oxidase (CO), or vesicular glutamate transporter 2 (VGLUT2). Six layers could be identified as the magnocellular layer external (ME), the magnocellular internal layer (MI), the parvocellular internal layer (PI), the koniocellular internal layer (KI), the koniocellular external layer (KE), and the parvocellular layer external (PE) in a ventral to dorsal sequence. Across four galagos with different ages, the CTB-green labeled terminals were located in the contralateral layers ME, KE, and PE of the dLGN, and the CTB-red labeled terminals were distributed in the ipsilateral layers MI, PI, and KI of the right hemisphere. Note that the CTB-green labeled terminals in the lateral portion of ME and PE in the adult galago Case 19–03 (n), likely represent adjacent inputs in layers ME and PE of the monocular segment of the LGN from the contralateral eye. The scale bar is 0.5 mm in (a-p).

In summary, the organization of the retinogeniculate projections remained largely similar at all developmental stages from at least P5 to adulthood. The retinal ganglion neurons from each eye projected to the ME, KE, and PE of the contralateral dLGN, and to the MI, KI, and PI of the ipsilateral dLGN. Variations in the densities and locations of the label in the LGN layers may reflect uneven uptake of the tracer in the retina.

Superior colliculus (SC)

After tracers were injected into each eye of infant, juvenile, and adult galagos, the distributions of labeled terminals in the SC were examined to reveal the patterns of the retinocollicular pathway during development. In one infant galago (P5, Case 19–14), the CTB-red injection in the right eye labeled boutons throughout the deeper, ventral tier of the uSGS in the right ipsilateral SC. The CTB-green injection was not transported in this case, leaving the more superficial tier of uSGS to the stratum zonale (SZ) largely free of labeling of CTB green boutons from the injection in the contralateral eye (Fig. 3a–d). Some patchy, less continuous labeling from the ipsilateral eye was seen in the lateral region of uSGS. In the left SC, a dense CTB-red labeled axonal projection from the contralateral right eye was seen in the most superficial tier of uSGS, with some slight spreading into the deeper, ventral sublayer of uSGS (see Fig. 4). Another infant galago (P20, Case 18–14) had effective injections in the left (CTB-green) and right (CTB-red) eyes. We observed a dense band of CTB-green labeled boutons in the superficial tier of uSGS of the right contralateral SC, and a wider, less dense band of CTB-red labeled boutons from the ipsilateral eye in the ventral uSGS tier (Fig. 3e–h). Some overlap of the two labeled bands in uSGS was noticed (Fig. 3h). In the right uSGS the CTB-red labeled band and CTB-green labeled band were distributed in a reversed arrangement. Such labeling pattern was also found in the SC in juvenile galagos (P79, Case 18–16), but the extent of overlap from the ipsilateral and contralateral eye projections to the SC was less evident (Fig. 3i–l). In one adult galago (9.6y, Case 18–11), different tracer injections in each eye again labeled two bands in the SC. In the uSGS of the right SC (Fig. 2m–p), the CTB-green labeled boutons from the contralateral eye were located in the superficial uSGS, and the CTB-red labeled boutons from the right eye were distributed in the ventral uSGS. The two labeled bands were clearly segregated (Fig. 3p).

Figure 3.

The organization of retinal ganglion projections to the superior colliculus (SC) in galagos at infant (P5, a-d; P20, e-h), juvenile (P79, i-l), and adult (4.3yr, m-p) stages. Nissl and CO stains revealed that the SC is differentiated into upper stratum griseum superficiale (uSGS), inferior stratum griseum superficiale (iSGS), stratum opticum (SO), stratum griseum intermediale (SGI), stratum album intermediate (SAI), and stratum griseum profundum (SGP) layers in a dorsal to ventral sequence (Kaas and Huerta, 1988). The CTB-red labeled terminals from the ipsilateral right eye are present in the lower level of uSGS of the ipsilateral SC, and the CTB-green labeled terminals from the contralateral eye are restricted to the upper level of uSGS. Note that the distribution of the CTB-red and CTB-green labeled layers tend to be more widespread and diffuse in earlier developmental stages (see the enlarged view in the insets d, h, l), but tend to be more segregated in adult galagos (p). Scale bar in (a) applies to (a-p). Scale bar in inset (d) applies to insets (h), (l), and (p).

Figure 4.

The distribution of retinal ganglion projections to the contralateral (left) and ipsilateral (right) SC after CTB-red was injected into the right eye of an infant galago (Case 19–14; P5). A dense CTB-red labeled axonal projection was seen in the most superficial tier of uSGS of the contralateral SC. In the ipsilateral SC, the CTB-red labeled terminals were located in the deeper, ventral sublayer of uSGS.

Overall, the present results illustrated that the SC receives retinal inputs from both eyes throughout the postnatal development period. Similar to the previous reports in adult galagos (Kaas and Huerta, 1988; Tigges and Tigges, 1970), the inputs from the ipsilateral eyes terminate in the ventral tier of the uSGS in the SC, and are often less dense and less continuous. The contralateral retinal inputs terminate in the more superficial tier of the uSGS with a denser distribution. Although the projections from both eyes to each SC are primarily separate, they appear to slighly overlap in early postnatal development, and become more segregated during the maturation to adulthood.

Pulvinar

Here we present results from four galagos of different ages. The organization of the pulvinar in galagos has not been as fully determined as in other primates. We used the three subdivisions of the pulvinar as defined by Turner et al. (2020) and Wong et al. (2009). The LPul and MPul are mostly located above the brachuim of the superior colliculus (bSC) with the LPul neurons more densely packed than the MPul in Nissl preparations. The IPul is mostly located below the brachuim but extends dorsally to adjoin MPul in the more medial and posterior portion of pulvinar. Parts of the IPul, possibly the posterior and centromedial subnuclei, are VGLUT2-immunoreactive, largely or completely as a result of activating inputs from the superior colliculus (Baldwin et al., 2013). However, VGLUT2 is also present in terminations from retinal ganglion cells (Balaram et al., 2011) to the extent that they exist in the pulvinars.

In one galago at P5 (Case 19–14; Fig. 5), the CTB-red injection in the right eye labeled some axonal boutons predominantly in the contralateral left pulvinar, with a smaller number in the ipsilateral right pulvinar. The majority of CTB-red labeled boutons were distributed in the ventromedial region of the caudal pulvinar, presumably the IPul (Fig. 5c, d), overlapping the larger VGLUT2-dense zone based on the inputs from the superior colliculus. The density of CTB-red labeled boutons decreased toward the more rostral contralateral pulvinar, but these labeled boutons remained in the more medial region of the pulvinar. Note that the CTB-red labeled terminals were densely distributed in the alternative layers of the LGN of the two sides. The same labeling pattern was observed in another infant galago that had the CTB-red injection in the right eye and the CTB-green injection in the left eye (P19, Case 19–07; Fig. 6). A small number of CTB-red labeled boutons were identified in the medioventral region of the left IPul, and overlapped the VGLUT2-dense zone in the more caudal sections. Sparse CTB-red and CTB-green labeled axonal boutons were present in the right pulvinar.

Figure 5.

The organization of retinal ganglion projection to the pulvinar of an infant galago (Case 19–14; P5) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. (a, b) Adjacent sections processed for VGLUT2 and Nissl were used to reveal the architectonic structures in the pulvinar. The most caudal pole of the IPul is mostly VGLUT2 immunoreactive, while the more rostral IPul and other subnuclei in the pulvinar are weak or not immunoreactive to VGLUT2. (c) In the left pulvinar, small numbers of CTB-red labeled terminals from the contralateral eye were found in the medioventral regions of the caudal pole, overlapping the VGLUT2-dense zone, and sparse CTB-red labeled terminals were identified in the more rostral pulvinar. (d) In the right pulvinar, again, small numbers of CTB-red labeled terminals from the ipsilateral eye were found in the medial region of the caudal pole and in the more rostral sector. No CTB-green labeled terminals from the ipsilateral eye were observed in this case because the CTB-green tracer did not transport. (e-h) Photomicrographs showing the CTB-red labeled terminals in selected sites in the pulvinar. Note that abundant CTB-red labeled terminals are present in layers of the adjacent dLGN of both sides. The scale bar in (a) and (b) applies to (c) and (d); (e) applies to (f-h). IPul, inferior pulvinar; MPul, medial pulvinar.

Figure 6.

The organization of retinal ganglion projections to the pulvinar of an infant galago (Case 19–07; P19) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. (a, b) Adjacent sections processed for VGLUT2 were used to reveal the architectonic structures in the pulvinar. (c) In the left pulvinar, a small number of CTB-red labeled terminals were found in the medial and ventral regions of the caudal sector, overlapping the VGLUT2-dense zone. We occasionally found CTB-red labeled terminals in the more rostral pulvinar. No CTB-green labeled terminals were found in the left pulvinar. (d) In the right pulvinar, sparse CTB-red and CTB-green labeled terminals are located in the ventral region of the caudal sector, mostly overlapping the shaded VGLUT2-dense zone. (e, f) Photomicrographs showing the CTB-red labeled terminals in selected sites in the pulvinar. Note that great numbers of CTB-green and CTB-red labeled terminals are present in alternating layers of the dLGN of both sides. The scale bar in (a, b) applies to (c) and (d).

The size of the retinal projection appears to decrease as the galagos mature. In one juvenile galago (P76, Case 18–16; Fig. 7), we were able to identify a small number of CTB-red labeled boutons in the most caudal end of the pulvinar on the contralateral left side, with few of them overlapping the VGLUT2-dense zone. No CTB-red or CTB-green labeled terminals were seen in the right pulvinar, although the adjacent LGN had a great number of terminals labeled by both tracers in the different layers. In adult galagos (see one example in Fig. 8), the pulvinars on both sides are free of labeled axonal boutons after the tracers were injected into the eyes. Yet, dense label was present in the dLGN and SC. In summary, our results indicate that a small retinopulvinar projection is present in the early developmental stages of prosimian galagos, but this pathway weakens as the galagos mature into juveniles and adults. The retinopulvinar projection is bilateral but predominantly contralateral, and especially connected to the medial region of the caudal pulvinar that is VGLUT2-immunoreactive.

Figure 7.

The organization of retinal ganglion projections to the pulvinar of a juvenile galago (Case 18–16; P79) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. (a, b) Adjacent sections processed for VGLUT2 and CO were used to reveal the architectonic structure in the pulvinar. (c) In the left pulvinar, we identified a small number of CTB-red labeled terminals in the lateral region of the caudal sector. No CTB-red and CTB-green labeled terminals were noted in the more rostral sector of the pulvinar in this case. The VGLUT2 positive zone is shaded. (d) In the right pulvinar, no CTB-red and CTB-green labeled terminals were identified. (e, f) Photomicrographs showing the CTB-red labeled terminals in selected sites in the pulvinar. Note that great numbers of CTB-green and CTB-red labeled terminals are present in the eye-specific layers of the dLGN of both sides. The scale bar in (a, b) applies to (c) and (d).

Figure 8.

The organization of retinal ganglion projections to the pulvinar of an adult galago (Case 19–03; 4.3 yrs.) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. (a, b) Adjacent sections processed for VGLUT2 and CO were used to reveal the architectonic structures in the pulvinar. (c) In the left pulvinar, no CTB-red or CTB-green labeled terminals were noted in the pulvinar. (d) Also in the right pulvinar, no labeled terminals were identified. Note that CTB-green and CTB-red labeled terminals are present in the eye-specific layers of the dLGN of both sides. The scale bar in (a, b) applies to (c) and (d).

Pregeniculate nucleus (PrG)

Our results indicate that the pregeniculate nucleus (PrG) receives direct retinal projections throughout the postnatal periods at least from P5 to adulthood in galagos (Fig. 9). Based on the descriptions in Tigges and Tigges (1970), we identified the PrG as the thalamic structure that lies dorsal and lateral to the rostral dLGN, and immediately rostral to the dLGN in Nissl preparations. In one infant galago, Case 19–14 (Fig. 9a, b), the alignment of the plotting from fluorescent sections with the adjacent Nissl sections revealed that the ventromedial region of PrG on the left side was heavily innervated by the CTB-red labeled axonal boutons from the tracer injected into the right eye. The CTB-red labeled terminals were also distributed at the layers in the adjacent dLGN. Similar to the findings in infant galagos, we found the CTB-red labeled terminals in the left PrG in older galagos. See results from three representative cases at P20, P79 and 4.3y in Fig. 9. While the CTB-green injections labeled fewer terminals, a small number of CTB-green labeled terminals were present in the dorsal region of PrG, forming a mosaic pattern with the CTB-red labeled terminals in one juvenile galago (Case 18–16; Fig. 9e, f). In summary, our results suggest that the PrG receives retinal ganglion projections predominately from the contralateral eye, with a much weaker retinal input from the ipsilateral eye in galagos from age P5 to adulthood.

Figure 9.

The distribution of retinal ganglion projections to the pregeniculate nucleus (PrG) in galagos aged P5 (a, b), P20 (c, d), P79 (e, f), and 4.3yr (g, h) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. Nissl stain is used to reveal the architectonic boundaries of the PrG that are located rostral and dorsal-lateral to the dLGN. In all galagos, the CTB-red labeled terminals are present in the expected locations of the ipsilateral PrG on the right side, especially the ventral region. We noted a small number of CTB-green labeled terminals from the contralateral in the PrG, which forms a complementary pattern with the CTB-red labeled terminals in the juvenile case (f).

Accessory optic system (AOS)

Here we showed the results of retinal projections to the LTN of the AOS in galagos at infant, juvenile, and adult ages. The LTN is the largest nucleus of the AOS. In one adult galago (Case 19–03), LTN extends along the adjoining midbrain, most caudally as a reducing narrow strip of neurons under the ventromedial margin of the medial geniculate nucleus (Fig. 10a, b), and most rostrally as a protruding group of neurons located on the superficial lateral surface of the cerebral peduncle (Fig. 10c, d; also see Cooper and Magnin, 1987; Tigges and Tigges, 1970). In four galagos from P5 to 4.3y, we observed a cluster of CTB-red labeled axonal terminals in the caudal location of LTN on the left hemisphere after the tracer was injected into the contralateral right eye (Fig. 10e–h). The labeled terminals extended into the rostral region of the LTN that is located as a bulge on the surface of cerebral peduncle (Fig. 10i–l). No clear CTB-green labeled terminals were seen in the territory of the LTN of the left hemisphere, although CTB-green was injected into the left eye in all animals. Overall, our results suggest that the AOS projection from the eye primarily terminate in the contralateral hemisphere.

Figure 10.

The distribution of retinal ganglion projections to the lateral terminal nucleus (LTN) of the accessory optic system (AOS) in galagos aged P5, P20, P79, and 4.3yr after CTB-green was injected into the left eye and CTB-red was injected into the contralateral right eye. As shown by an adult case (19–03) in Nissl (a, c) and VGLUT2 (b, d), LTN extends along the adjoining midbrain, most caudally as a narrow strip of neurons under the ventromedial margin of the medial geniculate nucleus (LTNc; a, b), and most rostrally as a bulging cluster of neurons located on the superficial lateral surface of the cerebral peduncle (LTNr; c, d). In galagos of all studied ages, clusters of CTB-red labeled terminals were identified in the contralateral LTNc (e-h) and extended to the LTNr (i-l). The scale bar in (a) applies to (b-d), in (e) applies to (f-h), and in (I) applies to (j-l). cp, cerebral peduncle; MGN, medial geniculate nucleus; SNR, substantia nigra reticular.

Pretectum

To reveal the patterns of retinal projections to the pretectum during the development in strepsirrhine galagos, we illustrated results from two representative galagos at infant (P5) and adult (4.3yr) ages after CTB-red was injected into the right eye, and CTB-green was injected into the left eye. Considering the poorly differentiated pretectum in primates, we used adjacent sections processed for Nissl, CO, or VGLUT2 to delineate the boundaries of the pretectal nuclei. In one infant galago, the CTB-red labeled axonal terminals from the right eye were distributed in the expected locations of nOT and OPT on the two sides. In the contralateral left pretectum, the majority of the labeled boutons are located in the dorsal region of OPT, and formed two clusters in the center and medial regions of nOT (Fig. 11). Sparse labeling was observed in the expected territory of APT. In the ipsilateral right pretectum, we found a smaller number of CTB-red labeled terminals in OPT and nOT. The labeled terminals were somewhat limited to the central region of OPT and the medial region of nOT. No labeling was seen in the APT and MPT. Labeled retinal projections from both eyes were also identified in the right pretectum in adult galagos, but appeared to be more comparable in number and intensity in Case 19–03 (Fig. 12). In the right pretectum, the CTB-green labeled terminals from the injection in the contralateral eye were diffusely distributed in the dorsal region of OPT, and formed a laminae in the nOT. The CTB-red labeled terminals from the injection in the ipsilateral eye were present in simiar locations in the OPT and nOT, but with a complementary pattern with the CTB-green labeled terminals from the ipsilateral eye. We did not find any labeled terminals in the APT and MPT.

Figure 11.

The distribution of retinal ganglion projection to the pretectal complex of an infant galago (Case 19–14; P5) after CTB-red was injected into the right eye. Nissl stain reveals the architectonic boundaries of the pretectal olivary nucleus (PTO), nucleus of the optic tract (nOT), anterior pretectal nucleus (APT), and medial pretectal nucleus (MPT). Abundant CTB-red labeled terminals from the contralateral eye are present in the dorsal region of the OPT and the mediocentral region of the nOT in the left pretectum. Sparse CTB-red labeled axonal terminals were identified in the APT. In the right pretectum, a smaller number of CTB-red labeled terminals from the ipsilateral eye are noted in the OPT and nOT.

Figure 12.

The distribution of retinal ganglion projections to the right pretectal complex of an adult galago (Case 19–03; 4.3yr) after CTB-green was injected into the left eye and CTB-red was injected into the right eye. (a-c) Adjacent Nissl, CO, and VGLUT2 sections were used to define the architectonic boundaries of the OPT, nOT, APT, MPT, and posterior pretectal nucleus (PPT). (d) The distribution of CTB-green labeled terminals from the contralateral eye in the OPT and nOT of the right pretectum. (e) The distribution of CTB-red labeled terminals from the ipsilateral eye in the OPT and nOT. (f) The merged view showing the complementary pattern of the CTB-green and CTB-red labeled terminals in the OPT and nOT. The scale bar in (a) applies to all. bSC, brachium superior colliculus; PLi, posterior limitans thalamic nucleus.

In summary, our results suggest that OPT and nOT in galagos received retinal projections from both eyes. The input from the contralateral eye is stronger than the ipsilateral input at an early developmental stage, but it tends to be proportionally equivalent to the input from the ipsilateral eye in adulthood. The retinal inputs to the three pretectal nuclei, either on the contralateral or ipsilateral side, if any, are sparse.

Discussion:

We studied the postnatal development of retinal projections in galagos to provide a better understanding of how the brains of these primates are similar and different from those of other extant primates. These types of observations then guide theories of what early primates were like, and how they changed in the various radiations. The logic of a comparative approach to do this is long standing, and has been outlined by others (e.g., Butler and Hodos, 1996; Kaas and Preuss, 1993; Strieder, 1998; Preuss, 2007). As we cannot directly study the organization and development of the brains of early primates, the “cladistic” approach is a useful way of reconstructing the likely features of brains of early primates (Hennig, 1996). This approach consists of studies on present members of the phylogenetic radiation of primates (or any taxa), and proposing that shared features across branches of a radiation are most likely features retained from a common ancestor. In addition, related findings can inform this process. We know from the fossil record that early primates were small, had small brains appropriate for their size, and had large, forward facing large eyes, suggesting a well developed visual system, and a nocturnal lifestyle (e.g., Kaas et al., 2022). Modern primates consist of strepsirrhines (galagos, lemurs) and haplorhines. The early haplorhines gave rise to tarsiers and anthropoids. Early anthropoids gave rise to platyrrhines (New World monkeys) and catarrhines (Old World monkeys, apes, and humans) (Preuss, 2007). Earlier classifications grouped strepsirrhines and tarsiers as prosimians, a term still in use. The fossil record supports the view that some extant primates have changed the least since early primates. Thus, the brains of galagos and some of the lemurs that occupy the “fine branch, nocturnal” niche of early primates may have changed the least, and human brains have changed the most. Although parts of this premise may be questionable, it is consistent with considerable comparative evidence.

The present results indicate that most of the retinal projections to subcortical structures are in place and well developed by birth in galagos, and similar results have been obtained in New and Old-World monkeys. Such results suggest that much of the maturation of the visual system in primates occurs before birth, and before functional vision. Although the retinal projections to part of the pulvinar are rather sparse in infant galagos, they may be denser before birth and they are exceptional in that they are reduced and apparently lost as galagos mature. The projection to the pulvinar appears to be greater in infant New World marmosets, but it also is reduced during maturation (e.g., Mundinano et al., 2018). Such a loss is unusual and somewhat puzzling. Thus, Mundinano and colleagues propose that the postnatal reduction with maturation of the pathway from retina to pulvinar to temporal cortex has a critical role in the early maturation of the middle temporal visual area (MT) and the dorsal stream of visual processing. Once this function is over, the retina to pulvinar pathway regresses. Given these results and interpretations, our following discussion has two main parts. First, we consider the evidence that most retinal projections are highly mature at birth in primates. Second, we consider the possibility that the retina to pulvinar projection regresses postnatally in primates and has a role in the early maturation of dorsal stream visual cortex.

The early development of retinal projections to subcortical visual structures

The two major targets of retinal projections are the dorsal lateral geniculate nucleus (dLGN) and the superior colliculus (SC). The present results indicate that the restrictions of retinal projections to laminar targets in these structures are in place by birth, and there is little or no apparent change during postnatal maturation. This early postnatal retinal projection pattern to the LGN of galagos (Fig. 2) is impressive in its resemblance to the mature pattern of lamination of six clearly separated pairs of layers, one for each eye, and three types of retinal ganglion cell projections corresponding to the pairs of magnocellular, parvocellular, and koniocellular layers (e.g., Kaas et al., 1978). Thus, three types of outputs from the retina sort themselves out by eye of origin and class of ganglion cells before birth. In galagos and other strepsirrhine primates, this includes two distinct K cell layers, which are absent in anthropoid primates, where K LGN cells are largely located between recognized cell layers and do not form distinct layers (Hendry and Reid, 2000).

The prenatal development of the LGN in galagos has not been studied.

What is now known about the prenatal development of the LGN in primates is almost completely from studies on macaque monkeys (Rakic, 1977a; 1977b). The migration of cells to form the LGN mass is early in development, about the time that the retinal ganglion cells emerge and send axons toward the LGN territory. The ganglion cells in the retina differentiate into the M and P types soon after they are generated, and initially, axons are sent past the LGN toward the superior colliculus. From around 50 to 75 days of prenatal development, the territory for P and M ganglion cell terminations are separated without divisions into layers from each eye, and this occurs even without retinal inputs (Rakic, 1981). At the early stages, the inputs from the two eyes overlap extensively within P or M cell territories. Retinal terminal arbors form first in the P LGN regions, and arbor development occurs first in both P and M regions that will represent central vision, and then in regions that will represent peripheral vision. Well before birth, the separate P and M layers for the two eyes emerge. As the K ganglion cell axons terminate on neurons between the LGN layers in macaques and other anthropoid primates, their prenatal development has not been studied. However, as spaces between layers form late as P and M layers for each eye emerge, K ganglion cell inputs may be late in segregating to the interlaminar zones.

The prenatal and early development of lamination patterns in the LGN have also been studied in tree shrews, cats and other carnivores, and rodents, with some similarities such as the formations of layers before eye opening, and the separation of W ( K like) layers from mixed X (P like) and Y (M like) cell layers, but without the separation of P (X) and M (Y) cell layers as in primates. Layers for the missing eye do not emerge if that eye is removed before eye specific layers emerge in monkeys, tree shrews, and mink (see Casagrande and Condo, 1988; Chapman, 2003).

The superior colliculus is the other major target of retinal projections. In adult galagos, the inputs from the contralateral eye terminate most superficially in the superficial grey of the superior colliculus, while the termination of the ipsilateral eye are separated across the colliculus just ventral to the inputs from the contralateral eye (Fig. 3; Compos-Ortega and Cluver, 1968; Tigges and Tigges, 1970; Symonds and Kaas, 1978). The more superficial terminations for the contralateral eye even have a discontinuity corresponding to the optic disc, indicating a location on the horizontal meridian some 20 degrees from the center of vision (Symonds and Kaas, 1978). The terminations from the eyes in the superior colliculus express the VGLUT2 protein, which is somewhat denser in the upper terminations from the contralateral eye than the lower terminations from the ipsilateral eye (Balaram et al., 2011), consistent with the evidence that the projection from the ipsilateral eye is weaker. Our postnatal results on the development of retinal projections to the superior colliculus indicate that the segregation of axons from the contralateral and ipsilateral eyes is complete as early as postnatal day 5, and likely emerge before birth and functional vision.

The retinal projections to the superior colliculus are somewhat different in galagos than in adult macaques (Hubel et al., 1975; Pollack and Hickey, 1979). Instead of an upper sublayer of the superficial grey with inputs from the contralateral eye over a ventral sublayer with inputs from the ipsilateral eye, as in galagos, macaques have a thicker layer in the superficial grey that is dominated by inputs from the contralateral eye that is subdivided by patches of inputs from the ipsilateral eye. The patches align to form rostrocaudal bands. The retinal projections to the superior colliculus in New World marmosets resemble those in macaques (Kwan et al., 2019; see Fig. 13 in Kaas and Huerta, 1988 for other primates and cats). Retinal projections in non-primate mammals vary, especially for those from the ipsilateral eye, which may range from very limited to approaching the density of those from the contralateral eye. The projections from the contralateral eye are more extensive due to inputs from the monocular segment of peripheral vision from the contralateral eye, and the projections of the complete retina of the contralateral eye to the superior colliculus (Lane et al., 1971; 1973). The prenatal development of the superior colliculus has been studied in macaque monkeys (Rakic, 1977). In the middle stages of prenatal development, the inputs from the two eyes are mixed in the superficial layer, but they become separate well before birth, indicating that visual experience is not necessary for the separation of inputs from the two eyes to occur.

Other terminations from the retina in our study of galagos included those to the pregeniculate nucleus, the accessory optic system, and the pretectum. All of these inputs were in place and adult-like by or before postnatal day five. The accessory optic system in galagos has been previously described by Tigges and Tigges (1969) and Compos-Ortega and Culver (1968). Similar retinal projections have been reported for another strepsirrhine primate, the mouse lemur (Cooper, 1986). As in other mammals, the projections are mainly from the contralateral eye. The accessory optic system is thought to correct for “retinal slip” that results from head and body movements (Simpson, 1984). The pretectal nuclei in galagos resemble those in other primates and in non-primate mammals (Hutchins and Weber, 1995) with the contralateral label being denser.

Retinal projections to the pulvinar

In contrast to the overall adult-like appearance of those retinal projections in newborn galagos and other primates, the projections of the retina to the pulvinar were sparse in newborn galagos and absent or nearly absent in adults. The evidence for projections from the retina to the pulvinar in other primates is limited, suggesting that such projections are sparse or absent. They have been described in macaques (O’Brien et al., 2001; Cowey et al., 1994), baboons (Campos-Ortega et al., 1970), and marmosets (Warner et al., 2010; Kwan et al., 2019). In adult macaques, the retinal projections were sparse and located in PIm, which projects to MT, and PIp, which projects to the MT surround (O’Brien et al., 2001). In adult marmosets, the rather sparse retinal inputs were mainly to the medial nucleus of the inferior pulvinar, PIm, with a few terminations in adjoining nuclei of the inferior pulvinar. Very few terminations from the ipsilateral retina were found in PIm. PIm projects to the middle temporal visual area (MT) (e.g., Warner et al., 2015). Projections of the retina to a part of the pulvinar have also been reported for tree shrews, squirrels, and cats (see Warner et al 2010 for review).

A surprising finding in marmosets was that the retinal projections to PIm were most dense when labeled in infants of eight days old (Warner et al., 2012). The projections remained dense for a marmoset just over two weeks of postnatal life, and then diminished as older marmosets approach maturity. This finding supported the proposal that an early retina to pulvinar to MT pathway is responsible for the early maturation of area MT. Histologically, MT appears to mature nearly as early as primary visual cortex (Turner et al., 2020; Bourne and Rosa, 2006). Thus, MT acts as a primary visual area, and the early activation of MT by the pulvinar allows an early development of other dorsal stream visual areas (Warner et al 2012). In galagos and marmosets, PIm appears to get few, if any, retinal inputs (present study; Mundinano et al., 2018), and PIm gets few or no inputs from the superior colliculus (Kaas et al 2019; Kaas and Lyon, 2007; Stepniewska et al., 2000). Thus, the activity in MT does not depend on inputs from the superior colliculus to PIm. Instead, MT is highly dependent on direct and indirect inputs from primary visual cortex, V1 (e.g., Collins et al., 2005). Therefor, the proposal is that the retina to PIm is a “transient visual pathway” that is still present at birth, and over several postnatal months the pathway from V1 “strengthens” its connections to MT as the retino-pulvinar pathway diminishes (Mundinano et al., 2018). The present findings of a very weak retina to pulvinar pathway in newborn galagos that disappears over maturation is consistent with this proposal, but the retina to pulvinar pathway at birth appears to be too weak to have a role in promoting the postnatal development of the dorsal stream via MT. Quite possibly, however, the retina to pulvinar pathway is much stronger prenatally, and the major impact on MT and the dorsal stream is already over by birth. This alternative seems somewhat possible as the segregation of eye specific pathways to the LGN and superior colliculus are already present at birth and the segregation is activity dependent but not visually dependent (Wong et al., 1993; Huberman et al., 2008; Ackman et al., 2012). Thus, before birth, the differing spontaneous activity waves in the retina of each eye activate targets such as the LGN and superior colliculus to mediate adult-like patterns of eye-specific layers. If the prenatal retina to pulvinar pathways were strong at these prenatal times, they could promote the early maturation of MT and the subsequent maturation of dorsal stream visual areas dependent on MT. Yet, architectonic features of MT develop early in MT of marmosets (Bourne and Rosa, 2006) and MT in galagos has inputs densely labeled with parvalbumin within days of birth that fades with maturation (Turner et al., 2020). Because of these uncertainties, more studies on the nature of the retina to pulvinar pathway are needed in different primates, both postnatally and prenatally. Studies are needed at the electron microscope level to confirm the existence of retino-pulvinar synapses, and neuronal recordings could establish the sources of activation of MT during postnatal development. A broader scope of such studies could reveal specific species variations and allow further inferences about evolutionary changes.

DATA AVAILABILITY STATEMENT:

The data that support the findings of this study are available from the corresponding author upon reasonable request.