Motor Systems

Early neurophysiological studies in the octopus nervous system revealed three major categories of nervous tissue that elicited motor patterns (Boycott, 1961). “Lower” motor centers of the arms and some lobes of the subesophageal brain contain motor neurons and generate responses in a restricted set of muscles. For example, the lateral pedal lobes of the subesophageal brain innervate the eye muscles and control the orientation of the pupil (Wells, 1960). Similarly, electrical stimulation in the chromatophore lobes causes dilation of restricted groups of the chromatophore components of the adaptive coloration system. Also included in this grouping are the vasomotor lobes, which reside at the most posterior end of the subesophageal brain and control the musculature of the blood vessels. “Intermediate” motor centers, located in the posterior subesophageal brain and the brachial lobe, can trigger more extensive muscle movements, such as contractions of the entire mantle, gut, or multiple arms.

Coordinated natural actions, such as swimming or locomotion, can only be generated by electrical stimulation of the “higher” motor centers of the supraesophageal brain. Electrical stimulation of the anterior basal lobe, for example, results in highly naturalistic turning motions of the head and eyes (Table 1). The majority of the inputs to the basal lobes comes from deep within the optic lobes, which itself is a “higher” motor center that processes visual information (Young, 1962).

Frontal-Vertical System

The supraesophageal brain also contains “silent lobes,” so-called because they do not elicit movements following electrical stimulation. In the adult octopus, these include the frontal lobes and the vertical lobe (Figure 2b). The frontal lobes stretch across the oral portion of the central brain (Figure 2b). The vertical lobes form oblong gyri extended along the oral-aboral axis on top of the basal lobes (Figure 2b). Each gyrus has its own distinct tube of neuropil, and antibody staining reveals distinct neurochemical properties for each vertical lobe gyrus (Figure 2c) (Shigeno & Ragsdale, 2015).

The vertical and frontal lobes form two different learning and memory systems (reviewed in Shigeno & Ragsdale, 2015): the inferior frontal system, dedicated to tactile learning and memory, and the superior frontal-vertical system for visual learning and memory. The two centers are largely separate, but their circuitries show a remarkable parallelism in organization (Figures 3a and b). For both modalities, information entering the central nervous system can arrive at the highest region of integration via parallel series of relays that traverse multiple lobes. These paths show striking similarity to the circuitry of the vertebrate cerebellum (Figure 3c). Experimental evidence suggests that these octopus memory centers engage in long-term potentiation, a cellular correlate of memory acquisition (Shomrat, Zarrella, Fiorito, & Hochner, 2008).

Open in new tab

Figure 3. Octopus learning and memory circuits contain parallel paths. Amacrine cell numbers are for Octopus vulgaris (Young, 1971) and granule cell numbers are for Mus musculus. (a). The tactile learning and memory system converges on the subfrontal lobe, with a long path that passes through the inferior frontal lobes and a short path (dotted line) through the posterior buccal lobe. (b). The visual learning and memory system converges on the vertical lobe. The long path passes through the superior frontal lobes, while the short path goes through the subvertical lobe. (c). Parallel long and short paths are a feature of learning and memory systems. In the vertebrate cerebellum, for example, sensorimotor information from the spinal cord reaches the deep cerebellar nuclei through parallel paths. The deep cerebellar nuclei send output back to the spinal cord through intermediaries, such as the red nucleus and the neocortex (not shown). Note that, although not diagrammed, many of these feed-forward connections are reciprocated.

The ultimate target of the tactile learning and memory system is the subfrontal lobe, which contains ~5 million tiny amacrine cells (Table 2, Figure 3). Tactile information from the arms arrives at the lateral inferior frontal lobe. Here, primary sensory neurons synapse onto relay neurons of the median inferior frontal lobe, which then project to the subfrontal lobe. Alternatively, tactile information can enter through the posterior buccal lobe, which also projects to the subfrontal lobe (not shown in Figure 3, but see legend).

The visual learning and memory system converges onto the vertical lobe (Table 3, Figure 3b). Visual information from the optic lobes projects to the lateral superior frontal lobe, which relays to the median superior frontal lobe. These neurons project to the vertical lobe. In the short path of the visual learning and memory system, information travels to the subvertical lobe by way of the lateral superior frontal lobe. The subvertical lobe then projects back to the optic lobe, but it can also feed into the vertical lobe (not indicated in Figure 3, but see Table 3).

Arms and the Axial Nerve Cords

Octopus arms and their suckers are multiplex sensorimotor organs that enable the octopus to explore new environments, manipulate objects with great dexterity, capture prey, and exchange gametes (Caldwell, Ross, Rodaniche, & Huffard, 2015; Hanlon & Messenger, 2018). The densely packed muscles of the arms, when co-contracted, provide rigidity in the absence of any skeleton (Kier, 2016). In each arm, these muscles enclose a bundle of nervous tissue extending directly from the subesophageal brain, called the axial nerve cord. Axons for the control of intrinsic muscles of the arm exit the axial nerve cord laterally. The ventral portion of the cord connects to a series of ganglia, with one ganglion for each sucker (Graziadei, 1971; Graziadei & Gagne, 1976). These sucker ganglia, subjacent to the axial nerve cord, are thought to contain circuits that generate the fine motor movements of the suckers and process incoming sensory information. Many details of axial nerve cord circuitry are as yet unresolved (Graziadei, 1971; Rowell, 1963), but its neural organization evokes the anatomy of the vertebrate spinal cord—that is, central nervous tissue with swellings along its length for appendage sensory-motor control and long-distance tracts to and from the brain.

Due to its elaborate array of muscles, the octopus arm can generate movements with almost infinite degrees of freedom. This presents a computational challenge for the neuromuscular system of the arms and its regulation by the central brain. In goal-directed behavior, the degrees of freedom in the arm can be reduced by the formation of “pseudo-joints.” The neural control of these pseudo-joints in activities such as reaching and grasping looks very similar to primate control of volitional movements (Sumbre, Fiorito, Flash, & Hochner, 2006; Sumbre, Gutfreund, Fiorito, Flash, & Hochner, 2001). The neuromusculature of the octopus arms has served as an important biological inspiration for the creation of soft robotics (Margheri, Laschi, & Mazzolai, 2012; Rus & Tolley, 2015).

Chemosensation

Behavioral studies suggest that chemical cues are detected by the suckers of the arms: upon contact with aversive chemical substances, octopuses withdraw their arms and suckers immediately. In chemotactile training experiments, octopuses were shown to be able to discriminate among sucrose, hydrochloric acid, and quinine (Wells, Freeman, & Ashburner, 1965). Octopuses can also distinguish different concentrations of salinity (Wells, 1963).

Despite these behavioral observations, the cellular and molecular substrates for octopus chemosensation have not yet been identified. In addition, the nature of chemosensory processing in the octopus remains largely unknown. In both vertebrates and flies, second-order olfactory sensory neurons form glomeruli with first-order sensory axons. The near-ubiquity of this circuitry in distantly related species would suggest that glomerular convergence is a fundamental organizing principle in olfactory processing. However, neural organization resembling glomeruli is conspicuously lacking in the octopus nervous system and chemosensory processing remains one of the least understood areas of octopus nervous system organization (Young, 1971).

Optic Lobes and the Visual System

The octopus’s highly developed visual system demonstrates how important sight is for every aspect of its life, including hunting, social communication, and adaptive coloration (Hanlon & Messenger, 2018). The camera-like eyes of the octopus are a classic example of convergent evolution with vertebrates. Octopus eyes have a cornea and a spherical lens that inverts the image of the world onto the retina. The light-absorbing photoreceptors that line the retina employ a single rhodopsin (Brown & Brown, 1958; Chung & Marshall, 2016). Processes from these photoreceptors form multiple optic nerves, which leave the orbital cavity and cross dorsoventrally before innervating the optic lobes, thereby reinverting the visual image onto the outer layers of the optic lobe.

The optic lobes are the largest lobes of the nervous system in every described cephalopod species, accounting for more neural mass than the central brain itself (Maddock & Young, 1987; Young, 1962). The outermost part of the optic lobes has a layered organization comprising outer and inner granule cell layers separated by a plexiform layer (shown for squid in Figure 4d). The majority of the optic nerve fibers terminates in the plexiform layer, making contact with second-order visual neurons. The granule cell layers contain amacrine cells of various sizes. The inner granular layer also features cells that give rise to efferent projections to the retina. Deep to these three layers, the radial layer has a columnar architecture reminiscent of the lamina cartridges and medulla columns of the insect visual system (Sanes & Zipursky, 2010). The outer layers of the optic lobes surround a medulla made up of neuronal processes with widely distributed “islands” of cells of no obvious histological distinction (Young, 1962). The medulla is involved in visual learning and memory as well as visually informed locomotor activities and dynamic body patterning (Liu & Chiao, 2017) (see the section “Decapodiforms”).

Vestibular Sensory and Oculomotor Control Systems

Many of the octopus’s complex behaviors, such as locating and tracking prey, depend on the integration of vestibular and visual information. The eyeballs of octopuses are under tightly regulated central control to preserve the horizontal orientation of the pupil with respect to gravity (Wells, 1960). In Octopus vulgaris, seven extraocular muscles are multiply innervated by seven oculomotor nerves that originate from the anterior part of the lateral pedal lobe of the subesophageal brain (Budelmann & Young, 1984; Young, 1971). The extraocular muscles allow the eyes to move forward, backward, upward, downward, clockwise, and anticlockwise. Aberration of this system has only been observed in senescent individuals and may reflect total neural degeneration at the end of life (Wang & Ragsdale, 2018).

Two neural pathways link each of the paired statocysts, which detect linear and angular accelerations, with the extraocular motor neurons. One path extends directly from the statocyst to the lateral pedal lobe. In the indirect pathway, neurons from the statocyst innervate higher integrative centers, including the anterior basal lobe and peduncle lobe, which then project to the lateral pedal lobe (Budelmann & Young, 1984). The basal and peduncle lobes use small cells and parallel fibers to combine equilibrium information from the statocysts with visual information arriving from the optic lobes, a cellular architecture evocative of the cerebellar cortex (Hobbs & Young, 1973; Young, 1976).

Neuroendocrine System

As in vertebrates, the neuroendocrine system of octopuses controls many important life history events and maintains homeostatic processes. The subpedunculate lobe of the supraesophageal brain targets the optic glands, so named because they sit atop the optic stalk between the optic lobes and central brain (Figure 2b and c). Rather than participate in visual processing, however, the optic glands secrete endocrine factors that act on the gonads and reproductive organs. These secretions are essential for sexual maturation, gametogenesis, and maternal care (Wells & Wells, 1959). Under physiological conditions, the female octopus ceases to feed while brooding her eggs and experiences a pronounced decay in body condition before dying. Removal of the optic glands drastically alters normal reproductive behaviors and semelparity. Glandectomized animals stop brooding, resume normal feeding, regain body weight, and live for an average of 5.75 months longer (Wodinsky, 1977). Secretions of the optic gland may target brain regions involved in the control of feeding and maternal care, but these putative hormones have not been identified. Together, the subpedunculate lobe, the optic glands, and the gonads serve an analogous function in the octopus to that of the hypothalamic-pituitary-gonadal axis in vertebrates (Minakata et al., 2009; Wang & Ragsdale, 2018; Wells & Wells, 1969).

Decapodiform Central Nervous System Anatomy

Much of the neuroanatomy for octopuses is shared among all coleoids (Hochner, 2012), but three major differences exist between decapodiform and octopod brains (Figure 4). First, lobes of the decapodiform brain are more widely distributed than those of the octopus. The superior buccal lobes, for example, are set at some distance from the central brain mass (Young, 1971). Second, axial nerve cords in decapodiforms are thin: arms and tentacles are dominated by muscle rather than neural tissue. Third, decapodiforms lack a well-differentiated inferior frontal system, and little is known about the molecular architecture of their frontal-vertical system. The vertical lobes comprise two parts fused along the midline, creating a single domain of neuropil (Young, 1979). In general, the decapodiform frontal-vertical lobes contain more neuropil and fewer cell bodies than those in octopuses, and show much less regional organization (Boyer, Maubert, Charnay, & Chichery, 2007).

Open in new tab

Figure 4. Squid nervous system anatomy. (a). Squid are principally pelagic decapodiforms. Their rigid mantle and swimming fins aid with fast locomotion. (b) and (b’). Stellate ganglion anatomy of Doryteuthis pealeii, courtesy of Paloma Gonzalez-Bellido and Trevor Wardill. See also Figure 5. Second-order fibers of
the squid escape system arrive at the stellate ganglion (SG) via the stellate connective (SC), a branch of the pallial nerve (PN). These second-order fibers form synapses with processes of third-order cells, which are located in the giant fiber lobe (GFL). Third-order fibers exit the stellate ganglion through one of the stellar nerves (asterisk). The giant axon, found in the last stellar nerve (SN), is the third-order cell with the greatest diameter. This dissection also shows the fin nerve (FN) and fin connective (FC), which control skin
coloration and are not part of the escape system per se. (c) and (c’). Dorsal view of the central brain and
optic lobes of Euprymna scolopes, specimen courtesy of Mark Mandel. The squid has vertical lobes (VL) that join at the midline and large optic lobes (OL). The statocysts (white arrowheads) remain attached to the cerebral cartilage. (d). Coronal view of the optic lobe of Euprymna scolopes, specimen courtesy of Bethany Rader. The outer layers of the squid optic lobe (1–3) follow octopus organization, but squid have an additional inner plexiform layer (4; Young, 1974). 1: outer granule cell layer; 2: outer plexiform layer; 3: inner granule cell layer; 4: inner plexiform layer. H&E-stained.

Anatomy of the Squid Escape System

Many squid are large pelagic animals that depend on jet propulsion. In the open water, they are susceptible to attack from predators. The giant fiber system provides a rapid mechanism for escape. The squid’s easily accessible nervous system and the unique anatomy of its giant axons enabled neurophysiologists to uncover the cellular and electrical mechanisms of resting membrane potentials and action potential propagation (Cole & Curtis, 1939; Hodgkin & Huxley, 1939). In many ways, the study of squid nervous systems engendered the emergence of modern cellular neurophysiology (Cole & Curtis, 1939; Curtis & Cole, 1938; 1940; 1942; Hodgkin & Huxley, 1939).

The squid escape system comprises first-, second-, and third-order giant fibers (Figure 5). A first-order giant neuron is located in each of the paired ventral magnocellular lobes. The axons of the two giant cells cross the midline and fuse together as the animal matures (Martin, 1969). This fusion ionically couples the two giant cells, ensuring that subsequent actions of the escape system are synchronized across the two sides of the body. The pair of first-order neurons forms connections with seven pairs of second-order neurons in the palliovisceral lobe (Young, 1939). Six pairs of these second-order neurons are motor neurons that directly innervate muscles of the head and funnel. The remaining pair of neurons forms the stellate connective (Miledi, 1972), which sends excitatory projections to the giant cell fibers of the stellate ganglia (Figures 4 and 5).

Open in new tab

Figure 5. Anatomy of the squid escape system. See also Figures 4b and b’. The escape system coordinates muscles of the fin and mantle on both sides of the body to facilitate fast jet-propulsion swimming (dotted line indicates the midline; for clarity, only one side of circuitry is detailed). First-order neurons arise in the ventral magnocellular lobe of the central brain (pink). The axons of these neurons cross and fuse at the midline. In the palliovisceral lobe, first-order neurons form connections with 7 second-order neurons (green). Six of these pairs innervate the retractor muscles of the head and funnel. The fibers of the remaining pair enter the stellate ganglion, where they innervate the processes of third-order cells (blue). The second- and third-order fibers form the giant synapse. The cell bodies of the 12–13 third-order neurons (only one schematized here) are located in the giant fiber lobe of the stellate ganglion. The fibers of each third-order cell leave the stellate ganglion and innervate the circular muscles of the mantle. Axons with greater diameters extend farther along the mantle.

The giant synapse between the second- and third-order neurons of the squid escape system is the largest yet described. An estimated 15,000 synaptic connections are found in each stellate ganglion between the second-order neuron and the 12–13 pairs of third-order neurons (Pumphrey & Young, 1938). These third-order cells are motor neurons that innervate circular mantle muscles. The fibers are graded in diameter and thereby in conduction velocity: thinner fibers terminate close to the stellate ganglion, and fibers with greater diameters reach the tip of the mantle (Pumphrey & Young, 1938). In this manner, a single pulse arriving via the giant fiber system results in a coordinated response from all muscle fiber targets (Otis & Gilly, 1990; Pumphrey & Young, 1938).

Functional Organization of the Cuttlefish Body Coloration System

Soft-bodied cephalopods have a unique ability to match the background by changing skin color and texture rapidly. These adaptive coloration displays include uniform, mottle, and disruptive camouflage patterns (Hanlon et al., 2009). The bulk of the quantitative research in body coloration has been conducted in cuttlefish, which use several dozen chromatic, textural, and postural components for adaptive coloration (Hanlon & Messenger, 1988). Many of these components are under independent neural control, equipping the cuttlefish with a modular palette from which to create an extensive repertoire of body patternings (Gonzalez-Bellido, Scaros, Hanlon, & Wardill, 2018).

The chromatophore organ is made up of a pigment-containing cell surrounded by radial muscle fibers (Dubas, Leonard, & Hanlon, 1986). Chromatophore patterns are initiated and controlled by the optic lobe: electrical stimulation of the optic lobe of Sepia officinalis produces whole-body patterns and chromatophore contractions (Chichery & Chanelet, 1976). When the optic tracts are bilaterally severed, the chromatophores contract and remain static (Young, 1971). The optic lobes send input to the lateral basal lobes, which then project to the chromatophore lobes of the subesophageal brain (Table 1). The anterior chromatophore lobes control chromatophore muscles in the head and arms, while the posterior chromatophore lobes are concerned with chromatophore muscles of the mantle and fins. In squid, separate neural circuits controlling the dilation of the chromatophores are separate from those regulating the iridophore structural coloration system (Gonzalez-Bellido et al., 2018; Gonzalez-Bellido, Wardill, Buresch, Ulmer, & Hanlon, 2014; Wardill, Gonzalez-Bellido, Crook, & Hanlon, 2012). The descending motor neurons that control iridophore activity are first routed through the stellate ganglion (Figure 4b’). It is currently unknown if the iridocytes themselves are activated by synapses (Gonzalez-Bellido et al., 2014) or indirectly through muscles (Wardill et al., 2012).

The “passing clouds” pattern is a particularly striking visual display with interesting implications for neural control. Passing clouds appear as a traveling burst of pigmentation. In cuttlefish, this dynamic skin pattern is often observed superimposed over a static color pattern (Laan, Gutnick, Kuba, & Laurent, 2014). This display is likely produced by wave-pattern-generating circuits with either weakly coupled, phase-locking oscillators or periodic circuit topologies. The identification of these circuits, either peripherally or in the chromatophore lobes, remains unresolved.