Orienting our view of the Superior Colliculus: Specializations and general functions

Abstract

The mammalian superior colliculus (SC) and its non-mammalian homologue, the optic tectum (OT) are implicated in sensorimotor transformations. Historically, emphasis on visuomotor functions of the SC has led to a popular view that it operates as an oculomotor structure rather than a more general orienting structure. In this review we consider comparative work on the SC/OT, with a particular focus on non-visual sensing and orienting, which reveals a broader perspective on SC functions and its role in species-specific behaviors. We highlight several recent studies that consider ethological context and natural behaviors to advance knowledge of the SC as a site of multi-sensory integration and motor initiation in diverse species.

Introduction

Reacting to the world

Orienting to environmental stimuli is fundamental to a wide variety of natural behaviors, such as prey capture, predator avoidance, and navigation. In the context of such behaviors, animals must interact with objects in their surroundings by orienting their sensory organs toward relevant stimuli. While the sensory organs animals use to probe the environment are species-specific (eye, whisker, ear, or mechanosensors on the body), the orienting responses across taxa share a common function, namely the collection of stimulus information to inform goal-directed actions. Here we consider the neural underpinnings of species-specific orienting behaviors, with a focus on sensorimotor transformations in the midbrain. Our review considers the functional organization of the mammalian superior colliculus (SC) and its non-mammalian homologue, the optic tectum (OT), for processing multi-modal sensory information and initiating orienting responses to salient stimuli. We highlight recent works that capitalize on natural behaviors in diverse species to show that animals, irrespective of sensory modality, perform similar computations to localize stimuli and produce appropriate orienting responses.

The SC as an orienting structure

The midbrain Superior Colliculus (SC) and the Optic Tectum (OT), operate in fundamental sensorimotor transformations that support species-specific orienting behaviors [1–5].

The SC has been intensively studied, with a seminal paper dating back to 1946, in which Spiegel and Scala identified the role of the SC in the vestibulo-ocular reflex. Other early works revealed the relationship between visual and oculomotor maps in the SC, laying the foundation for research focused on the role of the SC in oculomotor orienting in primates, cats and rodents [6–9]. Correspondence between visual maps in the superficial layers of the SC and eye movement maps in deeper layers were observed in the Rhesus monkey and cat [7,8]. Electrical stimulation experiments demonstrated that the direction and amplitude of eye movements depends on the locus and intensity of SC stimulation, and subsequent work revealed that SC stimulation also evokes pupil dilation [6,10–12], micro-saccades [13], and head movement [10]. Table 1 presents a non-exhaustive list of orienting behaviors evoked by SC/OT stimulation across species, listing species-specific behaviors, but emphasizing the causal role of this midbrain structure in orienting behaviors across taxa. Early studies of the SC also reported somatosensory and auditory maps [14–17] in snakes, owls, cats, and rodents, but the research community has largely continued to emphasize visuomotor orienting functions of the SC, even in mammalian species, such as mice and rats, whose natural behaviors depend heavily on other sensory modalities [18].

Table 1.

Summary table of orienting behaviors across species, elicited by SC stimulation

Species	Stimulation method	Animal behavioral state	Elicited behaviors	Reference
Cat	Strychnine combined with photic stimulation	Lightly anesthetized Head-fixed	Eye movements	Apter 1946
Cat	Electrical stimulation	Alert Neck-fixed	Saccades Head movements	Roucoux et al. 1980
Rhesus Monkey	Electrical stimulation	Alert Head-fixed	Saccades	Schiller & Stryker 1972
Rhesus Monkey	Electrical stimulation	Alert Head-fixed (head horizontal movement allowed)	Saccades Head movements	Schiller & Stryker 1975
Rhesus Monkey	Electrical stimulation	Alert Head-fixed	Saccades Pupil Dilation	Wang et al.2021
Bat	Electrical stimulation Chemical stimulation	Alert Head-fixed/Head-free	Pinna movements Head movements Vocalization Neck or Body movements	Valentine et al. 2002
Rat Hamster	Electrical stimulation	Lightly anesthetized Head-fixed Alert (n=3) Paralyzed(C-6 spinal cord transection)	Eye movements Pinna movements Vibrissae movements	McHaffie et al. 1982
Barn Owl	Electrical stimulation	Lightly anesthetized Head-fixed	Small eye-movements Pupil dilation	Netser et al. 2010
Barn Owl	Electrical stimulation	Alert Body restrained	Head movements	Du Lac et al. 1990
Lamprey	Electrical stimulation	Head-fixed	Eye movements Body movements Locomotor movement (stimulation length dependence)	Seitoh et al. 2007
Toad	Electrical stimulation	Alert Freely moving	Prey catching behavior: orienting, snapping Avoiding behavior	Matsushima et al. 1985

The SC/OT is a prime structure for comparative research, due to its highly conserved organization across vertebrates. This was noted by Heric and Kruger in 1966 during their investigation of the alligator OT: “anatomical similarities among vertebrates classes’’, including consistent laminar structure and retinal fibers projecting onto the SC/OT ([19]). Broadly, the SC/OT dorsal division receives retinal input and can be considered visuosensory, while the intermediate division receives information from other sensory modalities and the ventral division controls the initiation of orienting movements. Figure 1 highlights the consistent laminar organization across vertebrates by using a schematic representation of the structure in example species. For more details about the anatomical and functional divisions of the SC/OT, as well as the circuitry and computational processes underlying visually-guided orienting behaviors, see [20].

Figure 1.

Schematic representation of the SC (OT) for example species showing conserved laminar organization across taxa.

We write this review to call attention to exciting and important recent discoveries that highlight the role of the SC and OT in ethologically relevant orienting behaviors across species that rely on diverse sensory modalities. A comparative approach to the neural underpinnings of sensorimotor integration for species-specific orienting behaviors presents the opportunity to identify the mechanisms that are conserved across taxa, as well as species specializations, and to bring into focus open questions in the field.

Visual orienting

Sensorimotor functions of the SC were initially identified through studies of the oculomotor reflex and saccadic eye-movements. Subsequently, the field has been dominated by studies of SC visuomotor transformation, in animal models that either are visually dominant (such as macaque or cat) or even in animal models in which vision is a secondary modality (such as mouse). While this research yielded valuable working models of the SC’s 1) anatomical organization, 2) sensorimotor integration operations and 3) circuits underlying the initiation of orienting behaviors, studies tended to overlook the broader functions of SC across the animal kingdom.

Here we present new literature that considers visuomotor functions of the SC/OT in ethologically relevant contexts. Recent exciting results have put this question to the forefront of the community and invited us to redefine our framework [21]. An excellent example of the role of the SC/OT in guiding orienting behaviors is found in the lamprey. Lamprey are basal vertebrates, phylogenetically distant from mammals, with a well-developed OT, whose anatomy and function resembles that of other vertebrates (see Fig 1). Recent study of the lamprey OT in a reduced preparation shows that two distinct cell types in the deep layers contribute to distinct responses to visual stimuli. Prey-like visual stimuli drive activity in contralateral brainstem-projecting neurons and evoke activity corresponding to fictive orienting behaviors, or motor neuron activity indicative of a behavioral orienting response, while looming stimuli activate ipsilateral brainstem-projecting neurons, resulting in fictive escape behaviors. These stimulus selective responses are mediated by GABAergic interneurons in the deep tectal layer [22]). Similarly, recent studies in mice have begun targeting cell types responding to looming stimuli within the SC and have identified cells within the deep layers of the mouse SC that are likely candidates for producing highly selective responses to looming stimuli [23]. Simultaneously, specific cell types within the SC have been identified as being necessary for visual orienting and tracking behavior in both rats and mice [1,24] in natural contexts such as predator escape and prey hunting behaviors.

Auditory orienting in birds and bats

The Barn Owl (Tyto alba) hunts by using both visual and auditory cues generated by prey moving in the environment. The barn owl orients by turning its head to maximize stimulus localization, and microstimulation of the owl OT reveals topographic organization of motor fields ([25]; Fig. 2B). Because of the owl’s exceptional prey capture performance, which exploits highly sensitive visual and auditory localization, these animals have served as models of visual and auditory integration within the OT [26]. Even in the chicken, an avian species that is not an auditory specialist, robust connections between the OT and auditory processing structures are present, demonstrating that auditory localization is a function of the OT across species [27].

Figure 2.

Microstimulation of the SC/OT elicits species-specific orienting movements.

A. Eye movements in primates (Schiller and Stryker, 1972), Photo: David Raju - Own work, CC BY-SA 4.0 B. Head movements in owls (DuLac and Knudsen, 1990), Photo: Peter Trimming CC BY-SA 4.0, C. Pinna and head movements and echolocation call production in bats (Valentine et al., 2002), Photo: Brock Fenton and D. Vibrissa movements in rodents (Helmet and Keller, 2008) Photo: Clarice Diebold

While owls rely on visual and auditory stimuli generated by moving prey, echolocating bats localize food sources by emitting ultrasound signals and listening to echoes returning from objects in their surroundings ([28]). Coupled with the active production of sonar sounds, bats also move their external ears (pinnae) and head to align a stimulus with maximum sensitivity of the receiver [29]. Bat orienting movements of the head and ears facilitate sound reception, paralleling the function of eye movements of primates and cats to capture visual stimuli with maximum resolution. The SC in bats thus plays a role in auditory orienting, similar to visual orienting in other mammals.

The big brown bat SC shows functional organization that differs somewhat from other mammals: in contrast to primates, visual responses are weak or absent [30], and instead there are robust auditory responses in dorsal layers, which reinforces the notion that auditory representations of sound sources in bats guide motor responses [31]. It is noteworthy that most neural recording studies of the bat SC have been conducted in head-fixed, passively sensing animals (e.g. [32]). A recent report on SC activity in freely behaving bats calls into question the classic laminar organization of the SC. Wohlgemuth et al. (2018) took multichannel recordings from the SC of an echolocating bat tracking a moving target and discovered a co-mingling of auditory and premotor neurons throughout superficial and deep layers [5]. Previous electrophysiological recordings in the SC of the same insectivorous bat corroborate the lack of auditory spatial topography compared to other species and other modalities [33]. However, neurons do show 3D spatial tuning and are selective for object distance [30]. This may reflect the nature of auditory processing of acoustic scenes. Auditory responses depend on combinatorial effects of frequency, intensity, and timing, rather than the inherently topographic nature of visual input on the retina. Thus, in species with little reliance on visual input, topographical organization may be less integral to representing three dimensional spaces. In other bat species that rely on integrated sensory modalities, such as the omnivorous bat Phyllostomus discolor that uses both echolocation and visual orienting, auditory and visual space are mapped topographically. In this bat species, visual responses are restricted to the superficial layers and auditory maps are represented in the deep layers and these topographic maps are well aligned [34].

The functional role of the bat SC in sonar orienting has been characterized through stimulation and extracellular recording experiments. Electrical and chemical stimulation of the bat SC elicits sonar vocalizations, head movements, and adjustments of the pinnae ([35,36], see Table 1 for summary and Fig. 2C). Sinha and Moss [36] and Wohlgemuth, Kothari and Moss (2018) [5] report robust pre-vocal motor activity in the SC of freely echolocating bat, E. fuscus. Further, Sinha and Moss [36] discovered that the lead time of vocal premotor activity shortens as call duration decreases and rate increases. These findings illustrate some species-specific specializations for acoustic orienting by sonar.

In the first study of auditory responses of SC neurons to echoes from physical objects in a free-flying bat, Kothari et al. (2018) report 3D selectivity to the direction and delay of sonar objects [37]. Moreover, the 3D response areas of SC neurons are modulated by sonar-guided attention: Specifically, when bats adjust their echolocation calls to attend to objects, neurons show sharper tuning with best responses at shorter distances.

Tactile orienting in rodents

In rodents, whisking is used as a primary sensing modality, allowing for navigation and object exploration, even in the total absence of visual input [38]. Rodents use whisking to investigate fine-grained details of targets in their environment [39], and SC microstimulation elicits movements of vibrissae ([40]; see Fig. 2D). Somatosensation therefore plays a dominant role in rodent orienting and invokes specializations of the SC.

In rodents, afferent projections from the trigeminal nucleus of the brainstem [39], as well as feedback from the vibrissal somatosensory cortex (vS1), project to the intermediate and deeper layers of the SC [42]. In other animals that rely more heavily on somatosensation over vision, the SC shows marked architectural changes. For example, in naked mole rats and moles (Insectivora), both animals with very limited vision, the superficial visual layers of the SC are substantially reduced, while the intermediate and deep multimodal layers are well developed [43,44]. Despite the absence of a direct SC projection to vS1, activation of the SC enhances vS1 responses to whisker deflections in mice [45], suggesting that the SC plays a role in sensory detection through indirect projections, enhancing sensitivity and discrimination of tactile objects.

Proprioception in rodents and non-human primates

An important component of somatosensory driven behavior is precise control of head movements and proprioception of head position in relation to self and objects in the environment. Proprioception requires accurate encoding of head displacement along three axes (roll, pitch, and yaw). Neurons that accurately represent all three axes, as well as combinatorial vectors, have been identified in the intermediate layers of the mouse SC, and these neurons accurately predict head orienting movements [46]. This finding is particularly notable, as previous research in primates has shown that roll was not encoded in the SC [47]. This difference raises two hypotheses that warrant further investigation: 1) there is a major divergence in SC function in the rodent and primate lineages, or 2) previous primate studies did not allow animals the freedom of movement required to accurately measure the full range of head direction encoding. Both of these hypotheses highlight the need to carefully consider the neural mechanisms that evolved to support natural behaviors. Advances in recording technology are now enabling head-free recordings in monkeys, allowing new investigations of the role of the SC in guiding primate head movements [48,49]; however, current experimental paradigms still present inherent challenges in the interpretation of SC function, due to their limited nature: For example, reaching for a stimulus on a touch screen may present a very narrow view of SC activity, especially considering the intriguing role of the cell-specific SC circuitry highlighted in natural orienting behaviors in the mouse [1].

Multimodal orienting across species

The SC also serves as a site for multisensory integration [50]. Auditory and visual maps are aligned within the cat and primate SC, facilitating target localization [16,51], and auditory receptive fields in primates shift with eye position [52,53]. Barn owls have been historically used as research models to investigate the integration of visual and auditory information within the OT, and recent work has examined the circuitry of both integration, stimulus selection, and attention [54].

The SC has been implicated in spatial attention in both mammals and birds [55] and circuit mechanisms have been described in depth [56]. Recent studies of OT circuity reveal inhibitory feedback mechanisms that sharpen spatial networks in the OT of the owl [57]. These new findings unveil the circuitry and computations that take place in the OT and mediate stimulus selection for natural orienting behaviors.

In whisking rodents, in which auditory and visual information are augmented by somatosensory information, past work has investigated the convergence of these three modalities within the SC [49], but recent work has begun to dissect how cross-modal interactions in the SC develop. Cuppini et al. (2018) proposes a model of the SC that frames sensory input as innately competitive until sufficient experience forms non-competitive interactions between complementary input [58]. This framing is particularly useful when considering the role of the SC as an orienting structure, to enable rapid evaluation of multimodal stimuli for target selection. This model is supported by the finding that within intermediate layers of the SC in rats, very few neurons are activated by both visual and vibrissal stimulation, and more frequent are instances of cross-modal suppression of neuronal activity [59]. This raises interesting questions about the relative salience and aversiveness of somatosensory and visual inputs and their competing influences on orientation and attention. As demonstrated in the lamprey, the SC/OT is responsible for both orientation towards attractive targets as well as avoidance of aversive stimuli, with distinct circuitry devoted to these two opposed tasks. In an ethological context, somatosensation is primarily used by rats in foraging, burrowing, and navigation, while vision complements these functions, but is also associated with predator detection and avoidance behavior. The role of whisking in guiding the detection of attractive stimuli at close range and the role of vision in detection of aversive stimuli at a distance suggests an evolutionary explanation for competitive encoding of these two senses. An exciting avenue for future research would be to study the effect of additional sensory cues, either attractive or aversive stimuli, such as odorants, on modulating cross-modal interactions within the SC.

In situations where somatosensory and visual input serve complementary functions, learned associations may provide excitatory cross modal input [58]. In rodents, prey capture sequences are guided by visual and somatosensory inputs used together to detect, orient toward, and finally capture prey items. Orienting and capture behaviors arise from complex interactions of multisensory and motor circuits, supporting visual stimulus detection (WF neurons and NF neurons), target pursuit (NF neurons) [1] and capture behavior, mediated by somatosensory input [4]. This work demonstrates the role of the SC in linking visual and somatosensory stimulus detection and orienting responses. This work is exciting in particular for its use of ethologically relevant appetitive stimuli (live insects). It also suggests the importance of context and experience in shaping how sensory inputs are integrated and transformed to produce motor orienting responses.

Conclusions

Early work on the SC was dominated by animal models that rely primarily on vision to orient, establishing a framework to consider this midbrain hub an oculomotor structure. Comparative studies broaden this framework by revealing species specializations in the SC that reflect the dominant sensory modalities animals use to acquire information about the environment and premotor neurons that activate species-specific orienting movements. Comparative work thus uncovers the SC’s functions as an orienting structure that transforms visual, acoustic and/or somatosensory information to drive movement of the eyes, ears, head, whiskers, and in the case of bats, the production of echolocation calls. These broader functions suggest an overarching principle, namely the SC supports linkages between sensing and action to enhance the stimulus information that allows animals to assess and respond to biologically relevant events in the world.

Highlights.

• The SC is a layered sensorimotor structure, and its anatomical organization is highly conserved across taxa.

• Focus on the SC’s role in visuo-motor integration has detracted attention from its broad function, namely sensorimotor transformations for species-specific orienting behaviors mediating prey capture, predator avoidance and navigation.

• The SC has a conserved function across taxa in the execution of orienting behaviors to ethologically relevant stimuli.