Forebrain Development: Prosomere Model 315 Forebrain Development: Prosomere Model L Puelles, University of Murcia, Murcia, Spain model in which all sorts of detailed data on form and ã 2009 Elsevier Ltd. All rights reserved. structure down to cellular aspects can be systematically accumulated, organized, compared, and differentiated one from another in their mutual relationships. The basic model of the vertebrate brain is the concept of Definition the closed neural tube, from which adult brains emerge via differential morpho- and histogenesis. Since we The prosomeric model is a segmental structural model cannot know all from the beginning, morphological of the brain of vertebrates that explicitly holds that models essentially are reasonable and useful conjec- the brain is formed by an uninterrupted series of trans- tures about how many parts there are and how they verse subunits of the neural tube, generally called are patched together. Such models are periodically neuromeres. Among such subunits is a large rostral perfected over time, becoming in the case of brains forebrain unit – the secondary prosencephalon – that increasingly complex operational scaffolds based on encompasses hypothalamus, eyes, and telencephalon, accumulated data and a number of assumptions. There followed by three caudal forebrain or diencephalic is always the possibility of constructing better (or neuromeres (i.e., prosomeres), which are regarded as worse) models. being serially homologous with more-caudal neuro- A good model in essence should be parsimonious; meres, namely, a single midbrain mesomere, 11 hind- that is, it should identify a minimal set of characteris- brain rhombomeres, and the spinal myelomeres. It is tic parts or landmarks in the modeled system, which important to note that for descriptive purposes, the can be generally recognized and seem to encompass, model postulates in all vertebrates a morphogenetic or be able to explain, most if not all available struc- bending of the longitudinal axis of the tubular neural tural data. A good model also may delimit various primordium, most marked at the cephalic flexure, ‘unfilled’ conceptual domains, where new data should whose incurvation causes wedge-shaped deformation fit in (as the periodic table of chemical elements did of the topologically transverse cylindrical neuromeric when it was first formulated). Such predictive aspects sectors of the neural tube. These units share a set of of models are highly useful because they implicitly fundamental longitudinal zones (due to common dor- indicate which new questions might be meaningful soventral (DV) patterning processes) and therefore or how best to pose and answer them in practice. represent segments, that is, metameric developmental Simultaneously, models are instrumental in providing units (anteroposterior (AP) patterning). The common possible significance to any new, unexpected observa- causal background of the longitudinal zones establishes tion. Scientists in principle believe in and use a par- the property of metamery (i.e., serial homology) across ticular model as long as it seems to accommodate all neuromeres, irrespective of their differential molec- established knowledge, inspire significant research, ular identities and individual prospective adult fates and allow satisfactory incorporation of emerging and of the variable border properties of the cells sets of new data. Historical periods in which techno- found at the interneuromeric boundaries. Therefore, logical improvements produce radically novel sorts the prosomeric model visualizes all vertebrate brains of data are particularly critical for the survival of as segmented structures constructed along the same a model. Bauplan (same set of DV and AP developmental units). Models widely shared among a scientific commu- Orthogonal intersection of DV and AP boundaries in nity represent a scientific paradigm. In contrast to the neural tube wall defines a checkerboard pattern hypotheses and theories, paradigms are not meant of domains (histogenetic areas) in which specific prop- to be tested, since one must believe in one of them erties and finer regionalization phenomena appear and use it as if it represented the truth, in the very (shared or not among vertebrates). This makes the process of testing a hypothesis experimentally. A par- model useful for systematic descriptive neuroembryo- adigm comes dangerously close to becoming a dogma, logy, comparative neuroanatomy, and causal analysis a belief that wholly escapes criticism or doubt and is of conserved or variant brain morphogenesis. considered ascientific. Several models may coexist historically, sometimes because each one is perceived Characteristics to have different advantages, but usually due to lack of awareness that one of them is distinctly better than Why We Use Models the others, compounded with the human tendency to Structural neurobiology studies the form and func- persist irrationally in long-held beliefs. Nevertheless, tional inner structure of brains. This needs a conceptual models and paradigms eventually may be perceived as 316 Forebrain Development: Prosomere Model obsolete and be discarded by newer, less committed writings by von Kupffer, Hill, His, Neal, Palmgren, generations of scientists, particularly when they are Rendahl, Tello, and Vaage, among others, contain less manifestly unable to account for some data and explicit antecedents. A large-scale review of shared efforts to apply them lead to highly unparsimonious, neuromeric structural data collected for all vertebrate complicated lines of thought. Continued use of an lineages from agnatha to mammals was published by obsolete model, or mixed-up joint use of elements of von Kupffer at the turn of the twentieth century. different models, tends to obstruct the progress of Wilhelm His produced an alternative very influential science. neural model, though he certainly must have known Some neuroscientists wrongly think that they do the neuromeric views of von Kupffer and other con- not use a neural model. This means they simply are temporaries well. His defined the floor, basal, alar, and unaware of the model they are using. Frequently, roof plates, the alar-basal boundary (sulcus limitans some aspects they mistakenly regard as facts actually of His), the concept of isthmus, and the idea of neural are conjectures. Dogmatic conscious or unconscious tube morphogenetic deformation due to axial bend- belief in models is a condition that is prone to poor ing. This model was very influential because it under- thinking and poor science. Due to the great complex- pinned the first Nomina Anatomica in 1895, whose ity of the studied organ, brain science is a field where committee was presided over by His. The model was such interpretive malfunctioning is not uncommon. not explicitly neuromeric, though His’ concepts of axial bending and longitudinal zonation and most of Neuromeric Models his transverse boundaries, including those of the isth- The earliest morphological models of the brain were mus, clearly were consistent with neuromeric models based on the adult form of the human and animal (Figure 2). brains. This approach provided over time a rich set of neuroanatomical terms and conjectural meanings, Columnar Models many of which are now obsolete, although some old At the height of the prestige of neuromeric brain terms and concepts still persist in textbooks. During models, an important unrelated breakthrough resulted the late nineteenth century, and as a result of various from the analysis of functional components in the important advances such as microscopy, evolutionary cranial and spinal nerves. It was discovered that each theory, and cell theory, comparative anatomical and nerve component (motor or sensory fibers) either embryological knowledge of brains advanced enough originates from or projects on a distinct columnar to allow the initial formulation of developmental domain of the hindbrain or spinal cord. Separate brain models generally valid for all vertebrates. Devel- columns could be assigned to visceral and somatic opmental models also appeared for the entire body. nerve components. Afferent fibers usually bifurcate The first generally accepted developmental structural into ascending and descending branches that distrib- paradigm for brains was a segmental model of the ute widely within the corresponding column. In so neural tube, which appeared hand in hand with a doing, they do not respect the neuromeric bound- segmental model of the body and head of vertebrates. aries. These data were widely perceived as important, The axial skeleton was conceived as being segmented and they threw doubt on the neuromeric models, at into metameric vertebrae (with a number of units least for application to advanced embryos and adults, fused together in the sacrum and in the cranial basis). since the basic functional organization of the hind- The branchial apparatus also seemed segmented into brain and spinal cord seemed to be columnar and not serial branchial arches and slits. The brain and the set segmental, irrespective of the separate, more or less of spinal and cranial nerves were postulated to consist periodic nerve roots, and the peripheral dermatomes of a number of segmental units correlated one to one and myotomes. It was increasingly thought that with the vertebrae and/or branchial arches. maybe neuromeres were transient early embryonic The term ‘neuromere’ that was soon applied to these phenomena without impact in the mature brain, in transverse neural