Feature Extraction by Optic Lobe Units of Crustaceans

The visual information processing characteristics of crustacean CEs has been shown to be similar in many ways to that performed by nonflying insects. Crustacean CEs have, in general, seven retinula cells per ommatidium (Waterman, 1961). Crustaceans (e.g., crabs, lobsters, crayfish) are unique in that their CEs are at the ends of stalks that can be moved by the animal to track moving objects over a limited range, or be moved to a protected position if the animal is threatened. Insect CEs are, of course, fixed to their heads, which can move. As has been seen, certain diptera, however, have internal muscles in their heads that can warp the anterior portions of their CEs medially to track objects binocularly (Qi, 1989a, b).

Researchers studying crustacean CE vision have generally recorded from the optic nerve, which, by definition, runs down the eyestalk from the optic ganglia underlying each eye to the anteror-dorsal-lateral portion of the animal's protocere-brum. To make stable recordings, the eyestalk must be immobilized. The anatomy of certain crustacean optic nerves was described by Nunnemacher (1966). He examined cross sections of the optic nerve, counting fiber numbers and sizes and the number of facets per eye for nine different genera of decapod, specifically, Homarus americanus, Orconectes virilis, Pagurus longicarpus, Upogebia affinus, Emerita talpoida, Cancer borealis, and Uca pugilator. For example, the lobster Homarus had an average of 12,000 facets/eye. Thus, 84,000 retinula cell fibers projected into the lamina ganglionaris. Each Homarus optic nerve contains fibers from (or to) the four optic ganglia. Nunnemacher made histograms of fiber number vs. diameter range; for example, Homarus had 65,000 fibers of 0.15 to 1.0 |m diameter, 3266 of 1 to 1.5 |im diameter, 1625 of 1.5 to 3 |im diameter, 555 of 3 to 6 |im diameter, 48 of 6 to 9 |im diameter, a total of 70,494 in one lobster optic nerve. In making recordings of optic nerve activity, it is obvious that the 603 axons in the 3 to 9 | m diameter range will have the larger recordable spikes.

The fiddler crab, Uca pugilator, examined by Nunnemacher was unusual in that it had giant axons in its optic nerve. The Uca CE had about 9100 facets. In its optic nerve, the fibers were distributed 18,384 of 0.15 to 1.0 |m diameter, 665 of 1 to 1.5 |m diameter, 521 of 1.5 to 3 |m diameter, 106 of 3 to 6 |m diameter, 18 of 6 to 9 | m diameter, 4 of 9 to 12 | m diameter, and 4 of 12 to 24 | m diameter, for a total of 19,701 fibers. Some afferent information must be very important to require the high spike conduction velocities belonging to such large fibers.

Insects have no nerve tract analogous to the crustacean optic nerve. Afferent signals from the insect OL leave it through a large number of separate tracts that project into the protocerebrum, subesophageal ganglion, etc. To find all the afferent tracts grouped together in the optic nerve is an advantage when working on CE vision in crustaceans. On the other hand, the chitinous exoskeleton protecting the base of the crustacean CE makes single-unit recording from OL units more difficult than with insects.

Glantz (1973) recorded from optic nerve units in the hermit crab, Pagurus pollicarus. Basic ON/OFF responses to general illumination and illumination of small numbers of facets with a 100-| m-diameter spot led to the basic classification of units by their responses to changes in illumination. It is not surprising that he found tonic on, tonic off, phasic on, phasic off, and phasic on/off units in the optic nerve. Glantz also tested qualitatively for motion and directional sensitivity using white or black disks moved against a contrasting background with various velocities. Of nine phasic on units tested for directional sensitivity, Glantz found only four that exhibited null direction responses. The others showed only slight to moderate differences to different directions of movement. One unit had up as a null direction, and the preferred direction was somewhere between down and forward. A phasic off unit was tested with a black spot object. It adapted rapidly to repeated motions. The response was more prolonged when the spot was given jerky motions. There was no directional sensitivity. (This behavior is reminiscent of the locust DMD units; see Section 5.4.1.) Phasic on/off units responded to moving objects with no directional sensitivity, and less habituation than phasic off units. No information was given by Glantz on object size, contrast, or velocities used. Apparently moving, contrasting edges were not tried.

A more-detailed study of crustacean CE vision was done by Waterman, et al. (1964) on the optic nerve of the crab Podophthalmus vigil. The crab and a schematic of its eye, optic ganglia, and optic nerve are shown in Figure 5.4-9. They found four major classes of interneuron in the optic nerve: (1) Afferent visual interneurons of ipsilateral origin. (2) Efferent visual interneurons of contralateral origin. (3) Efferent interneurons carrying mixed mechanoreceptor information including gravity vector signals from the ipsilateral statocyst. (4) "Afferent as well as efferent interneurons carrying mixed ipsilateral or contralateral visual and body mechanoreceptor information." The optic nerve also carried efferent motor signals to the distal eyestalk muscles, and afferent mechanoreceptor (proprioceptor) signal from the eyestalk distal joints.

Waterman et al. (1964) examined in detail the visual properties of 116 afferent visual units from group 1 above. These units had the following summary properties:

1. They had large receptive fields (30° to 180° or more) and are estimated to contain 300 to 104 facets. Thus, the outputs of many retinula cells (hence, dioptric units) are combined to determine optic nerve outputs.

2. There were rapidly adapting, novel movement units that behaved similar to locust and grasshopper DCMD units. These units exhibited area habit-uation similar to DCMD units, where repeated stimulation by object movement over one area of the eye quickly caused adaptation of the response; moving the object to a different area of the eye restored the response, which would then adapt. No directional sensitivity was noted for this class of unit. Waterman et al. did not test this unit for size preference or limiting resolution; this was a pity, because their narrative suggested that it did respond to high spatial frequencies: "Thus a unit which had reduced or ceased its response to movement of a particular target (such as a fine needle point) (italics added) in a localized part of its field would respond strongly again when the same stimulus was transferred to another, recently unstimulated area within its field."

Eyestalk Hermit Crab

FIGURE 5.4-9 (A) Anterior view of the crab Podophthalmus vigil. Podophthalmus has unusually long eyestalks supporting its two CEs. (B) Schematic section through the right distal eyestalk segment showing the ommatidia, optic ganglia, and optic nerve fibers that run through the eyestalk. Unlike insects, it is relatively easy to record from the efferent and afferent fibers running to the optic ganglia. (From Waterman, T.H. et al., J. Cell Comp. Physiol., 63(2): 135, 1964. With permission from The Wistar Institute, Philadelphia.)

FIGURE 5.4-9 (A) Anterior view of the crab Podophthalmus vigil. Podophthalmus has unusually long eyestalks supporting its two CEs. (B) Schematic section through the right distal eyestalk segment showing the ommatidia, optic ganglia, and optic nerve fibers that run through the eyestalk. Unlike insects, it is relatively easy to record from the efferent and afferent fibers running to the optic ganglia. (From Waterman, T.H. et al., J. Cell Comp. Physiol., 63(2): 135, 1964. With permission from The Wistar Institute, Philadelphia.)

3. Movement units that exhibited directional sensitivity (DS) were found. No systematic catalog of preferred directions was made. However, their narrative suggested that most preferred directions of objects were front to rear, or rear to front in the horizontal plane. Although a variety of test objects were listed under "methods," the impression is left that most DS units were tested with a vertical black strip or a 45° white spot object. Quantitative information was given in one figure about optimum object speed for two units. It was not clear whether these were DS units, however.

4. "Mixed modality (afferent) interneurons" were also found (see multimodal units in the Romalea OL). There units combined visual and mechanore-ceptor sensitivity to touch of the body or legs. Presumably, pure mecha-noreceptor efferent information is "mixed" in the optic ganglia with visual signals, then sent back to the brain.

5. ON/OFF units with different levels of response to ON and OFF, and ON-sustaining units were also found. Note that any such unit responding to changes of illumination in its receptive field will also fire in response to a moving, contrasting object in its RF having the correct size and polarity.

Finally, Waterman et al. examined some Podophthalmus ON responses for specificity to small, moving black spots (in this case, 5.7° diameter). Several units were found exhibiting size preference for this small spot and having little response to larger spots. It was not made clear what other unit response category these units fell into, however.

In a second, companion paper, Wiersma et al. (1964) examined the optic nerve responses in the crab Podophthalmus to visual stimulation of the contralateral eye. The ipsilateral eye was cut off distal to the recording site, eliminating any responses of ipsilateral origin. Most of the units found gave responses similar to the ipsitateral fibers studied by Waterman et al. (1964). In fact, the authors claimed that the interneurons recorded from normally carried two-way traffic; i.e., signals arising from a visual unit in the right eye was sent to the left eye, and vice versa, on the same fiber; a rather amazing duplexing of visual information. Wiersma et al. found units that appear to behave similarly to Romalea multimodal units; they called these "fast movement fibers." Their "slow movement fibers" apparently had size preference for small, dark objects. They had true directional sensitivity, but curiously, adapted quickly to repeated target movement over a designated part of the contralateral eye. Response was refreshed for the same movement over another part of the eye. The usual mix of sustaining ON fibers and phasic OFF units were also observed.

Much more information could have been forthcoming about feature extraction had these workers been more systematic and quantitative on their tests of object size, directional preference and speed preference. The two most amazing facts presented by these studies is the extent of interoptic lobe information transfer and the high amount of multimodal traffic observed. Interoptic lobe information transfer and multimodal responses have also been observed in insect OLs, and appears to be a design feature of CE visual systems. It is not known whether duplex traffic (i.e., L ^ R and R ^ L on the same, decussating interneuron fiber) occurs in insect visual systems.

Wiersma and Yamaguchi (1967) recorded from single optic nerve units in the crayfish Procambarus clarki. Several interesting response properties emerged from this study that had not been seen in the Podophthalmus work. First, no DS movement units were reported. Second, jittery-movement (JM) fibers were observed that responded to novel, quick, movements of a contrasting object anywhere over the ipsilateral eye. JM fibers adapted quickly, and moving the object over a new area of the RF restored the response. Again, the JM fiber of the crayfish optic nerve appears to be like the grasshopper DMD units. An important observation made by Wiersma and Yamaguchi was that voluntary or forced movements of the eyestalk inhibited the JM unit responses, and if it was adapted, canceled the adaptation.

The other class of ON units of major interest in the crayfish were the space constant (SC) fibers. There are four SC units in a crayfish's optic nerve: Two behave like tonic ON units (sustaining fibers), one like a JM unit, and the fourth responding to rapidly approaching objects (a looming operator?). Space constancy is a property mostly derived from the animal's statocysts (gravity vector sensors). The neural outputs from the statocysts, and perhaps signals from other proprioceptors monitoring limb loading, are sent to the CNS, thence to the OLs. In the OLs, this information acts to maintain the receptive fields of the SC units so that they always maintain the same relative position with respect to the Earth's gravity vector regardless of the animal's roll and pitch. For example, the RF of an SC unit is sector of a hemisphere between 10 and 2 o'clock (Figure 5.4-10) when the animal is resting normally at zero roll and pitch. When the animal is rolled +90° (right-side down), the SC unit RF now occupies a new set of ommatidia so that it still is between 10 and 2 o'clock. The new RF is actually between 1 and 5 o'clock on the eye. Some very interesting neural switching or gating must take place for this to happen. Space constancy is still a neurophysiological enigma.

Dorsal Dorsal

FIGURE 5.4-10 Diagram illustrating the curious property of space constancy. See text for discussion.

Ventral Ventral

Front view of eye

FIGURE 5.4-10 Diagram illustrating the curious property of space constancy. See text for discussion.

Was this article helpful?

0 0
Peripheral Neuropathy Natural Treatment Options

Peripheral Neuropathy Natural Treatment Options

This guide will help millions of people understand this condition so that they can take control of their lives and make informed decisions. The ebook covers information on a vast number of different types of neuropathy. In addition, it will be a useful resource for their families, caregivers, and health care providers.

Get My Free Ebook


Post a comment