Introduction

The zebrafish (Danio rerio; Brachydanio rerio in older literature) has become a powerful model system to study genetic mechanisms of vertebrate development and disease. Much of the current success can be traced back to the pioneering work of George Streisinger and colleagues at the University of Oregon. Like many of his peers, Streisinger had an acclaimed research programme on phage genetics but sought a eukaryotic system to expand further the known roles of genes in biological processes. Whereas Seymour Benzer focused his efforts on Drosophila and Sydney Brenner (Brenner, 1974) adopted the nematode worm, Streisinger, a fish hobbiest, turned his efforts towards the zebrafish (Streisinger et al., 1981; Chakrabarti et al., 1983; Walker and Streisinger, 1983; Grunwald and Streisinger, 1992). Streisinger first recognized many of the oft-cited advantages for the use of zebrafish as a genetic model (Mullins and Nusslein-Volhard, 1993; Driever et al., 1994; Solnica-Krezel et al., 1994). Zebrafish, small freshwater teleosts, are easily adapted to the laboratory setting and can be maintained in a relatively small space. The fish typically reach sexual maturity in 3 to 4 months, and a breeding pair of fish can produce >200 fertilized eggs per mating. Fertilization is external, and the egg and embryo are transparent, facilitating visual identification of morphogenetic movements and organogenesis with a standard dissecting microscope. Development is rapid; by 24 hours post-fertilization (hpf) all of the major organ systems have formed and spontaneous muscle flexures soon begin.

Introduction

The study of regeneration in the vertebrate began with the pioneering experiments of Claude Bonnet in 1781. He found that if part of the eye of an adult newt (Triturus cristatus) was removed, a smaller, but complete, eye was regenerated within a few months. All of the ocular tissues, including the cornea, lens and retina, were capable of regenerating. Subsequent work by biologists, working primarily in the 1800s and early 1900s, characterized many critical features of the regeneration process in the eye. The molecular basis for this remarkable process is still not understood. However, recent progress in eye development research has spurred new lines of investigation into this question. In this review, we briefly discuss highlights of historical work and then focus on recent experiments in a variety of species that illustrates the complexities of the questions being investigated today.

A brief history of retinal regeneration

One of the first questions that arose historically concerning retinal regeneration in newts was the nature of the cells that provided the regenerated tissue. Early studies argued that a ring of cells at the peripheral retinal margin, what is now most commonly called the ciliary margin zone (CMZ), was the primary source of regenerated retina (Colucci, 1891 (cited in Keefe, 1973d); Fujita, 1913). Later studies confirmed the CMZ as a source of regeneration, but also demonstrated that the retinal pigmented epithelium (RPE) could regenerate neural retina in the posterior eye (Wachs, 1914, 1920).

Introduction

In addition to intrinsic control mechanisms (see Chapter 5 and Cepko et al., 1996), the production of neurons by progenitor cells and the determination of their fate are regulated via an array of diffusible factors, two families of which are considered in this chapter: neurotransmitters and neurotrophins. Neurotrophins are now known to play an essential role in both the formation and the maintenance of the nervous system throughout development and adult life. There is growing evidence that besides their role as molecules mediating communication between nerve cells in the mature nervous system, a variety of both slow and fast neurotransmitters also play important roles during neuronal development. This chapter reviews recent evidence that demonstrates that a number of non-synaptic neurotransmitter release mechanisms, together with many neurotransmitters and their receptors, are present in the developing retina prior to the onset of synapse formation and that these early neurotransmitters act to modulate a range of events in neural development. Their precise mechanisms of action are still being elucidated but, as described here, the ability to modulate [Ca2+]i is one feature common to these early neurotransmitter systems, and is thought to underlie a number of their developmental actions. It is becoming clear that both neurotransmitters and neurotrophins play important regulatory roles in the early stages of retinal development, including the modulation of proliferation, differentiation, cell survival and circuit formation.

Introduction

Like most parts of the CNS, retinal cells are generated some distance from where they will ultimately reside. Migration to the correct place at the right time is vital for their ability to make appropriate synaptic connections and function normally. Understanding how each of the seven retinal cell types migrate to their appropriate layer is critical to understanding how this CNS structure becomes organized during development.

The entire retinal neuroepithelium is a proliferative zone early in development. Retinal neuroepithelial cells with cytoplasmic processes that extend from the outer limiting membrane (OLM) to the inner limiting membrane (ILM) engage in interkinetic nuclear migration, a process by which their nuclei migrate within the cytoplasm, undergoing different phases of the cell cycle at different depths within the neuroepithelium (see Chapter 3). Thus, neuroepithelial cells in S-phase have their nuclei positioned near the ILM, and they enter M-phase at the OLM. Consequently, following a final mitotic divison, when cells leave the cell cycle they do so adjacent to the OLM. Newborn postmitotic cells therefore need to migrate varying distances to take up residence in one of the three prospective cellular layers. Cells destined for the ganglion cell layer (GCL), for example, have comparatively longer distances to travel than rod and cone photoreceptors. Birthdating studies in diverse species have shown that the first cohort of cells to become postmitotic are ganglion cells (Prada et al., 1991; Rapaport et al., 1996, 2004; Hu and Easter, 1999) (see Chapter 3).

Introduction

The sheet of retinal neuroepithelial cells resulting from the specification of the eye field is transformed into a layered array of differentiated cells by the simultaneous processes of cell division, apoptosis, differentiation and migration. The production of the six major cell types, with their multiple subtypes, in the correct numbers and at the appropriate time is essential for normal development. The retina has been studied extensively as a model for cell determination in the vertebrate nervous system for a number of reasons. It is easily accessible to genetic and embryological manipulations in vivo because of its position and large size and can also be studied in vitro because cells in retinal explant cultures faithfully follow in vivo differentiation programmes. Numerous genes involved in cell determination do not affect other processes and their disruption does not cause early lethality. The different major cell types can be readily distinguished by their laminar position, their distinct morphologies and by cell-specific markers. The persistence of a proliferating ciliary marginal zone in amphibians, fish and avian species provides a model that recapitulates embryonic proliferation and differentiation and facilitates the examination of gene expression and function (Perron et al., 1998).

Retinal progenitors are multipotent and vary greatly with respect to their clonal compositions, both in terms of the cell types produced and the number of progeny.

Introduction

Although the newborn retina is highly active, with spontaneous waves propagating across the amacrine and the ganglion cell layers every few minutes (see Chapter 13), at that time it is not yet possible to elicit light responses in retinal ganglion cells (RGCs). This lack of responsiveness to light is due to the immaturity of the vertical synaptic pathway between photoreceptors and RGCs provided by bipolar cells (BCs), despite the fact that lateral connections in the inner retina are already well established (see Chapter 13). Moreover, rod and cone opsins are not yet functional at birth. In mouse for example, ultraviolet cone opsin does not appear until postnatal day (P)1, rod opsin until P5 and green cone opsin until P7 (Tarttelin et al., 2003). Hence, RGCs become visually responsive only shortly before eye opening (around P10 in rabbit; Masland, 1977; Dacheux and Miller, 1981a, b; P7 to P10 in cat; Tootle, 1993; P12 in mouse; Sekaran et al., 2005). Humans and other primates, on the other hand, are born with their eyes open and although primate vision is poor at birth a newborn human infant is capable of tracking visual stimuli (Teller, 1997).

This chapter reviews the earliest light responses that can be detected in the developing retina. New studies show that the newborn retina is actually not insensitive to light and this will be considered in the first part of the chapter.

Introduction

Interest in programmed cell death (PCD) emerged over a century ago (reviewed in Clarke and Clarke, 1996), and such naturally occurring cell death in the developing nervous system has been extensively documented (Oppenheim, 1991 for review). More recently, the concept of PCD has been the subject of some controversy mainly due to the overwhelming interest in one of its forms, apoptosis (Sloviter, 2002). For the purpose of this chapter, PCD is defined simply as a sequence of events based on cellular metabolism that leads to cell destruction (Lockshin and Zakeri, 2001; Guimarães and Linden, 2004), without commitment to particular morphological types.

Programmed cell death has been identified using a variety of techniques, though each of them is prone to errors when estimating the magnitude of cell loss. Estimating the size of the population based on counts of axons in developing nerves or tracts may be confounded by the simultaneous occurrence of both cell death and axonal ingrowth, and by the transient contaminating presence of other axonal populations. Estimates based on cell counts may be influenced by the continuous migration of differentiating cells into spatially delimited cell populations, as well as by the inclusion of other types of cells that are not so readily discriminable at earlier developmental stages. And while great progress has been made in understanding the molecular mechanisms of apoptosis in the last decade, multiple alternative pathways of PCD add a further degree of complexity in understanding developmental cell death and estimating its magnitude.

Introduction

In the past half-century the field of biology has witnessed a burgeoning of understanding of the biochemistry, molecular and cell biology of cell signalling. More recently, a significant effort was made to focus the techniques and concepts of biology to a mechanistic understanding of the nervous system. Within the area of development, perhaps the cardinal question has been how to signal immature cells to form the diverse organs, tissues and differentiated cells of the body – a particularly challenging question in the nervous system given the great diversity of cell types to be made. Because of its combination of diverse cell types within a highly structured tissue the vertebrate retina has served as an important model tissue in pursuit of answers to such questions. Specifically, the retina displays a laminar cytoarchitecture, and seven cell types that are largely confined to one of three laminae. These include receptors (rod and cone photoreceptors), short and long projection neurons (bipolar and retinal ganglion cells, respectively), local circuit neurons (horizontal and amacrine cells) and glia (Müller cells). The constancy of retinal structure and cell types across vertebrates allows cross-species comparisons to be readily made. Further, almost all retinal cell types exhibit multiple levels of differentiation. For example, there are several subtypes of ganglion cells or amacrine cells based on morphology, transmitter content, synaptic connectivity, etc. Thus, explanation of determination and differentiation can be sought at multiple levels of specificity.

Introduction

The macula lutea (‘yellow spot’), located towards the posterior pole of the human retina was identified grossly in the late eighteenth century, and the fovea centralis – located approximately at the centre of the macula – was first described by Soemmerring in 1795. It was H. Müller who produced the first histological description of the human fovea (see Polyak, 1941), along with identification of the retinal layers and a correct analysis of their general place in the retinal circuitry. Knowledge of its anatomical organization was greatly expanded by the Golgi impregnation work of Cajal (1893) and Polyak (1941), which indicated that foveal cones give rise to highly specialized circuits.

Some of the earliest studies of developing primate retina were carried out by Chievitz (1888), Magitot (1910) and Bach and Seefelder (1911, 1912, 1914), from whose work some illustrations are reproduced later in this chapter. These early studies were greatly expanded on by Ida Mann in the first part of the twentieth century and published in monograph form in 1928 (see Mann, 1964, 2nd edition). Little further work was published on foveal development until Hendrickson and Kupfer's study (1976) showing photoreceptor displacement towards the developing fovea. This most recent period of analysis of primate (including human) retinal development (1976 to present) has taken place in the context of an expanding body of knowledge, gleaned largely from investigation of retinal development in non-primate species – much of which is covered in other chapters of this volume – including gene regulation of eye and retinal formation, retinal cell generation, development of target visual nuclei, guidance mechanisms, the importance of neuronal acitivity and the significance of apoptosis.

The “lack of invariance problem” and multisensory speech perception

In speech there is a many-to-many mapping between acoustic patterns and phonetic categories. That is, similar acoustic properties can be assigned to different phonetic categories or quite distinct acoustic properties can be assigned to the same linguistic category. Attempting to solve this “lack of invariance problem” has framed much of the theoretical debate in speech research over the years. Indeed, most theories may be characterized as to how they deal with this “problem.” Nonetheless, there is little evidence for even a single invariant acoustic property that uniquely identifies phonetic features and that is used by listeners (though see Blumstein and Stevens, 1981; Stevens and Blumstein, 1981).

Phonetic constancy can be achieved in spite of this lack of invariance by viewing speech perception as an active process (Nusbaum and Magnuson, 1997). Active processing models like the one to be described here derive from Helmholtz who described visual perception as a process of “unconscious inference” (see Hatfield, 2002). That is, visual perception is the result of forming and testing hypotheses about the inherently ambiguous information available to the retina.

Introduction

Rizzolatti and Arbib (1998) argue in their exposition of the Mirror System Hypothesis that brain mechanisms underlying human language abilities evolved from our non-human primate ancestors' ability to link self-generated actions and similar actions of others (see Arbib, Chapter 1, this volume). On this view, communicative gestures emerged eventually from a shared understanding that actions one makes oneself are indeed like those made by conspecifics. Thus, what the self knows can be enriched by an understanding of the actions and aims of others, and vice versa. From this perspective, the origins of language reside in behaviors not originally related to communication. That is, this common understanding of action sequences may provide a “missing link” to language.

In answering the question “What are the sources from outside the self that inform what the child knows?”, the basic idea is that negotiating a shared understanding of action grounds what individuals know in common, including foregrounding the body's part in detecting that the actions of the self are “like the other.” Given this footing, what then might the evolutionary path to language and the ontogeny of language in the child have in common? This perspective roots the source of the emergence of language in both as arising from perceiving and acting, leading to gesture, and eventually to speech.

I report here on an ongoing research program designed to investigate how perceiving and acting inform achieving a consensus or common understanding of ongoing events hypothesized to underlie communicating with language.

Introduction

One topic that is strikingly pervasive across the cognitive sciences is that of Theory of Mind, referring to the abilities that people have in reasoning about their own mental states and those of others. It is the set of Theory of Mind abilities that enable people to reflect introspectively on their own reasoning, to empathize with other people by imagining what it would be like to be in their position, and to generate reasonable expectations and inferences about mental states and processes.

Although there are inherent difficulties involved in investigating behavior that is largely unobservable, a relatively sophisticated understanding of Theory of Mind abilities has emerged through the synthesis of widely disparate sources of evidence. This evidence suggests that Theory of Mind abilities progressively develop in children and adults (Happé et al., 1998; Wellman and Lagattuata, 2000), are degraded in people diagnosed with the illness of autism (Baron-Cohen, 2000), have a relationship to localized brain regions (Happé et al., 1999; Frith and Frith, 2000), and are a uniquely human cognitive faculty not available to other primates, e.g. chimpanzees and orangutans (Call and Tomasello, 1999). This last contribution to our understanding of Theory of Mind suggests that these abilities must have arisen in the human lineage only after a split from that of chimpanzees some 6–8 million years ago (but see Stanford, this volume, for a review of dissenting opinions).

Introduction: a mirror system perspective on grasp development

Neonates and young infants are innately compelled to move their arms, the range of possible spontaneous movements being biologically constrained by anatomy, environmental forces, and social opportunity. Over the first 9 postnatal months, reaching movements are transformed as infants establish an array of goal-directed behaviors, master basic sensorimotor skills to act on those goals, and acquire sufficient knowledge of interesting objects to preplan goal-directed grasping. In monkeys, it appears that the neural circuit for control of grasping also functions to understand the manual actions of other primates and humans (Arbib, Chapter 1, this volume). Within the grasp circuitry, “mirror neuron” activity encodes both the manual actions executed by the monkey and the observed goal-directed actions of others. Recent imaging studies on humans indicate that a mirror neuron network may exist in humans linking observation and execution functions. However, the link between grasp development and mirror system development is widely unexplored. To address this, we will build models based both on behavioral data concerning the course of development of reaching in human infants and on neurophysiological data concerning mirror neurons and related circuitry in macaque monkeys.

In humans, the foundation for reaching may begin as early as 10–15 weeks of fetal development when fetuses make hand contact with the face and exhibit preferential sucking of the right thumb (de Vries et al., 1982; Hepper et al., 1991).

Introduction

The Mirror System Hypothesis (MSH), described in Chapter 1, asserts that recognition of manual actions may ground the evolution of the language-ready brain. More specifically, the hypothesis suggests that manual praxic actions provide the basis for the successive evolution of pantomime, then protosign and protospeech, and finally the articulatory actions (of hands, face and – most importantly for speech – voice) that define the phonology of language. But whereas a praxic action just is a praxic action, a communicative action (which is usually a compound of meaningless articulatory actions; see Goldstein, Byrd, and Saltzman, this volume, on duality of patterning) is about something else. We want to give an account of that relationship between the sign and the signified (Arbib, this volume, Section 1.4.3).

Words and sentences can be about many things and abstractions, or can have social import within a variety of speech acts. However, here we choose to focus our discussion by looking at two specific tasks of language in relation to a visually perceptible scene: (1) generating a description of the scene, and (2) answering a question about the scene. At one level, vision appears to be highly parallel, whereas producing or understanding a sentence appears to be essentially serial. However, in each case there is both low-level parallel processing (across the spatial dimension in vision, across the frequency spectrum in audition) and high-level seriality in time (a sequence of visual fixations or foci of attention in vision, a sequence of words in language).

Introduction

In this chapter I explore two qualities of the mirror neuron system that are critical for the evolution of tool use and language, central characteristics of human culture. The two characteristics of the mirror system are: (1) the ability of the system to respond both to one's own act and to the same act performed by another and (2) the system's selective response to intentional or goal-directed action (Fogassi et al., 2005). The ability to respond neurally both to one's own act and to the same act performed by another constitutes the neural foundation of imitation on the behavioral level (Iacoboni et al., 1999) and of repetition on the linguistic and cognitive levels (Ochs (Keenan), 1977). The selective response of the mirror neuron system to goal-directed action constitutes the neural facilitation of goal-directed action on the behavioral level and of intentionality on the cognitive level (Greenfield, 1980). My purpose is then to demonstrate the importance of these neurally grounded behavioral competencies for the evolution and ontogenetic development of two key aspects of human culture, tool use and language. In so doing, my larger goal is to contribute to understanding the neural underpinnings for the ontogeny and phylogeny of human culture.

In order to provide data on phylogeny, I draw upon my own research and that of others to compare chimpanzees (Pan troglodytes), bonobos (Pan paniscus), and humans (Homo sapiens). The Pan line and the hominid line diverged in evolutionary history approximately 5 million years ago (Stauffer et al., 2001).