Update TRENDS in Cognitive Sciences Vol.8 No.

7 July 2004

| Research Focus

Monkey homologues of language areas:

computing the ambiguities
Terrence Deacon
Department of Anthropology and Helen Wills Neuroscience Institute, University of California, Berkeley, USA

The ‘language-readiness’ of human brains most prob- to be concordant with one another in brain structure
ably resulted from modification of structures present in comparisons.
non-human primate brains, but identifying such homo- The problem of non-local influences on tissue differen-
logues and the nature of their modifications has been tiation is most vividly exemplified with respect to axonal
highly problematic. In a recent article, Arbiband Bota connectivity. The growth of axons from one brain region
suggest that these problems can be overcome using a to another non-contiguous region creates a context for
neuroinformatics approach. But its assumptions ignore discrete cell – cell interactions at a distance, including self-
many non-local, activity-dependent, regressive, and organizing, regressive (apoptosis and axonal elimination)
allometric effects of neurodevelopment that violate and activity-dependent modifications of cytoarchitecture
assumptions of classic homology. What if these effects and function [4 –6]. Even phylogenetically unprecedented
are what matter most? connection topographies can emerge de novo as a result
of afferent changes that alter competitive interactions
Explaining the neurological basis for language ultimately between axonal projections [7]. These are not mere
hinges on determining how human brains differ from non- laboratory curiosities. Corresponding patterns of natur-
human primate brains. In their article reviewing the ally evolved functional-connectional shifts have been
debates over the location, function and primate homo- identified in species with specializations that are analo-
logues of Broca’s language area, Michael Arbiband Mihail gous to some of these manipulations [8] and differential
Bota [1] argue that analyzing homological correspon- expression of regressive modifications also might correlate
dences between human and non-human primate brains, with size [2]. Indeed, many non-local interaction effects are
using a neuroinformatics database, can resolve many long- sensitive to size (i.e. allometry; see below). Despite decades
standing questions about the ‘language-readiness’ of of intensive investigation of the importance of these
human brains. The concept of anatomical correspondence secondary developmental modifications, however, only a
via common inherited information is not problematic in few researchers have seriously considered their role in
itself, but establishing homologies a posteriori in living brain evolution and their consequences for homology
species, on the basis of anatomical similarities, poses [9– 12]. Considering that the major motivation for adopt-
several difficulties, especially where nervous tissues are ing this informatics approach is to aid analysis of a
involved [2 – 3]. Many confounding factors can contribute quantitatively deviant brain exhibiting a non sequitur
to superficial similarities or obfuscate structural resem- functional specialization (language), departures from the
blances between brains, irrespective of phylogeny, limiting standard primate brain ‘bauplan’ is the kind of non-
the precision of homological assessment. This highlights homology we might expect to discover. Although not
the need for a computational approach able to compensate explicitly encoded in the NeuroHomology DataBase
for diversity of criteria, methodology and nomenclature by (NHDB), non-local developmental effects should appear
exploiting intrinsic redundancies implicit in the data. as non-concordant relationships between hodology (con-
nection patterns) and location.
The homology puzzle
Although homology is often discussed as a comparison of Avoiding the ‘fallacy of simple location’
adult-to-adult body structures, the anatomical similarities The classic descriptions of language functions attributed
and topologies that constitute homologous relationships to Broca’s and Wernicke’s areas, respectively, have long
emerge from patterns of gene expression and cell inter- been subjects of debate, fueled by uncertainties of
actions during development. In most regions of the body, physiology and localization. Homological comparisons to
structural differentiation is largely a reflection of spatially monkey cortex offer a way to deconstruct function and
proximate cell – cell interactions, but in the brain, cellular structure of these areas, taking advantage of greater
migration, axonal growth, and activity-dependent cull- laboratory access to physiological information in monkeys.
ing of neurons and connections superimpose many non- Arbiband Bota focus on two theories of language function
local influences on this basic logic. So many standard for which the homological mapping of Broca’s area to
monkey homologues plays a central role. Their favored
criteria for homology assessment cannot be relied upon
theory – based on the discovery of ‘mirror’ neurons in the
Corresponding author: Terrence Deacon ([email protected]). macaque ventral premotor cortex (area F5) – proposes
Available online 7 June 2004 that this specific region and the functions ascribed to these
www.sciencedirect.com
 Update TRENDS in Cognitive Sciences Vol.8 No.7 July 2004 289

neurons constitutes the homologue to Broca’s area (which in a premotor region (area F5) that is considered to be the
they locate to Brodmann’s area 44) [13]. A competing theory, cytoarchitectonic homologue to BA44 in the human brain;
proposed by Aboitiz and Garcia, argues instead that the and (4) many researchers have suggested that BA44 is the
functions attributed to Broca’s area are more distributed anatomical substrate of the functions associated with
between multiple frontal motor and prefrontal areas [14]. Broca’s aphasia. On the basis of these inferential steps
Contrasting these theories on anatomical homology criteria, Rizzolatti and Arbib [13] hypothesized that mirror neurons
to the exclusion of other forms of data (e.g. aphasiology and probably provided an important ‘pre-adaptive’ substrate
imaging), inadvertently shows how important these other that was eventually recruited for language function in
kinds of data are to assessing homology. human evolution. But mirror neurons do not provide
The identification of a region of monkey brain that is macaques and other non-human primates with language
homologous to Broca’s area is limited by any ambiguities of abilities, and (as noted above) it is not at all clear that lesions
its structural and functional delineation in humans. The destroying BA44, the putative mirror-neuron-containing
determination of the substrate for Broca’s aphasia and the region, in human patients results in persistent language
corresponding language function has been controversial impairment. Mirror-neuron functions might have been
throughout the 20th century [14]. Despite common belief, recruited to aid language-readiness, but we should not
considerable neuropsychological evidence suggests that necessarily expect them to be the locus of a distinctive
damage confined to BA44 is not the sole and probably not language adaptation. To explain that we must look for
the primary cause of the aphasic syndrome that bears something different.
Broca’s name [15–16]. Many adjacent regions are currently
thought to be involved in the symptomatology of Broca’s Scaled-up homologies
aphasia [17], including cortical areas 45, 47, the motor face One final issue of homology should have been more
strip, and the anterior insula, as well as deep white-matter cautiously explored and more explicitly implemented in
tracts and parts of the basal ganglia. Even the preserved the NHDB. Because of the disproportionate size of the
brain of Broca’s first celebrated patient was recently human brain, any investigation of its evolution must
reanalyzed by CAT scan and MRI, and was found to include address questions of allometry (i.e. differential scaling of
damage to all these structures and more [18]. This structures or functions correlated with differences in size).
heterogeneity of frontal language processing is consistent To ignore this misses the most widely investigated feature
with tracer studies in non-human primate brains that show of human brains, and one that is crucial to issues of
auditory projections terminating anterior to premotor cortex establishing homology. Allometric analyses identify the
in prefrontal areas, and parietal and multimodal projections fraction of quantitative change in a given structure that is
terminating in posterior premotor areas [19]. It is also ‘expected’ on the basis of overall size increase and what
consistent with electrophysiological studies demonstrating fraction deviates from this expectation. This can involve
that ventral prefrontal areas contain auditory responsive cytoarchitectonic features as well as gross morphology.
cells [20] whereas ventral premotor cortex contains the Affected cytoarchitectonic features include relative
visuo-manually responsive ‘mirror’ cells [13]. Electrical neuronal size, cell morphology, neural/glial ratio, lamin-
stimulation, aphasia studies, and in vivo imagery of ation, myelination, ratio of gray matter to white matter,
language processing tasks in humans have also consistently and ratio of koniocortex (‘primary’) to eulaminate cortex
shown that left ventral prefrontal areas are linked with (‘association’). All these vary systematically with respect
mnemonic and auditory components of language tasks to absolute brain size, and largely irrespective of common
and that ventral premotor, motor, and insular areas of ancestry, and can also be highly divergent in closely
cortex are linked with motoric components of language tasks related species that differ significantly in size [2,23 – 24].
[16–17,21–22]. These considerations suggest that a NHDB The comparatively large gap in human versus monkey
that included these multiple sources of data would be as brain size inevitably involves architectonic deviations
likely to deconstruct the very conception of Broca’s area, as to that are crucial to establishing homologies, particularly
delineate a simple structure-by-structure homology in the increased differentiation and subdivision of cortical areas.
macaque brain. Considering the importance of this factor it is curious that
The whole point of these homology assessments is to Arbiband Bota reiterate a simple but misleading compari-
distinguish between language recruitments (antecedent son between a large heterogeneous region and one of its
systems recruited for this unprecedented function) and subregions in non-human and human brains.
language adaptations (deviations from antecedent forms To support their focus on human BA44 and macaque area
reflecting adaptation). The importance of maintaining this F5, they cite an MRI-based study showing that ‘frontal’
distinction is well exemplified by the discussion of the ‘mirror cortex in humans is not allometrically deviant in size from
system’ hypothesis for language. Evidence from macaque that in other apes [25]. They agree with its authors that these
brains suggested a homology linkage to frontal language data contradict previous studies showing that human
functions via the following chain of inferences: (1) mirror prefrontal cortex is disproportionately large (reviewed in
neurons exhibit what might be called multi-perspectival [12]). However, this interpretation involves a clear violation
responsivity to multi-modal (primarily visuo–manual) of one of the criteria that the NHDB is intended to control
stimuli; (2) both imitation of speech and reciprocally under- for: comparisons that equate hierarchically nested homol-
stood referential communication involve perspective shift- ogues. Differences in methodology contributed to this
ing, which is superficially analogous to a ‘mirror’ response; error. MRI data have insufficient resolution for cytoarchi-
(3) these cells are located within the macaque frontal cortex tectonic comparison, but cytoarchitecture is necessary for
www.sciencedirect.com
290 Update TRENDS in Cognitive Sciences Vol.8 No.7 July 2004

determining the boundaries of prefrontal cortex. The MRI blind mole rat Spalax ehrenbergi: A WGA-HRP tracing study.
analysis relied on a gross morphological feature – the central Neuroreport 5, 2697– 2701
9 Ebbesson, S. (1984) Evolution and ontogeny of neural circuits. Behav.
sulcus – to distinguish ‘frontal’ from non-frontal cortex,
Brain Sci. 7, 321 – 366
lumping together prefrontal, orbital, premotor, motor, 10 Wilczynski, W. (1984) Central neural systems subserving a homo-
cingulate and anterior insular cortex (plus some other plasous periphery. Am. Zool. 24, 755– 763
structures). But ‘frontal’ cortex delimited this way is 11 Finlay, B. et al. (1987) Regressive events in brain development and
hierarchically much more inclusive than prefrontal cortex scenarios for vertebrate evolution. Brain Behav. Evol. 30, 102 – 117
12 Deacon, T. (1997). The Symbolic Species: The Coevolution of Language
alone, and the allometric predictability of the whole sector
and The Brain, W.W. Norton
cannot be assumed to be inherited by this one part. Although 13 Rizzolatti, G. and Arbib, M.A. (1998) Language within our grasp.
Arbiband Bota missed catching this dishomology, or the Trends Neurosci. 21, 188 – 194
misunderstanding it suggests, the NHDB should be specifi- 14 Aboitiz, F. and Garcia, V. (1997) The evolutionary origin of the
cally designed to avoid such errors, even if glossed over in the language areas in the human brain: a neuroanatomical perspective.
original research reports. Brain Res. Brain Res. Rev. 25, 381 – 396
15 Lecours, A-R. et al. (1983) Aphasiology, Baillière Tindall
Regimenting the comparative use of the many types of
16 Mohr, J. et al. (1978) Broca aphasia: pathologic and clinical. Neurology
homology criteria is an essential step in explaining the 28, 311 – 324
basis for our language-readiness. Arbiband Bota could do 17 Dronkers, N. (2000) The pursuit of brain-language relationships.
the field a great service by providing a more sophisticated Brain Lang. 71, 59 – 61
informatics approach to homology criteria, but it will not 18 Signoret, J-L. et al. (1984) Rediscovery of Leborgne’s brain: anatomical
description with CT scan. Brain Lang. 22, 303 – 319
merely be a matter of data collection and tabulation. Much
19 Deacon, T. (1992) Cortical connections of the inferior arcuate sulcus
work remains to be done before neural homologies are cortex in the macaque brain. Brain Res. 573, 8 – 26
truly unambiguous. 20 Romanski, L. et al. (1999) Dual streams of auditory afferents target
multiple domains in the primate prefrontal cortex. Nat. Neurosci. 2,
References 1131– 1136
1 Arbib, M. and Bota, M. (2003) Language evolution, neural homologies 21 Paulesu, E. et al. (1997) Functional heterogeneity of left inferior
and neuroinformatics. Neural Netw. 16, 1237 – 1260 frontal cortex as revealed by fMRI. Neuroreport 8, 2011 – 2016
2 Deacon, T. (1990) Rethinking mammalian brain evolution. Am. Zool. 22 Newman, S. et al. (2003) Differential effects of syntactic and semantic
30, 629 – 705 processing on the subregions of Broca’s area. Brain Res. Cogn. Brain
3 Striedter, G. (1998) Stepping into the same river twice: homologues as Res. 16, 297 – 307
recurring attractors in epigenetic landscapes. Brain Behav. Evol. 52, 23 Deacon, T. (2000). Heterochrony in brain evolution: cellular versus
218 – 231 morphological analyses. In Biology, Brains, and Behavior: The
4 Frost, D. (1981) Orderly anomalous retinal projections to the medial Evolution of Human Development (Parker, S. et al., eds), pp. 41 – 88,
geniculate, ventrobasal and lateral posterior nuclei of the hamster. SAR Press, Santa Fe
J. Comp. Neurol. 203, 227 – 256 24 Roth, G. et al. (1997) Genome size, secondary simplification, and the
5 Sur, M. et al. (1988) Experimentally induced visual projections into evolution of the brain in salamanders. Brain Behav. Evol. 50, 50– 59
auditory thalamus and cortex. Science 242, 1437– 1441 25 Semendeferi, K. et al. (1997) The evolution of the frontal lobes: a
6 Stanfield, B. and O’Leary, D. (1985) Fetal occipital cortical neurons volumetric analysis based on three-dimensional reconstructions of
transplanted to the rostral cortex can extend and maintain a magnetic resonance scans of human and ape brains. J. Hum. Evol. 32,
pyramidal tract axon. Nature 313, 135 – 137 375– 388
7 Law, M. and Constantine-Paton, M. (1981) Anatomy and physiology of
experimentally produced striped tecta. J. Neurosci. 1, 741 – 759 1364-6613/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
8 Doron, N. and Wollberg, Z. (1994) Cross-modal neuroplasticity in the doi:10.1016/j.tics.2004.05.006

| Research Focus Response

Response to Deacon: Evolving mirror systems:

homologies and the nature of neuroinformatics
Michael A. Arbib1,2 and Mihail Bota2
1
Department of Computer Science, University of Southern California, Los Angeles, CA 90089-2520, USA
2
Department of Neuroscience, University of Southern California, Los Angeles, CA 90089-2520, USA

We are delighted that our article [1] has received a Evolving the language-ready brain
thoughtful response from Terry Deacon [2]. However, his Deacon [2] asserts that ‘Mirror neuron functions may have
discussion of homologies ignores key elements of our views been recruitedto aid language readiness, but we should not
on the evolution of brain mechanisms supporting language necessarily expect them to be the locus of a distinctive
and does not fully address the details of our approach to language adaptation. To explain that we must look for
neuroinformatics, the NeuroHomology Database (NHDB). something else that is different’, as if this were a critique of
our approach. However, we said explicitly that Arbib [3]
Corresponding author: Michael A. Arbib ([email protected]). has amplified the approach of Rizzolatti and Arbib [4] to
Available online 10 June 2004 hypothesize seven evolutionary stages.
www.sciencedirect.com