Content uploaded by Sinziana Pop
Author content
All content in this area was uploaded by Sinziana Pop on Nov 21, 2022
Content may be subject to copyright.
Nature Reviews Neuroscience
nature reviews neuroscience https://doi.org/10.1038/s41583-022-00644-y
Review article Check for updates
Evolution of central neural circuits:
state of the art and perspectives
Ruairí J. V. Roberts1,2, Sinziana Pop 1,2 & Lucia L. Prieto-Godino 1
Abstract
The wide variety of animal behaviours that can be observed today arose
through the evolution of their underlying neural circuits. Advances in
understanding the mechanisms through which neural circuits change
over evolutionary timescales have lagged behind our knowledge of
circuit function and development. This is particularly true for central
neural circuits, which are experimentally less accessible than peripheral
circuit elements. However, recent technological developments —
including cross-species genetic modications, connectomics and
transcriptomics — have facilitated comparative neuroscience studies
with a mechanistic outlook. These advances enable knowledge from
two classically separate disciplines — neuroscience and evolutionary
biology — to merge, accelerating our understanding of the principles
of neural circuit evolution. Here we synthesize progress on this
topic, focusing on three aspects of neural circuits that change over
evolutionary time: synaptic connectivity, neuromodulation and
neurons. By drawing examples from a wide variety of animal phyla,
we reveal emerging principles of neural circuit evolution.
Sections
Introduction
Open questions in circuit
evolution
Evolutionary processes
Evolution of circuit
connectivity
Evolution of neuromodulation
Evolution of neurons
Emerging principles
Future perspectives
1The Francis Crick Institute, London, UK. 2These authors contributed equally: Ruairí J. V. Roberts, Sinziana Pop.
e-mail: lucia.prietogodino@crick.ac.uk
Nature Reviews Neuroscience
Review article
how evolutionary forces tinker with brain circuits can provide us with
additional insights into how neural networks assemble and function.
Here we describe what the field has learned about the evolution
of central neural circuits. There are a number of factors that must
be considered when selecting which species to study to gain insight
into circuit evolution (Box1). However, studies that compare closely
related and/or experimentally accessible species have been especially
informative4–6, particularly when facilitated by recent technological
developments (including those that enable genetic modification of a
wide range of species, connectomics and single-cell transcriptomics).
The comparison of such carefully chosen species has identified evolution-
ary patterns at the cellular, synaptic and molecular levels and, on occasion,
allowed us to move beyond correlation to establish causal links between
neural and behavioural changes during evolution7–9. Wetherefore focus
Introduction
Animals have conquered almost every corner of our planet by adapting
their physiology and behaviour to different environmental conditions.
Historically, neuroscientists have taken advantage of some of these
exquisite adaptations to gain insights into how neural circuits work.
Classic examples of this approach include the discovery of the neural
motifs thatimplement sound localization in barn owls
1
and of the prin-
ciples of developmental wiring in frogs
2
. Thus, the use of diverse organ-
isms in neuroscience has given us a broad perspective of how brains
develop and function, helping us toidentify general organizational
principles of neural circuits
3
. When considering these principles, it is
essential to keep in mind that neural circuits are fruits of the evolution-
ary process, and thus the features that we observe have been shaped by
the evolutionary paths that were available. Therefore, understanding
Box 1
Choosing which species to study
Which species should we work with to gain the best insights into
the evolution of neural circuits? This question has no single answer
because the key to unravelling general principles is to work as a
community on diverse species. Here we highlight the factors that
make particular species useful models in the study of neural circuit
evolution and illustrate the advantages, disadvantages and trade-os
of dierent choices.
Phylogenetic relationships
Phylogenetic relationships are a key consideration when one
is choosing species to understand neural circuit evolution. For
example,if we want to uncover dierences at the level of synaptic
connectivity between homologous neurons, it might be best to
choose closely related species with divergent behaviours10,11,13,50,51.
However, if we aim to understand the evolutionary mechanisms
behind the generation of novel cell types, species separated by
longer evolutionary distances might be a better choice42. Thus, we
endorse the proposal to develop the use of ‘model clades’134, deined
as species groups whose phylogenetic relationships make their
comparisons particularly convenient or exciting and that also display
other features (as detailed below) that make their nervous systems
accessible for study.
Behavioural dierences
To unravel the neural circuit mechanisms by which diverse behaviours
emerge, we need to study species with clear behavioural dierences.
To that end, it is key to be able to precisely quantify behavioural
changes to make accurate comparisons. Frustratingly, those species
that display striking behavioural features are often also those that are
hard to keep in the laboratory and to experimentally manipulate. It is
therefore encouraging to see recent developments in tracking software
that seeks to classify behavioural states in more natural settings142.
Proximity to established laboratory animals
Choosing species that are closely related to classic laboratory
animals has many advantages. For one, the laboratory species —
for which we have a lot of information — can be used as one point
of comparison, reducing the groundwork. Furthermore, it is often
possible to transfer technical tools across closely related species,
albeit with a substantial amount of work. Thus, choosing species
closely related to well-established laboratory animals can be the
quickest entry into evolutionary studies.
Neural circuitry accessibility
Not all species are equal in terms of how easy it is to study their
central neural circuits. Generally, the larger the brain, the harder it is,
but there are other considerations. Small brains with relatively few
neurons allow high-resolution neural circuit reconstruction using
electron microscope (EM)-based connectomics approaches50,143,144.
While it is currently impossible to image the entire central neural
circuits of larger animals (more than about 1 m3)at this resolution,
reconstructions of subcircuits can be informative77, and connectomic
technologies (both EM and not EM based) are developing rapidly.
Another important aspect to consider is how easy it is to probe
neuronal function. While species with fewer but large neurons enable
easy electrophysiological access10, other systems oer transparency
and genetic accessibility (see below), enabling access to functional
information through optical physiology.
Genetic accessibility
Genetic accessibility is key for two reasons. First, it enables the
introduction of genetic tools that can facilitate the study of neural
circuits. Even more importantly, it can support the ultimate goal of
demonstrating causality (that is, showing that the identiied cross-
species dierences indeed underlie the observed behavioural
phenotypes), which is very diicult without direct genetic access
to manipulate neural circuits in behaving animals.
Some species groups meet most or all of these conditions, making
them good systems for the study of neural circuit evolution. These
include Drosophila species, the surface and cave forms of the ish
Mexican tetra, the clade formed by the nematodes Caenorhabditis
elegans and Pristionchus paciicus, nudibranchs, Peromyscus mice,
species of weakly electric ish, Xenopus frogs and mosquito species.
Nature Reviews Neuroscience
Review article
this Review on examples of studies of closely related and/or experimen-
tally accessible species that we believe have given us important insights
into the mechanistic bases of central neural circuit evolution.
Throughout this Review it is essential to keep in mind that the
effects of evolution can be observed at several phenotypic levels, such
as behavioural output, network configuration and neuronal molecu-
lar composition. Although changes in behaviour are often the most
obvious outcome of neuronal evolution, it is important to examine
all of these levels in an integrated fashion to gain insights into how
evolutionary forces shape neural circuits. For example, as discussed
later, behaviours might remain conserved while their underlying neural
networks evolve
10
, and it is therefore only by simultaneously investi-
gating both levels (behaviour and neural circuitry) that we can extract
principles of brain evolution.
We begin by outlining some of the key questions that the field and
this Review hope to address, and we follow this with a brief introduc-
tion to evolutionary processes and their application to central neural
circuits. Next, we structure a discussion of these important issues
around three aspects of neural circuits that evolution can affect: syn-
aptic connectivity, neuromodulatory systems and neurons. Finally, we
consider the extent to which our accumulated knowledge can answer
the questions posed at the outset.
Open questions in circuit evolution
A key goal of evolutionary neuroscience is to go beyond a description
of what has changed in the brains of diverse animals over time to draw
general principles and try to understand the mechanisms by which
neural circuits and behaviours change. To achieve this, a few key ques-
tions can help us toset up a framework for the study of the evolution
of neural circuits:
Question 1: Is the sensory periphery more evolvable than central
circuits? Owing to their accessibility, peripheral sensory circuits have
been the focus of most studies into neuronal evolution
11–15
. This has
created a sampling bias that might give the unfounded impression that
the periphery is more evolvable than downstream central circuitry.
The best way to address this question is to examine a wide range of
central neural circuits in animals separated by differentphylogenetic
distances.
Question 2: Do some types of evolutionary change happen more
frequently than others? The complex and interconnected neural cir-
cuits of the brain can change in multiple ways, raising the question
of whether some changes are more likely to occur than others. For
example, do brains evolve primarily through the addition of new cell
types or through the rewiring of existing ones? How common is the
evolution of neuronal function without circuit rewiring? A key goal
of the field should be to understand whether some aspects of neural
circuits change more often over evolutionary time or are more likely
to be targeted by evolution under certain conditions.
Question 3: Do neural circuits have ‘hot spots’ for evolutionary
change, or are changes homogeneously distributed across the net
-
work? If we were to examine all central circuit elements across many
species and over a range of evolutionary distances, would we find that
changes are equally likely to occur anywhere within the network? Alter-
natively, are there some neuronal populations or circuit elements that
Glossary
Alleles
Dierent versions of a DNA sequence
found in a particular genomic location.
Alternative splicing
The process that, during pre-mRNA
processing, rearranges exons (coding
sequences) and thus allows a single
gene to encode multiple proteins.
Clades
Groups of organisms that are derived
from a common ancestor.
Comparative studies
Research work where two or more
species are compared side by side to
identify dierences between them.
Connectomics
The generation of comprehensive
synaptic resolution maps of the
connections among the neurons
of an organism’s nervous system.
Corollary discharge
Internal activity generated by nervous
systems that carries information about
a motor command and is used to
estimate the outcomes of executed
movements (for example, to help
distinguish between self-generated and
externally induced sensory input).
Evolvability
Here, the ability of particular features of
a system to facilitate change.
Extant species
Species that can be found alive today,
as opposed to species that were
extinguished at some time in the past.
Functomics
High-throughput investigation of
neuronal circuit function.
Homologues
Structures that share a common
ancestral origin.
Neural circuit blueprint
The assembly of characters that
determine the general organization of a
brain, including the number and/or type
of neurons, organization of cell bodies
and mesoscale connectivity.
Phylogenetic distances
The amounts of time thatelapsed
between the divergence of pairs of
species from their most recent common
ancestor.
Pleiotropy
The phenomenon in which a gene
aects multiple unrelated phenotypic
traits. In evolution, pleiotropy is
considered to limit the potential for
change because modiications in a
gene could simultaneously aect
one trait in an adaptive way but cause
maladaptive changes in another.
RNA editing
The process that modiies speciic
nucleotides of RNAs potentially altering
their function.
Retroelement
Genome sequences that can be
transcribed into RNA, can be reverse
transcribed into DNA and can be
inserted at another genome location.
Single-cell transcriptomics
The study of the gene expression of
individual cells.
Terminal dierentiation genes
Genes that confer cellular properties
through their function, rather than
inluencing the expression of other
genes.
Transcription factor genes
Genes that function by regulating the
expression of other genes.
Wiring economy theory
A theory that proposes that nervous
systems have evolved to minimize
neurite length to minimize the energetic
cost of the maintenance of neuronal
extensions.
Nature Reviews Neuroscience
Review article
are more likely to accumulate a disproportionate amount of change?
For the evolution of sensory receptor proteins, in which a similar ques-
tion has been addressed, researchers have found that a few amino acid
positions within a protein can act as hot spots, changing repeatedly
and convergently to modify receptor sensitivity as animals adapt to
different environments13,16–19. A goal for evolutionary neuroscience is
to reach a similar level of understanding for central circuits.
Question 4: Do all central circuits evolve according to similar
rules, or does theneural circuit blueprint influence how they evolve?
For example, are there any fundamental differences in how brains
evolve across different animalclades? Do features such as brain size or
neuronal redundancy (that is, the number of neurons of each neuron
type) influence how evolvable a circuit can be? Understanding whether
certain features favour certain evolutionary outcomes will bring us
closer to understanding why central circuits function the way they do.
Evolutionary processes
With the aim of bridging the gap between circuit neuroscience and
evolutionary biology, we introduce some basic evolutionary concepts,
including the main evolutionary forces that influence neural circuits
and some common evolutionary outcomes. For a more thorough
introduction to evolutionary biology, we direct the reader to refs.20,21.
a Evolutionary forces
b Evolutionary outcomes
DriNatural selection
Available genetic diversity
Mutation Migration
Random
sampling
Divergent Convergent Parallel
c Changes in neural circuit elements
Connectivity Neuromodulation Neurons
Example: Neuronal co-option in moths Example: Maternal care in frogs Example: Neuronal microexon
splicing programme
AL
SEZ
Ventral
nerve cord
Butterfly
interneuron
Moth
interneuron
Maternal care
POA
South American frog Madagascan frog
Common ancestral state
eMIC
Regulation of neuronal
development and function
Changes in
regulation
eMIC eMIC
Changes in target
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 4 5
Change in the expression
of neuromodulators, their
receptors or degrading
enzymes
Change to neuron
physiological properties
Change to number and
type of neurons
Neuromodulator
Degrading enzymes
Receptor
Nature Reviews Neuroscience
Review article
Evolutionary forces
Among the numerous forces that shape evolution, four of the most
central are natural selection, mutation, gene flow (also known as migra-
tion) and genetic drift (Fig.1a). The relative contribution of each of
these processes to the emergence of between-species differences is
an outstanding question within the field of evolutionary biology22–24.
These forces are not mutually exclusive and interact with each other:
gene mutation, gene flow and genetic drift together shape the genetic
diversity that exists within a population, providing the raw material on
which natural selection operates.
Natural selection itself encompasses multiple forms of selec-
tion. Most people associate evolution with positive selection: the
increase, within a population, ofalleles underlying traits that confer a
reproductive advantage to the individuals harbouring them (Fig.1a).
Another form of selection that many will be familiar with is that which
removes alleles harbouring harmful mutations from the population.
This is referred to as purifying (or negative) selection and enables
the conservation of essential traits across evolution. A third form of
selection is balancing selection, in which multiple alleles are actively
maintained in a population. This may occur, for example, when differ-
ent alleles confer a selective advantage at different times of the year.
While these types of selection may account for most of the examples
discussed in this Review, it is important to note that selection can take
numerous other forms and can generate different patterns of genetic
variation within a population.
Mutations span a range of sizes and types, from single-nucleotide
changes and small insertions or deletions to large-scale rearrange-
ments (including duplications, deletions, inversions, chromosomal
fusions and translocations). Considerable attention has been paid to
mutations that generate gene duplications, as these events have long
been considered a primary source of novel molecular functions
25–27
.
The most likely outcome of gene duplication is that one copy of the
gene will be rendered non-functional (a pseudogene) through theaccu-
mulation of disruptive mutations, while the other copy will keep its
ancestral function through purifying selection. However, an alternative
outcome is that one of the duplicated copies accumulates mutations
that endow it with a novel function or expression pattern that provides
a selective advantage, a process known as duplication and divergence.
Intriguingly, comparative single-cell sequencing in cerebral organoids
derived from humans, chimpanzees and macaques showed an over-
representation of recently duplicated genes among those genes that are
differentially expressed in humans when compared with chimpanzees
and macaques, suggesting that the human brain may have evolved
through gene duplication and divergence28 (see later).
Genetic diversity is also influenced by gene flow (that is, the move-
ment of genes into or out of a population as a result of the migration of
individuals between populations) and genetic drift (changes in allele
frequencies caused by the random sampling that occurs in the produc-
tion of each generation) (Fig.1a). Alleles can become fixed by drift,
especially in small populations (where the impact of chance during sam-
pling is higher) and when the alleles do not have strongly deleterious
effects. Genetic drift can explain why non-adaptive traits can evolve in
a population, which can otherwise appear counterintuitive.
These clearly defined genetic-centred evolutionary concepts are
sometimes borrowed in the evolutionary neuroscience literature to
describe patterns of change observed at the circuit level. However, it
is important to remember that these parallelisms, while superficially
intuitive, do not hold at the deeper mechanistic level. For example,
inspired by the genetic duplication and divergence model, it has been
noted that entire neural circuits can duplicate and diverge during evolu-
tion29. Examples of this process include the duplication and divergence
of basal ganglia nuclei from the basic blueprint present in lamprey-like
ancestors to the complex system found in mammals
29,30
, the generation
of song circuits through the duplication and divergence of motor con-
trol circuits in birds29,31 and the evolution of vertebrate retinal circuits
through the duplication of simpler circuit motifs present in the eye of
their urbilaterian ancestor
32,33
. Other cross-species comparisons have
gathered evidence in support of this mode of evolution for the olfactory
system11,34–36 and the cerebellar nuclei37. While it is important to keep
in mind that the genetic underpinnings of such circuit duplication and
divergence may not be a genetic duplication and divergence, it does
provide a useful analogy to describe a commonly observed pattern
Fig. 1 | Overview of evolutionary forces and outcomes. a, Illustration of four
key evolutionary forces: natural selection, migration, mutation and drift. Each
panel shows a diagram of the effects of the force at the population level (light
grey and dark grey circles represent different alleles of the same gene). In natural
selection, alleles that confer a reproductive advantage increase in frequency
in the population over subsequent generations. The available genetic diversity
on which natural selection can act is dictated by three main processes. Gene
mutation results in the de novo appearance of new alleles in the population. Gene
flow (migration) describes the flow of new alleles into a population from another
population135. In genetic drift, changes occur in allele frequencies in a population
due to random sampling. b, Illustration of the main evolutionary outcomes
referred to in this Review: divergent, convergent and parallel. Each panel contains
a schematic of the evolutionary outcome (the differently coloured hexagons
represent different animal features or traits) and an example from the evolution
of neural circuits below. In divergent evolution, two ancestral species that shared
a common trait took different evolutionary paths, and their descendants now
display differences in this trait. For example, an ancestral mechanosensory
interneuron that projects from the ventral nerve cord to the suboesophageal
zone (SEZ) evolved in moths (red) to provide feedback to the olfactory system
via the antennal lobe (AL), allowing it to compensate for the turbulence created
by their wing beating. Closely related butterfly species, whose wing beating does
not interfere with odour-plume tracking, did not evolve this trait40. In convergent
evolution, two species whose ancestors had different traits have evolved a
common trait independently. An example is the evolution of maternal care in two
species of frogs that inhabit South America (Oophaga sylvatica) and Madagascar
(Mantella laevigata) through modifications to different cell types in the same
brain structure, the preoptic area of the hypothalamus (POA)41. Parallel evolution
starts from a common ancestral state (either in a common ancestor or in separate
ancestors) from which two species independently evolve (that is, without gene
flow between the species) a similar trait that is different from the ancestral
state. As an example, a protein domain (eMIC) that is a master regulator of the
neuronal microexon splicing programme first arose in bilaterians136, where it was
thought to control general neuronal functions. Subsequently, vertebrate and
invertebrate lineages evolved the microexon splicing programme in parallel, and
in both cases the programme now regulates neuronal function and development.
However, the genes regulated by eMIC differ between mice and flies, and the
regulatory network for eMIC expression in neurons has also evolved across the
two lineages49. c, Illustration of the main ways in which neural circuits can evolve.
Changes may occur in connectivity among homologous neurons (left, triangles
with the same number represent homologous neurons), in components of the
neuromodulatory system (centre) and in the neurons themselves (that is, in their
physiological properties, type or number) (right). Evolutionary changes are
highlighted in red.
Nature Reviews Neuroscience
Review article
of evolution at the circuit level. Similarly, potentially confusing is the
use of the word ‘drift’. In contrast to genetic drift, ‘neural circuit drift’
refers to evolutionary changes in neural circuit configurations that
do not alter their functional output. Neural circuit drift is consistent
with modelling studies showing that different synaptic connectivity
and activity patterns within a neural circuit can produce identical
network output38.
Evolutionary outcomes
At the phenotypic level, we can classify evolutionary outcomes into
three main types (although there are others): divergent, convergent
and parallel (Fig.1b). The evolutionary forces mentioned in the previous
subsection can all generate any of these outcomes. It is important to
reflect on the level of analysis under study (behavioural, neural circuit,
cellular and genetic) when one is considering these outcomes. For
example, when starting from ancestral species that do not display a
particular behaviour, two descendant species can evolve the same
behaviour (we refer to this as ‘convergent evolution at the behav-
ioural level’; see later). The neural circuit modifications underlying
the behavioural change in each species could be the same or could
differ. In the latter case, the behaviourally convergent change would
have arisen through divergent evolution at the neural circuit level.
Thus, to gain mechanistic insights into how neural circuits evolve, we
must look across multiple levels of analysis. It is also essential to keep
in mind that evolution is a science of inference: as we do not have direct
access to the behaviour and neural circuits of ancestral species, we infer
their phenotype by examiningextant species. The accuracy of these
inferences increases as the number of species examined increases
(Supplementary information).
Divergent evolution — in which two species evolving from the same
ancestor become different from each other — is probably what most
people have in mind when they think of evolution. One type of diver-
gent evolution that is worth singling out is co-option(also referred to
as exaptation), in which a trait is selected because it gains a function
that is different from the function for which it originally evolved. This
enables rapid evolutionary change through reuse of already existing
structures39. An example is found in moths, where a pre-existent mecha-
nosensory interneuron was co-opted to provide acorollary discharge
to the olfactory system, enabling moths to compensate for their own
wing beating during odour tracking40 (Fig.1b).
Convergent evolution is the process by which two or more organ-
isms whose ancestors exhibited different traits evolve the same trait
independently of each other (Fig.1b). An example is the independent
evolution of nursing behaviour in female poison frogs of different
species from Ecuador and Madagascar. Interestingly, this behaviour is
associated with increased activity in the preoptic area in both species,
but the neuronal populations activated are different41. Therefore, the
behaviour evolved convergently through divergent evolution at
the neuronal circuitry level. Another example is the convergent evolu-
tion of the function, circuitry and molecular make-up of the mammalian
neocortex and the dorsal ventricular ridge of reptiles and birds, as
recently suggested by single-cell transcriptomics studies42,43.
Parallel evolution is the process by which two or more organisms
whose ancestors shared a particular trait independently evolve the
same trait (this being different from the ancestral trait). Therefore,
unlike convergent evolution, the species share evolutionary ‘start’ and
‘end’ points, but the start point and the end point are different from
each other44–46 (although see refs.46–48 for alternative definitions and a
debate in the evolutionary field on how to distinguish parallel evolution
and convergent evolution) (Fig.1b). An example is the evolution in
vertebrates and invertebrates of a neuron-specific splicing programme
that regulates central neuron function and behaviour49.
Keeping in mind these evolutionary processes, we now consider
the elements of neural circuits that evolutionary forces shape: con-
nectivity, neuromodulation and the neurons themselves (Fig.1c).
Although these divisions are useful for the purposes of this Review, it
is important to note that they have been introduced for simplicity and
that many of the mechanisms of neuronal evolution discussed later
herein can and do intermingle. Finally, because we focus on neural
circuit changes, the examples discussed mostly illustrate divergent
evolution at the neural circuit level.
Evolution of circuit connectivity
Changes in synaptic connectivity between neurons — either in the
number of synapses or in the type andnumber of neurons that are
connected — seems intuitively to be one of the easiest ways to modify
neural circuits as it allows the conservation of a general blueprint for
circuit development with only subtle modifications required at the
last step (synaptogenesis). A hypothesis is that this mode of evolu-
tion could be prevalent when the general organization of neural cir-
cuits is constrained, for example to remain within a particular size
or keep to an energy budget, or because it might require few genetic
changes with lowpleiotropy. The functional outcome of these diver-
gent circuit modifications can be conserved, divergent, parallel or
convergent (Fig.2).
Neural circuit rewiring producing divergent behaviours
The study of species that are related to classical laboratory ani-
mals but display different behaviours (Box1) has shed light on the
extent of evolutionary neural circuit rewiring and the causal links
between these changes and particular cross-species behavioural
differences.
The small size and numerical simplicity of the nematode nerv-
ous system has enabled a comparative connectomics approach to
be applied to two distantly related species with divergent feeding
behaviours50,51: Caenorhabditis elegans, a bacterial feeder, and Pristion-
chus pacificus, which preys on other nematodes (Fig.2a). Systematic
reconstruction of the neuronal circuits involved in chemosensation and
feeding50,51 revealed that — while neuron number, cell body positions
and axonal branching are conserved — both circuits have undergone
major synaptic rewiring during their evolution(although see ref.
52
).
For example, a pair of chemosensory neurons that have different che-
mosensory responses in C. elegans53 are connected by gap junctions
in P. pacificus, breaking this functional asymmetry. Consistent with
this, the microRNA that regulates this asymmetry in C. elegans54 is not
encoded in the P. pacificus genome
51
. In addition, in P. pacificus the
muscles involved in motor predatory action are innervated by motor
neurons whose C. eleganshomologues are interneurons. The latter
do not form synapses with muscles, hinting at a higher level of motor
regulation in the predatory species50.
The observed synaptic rewiring with conserved axonal branching in
these worm species suggests that the evolution of these neural circuits
relied on changes in synaptogenesis, possibly as a result of modifications
in surface molecules determining the synaptic specificity of partner
neurons
51
. This mode of evolution may have been favoured in this case
because surface molecules areencoded byterminal differentiation genes
whose changes are thought to have low pleiotropy55. Another driving
force might have been energy conservation: wiring economy theory
56,57
Nature Reviews Neuroscience
Review article
postulates that neural systems have evolved efficient wiring strategies
that minimize neurite length. Therefore, the cell body position of these
neurons might be constrained to that of an efficiently wired ancestral
state based on long-range projections.
This example highlights how the detailed study of neural circuits
in species with divergent behaviours can shed light on the evolution
of connectivity through changes in synaptic specificity. However,
this is insufficient to reveal the evolutionary forces that led to these
connectivity changes. Addressing this point is likely to require com-
bining furthercomparative studies including more species with eco-
logical information and a theoretical framework. Another goal for
the future will be to identify the molecular factors underlying the
rewiring. Here the generation of a single-cell transcriptomics data-
set for P. pacificus will be instrumental. Two recent studies com-
bined the C. elegans connectome with single-cell gene expression
data
58,59
and identified potentially novel molecules determining syn-
aptic specificity. Applying this framework in a comparative context
could reveal the molecular underpinnings of neural circuit changes
across species.
The small networks of identifiable neurons in nematodes can
greatly improve our understanding of how neural circuits evolve;
however, it is possible that more-complex brains might evolve via other
paths. It is, therefore, important to consider how general the evolution of
behaviour through rewiring of homologous neurons is. The study of the
evolution of sexual behaviours across drosophilid fly species provides
some evidence for the generality of this mechanism, and illustrates how
it interacts with other types of changes in neural circuits.
Drosophilid fly males evaluate potential partners by tasting the
cuticular hydrocarbons on the abdomen of females, and then sing by
vibrating their wings to increase female receptivity60. Both the cuticular
chemical cues and the male singing patterns are highly variable across
species, providing a model to study how sensory and motor circuit
components evolve. Changes in the sensory system have been best
characterized by comparing three taste pathways that mediate partner
acceptance or rejection in the males of Drosophila melanogaster and
Drosophila simulans. Of these pathways, one has been shown to be
conserved (the Ppk25 pathway)
61
, one is presumed to have diverged
through changes in peripheral neuron receptor sensitivity (the Gr32a
pathway)4,61,62 and the third is conserved at the periphery but has
undergone central circuitry changes (the Ppk23 pathway)
4
(Fig.2b).
Thus, different pathways within the same sensory system can undergo
different evolutionary trajectories.
Together with other studies
62,63
, this example suggests that periph-
eral and central circuits are equally likely to be targeted by evolu-
tion and, furthermore, that targeting peripheral and central circuits
together may generate morerobust behavioural shifts than individual
modifications. These changes can happen ‘in parallel’ (with some
pathways changing in the periphery and others changing centrally as
in the example above) or ‘in series’ (with multiple changes occurring
within the same pathway13) (Fig.2b). Of the pathways described above,
the Ppk23 pathway evolved through changes in the connectivity of
central homologous neurons. Specifically, there was a shift in the
balance of excitatory and inhibitory inputs onto P1, a male-specific
cluster of neurons that integrate multisensory signals and trigger
the initiation of courtship
4
. It has been proposed that neurons such
as P1 neurons — sites of sensory convergence that can directly trigger
behaviours60 — might act as evolutionary hot spots because their key
role within neural circuits enables subtle modifications to have large
behavioural effects. At the mechanistic level, the exact nature of the
synaptic changes onto P1 remains unclear. One hypothesis is that there
is increased synaptic connectivity from inhibitory neurons onto P1,
which may have been mediated through small synaptic modifications
such as changes in the molecules involved in partner-specific synapse
recognition64, an increase in the synaptogenic potential of the inhibi-
tory presynaptic neurons65 or differential postsynaptic control of the
balance of excitatory and inhibitory synaptic input66.
Further evidence for neural circuit evolution through rewiring
of homologous neurons comes from the motorcircuits involved in
drosophilid courtship song production. The differences in song pat-
tern between the males of Drosophila yakuba and Drosophila santomea
are caused by changes in the circuits downstream of a descending
interneuron called ‘pIP10’, which displays subtle differences in axonal
arborization across species6.
Another intriguing instance of neuronal rewiring is that which pre-
sumably occurred in the evolution of our own species. A study in mice
showed that misexpression of a human-specific gene in mouse cortical
pyramidal neurons increased the number of inputs that they received
from local excitatory (but not inhibitory) neurons and from long-range
projections. Importantly, this connectivity changes increased the sen-
sory encoding properties of the pyramidal neurons and augmented the
learning ability of the animals in a cortex-dependent task67. Interestingly,
the gene in question arose through duplication and divergence of a
mammalian ancestral gene at the time of the emergence of the hominid
lineage, before the major increase in brain size that happened in modern
humans (see later), suggesting that a change in neocortical connectivity
might have increased cognitive abilities in ancient hominids68.
Evidence of rewiring of homologous neurons is sparse outside
these examples because it is difficult to interrogate central circuits at
the required resolution. However, prospects for finding connectivity
changes among homologous neurons across vertebrate species might
come from single-cell sequencing studies. For example, such studies
have found that most inhibitory cortical interneuron types are broadly
conserved across vertebrates from lampreys to humans30,42,43,69–71.
Interestingly, despite being broadly homologous, interneuron types
have diversified, displaying differences in gene expression across spe-
cies69, some of which might regulate fundamental differences in their
synaptic connectivity. However, we remain far from being capable of
bridging the gaps between the identification of homologous neuronal
populations, the characterization of their rewiring across species and
a full understanding how this influences circuit function, let alone
behaviour. A good place to start addressing these questions in verte-
brates is the retina, because we know so much about its cell types and
synaptic connectivity
72,73
and because many retinal circuits can directly
influence behaviour
74–76
. Furthermore, it is possible to identify at least
some homologous retinal cell types across species and look at their
synaptic connectivity (see ref.77, but also see refs.78–83). Therefore, the
time is right to exploit these well-known vertebrate circuits through
cross-species comparisons72.
Neural circuit rewiring producing similar behaviours
As noted already, two species can convergently evolve the same behav-
iour from ancestors that did not display such behaviour (with the cir-
cuits underlying this behaviour across species evolving convergently or
divergently). Alternatively, a behaviour present in a common ancestor
can remain conserved in two descendant species as its underlying neu-
ral circuits diverge either through neuronal circuit drift or by natural
selection (Fig.2c). These evolutionary outcomes are well illustrated
by studies of the swimming central pattern generators of sea slugs.
Nature Reviews Neuroscience
Review article
Two species, Dendronotus iris and Melibe leonine, which descended
from a common ancestor that is predicted to have had a left–right
swimming pattern also display similar left–right swimming behaviour.
However, this behaviour is controlled by homologous neurons with
divergent wiring, a configuration that seems to have evolved through
neuronal drift84 (Fig.2d). Some other sea slugs swim instead through
flexions in their dorsoventral axis. The neural circuits underlying this
behaviour have been characterized in detail for two species: Tritonia
diomedea and Pleurobranchaea californica. Both use different circuits
composed of partially overlapping homologous neurons
85
, and phy
-
logenetic reconstruction suggests that the behaviours might have
evolved through convergent evolution from non-swimmer ancestors10,86
(Fig.2d). Finally, given that most nudibranchs are non-swimmers, and
that swimmer species are distributed across the phylogeny, it appears
likely that left–right swimming behaviours (in D. iris and M. leonine)
versus the dorsoventral swimming behaviours (in T. diomedea and
P. californica) evolved independently (and this is thus an example of
divergent evolution). In line with this hypothesis, the circuit elements
for dorsoventral swimming do not overlap with the neurons involved in
left–right swimming, suggesting that different neuronal populations
were independently recruited in the evolution of these two swimming
behaviours10,86.
As discussed above, behaviours can remain conserved while their
underlying networks diverge through neuronal circuit drift, but this
is not the only explanation for divergent circuitry in the presence of
conserved network output. Natural selection for a given output in the
face of other changing features might drive the evolution of divergent
neuronal networks. For example, all mammalian retinas are thought
to compute object velocity
87–89
. However, the eyes of different species
differ largely in size, which means that an object moving at a given
velocity traverses their retinas at different absolute speeds (Fig.2e).
Therefore, to maintain an accurate computation of absolute speed,
the retinal networks of animals with different eye sizes must have
changed. Indeed, connectomic reconstruction of the circuits in mice
Divergent network and behaviour
Divergent network and conserved output
a b
c d e
C. elegans
bacterial feeder
P. pacificus
predatory feeder
Neurons Muscles Other outputs
Sexual circuitry Olfactory circuitry
D. melanogaster
D. simulans
D. sechellia
D. yakuba
D. melanogaster
D. simulans D. melanogaster D. sechellia
D. melanogaster
female
pheromone
D. melanogaster
female
pheromone
P1 neuron
mAL
neuron
Ppk23 taste
neuron
vAB3 neuron
Stimulates courtship
behaviour
Inhibits courtship
behaviour
Or22a
olfactory
neuron
Projection
neuron
Ir75b olfactory
neuron Altered
receptor
sensitivity
Altered
morphology
Neural circuit dri Convergent evolution Adaptation to changing body size
Tritonia diomedea
Pleurobranchaea californica
Dorsoventral alternation swimmingLe–right alternation swimming
L-Si1 R-Si1
L-Si2 R-Si2
L-Si1 R-Si1
L-Si2 R-Si2
DSI
VSI C2
Melibe leonine
Dendronotus iris
Mouse Rabbit
Inhibitory synaptic inputs
AS A3
IVS A1
A10
t1
t1
t1t2
t2
t2
dsmall dbig
d
Speed is fast Speed is slow
25° 25°
Pleurobranchaea
Excitatory synapse
Inhibitory synapse
Electrical synapse
Nudipleura
Nudibranchia
Tritonia DV
DV
Dendronotus LR
Melibe LR
Arminia NS
Archidoris NS
Doto NS
RabbitMouse
t2
– t1
d
Speed =
Altered
connectivity
Nature Reviews Neuroscience
Review article
and rabbits suggests that this issue has been solved by the evolution
of divergent synaptic connectivity in the neural circuits performing
this computation. Specifically, a neuron type essential for computing
direction selectivity, the starburst amacrine cell, receives inhibitory
input onto the proximal region ofits dendrites in mice, but across the
length of the dendrites in rabbits. This synaptic reconfiguration alters
the velocity tuning of these neurons in a way that compensates for the
differences in eye size77.
The examples presented here show that synaptic rewiring of con-
served cell types is an important mechanism in the evolution of neural
circuits. An open question, however, is how this rewiring is encoded at
the level of the genome. Furthermore, it will be essential to understand
how prevalent this mechanism is compared with the mechanisms
explored later herein, how different evolutionary mechanisms are
integrated and what determines the path taken by evolution.
Evolution of neuromodulation
Neuromodulators are signalling molecules that, unlike neurotrans-
mitters, can act at a distance and do not need to be released into
the synaptic cleft to be effective. They can modify neuronal excitability,
synaptic strength and even the polarity (that is, whether a synaptic
input is excitatory or inhibitory) conferred by neurotransmitters90.
Thus, neuromodulators can act as master regulators by, for example,
changing the functional connectivity of neuronal circuits. This might
make neuromodulatory systems particularly evolvable because rela-
tively small changes— to the neurons producing a given neuromodu-
lator and/or to those expressing receptors for it — have the potential
to produce large functional changes. For instance, the novel expres
-
sion of a neuromodulator receptor in a neuronal population could
‘rewire’ its functional connectivity to those brain areas that release
the neuromodulator without the need to form energetically expensive
long-range neuronal extensions. The evolution of neuromodulatory
systems has been reviewed elsewhere
91,92
, and we therefore focus here
on two examples that illustrate the multiple mechanisms by which
neuromodulatory systems can evolve (Fig.3).
Changes in neuromodulation have played recurrent roles in the
evolution of social behaviours in rodents. One of the best studied
examples is the differences between polygamous meadow voles and
the closely related monogamous prairie voles. The expression pattern
of the neuromodulator argininevasopressin (AVP) is conserved in the
brain of these two species. However, the expression pattern of the AVP
receptor has diverged in prairie voles through changes to the proximal
Fig. 2 | Evolution of neural circuits through connectivity changes. a, An
example of divergent evolution of neural circuits through connectivity changes
leading to divergent behaviour. The diagrams show a set of homologous cells in the
pharyngeal circuits of two nematode species, Caenorhabditis elegans (a bacterial
feeder) and Pristionchus pacificus (a predatory feeder)50. Synaptic connectivity —
represented by lines that are coloured to match the postsynaptic cell and with
a thickness that reflects the number of synaptic connections between cells —
has markedly diverged across both species. For example, neuron I2 (indicated
by the arrow) is an interneuron in C. elegans but synapses with several pharyngeal
muscles in P. pacificus. b, Another example of divergent evolution of both
circuits and behaviour, illustrating the evolution of peripheral versus central
parts of neural circuits. The left panel shows the elements of circuits governing
sexual behaviour that have been identified to change between two Drosophila
species. Drosophila melanogaster and Drosophila simulans males court
females of only their own species. The sensory neurons that express the Ppk23
ion channel sense the same compound (a D. melanogaster female-specific
pheromone) in both species. In D. melanogaster males, activation of these
neurons induces courtship behaviour towards conspecific females, whereas in
D. simulans it inhibits interspecific courtship. This difference is due to changes
in the circuits upstream of P1 neurons, which are responsible for sensory
integration. Specifically, there is a difference in the balance of excitation and
inhibition that these neurons receive in response to the activation of Ppk23-
expressing neurons4. The right panel shows evolutionary changes that have
occurred in parallel (affecting different pathways) and in series (affecting
the same pathway) in drosophilid olfactory pathways. Drosophila sechellia
diverged from its relatives D. simulans and D. melanogaster as it evolved a
specialized preference for noni fruits, which are toxic to the other two species.
This specialization involved changes within multiple sensory pathways. These
included the Ir75b pathway, in which the olfactory receptor changed sensitivity
but there were no detectable associated changes to the central circuits11, and
the Or22a pathway, in which both the sensitivity of the olfactory receptor
and the morphology of second-order projection neurons evolved in D. sechellia.
The morphological change involves an additional branching of the axonal
terminals (pictured in red), which is hypothesized to synapse with different
downstream partner neurons13. c–e,Examples of cases in which behaviours are
conserved across species while the underlying circuit connectivity diverges.
Part c shows an example of neural circuit drift. The sea slugs Melibe leonine
and Dendronotus iris both display left–right alternation in their swimming
pattern. While the same set of homologous neurons (Si1 and Si2, in the right
and left hemis pheres; that is, L-Si1 and R-Si1, and L-Si2 and R-Si2) controls this
behaviour across both species, the connectivity among them, illustrated in
the circuit diagrams, has diverged. The thickness of the lines indicates the
strength of the connections as measured electrophysiologically84. Also shown
are the phylogenetic relationships between these species and some of their
relatives as well as information on their swimming patterns10,86. Part d shows an
example of convergent evolution. The sea slugs Pleurobranchaea californica and
Tritonia diomedea display a dorsoventral swimming pattern; however, in this
case their phylogeny suggests that they acquired this behaviour via convergent
evolution because many of their relatives do not swim. In both species,
swimming behaviour is controlled by partially overlapping sets of homologous
neurons (shown in the same colour across species)10,137, but these are connected
differently. The IVS neuron has been hypothesized to exist in P. californica on the
basis of electrophysiological recordings of other neurons, but has never been
found (indicated by the dashed lines). Connections with both an arrowhead and
a blunt end represent mixed inhibitory and excitatory synapses. Part e illustrates
changes in connectivity that have evolved via natural selection to maintain a
conserved function in face of a changing feature, in this case a difference in eye
size. This example differs from circuit drift because the connectivity changes
were driven by the need to adapt to a changing feature of the animal, rather than
occurring by chance. The top panel shows how the different sizes of the mouse
and the rabbit eyes affect the speed at which an image travels across the retina for
the same displacement in the outside world. The bottom panel shows a schematic
representation of the distribution of input synapses onto the dendritic arbour
of a starburst amacrine cell (a key retinal interneuron for the computation of
direction selectivity) in mice and rabbits. In mice, inhibitory synaptic inputs are
restricted to the proximal portion (near the cell body) of the dendrites, while in
rabbits they are more homogeneously distributed across the dendritic arbour77.
The higher degree of segregation in mice generates more robust direction
tunning for objects traversing the retina more slowly, in line with the slower
relative displacement of external objects on their smaller eyes. d, distance; DV,
dorsoventral swimmers; LR, left–right swimmers; NS, non-swimmers; t, time;
IVS, AS, A1, A3, A10, DSI, VSIand C2 are the names of neurons of the central
pattern generator circuits of the illustrated nudibranches. Part a adapted with
permission from ref.50, Elsevier. Parts c and d adapted with permission from ref.10,
PNAS. Part e adapted from ref.138 and ref.77, Springer Nature Ltd.
Nature Reviews Neuroscience
Review article
a Change in receptor expression
1 2
3 4
1 2
3 4
Neuromodulator
Degrading enzymes
Receptor
1 2
3 4
1
2
3 4
1
b Change in neuromodulator expression
c Complex changes: increased neuromodulatory neurons and decreased catabolism
1 2
3 4
1 2
3 4
Avpr1a
Meadow vole (polygamous) Prairie vole (monogamous)
Avp
Deer mouse (polygamous) Oldfield mouse (monogamous)
Surface fish Cave fish
Expression level of AVP
Expression level of AVPR1A
Level of serotonin
Ventral
pallidum
Hypothalamus
Lateral
septum
Decreased
expression in
hypothalamus
Nest building
Avpr1a
Changes in proximal
regulatory region
Avp
Unknown changes
in regulatory region
Hypothalamus
Increased
expression
in ventral
pallidum
MAO mutation (reduced
serotonin catabolism)
More serotonin
neurons
Brain volume
differences
• Solitary feeding
• Decreased sleep
• Eye loss
Schooling behaviour
Fig. 3 | Evolution of neural circuits through changes in neuromodulation.
Evolutionary changes to neuromodulatory systems may involve changes in
the expression of receptors or neuromodulators or more complex changes.
To the left in every panel there is a schematic illustrating these processes. The
numbered triangles represent neurons or neuronal populations, with the same
number indicating homologous neurons. a, An example of evolution through
changes in the expression of receptors for neuromodulators. Differences in
the social behaviour of polygamous meadow voles and monogamous prairie
voles are due, at least in part, to the evolution of differences in the expression
of argininevasopressin V1a receptor (AVPR1A), including increased expression
in the ventral pallidum and decreased expression in the lateral septum in
prairie voles. These expression changes are due to modifications in the
proximal enhancer of the receptor93,94. b, An example of evolution through
changes in expression of neuromodulators. Differences in nest-building
behaviour (associated with parental care) between deer mice and oldfield
mice rely on decreased hypothalamic expression of the neuromodulator
argininevasopressin (AVP)95. c, An example of evolution through more
complex changes in neuromodulatory systems. Cave fish populations of
Mexican tetra have evolved extreme behavioural modifications, together
referred to as ‘cave behavioural syndrome’, which include loss of schooling
behaviour, solitary feeding and reduced sleep. These changes rely on multiple
modifications in diverse neuromodulatory systems, the most prominent of
which is the serotonin system. The changes include increases in the number
of serotonergic neurons in certain hypothalamic nuclei96,139 and, in some
populations, a mutation in an enzyme that catabolizes serotonin7. Together
these increase serotonergic levels in the brain of cave fish populations.
MAO, monoamine oxidase.
Nature Reviews Neuroscience
Review article
enhancer of its encoding gene. Importantly, this expression shift has
been causally linked to the behavioural change from polygamy to
monogamy93,94 (Fig.3a).
Intriguingly, the evolution of the AVP pathway was also linked to
divergent parental behaviour in an unbiased quantitative trait locus
study (see Box2) examining another pair of rodent species — the pro-
miscuous deer mouse and the monogamous oldfield mouse; however,
in this case, changes in the expression of AVP itself underlie a change in
nest-building behaviour
95
(Fig.3b). Thus, changes in either neuro-
modulator expression or neuromodulator receptor expression can
contribute to the evolution of divergent social behaviours.
In some cases, multiple changes in different aspects of neuromodu-
lation can work together to generate extremely divergent behavioural
phenotypes. The evolution of cave forms of the fish Mexican tetra from
related surface populations involved a suite of behavioural modifica-
tions that have been linked to diverse changes in neuromodulatory
systems. These changes include expansions in the number of neurons
in certain hypothalamic nuclei, leading to increased release of the
neuromodulator serotonin — which has been linked to the reduced
aggression of cave forms
96
. Additionally, some cave fish populations
independently evolved a hypomorphic variant of the gene encoding
the enzyme monoamine oxidase, which reduces the catalysis of sero-
tonin, thereby effectively increasing serotonin concentration
7
(Fig.3c).
The interplay between modulatory systems is evident from studies of
the reduced sleep phenotype of cave fish; inhibition of either hypo-
cretin signalling97 or dopaminergic signalling98 can partially restore
Box 2
A geneticist’s approach to neuronal evolution
Studying the evolution of neural circuits and behaviour can be
approached from a neurocomparative perspective that starts by
asking how neural circuits dier and follows this with a search
for the genetic basis of those changes. However, another way to
approach the problem is to ind genetic loci underlying behavioural
dierences and then identify their place of action within the brain.
Four main approaches have been used:
Quantitative trait locus analysis
In quantitative trait locus (QTL) analysis, two parental populations
(from the same or dierent, but interbreedable, species) with a
phenotypic dierence are crossed with each other, and their progeny
are backcrossed to the parental population or to themselves.
Through recombination, the genomes of the two parental
populations are scrambled together in the gametes of the hybrid
progeny. By simultaneously examining the phenotypes of individuals
of the backcrossed progeny and sequencing their genomes, one can
associate the presence of particular parental genetic loci with the
phenotype of interest. The challenge with this approach is to narrow
down the large genetic regions often revealed by QTL analysis to
speciic genes and to test their function. Limitations of QTL analysis
include the requirement for a large sample size, the need for easily
quantiiable phenotypes, the ability to experimentally cross parental
populations (not feasible for large animals or humans) and the fact
that it evaluates only alleles present in the two parental populations
(ignoring other variants in the natural population). Two examples
in which QTL analysis uncovered the genetic basis of behavioural
evolution are the study of cross-species ly courtship behaviours5
and the evolution of parental care in deer mice95.
Genome-wide association studies
This method also aims to ind genetic variants underlying a speciic
and quantiiable phenotypic dierence and requires a large sample
size. Unlike QTL analysis, genome-wide association studies (GWAS)
rely on existing variation in natural populations to link phenotype
to genotype, and typically uncover individual unlinked genes or
short fragments of nucleotides. Another advantage of GWAS is that,
once an individual or isogenic line has been sequenced, it is usually
possible to perform further phenotyping (see examples of studies
in Drosophila melanogaster145–147). However, drawing meaningful
associations between genetic variants and phenotypes can be
diicult, and requires a multitude of computationally expensive
statistical tests with corrections to exclude false positives. A great
example of the application of GWAS was the identiication of the
genetic underpinnings of gait diversity in Icelandic horses148.
Study of candidate evolutionary genes
Another method to identify genes underlying neural circuit evolution
is to use comparative genomics and population genetics to identify
loci with signatures of evolutionary change. For example, a study in
D. melanogaster curated a list of brain-expressed genes that arose
less than 25 million years ago. The study authors found a burst of
‘brain genes’ formed between three million years ago and six million
years ago and showed that they have strong signatures of positive
selection. Follow-up functional studies on two of these new genes
showed that they were important for foraging behaviour149. Similar
approaches have been used to understand some of the unique
features of the human brain66,68,150–153.
Experimental evolution
Animals can be directly evolved by exposing subsequent
generations of one population to new environmental conditions. By
keeping a sibling population in the initial ancestral conditions, one
can compare the phenotypic and genetic traits of the siblings154.
Traditionally, this has been used to ind the conditions under which
a trait is evolvable155. However, some studies have combined
experimental evolution with transcriptomic analysis of the evolved
populations156,157. More recently, experimental evolution and whole-
genome sequencing were brought together in animals in a method
coined ‘evolve and resequence’158–160. However, experimental
evolution also has many caveats, the most important of which
include the large scale of the experiments and the strong eect of
the genetic diversity of the starting population. For a more detailed
discussion of this topic, we refer the reader to refs.155,160.
Nature Reviews Neuroscience
Review article
Photoreceptor disc
1
2
1
2
1 1
1 1
a Duplication and divergence
b Change in RNA editing
1 1
d Change in gene expression
c Change in splicing
All vertebrates All teleost fish
Nav1.4
muscle expression
N C
Nav1.4b
muscle expression
Nav1.4
duplication Nav1.4a
muscle expression
Electric fish
Mormyroidea and Gymnotiformes
Nav1.4b
muscle expression
Nav1.4
divergence Nav1.4a
electric organ
expression
Nautiloids
Nautiloids
Coleoids
Coleoids
Kv2 potassium
channel
RNA editing
Exon inclusion
Slowpoke gene
D. mauritiana
N
C
D. simulans
Exon exclusion
N
C
Retroelement
ZebrafishPeripheral retina
Central retina
Primates
Opsin
Transducin
Peripheral retina Central retina
UV cones Green/red cones
Increased
behavioural
complexity
Altered courtship
song frequency
Nature Reviews Neuroscience
Review article
surface-like sleep patterns in cave fish. Furthermore, a recent single-
cell sequencing study of the hypothalamus of surface and cave forms
found that while most neuronal clusters are common between them, the
clusters that are the most transcriptionally different contain neuropep-
tidergic cells99. Interestingly, that study also found cave-lineage specific
gains and losses of cellular clusters when compared with the surface
morph, despite their close relatedness, pointing to an accelerated rate
of evolution in the brain of cave fish99.
Evolution of neurons
Neurons are the building blocks of neural circuits, and a change in the
neurons themselves is therefore perhaps the most fundamental way
in which neural circuits can evolve.
Changes in neuronal physiology
The physiological properties of neurons underlie the computations that
they perform and are determined, in part, by the ion channels present
in the membranes of different neuronal compartments. Changes at
the level of neuronal physiology are potentially an ‘easy’ way for evo-
lution to alter neuronal circuits, given the low pleiotropy that arises
due to the modular nature of ion channels (which allows functional
diversification through alternative splicing and altered subunit com-
position) and their compartmentalized expression in specific neurons
(or even neuronal compartments). A number of detailed comparative
analyses have shown correlation, and sometimes causation, between
alterations in neuronal physiological properties — including changes
in the expression of ion channels, channel properties or components
of downstream signalling pathways — and behavioural changes across
species. Examples are drawn from fish, frogs, cephalopods, primates
and flies
5,100–102
, and are therefore suggestive of principles universal to
neuronal evolution (Fig.4).
Initial evidence for this mode of evolution came from electrophysi-
ological studies. Differences in the courtship song of two species of
frog, Xenopus laevis and Xenopus petersii, have been associated with
differences in the intrinsic properties of neurons in the hindbrain cir-
cuits that generate vocalization patterns100,101. Changes in the intrinsic
properties of central neurons were also associated with intraspecific
variation in foraging behaviour among the larvae of D. melanogaster
103
.
In this case, ion channel expression seems to be regulated by the protein
kinase Foraging (For)
104
. Two alterative alleles of the for gene are main-
tained in the population through balancing selection, leading to larvae
displaying two alternative foraging strategies105. Subsequent studies
found that for is also associated with several behavioural phenotypic
differences in adult D. melanogaster
106
and other species
107–109
. The
mechanism of action of for in central neural circuits remains unclear,
but recent evidence suggests that it might be, at least in part, mediated
by glia
110,111
, underscoring the often forgotten role of glia in neuronal
function, and perhaps also in neural circuit evolution.
Neuronal physiology can also evolve through the duplication and
divergence of the genes that encode specific ion channels (Fig.4a).
Although there are currently no reported examples of this mechanism
in the evolution of central circuits, its prevalence in the evolution of
peripheral sensory
11
and motor circuit
102
components makes it likely
to also be a mechanism in the evolution of central circuits. An exam-
ple of its contribution to peripheral circuit evolution is provided by
the emergence of electrogenesis in two families of weakly electric
fish that has occurred through the duplication and divergence of the
voltage-gated ion channel Nav1.4 (refs.102,112,113) (Fig.4a).
Evolution can also diversify animals’ repertoire of ion channels
— and other proteins — through changes inRNA editing. For example,
coleoid cephalopods, such as octopuses, cuttlefish and squid, display
a dramatic increase in RNA editing within the nervous system when
compared with their nautiloid relatives, which is thought to have con-
tributed to their behavioural diversification. Indeed, coleoid-specific
RNA editing of the delayed rectifier potassium channel Kv2 has been
shown to alter its inactivation and closing rates
114
(Fig.4b). Beyond this
example, nothing is known about how individual editing sites affect
neural circuits, yet their prevalence in genes known to be important
Fig. 4 | Evolution of neuronal circuits through changes in neuronal
physiology. Aspects of neuronal physiology might evolve easily owing to the
compartmentalized nature of ion channels. Such evolutionary changes can
come about in various ways, including changes in the expression of alleles
of ion channel genes within a population, duplication and divergence of ion
channel genes, changes in RNA editing, changes in ion channel splicing and
changes in the expression of ion channels or other molecules determining
neuronal physiology. To the left in each panel there is a schematic illustrating
these processes. The numbered triangles represent neurons or neuronal
populations, with the same number indicating homologous neurons.
a, Like other proteins, ion channels can evolve through gene duplication and
divergence, whereby the gene encoding the channel is duplicated and one of
the copies then evolves to produce a protein with new functionality and/or a
new expression pattern. Although there are currently no reported examples of
this mechanism contributing to the evolution of central circuits, its prevalence
in the evolution of both sensory and motor systems hints at its generality in the
evolution of neural circuits. The ion channel Nav1.4, present in all vertebrates,
was duplicated in all teleost fish, where it is expressed in muscle. In two families
of weakly electric fish, one of the two duplicated channels (Nav1.4a) altered
its expression pattern (becoming expressed in the electric organ and losing its
expression in the muscles) and evolved its protein sequence in domains that
alter channel innactivation140, enabling the diversification of communication
signals. The other ion channel orthologue, Nav1.4b, maintained its essential
muscle functions in both families102. Interestingly, in both families these changes
occurred via convergent evolution. b, RNA editing enables the functional
diversification of ion channels (and other proteins) without duplications or other
changes to their genomic sequence. In cephalopods, behavioural complexity is
associated with increased RNA editing in the nervous system; coleoids, which
display greater behavioural complexity than their simpler nautiloid relatives,
also exhibit increased RNA editing114. The observed RNA editing has functional
consequences. For example, editing of the Kv2 potassium channel in coleoids
alters its kinetics. This is illustrated in the middle panel, which shows differing
membrane depolarizations upon a fixed current injection in oocytes expressing
a Kv2 channel with (red trace) and without (grey trace) the modification that is
introduced by RNA editing in coleoids. However, how the functional changes
in ion channel properties could lead to changes in central circuits that govern
increased behavioural complexity remains unknown114. c, Changes to alternative
splicing have also been shown to be key in the evolution of neural circuits. For
example, differences in the splicing of RNA encoding ion channels can confer
novel properties on the channels. Drosophila mauritiana and Drosophila
simulans display differences in the frequency of their courtship song. These
differences are caused by the insertion of a retroelement in the gene that encodes
the potassium ion channel Slowpoke that alters the inclusion of a small exon
in the final transcript5. Interestingly, slowpokeis widely expressed throughout
the central nervous system of flies, and its deletion in D. simulans produces
dramatic phenotypes, as opposed to the subtle and specific behavioural changes
induced by the evolutionary selected modification5. d, Evolution may also
involve changes in the expression of key physiological components. For example,
both zebrafish and primates have evolved to exhibit increased sensitivity of
the cones in the central part of their retina through what are thought to be
convergent changes in the expression of transducin, a protein involved in
phototransduction76,118. Although the ancestral state is not known in this case,
in fish, central retina UV cones have a light-biased photoresponse, while in the
primate fovea red/green photoreceptors have slower kinetics, both strategies
increase signal integration in the cones of the central retinas of each animal76,118.
C, carboxy terminus; N, amino terminus. The trace in part b is adapted, with
permission, from ref.114, Elsevier. The traces in part d are adapted, with permission,
from refs.76,118, Elsevier.
Nature Reviews Neuroscience
Review article
for neural circuit function and assembly, such as those encoding
protocadherins114,115, suggests widespread roles across coleoid
central circuits.
Although the compartmentalized expression and functional
specialization of many channels increase theirevolvability, some
channels are more widely expressed and therefore likely to be highly
pleiotropic. Surprisingly, a study in flies showed that subtle changes
in one such pleiotropic channel underlie the behavioural evolution of
divergent male courtship songs in D. simulans and Drosophila mau-
ritiana. Quantitative trait locus analysis (Box2) followed by elegant
genetic mapping experiments revealed the cause of the behavioural
divergence to be aretroelement insertion in an intron of the slowpoke
gene, which encodes a calcium-activated potassium channel
5
. The
insertion reduces the inclusion of the nearby exon, suggesting that
the behavioural evolution was caused by modifications in alternative
splicing
5
(Fig.4c). The specific neurons affected by this change are
yet to be identified. Intriguingly, slowpoke is widely expressed in the
fly nervous system, and its deletion in D. simulans leads to severely
disrupted song patterns, unlike the subtle differences caused by its
disruption by theretroelement insertion5. This highlights how the evo-
lution of the regulation of a pleiotropic channel can result in discrete
behavioural changes.
The physiological properties of neurons might also be altered by
changes in other molecular components. In vertebrate vision, pho-
totransduction is one key target for such molecular tuning. In addition
to changes in opsins — the light-receptor proteins present in retinal
photoreceptors (reviewed elsewhere
116,117
), one hot spot for evolution-
ary change appears to be transducin, the messenger that links the
opsin’s photoresponse to downstream effectors. In both primates
118
and zebrafish
76
, the expression levels of transducin differ between
the cone photoreceptors located centrally and peripherally within
the retina. This expression pattern must have arisen independently
in the two lineages as it occurs in different cone types in each species.
Intriguingly, these expression differences correlate with functional
strategies to aid signal integration: slower kinetics in the primate foveal
cones118 and a light-biased photoresponse in zebrafish UV cones76
(Fig.4d). Interestingly, zebrafish central UV cones alsoshow slowed
kinetics like the primate red/green cones; however, in this case, this
effect is achieved through feedback from horizontal cells76. This is thus
a great example of how convergent evolution can generate similarly
functioning neurons through diverse mechanisms.
Changes in the numbers and types of neurons
One dramatic way in which neural circuits can evolve is through changes
in their neuronal composition. This can involve changes in the number
of neurons of a particular type, the evolution of new types of neurons
or a combination of both.
The evolutionary emergence of new neuron types is difficult to
track because it is hard to detect an unknown neuron type for which we
have no markers. It is also unclear when a neuron present in an ancestor
has diverged enough from its original characteristics to be considered
a novel type, nor it is straightforward to track the ancestry of novel neu-
ron types (Supplementary information). However, single-cell sequenc-
ing is proving to be a particularly useful method to identify neuron-type
innovations across species. For example, a recent article reported the
discovery of a novel neuron type present only in the primate striatum
71
that expresses unique combinations of transcription factors and termi-
nal differentiation genes and is extremely abundant71. This evolutionary
pattern is also evident in the eye: single-cell transcriptomics studies
of the retina of chickens, mice, macaques and humans have revealed
conservation of the main retinal classes, but divergence of the types
within classes
83
. A study taking a similar approach by comparing the
hypothalamus of Mexican tetra and zebrafish found that species-
specific cell types exhibited enriched expression of species-specific
genes, indicating that cellular novelty was driven by the evolutionary
emergence of new genes99. Furthermore, that study also showed that
terminal differentiation genes are more conserved thantranscription
factor genes in homologous cell types, suggesting genetic drift of
the gene regulatory networks. However, it is also possible that these
neuronal populations are not homologous but have convergently
evolved to express similar sets of terminal differentiation genes. This
highlights the need to identify homologous cell types on the basis of
multiple features99 (Supplementary information).
Cell numbers might evolve through changes in proliferation or cell
death. An example of the former mechanism is provided by the expan-
sion in the number of flight interneurons that has occurred during the
evolution of winged insects
119
. Similarly, the neocortical expansion
of the primate brain has been proposed to have occurred mostly
through the expansion of progenitor cells120. Comparative transcrip-
tomics analysis of neocortical progenitors in mice and humans uncov-
ered a hominid-specific gene that arose through partial duplication
of an ancestral gene and that is likely to have contributed to this evo-
lutionary expansion
121
. Interestingly, the key functional difference
between the duplicated hominid gene — which experimentally does
not induce progenitor proliferation when inserted into mice — and
the modern human gene — whose expression in mouse developing
neocortex induces an increase in progenitor proliferation — is a single-
nucleotide change that alters splicing122. This provides another example
of evolution of central brain circuits through gene duplication and
divergence, and highlights the dramatic evolutionary consequences
that can emerge from single-nucleotide changes (Fig.5a).
Changes in the patterns of developmental cell death seem to under-
lie the evolution of locomotor and olfactory circuits in insects
123,124
.
Peripheral olfactory circuits in insects have different numbers of olfac-
tory sensory neurons, and in D. melanogaster artificially blocking
developmental cell death in the sensory neuron precursors leads to
the formation of novel olfactory sensory neurons with characteris-
tics akin to those found in other species
124,125
(Fig.5b). Interestingly,
these neurons integrate into existing central neural circuits, induc-
ing the creation of novel central structures (new glomeruli within
the first olfactory relay centre). Together with findings from other
studies in flies
126
and mice
127
, this highlights how central neural cir-
cuit plasticity can accommodate changes to presynaptic or post-
synaptic neuronal populations, and in turn potentially facilitate
their evolution.
Functionally testing the behavioural consequences of changes
in neuron numbers across species is challenging because it requires
specific developmental manipulations that do not interfere with overall
brain development. Elegant experiments in Mexican tetra cave fish have
shown that an evolutionary increase in sonic hedgehog signalling dur-
ing early development had adaptive pleiotropic effects that led to eye
loss, expanded neuronal populations in selected hypothalamic nuclei
and reduced aggression compared with their surface-dwelling cous-
ins96. For some of the hypothalamic nuclei, the developmental effect
of sonic hedgehog is mediated via transcription factors of the LHX
family9. For example, the LHX9 expression pattern is expanded in cave
fish compared with surface forms, and its experimental depletion leads
to fewer hypocretin neurons in the hypothalamus and behavioural
Nature Reviews Neuroscience
Review article
activity levels identical to those of surface fish9. In the future, it will be
important to establish similar causal links between behavioural evolu-
tion and changes in neuron number and/or neuronal composition for
other clades.
Thus, despite their initial apparent complexity, changes in neu-
ron number and/or neuron types can result from simple genetic
modifications, making them, in principle, no less likely to evolve than
changes in connectivity or neuromodulation. How new cells integrate
within existing circuits and whether there are any general rules that
govern the interaction of evolutionary novel neuron types, connec-
tivity changes and neuromodulatory shifts are open questions for
the future.
a Increased progenitor proliferation
b Decreased developmental cell death
1 2 1 2
Mice
Chimpanzees
Ancient
hominids
Modern
humans
Denisovans
Neanderthals
Progenitor
cell
Neurons
Drosophila
Olfactory
sensory
neurons
Mosquito Drosophila + blockade of developmental
cell death
ARHGAP11A
Gene duplication
• ARHGAP11A
• ARHGAP11B
C>G mutation in
ARHGAP11B
ARHGAP11A
Arhgap11a
ARHGAP11B
ARHGAP11B
ARHGAP11A
Sensillum
Central brain
antennal lobe New glomerulus
in central brain
Extra
neuron
C>G leads to
differential splicing
Increased proliferation
of basal progenitors
Increased
self-renewal
Increased
production
of progeny
Fig. 5 | Evolution through changes in neuron number. Changes in neuron
numbers can come about through changes in neuronal proliferation and/or
developmental cell death, as illustrated in the schematics on the left of each
panel. a, In evolution, new neurons can arise through increased neuronal
proliferation, whereby progenitor cells divide faster or for longer periods to
self-renew or generate more progeny (that is, neurons). An example of this is the
expansion of the primate neocortex, which happened (at least in part) through
an increased proliferation capacity of basal progenitors (a subset of neural
progenitors that divide at a basal location within the developing cortex and that
have been associated with the evolutionary expansion of the neocortex)120,122,141.
The ARHGAP11A gene is partially duplicated in the hominid lineage to generate
a copy known as ARHGAP11B (ref.121). The ancestral ARHGAP11B copy present in
ancient hominids does not increase basal progenitor proliferation118. However, in
the lineage leading to modern humans, Neanderthals and Denisovans, this copy
underwent a C>G mutation that generated a novel splice site. This in turn altered
the carboxy-terminal domain of the protein, changing its function, leading to
increased proliferation of basal progenitors121,122. The phylogenetic tree shows
relative relationships, but the lengths of the arms do not indicate phylogenetic
distances. b, Increased neuron number may also arise through changes in
developmental cell death. Insect olfactory sensory neurons are organized inside
sensilla. Each sensillum can host up to four neurons, but one or two usually die
through programmed developmental cell death. All sensilla in adult Drosophila
melanogaster maxillary palps contain two neurons. In contrast, all sensilla in the
mosquito Aedes aegypti maxillary palps contain three neurons, one of which
senses CO2. Blocking developmental cell death in D. melanogaster leads to some
sensilla in the maxillary palps developing three neurons, including one neuron
with features similar to those of the mosquito CO2-sensing neuron. Interestingly,
these new neurons integrate into existing central neural circuits, inducing the
creation of novel central structures (glomeruli)124.
Nature Reviews Neuroscience
Review article
Emerging principles
Recent technological developments are enabling us to start addressing
how central neural circuits evolve at the genetic, molecular, cellular and
network levels, and it is becoming clear that all of the mechanisms
and principles highlighted in this Review are at play. In addition,
some partial answers to the questions we posed at the beginning of
the Review are beginning to emerge (see below). However, there are
still many open questions that will be answered only by the study of
complete neural circuits from the sensory periphery to motor output
and by the establishment of experimental tools in additional species
at key phylogenetic distances (Box1).
Question 1: Is the sensory periphery more evolvable than central
circuits? Despite long-standing beliefs that central neural circuits
are less evolvable than their peripheral counterparts92,128,129, it is obvi-
ous from the examples discussed in this Review that central neural
circuits change, sometimes over very short evolutionary times
4,9,99
.
Furthermore, these central changes can emerge through the simplest
of genetic modifications7,122, an indication that central circuits are more
evolvable than initially expected. As seen in the evolution of sexual
circuitry in drosophilids, it is likely that both peripheral circuits and
central circuits evolve as species change their behavioural output.
Whether one of these changes is more prevalent than the other will
be answered only by broadening the work in this areaby studying a
larger number of complete circuits, from the sensory periphery to
motor output. Small animals, such as drosophilids and nematodes, are
particularly well positioned to meet this challenge (Box1).
Question 2: Do some types of evolutionary change happen more
frequently than others? We currently lack enough data to address the
question of whether some elements of neural circuits change more
often than others over evolutionary time, or whether certain elements
are selected under certain conditions but not others. Although changes
in neuron number and type might intuitively seem to be genetically
complex, and therefore relatively unlikely to occur, we have shown that
such changes can be mediated by single-nucleotide substitutions
122
.
Another feature of neural circuits that might facilitate their evolution
via what might otherwise be expected to be unlikely mechanisms is their
plasticity, which can enable modifications to individual components
within a complex network to persist by facilitating their accommoda-
tion until further genetic modifications make these stable. Evidence
supporting this idea comes from developmental experiments in both
vertebrates and invertebrates which have shown remarkable flexibility
in the wiring of their neural circuits upon perturbation123,124,126,127,130–132.
Question 3: Do neural circuits have hot spots for evolutionary
change, or are changes homogeneously distributed across the net
-
work? This question will be best answered by future studies that are
able to completely and systematically evaluate entire neural circuits by
combining comparative connectomics and functomics (for example,
see the Retinal Functomics website). Here again, animals with small,
accessible brains promise to provide the first insights (Box1). So far,
however, evidence suggests that indeed hot spots for change might
exist. Perhaps the clearest example is the repeated involvement of the
vertebrate preoptic area in the evolution of parental behaviours41,93–95.
Another suggested hot spot is the P1 neurons present in the courtship
circuit of flies4. The central position of these neurons as information
integrators and the direct influence of their output on behaviour means
that changes in their physiology or connectivity can profoundly affect
information processing.
Question 4: Do all central circuits evolve according to similar
rules, or does the neural circuit blueprint influence how neural circuits
evolve? A related question is how much can we learn about the evo-
lution of neural circuits when studying species with smaller brains.
Smaller brains, such as those of Drosophila species and nematodes, are
more stereotyped and develop from lineagesthat are more fixed than
those ofvertebrate brains, whose development is thought to be more
flexible. Furthermore, smaller brains have lower neuronal redundancy.
Thus, it is reasonable to question whether certain evolutionary paths
are facilitated in one kind of brain versus the other. For instance, work
in nematodes has shown extensive rewiring of neural circuits without
changes in neuronal composition: could this be due to their numerical
simplicity and fixed lineages? Work in insect nervous systems, which,
although numerically more complex, also develop through stereo-
typed neuronal lineages, suggests that this might not be the case. The
number and type of neurons generated by insect progenitors have
been shown to be evolutionary labile123,124,133, indicating that develop-
mental stereotypy does not constrain the generation of evolutionary
novel neuron types in insect brains. More work is needed to under-
stand whether certain circuit features favour particular evolutionary
mechanisms, but so far similar patterns have been observed across the
animal kingdom.
Future perspectives
Beyond answering the fascinating question of how central circuits
evolve, understanding their evolutionary history (Supplementary infor-
mation) can help us todistinguish, for example, whether their similari-
ties are conserved or have evolved convergently. This has particularly
important consequences for our understanding of which features of
central circuits are adaptive, and thus have evolved independently
multiple times, which are the fruit of developmental or functional
constraints or have been maintained through purifying selection, and
which have diverged purely due to neuronal drift while the selectable
neuronal output has remained conserved.
To keep progressing in our understanding of neural circuit
evolution, we must combine modern techniques such as genetic
mani pulation, connectomics, functional imaging and single-cell tran-
scriptomics in a wide variety of species that are separated by particular
phylogenetic distances (see Box2 and ref.
134
). It will also be impor-
tant to bring these methods together with ecological and population
genetics studies to understand the factors that drive behavioural
evolution. Going forward, a truly multidisciplinary approach will be
the best way to understand how neural circuits in the brain change over
evolutionary time.
Published online: xx xx xxxx
References
1. Knudsen, E. I. & Konishi, M. Mechanisms of sound localization in the barn owl (Tyto alba).
J. Comp. Physiol. 133, 13–21 (1979).
2. Sperry, R. W. Eect of 180 degree rotation of the retinal ield on visuomotor coordination.
J. Exp. Zool. 92, 263–279 (1943).
3. Laurent, G. On the value of model diversity in neuroscience. Nat. Rev. Neurosci. 21,
395–396 (2020).
4. Seeholzer, L. F., Seppo, M., Stern, D. L. & Ruta, V. Evolution of a central neural circuit
underlies Drosophila mate preferences. Nature 559, 564–569 (2018).
Landmark article showing that species-speciic pheromone responses can emerge
through conservation of peripheral sensory neuron responses and changes in central
circuitry.
5. Ding, Y., Berrocal, A., Morita, T., Longden, K. D. & Stern, D. L. Natural courtship song
variation caused by an intronic retroelement in an ion channel gene. Nature 536,
329–332 (2016).
6. Ding, Y. et al. Neural evolution of context-dependent ly song. Curr. Biol. 29, 1089–1099.e7
(2019).
Quantitative trait locus study identifying the genetic bases of divergence in courtship
songs between two Drosophila species.
Nature Reviews Neuroscience
Review article
7. Elipot, Y. et al. A mutation in the enzyme monoamine oxidase explains part of the
Astyanax caveish behavioural syndrome. Nat. Commun. 5, 3647 (2014).
8. Yoshizawa, M., Gorički, Š., Soares, D. & Jeery, W. R. Evolution of a behavioral shift
mediated by supericial neuromasts helps caveish ind food in darkness. Curr. Biol. 20,
1631–1636 (2010).
9. Alié, A. et al. Developmental evolution of the forebrain in caveish, from natural variations
in neuropeptides to behavior. Elife 7, e32808 (2018).
10. Newcomb, J. M., Sakurai, A., Lillvis, J. L., Gunaratne, C. A. & Katz, P. S. Homology and
homoplasy of swimming behaviors and neural circuits in the Nudipleura (Mollusca,
Gastropoda, Opisthobranchia). Proc. Natl Acad. Sci. USA 109, 10669–10676 (2012).
Review article on the evolution of swimming circuits in nudibranchs.
11. Prieto-Godino, L. L. et al. Evolution of acid-sensing olfactory circuits in drosophilids.
Neuron 93, 661–676 (2017).
12. Prieto-Godino, L. L. et al. Olfactory receptor pseudo-pseudogenes. Nature 539, 93–97 (2016).
13. Auer, T. O. et al. Olfactory receptor and circuit evolution promote host specialization.
Nature 579, 402–408 (2020).
14. Hart, N. S. Vision in sharks and rays: opsin diversity and colour vision. Semin. Cell Dev.
Biol. 106, 12–19 (2020).
15. Yokoyama, S. & Yokoyama, R. Adaptive evolution of photoreceptors and visual pigments
in vertebrates. Annu. Rev. Ecol. Syst. 27, 543–567 (1996).
16. Prieto-Godino, L. L., Schmidt, H. R. & Benton, R. Molecular reconstruction of recurrent
evolutionary switching in olfactory receptor speciicity. Elife 10, 69732 (2021).
17. Baldwin, M. W. et al. Evolution of sweet taste perception in hummingbirds by
transformation of the ancestral umami receptor. Science 345, 929–933 (2014).
18. Marques, D. A. et al. Convergent evolution of SWS2 opsin facilitates adaptive radiation of
threespine stickleback into dierent light environments. PLoS Biol. 15, e2001627 (2017).
19. Bowmaker, J. K. Evolution of vertebrate visual pigments. Vis. Res. 48, 2022–2041 (2008).
20. Herron, J.C. & Freeman, S. Evolutionary Analysis 5th edn (Pearson Education, 2015).
21. Futuyma, D. J. Evolution 3rd edn (Sinauer Associates, 2013).
22. Lynch, M. & Hill, W. G. Phenotypic evolution by neutral mutation. Evolution 40, 915 (1986).
23. Zhang, J. Neutral theory and phenotypic evolution. Mol. Biol. Evol. 35, 1327–1331 (2018).
24. Arguello, J. R. et al. Extensive local adaptation within the chemosensory system
following Drosophila melanogaster’s global expansion. Nat. Commun. 7, 11855 (2016).
25. Long, M., Betrán, E., Thornton, K. & Wang, W. The origin of new genes: glimpses from the
young and old. Nat. Rev. Genet. 4, 865–875 (2003).
26. Long, M., Vankuren, N. W., Chen, S. & Vibranovski, M. D. New gene evolution: little did we
know. Annu. Rev. Genet. 47, 307–333 (2013).
27. Lynch, M. & Conery, J. S. The evolutionary fate and consequences of duplicate genes.
Science 290, 1151–1155 (2000).
28. Pollen, A. A. et al. Establishing cerebral organoids as models of human-speciic brain
evolution. Cell 176, 743–756 (2019).
29. Tosches, M. A. Developmental and genetic mechanisms of neural circuit evolution. Dev.
Biol. 431, 16–25 (2017).
30. Suryanarayana, S. M., Pérez-Fernández, J., Robertson, B. & Grillner, S. The evolutionary
origin of visual and somatosensory representation in the vertebrate pallium. Nat. Ecol.
Evol. 4, 639–651 (2020).
31. Chakraborty, M. & Jarvis, E. D. Brain evolution by brain pathway duplication. Philos. Trans.
R. Soc. B Biol. Sci. 370, 20150056 (2015).
32. Arendt, D. The evolution of cell types in animals: emerging principles from molecular
studies. Nat. Rev. Genet. 9, 868–882 (2008).
33. Nilsson, D.-E. & Arendt, D. Eye evolution: the blurry beginning. Curr. Biol. 18, 1096–1098
(2008).
34. Silbering, A. F. et al. Complementary function and integrated wiring of the evolutionarily
distinct Drosophila olfactory subsystems. J. Neurosci. 31, 13357–13375 (2011).
35. Couto, A., Alenius, M. & Dickson, B. J. Molecular, anatomical, and functional organization
of the Drosophila olfactory system. Curr. Biol. 15, 1535–1547 (2005).
36. Ramdya, P. & Benton, R. Evolving olfactory systems on the ly. Trends Genet. 26, 307–316
(2010).
37. Sukhum, K. V., Shen, J. & Carlson, B. A. Extreme enlargement of the cerebellum in a clade of
teleost ishes that evolved a novel active sensory system. Curr. Biol. 28, 3857–3863.e3 (2018).
38. Marder, E. & Prinz, A. A. Modeling stability in neuron and network function: the role
of activity in homeostasis. Bioessays 24, 1145–1154 (2002).
39. McLennan, D. A. The concept of co-option: why evolution often looks miraculous.
Evol. Educ. Outreach 1, 247–258 (2008).
40. Chapman, P. D. et al. Co-option of a motor-to-sensory histaminergic circuit correlates
with insect light biomechanics. Proc. R. Soc. B Biol. Sci. 284, 20170339 (2017).
41. Fischer, E. K. et al. Mechanisms of convergent egg provisioning in poison frogs.
Curr. Biol. 29, 4145–4151.e3 (2019).
Study on poison frogs showing that two dierent species evolved egg provisioning
convergently by independently recruiting the same brain area — the preoptic area —
but that the activity patterns in this area are dierent across the two species.
42. Tosches, M. A. et al. Evolution of pallium, hippocampus, and cortical cell types revealed
by single-cell transcriptomics in reptiles. Science 360, 881–888 (2018).
43. Colquitt, B. M., Merullo, D. P., Konopka, G., Roberts, T. F. & Brainard, M. S. Cellular
transcriptomics reveals evolutionary identities of songbird vocal circuits. Science 371,
eabd9704 (2021).
Single-cell transcriptomics study of the songbird motor pathway suggesting that the
mammalian neocortex and the dorsal ventricular ridge of birds and reptiles are not
homologous.
44. Storz, J. F. Causes of molecular convergence and parallelism in protein evolution.
Nat. Rev. Genet. 17, 239–250 (2016).
45. Magalhaes, I. S. et al. Intercontinental genomic parallelism in multiple three-spined
stickleback adaptive radiations. Nat. Ecol. Evol. 5, 251–261 (2021).
46. Stuart, Y. E. Divergent uses of “parallel evolution” during the history of The American
Naturalist. Am. Nat. 193, 11–19 (2019).
47. Arendt, J. & Reznick, D. Convergence and parallelism reconsidered: what have we
learned about the genetics of adaptation? Trends Ecol. Evol. 23, 26–32 (2008).
48. Elmer, K. R. & Meyer, A. Adaptation in the age of ecological genomics: Insights from
parallelism and convergence. Trends Ecol. Evol. 26, 298–306 (2011).
49. Torres-Méndez, A. et al. Parallel evolution of a splicing program controlling neuronal
excitability in lies and mammals. Sci. Adv. 8, eabk0445 (2022).
50. Bumbarger, D. J., Riebesell, M., Rödelsperger, C. & Sommer, R. J. System-wide rewiring
underlies behavioral dierences in predatory and bacterial-feeding nematodes. Cell
152, 109–119 (2013).
Electron microscopy-based connectomic reconstruction of the motor pharyngeal
system of the nematode P. pacificus and comparison with that of C. elegans.
The study reveals extensive rewiring of homologous neurons across the two species.
51. Hong, R. L. et al. Evolution of neuronal anatomy and circuitry in two highly divergent
nematode species. Elife 8, e47155 (2019).
52. Cook, S. J., Crouse, C. M., Hall, D. H., Emmons, S. W. & Hobert, O. The connectome of the
Caenorhabditis elegans pharynx. J. Comp. Neurol. 528, 2767–2784 (2020).
53. Suzuki, H. et al. Functional asymmetry in Caenorhabditis elegans taste neurons and its
computational role in chemotaxis. Nature 454, 114–117 (2008).
54. Johnston, R. J. & Hobert, O. A microRNA controlling left/right neuronal asymmetry in
Caenorhabditis elegans. Nature 426, 845–849 (2003).
55. Chesmore, K. N., Bartlett, J., Cheng, C. & Williams, S. M. Complex patterns of association
between pleiotropy and transcription factor evolution. Genome Biol. Evol. 8, 3159–3170
(2016).
56. Pérez-Escudero, A. & De Polavieja, G. G. Optimally wired subnetwork determines
neuroanatomy of Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 104, 1–6 (2007).
57. Chen, B. L., Hall, D. H. & Chklovskii, D. B. Wiring optimization can relate neuronal
structure and function. Proc. Natl Acad. Sci. USA 103, 4723–4728 (2006).
58. Barabási, D. L. & Barabási, A. L. A genetic model of the connectome. Neuron 105,
435–445 (2020).
59. Seth, R. et al. Molecular topography of an entire nervous system. Cell 184, 4329–4347.e23
(2021).
60. Auer, T. O. & Benton, R. Sexual circuitry in Drosophila. Curr. Opin. Neurobiol. 38, 18–26 (2016).
61. Ahmed, O. M. et al. Evolution of mechanisms that control mating in Drosophila males.
Cell Rep. 27, 2527–2536.e4 (2019).
62. Fan, P. et al. Genetic and neural mechanisms that inhibit Drosophila from mating with
other species. Cell 154, 89–102 (2013).
63. Khallaf, M. A. et al. Mate discrimination among subspecies through a conserved olfactory
pathway. Sci. Adv. 6, eaba5279 (2020).
64. Hong, W., Mosca, T. J. & Luo, L. Teneurins instruct synaptic partner matching in an
olfactory map. Nature 484, 1–9 (2012).
65. Martin-Pena, A. et al. Age-independent synaptogenesis by phosphoinositide 3 kinase.
J. Neurosci. 26, 10199–10208 (2006).
66. Fossati, M. et al. SRGAP2 and its human-speciic paralog co-regulate the development
of excitatory and inhibitory synapses. Neuron 91, 356–369 (2016).
67. Schmidt, E. R. E. et al. A human-speciic modiier of cor tical connectivity and circuit
function. Nature 599, 640–644 (2021).
68. Charrier, C. et al. Inhibition of SRGAP2 function by its human-speciic paralogs induces
neoteny during spine maturation. Cell 149, 923–935 (2012).
69. Tosches, M. A. & Laurent, G. Evolution of neuronal identity in the cerebral cortex.
Curr. Opin. Neurobiol. 56, 199–208 (2019).
70. Bakken, T. E. et al. Comparative cellular analysis of motor cortex in human, marmoset
and mouse. Nature 598, 111–119 (2021).
71. Krienen, F. M. et al. Innovations present in the primate interneuron repertoire. Nature
586, 262–269 (2020).
Single-cell transcriptomics study of the neocortex, hippocampus and striatum of ferrets,
mice, marmosets, macaques and humans. The study uncovers a human-speciic cell type.
72. Baden, T. Vertebrate vision: lessons from non-model species. Semin. Cell Dev. Biol. 106,
1–4 (2020).
Editorial in a special issue on the evolution of retinal circuits across vertebrates.
73. Schwartz, G. Retinal Computation (Elsevier, 2021).
74. Kim, T., Shen, N., Hsiang, J. C., Johnson, K. P. & Kerschensteiner, D. Dendritic and parallel
processing of visual threats in the retina control defensive responses. Sci. Adv. 6,
eabc9920 (2020).
75. Kölsch, Y. et al. Molecular classiication of zebraish retinal ganglion cells links genes
to cell types to behavior. Neuron 109, 645–662 (2021).
76. Yoshimatsu, T., Schröder, C., Nevala, N. E., Berens, P. & Baden, T. Fovea-like photoreceptor
specializations underlie single UV cone driven prey-capture behavior in zebraish. Neuron
107, 320–337.e6 (2020).
77. Ding, H., Smith, R. G., Poleg-Polsky, A., Diamond, J. S. & Briggman, K. L. Species-speciic
wiring for direction selectivity in the mammalian retina. Nature 535, 105–110 (2016).
Study in mice and rabbits showing that their retinal direction selectivity circuits are
dierentially wired in a way that compensates for the dierent eye sizes of the two
species.
Nature Reviews Neuroscience
Review article
78. Baden, T. & Osorio, D. The retinal basis of vertebrate color vision. Annu. Rev. Vis. Sci. 5,
177–200 (2019).
79. Lamb, T. D., Collin, S. P. & Pugh, E. N. Evolution of the vertebrate eye: opsins,
photoreceptors, retina and eye cup. Nat. Rev. Neurosci. 8, 960–976 (2007).
80. Nilsson, D. E. The diversity of eyes and vision. Annu. Rev. Vis. Sci. 7, 19–41 (2021).
81. Yoshimatsu, T. et al. Ancestral circuits for vertebrate color vision emerge at the irst
retinal synapse. Sci. Adv. 7, eabj6815 (2021).
82. Euler, T., Haverkamp, S., Schubert, T. & Baden, T. Retinal bipolar cells: elementary
building blocks of vision. Nat. Rev. Neurosci. 15, 507–519 (2014).
83. Yamagata, M., Yan, W. & Sanes, J. R. A cell atlas of the chick retina based on single-cell
transcriptomics. Elife 10, e63907 (2021).
84. Sakurai, A. & Katz, P. S. Artiicial synaptic rewiring demonstrates that distinct neural
circuit conigurations underlie homologous behaviors. Curr. Biol. 27, 1721–1734.e3 (2017).
Study suggesting that homologous circuits for swimming behaviour in two
nudibranch species evolved through neuronal drift.
85. Jing, J. & Gillette, R. Central pattern generator for escape swimming in the notaspid sea
slug Pleurobranchaea californica. J. Neurophysiol. 81, 654–667 (1999).
86. Gunaratne, C. A., Sakurai, A. & Katz, P. S. Variations on a theme: species dierences in
synaptic connectivity do not predict central pattern generator activity. J. Neurophysiol.
118, 1123–1132 (2017).
87. Borst, A . & Euler, T. Seeing things in motion: models, circuits, and mechanisms. Neuron
71, 974–994 (2011).
88. Kim, Y. J. et al. Origins of direction selectivity in the primate retina. Nat. Commun. 13,
2862 (2022).
89. Patterson, S. S. et al. Conserved circuits for direction selectivity in the primate retina.
Curr. Biol. 32, 2529–2538 (2022).
90. Johnson, B. R., Peck, J. H. & Harris-Warrick, R. M. Dierential modulation of chemical
and electrical components of mixed synapses in the lobster stomatogastric ganglion.
J. Comp. Physiol. A 175, 233–249 (1994).
91. Katz, P. S. & Lillvis, J. L. Reconciling the deep homology of neuromodulation with the
evolution of behavior. Curr. Opin. Neurobiol. 29, 39–47 (2014).
92. Bendesky, A. & Bargmann, C. I. Genetic contributions to behavioural diversity at the
gene–environment interface. Nat. Rev. Genet. 12, 809–820 (2011).
93. Lim, M. M. et al. Enhanced partner preference in a promiscuous species by manipulating
the expression of a single gene. Nature 429, 754–757 (2004).
94. Young, L. J., Nilsen, R., Waymire, K. G., MacGregor, G. R. & Insel, T. R. Increased ailiative
response to vasopressin in mice expressing the V1a receptor from a monogamous vole.
Nature 400, 766–768 (1999).
95. Bendesky, A. et al. The genetic basis of parental care evolution in monogamous mice.
Nature 544, 434–439 (2017).
Quantitative trait locus study investigating the genetic bases of parental care in mice.
The study uncovers the role of AVP in the evolution of nest-building behaviour.
96. Elipot, Y., Hinaux, H., Callebert, J. & Rétaux, S. Evolutionary shift from ighting to foraging
in blind caveish through changes in the serotonin network. Curr. Biol. 23, 1–10 (2013).
97. Jaggard, J. B. et al. Hypocretin underlies the evolution of sleep loss in the Mexican
caveish. Elife 7, e32637 (2018).
98. Duboué, E. R., Borowsky, R. L. & Keene, A. C. β-adrenergic signaling regulates
evolutionarily derived sleep loss in the Mexican caveish. Brain. Behav. Evol. 80, 233–243
(2012).
99. Shafer, M. E. R., Sawh, A. N. & Schier, A. F. Gene family evolution underlies cell-type
diversiication in the hypothalamus of teleosts. Nat. Ecol. Evol. 6, 63–76 (2022).
Single-cell transcriptomics study comparing the hypothalamus of zebraish with that
of surface and cave forms of Mexican tetra. The indings suggest that in homologous
cell types, terminal dierentiation genes can remain conserved while the underlying
transcription factor regulatory network evolves.
100. Barkan, C. L., Kelley, D. B. & Zornik, E. Premotor neuron divergence relects vocal
evolution. J. Neurosci. 38, 5325–5337 (2018).
101. Barkan, C. L., Zornik, E. & Kelley, D. B. Evolution of vocal patterns: tuning hindbrain
circuits during species divergence. J. Exp. Biol. 220, 856–867 (2017).
102. Carlson, B. A. & Gallant, J. R. From sequence to spike to spark: evo-devo-neuroethology
of electric communication in mormyrid ishes. J. Neurogenet. 27, 106–129 (2013).
103. Osborne, K. A. et al. Natural behavior polymorphism due to a cGMP-dependent protein
kinase of Drosophila. Science 277, 834–836 (1997).
104. Renger, J. J., Yao, W. D., Sokolowski, M. B. & Wu, C. F. Neuronal polymorphism among
natural alleles of a cGMP-dependent kinase gene, foraging, in Drosophila. J. Neurosci. 19,
RC28 (1999).
105. Fitzpatrick, M. J., Feder, E., Rowe, L. & Sokolowski, M. B. Maintaining a behaviour
polymorphism by frequency-dependent selection on a single gene. Nature 447, 210–212
(2007).
106. Mery, F., Belay, A. T., So, A. K. C., Sokolowski, M. B. & Kawecki, T. J. Natural polymorphism
aecting learning and memory in Drosophila. Proc. Natl Acad. Sci. USA 104, 13051–13055
(2007).
107. Ben-Shahar, Y., Robichon, A., Sokolowski, M. B. & Robinson, G. E. Inluence of gene action
across dierent time scales on behavior. Science 296, 741–744 (2002).
108. Ingram, K. K., Oefner, P. & Gordon, D. M. Task-speciic expression of the foraging gene in
harvester ants. Mol. Ecol. 14, 813–818 (2005).
109. Hong, R. L., Witte, H. & Sommer, R. J. Natural variation in Pristionchus paciicus insect
pheromone attraction involves the protein kinase EGL-4. Proc. Natl Acad. Sci. USA 105,
7779–7784 (2008).
110. Allen, A. M. & Sokolowski, M. B. Expression of the foraging gene in adult Drosophila
melanogaster. J. Neurogenet. 35, 192–212 (2021).
111. Dason, J. S. & Sokolowski, M. B. A cGMP-dependent protein kinase, encoded by the
Drosophila foraging gene, regulates neurotransmission through changes in synaptic
structure and function. J. Neurogenet. 35, 213–220 (2021).
112. LaPotin, S. et al. Divergent cis-regulatory evolution underlies the convergent loss of
sodium channel expression in electric ish. Sci. Adv. 8, eabm2970 (2022).
113. Arnegard, M. E., Zwickl, D. J., Lu, Y. & Zakon, H. H. Old gene duplication facilitates origin
and diversiication of an innovative communication system - twice. Proc. Natl Acad.
Sci. USA 107, 22172–22177 (2010).
114. Liscovitch-Brauer, N. et al. Trade-o between transcriptome plasticity and genome
evolution in cephalopods. Cell 169, 191–202 (2017).
115. Wang, Z. Y. & Ragsdale, C. W. Cadherin genes and evolutionary novelties in the octopus.
Semin. Cell Dev. Biol. 69, 151–157 (2017).
116. Lamb, T. D. Evolution of vertebrate retinal photoreception. Philos. Trans. R. Soc. B
Biol. Sci. 364, 2911–2924 (2009).
117. Baden, T. & Euler, T. Early vision: where (some of) the magic happens. Curr. Biol. 23,
R1096–R1098 (2013).
118. Sinha, R. et al. Cellular and circuit mechanisms shaping the perceptual properties of the
primate fovea. Cell 168, 413–426 (2017).
119. Truman, J. W. & Ball, E. E. Patterns of embryonic neurogenesis in a primitive wingless
insect, the silverish, Ctenolepisma longicaudata: comparison with those seen in lying
insects. Dev. Genes Evol. 208, 357–368 (1998).
120. LaMonica, B. E., Lui, J. H., Wang, X. & Kriegstein, A. R. OSVZ progenitors in the human
cortex: an updated perspective on neurodevelopmental disease. Curr. Opin. Neurobiol.
22, 747–753 (2012).
121. Florio, M. et al. Human-speciic gene ARHGAP11B promotes basal progenitor
ampliication and neocortex expansion. Science 347, 1465–1470 (2015).
122. Florio, M., Namba, T., Paabo, S., Hiller, M. & Huttner, W. B. A single splice site mutation in
human-speciic ARHGAP11B causes basal progenitor ampliication. Sci. Adv. 2, e1601941
(2016).
Work performing ancestral molecular reconstruction of a hominid-speciic gene
and misexpressing it in mice to show that a single-nucleotide substitution in this
gene in the human lineage conferred it with its current function in neuronal progenitor
ampliication.
123. Pop, S. et al. Extensive and diverse patterns of cell death sculpt neural networks in
insects. Elife 9, e59566 (2020).
124. Prieto-Godino, L. L. et al. Functional integration of “undead” neurons in the olfactory
system. Sci. Adv. 6, eaaz7238 (2020).
125. Cayirlioglu, P. et al. Hybrid neurons in a microRNA mutant are putative evolutionary
intermediates in insect CO2 sensory systems. Science 319, 1256–1260 (2008).
126. Prieto-Godino, L. L. L. L., Diegelmann, S. & Bate, M. Embryonic origin of olfactory circuitry
in Drosophila: contact and activity-mediated interactions pattern connectivity in the
antennal lobe. PLoS Biol. 10, e1001400 (2012).
127. Willett, R. T. et al. Cerebellar nuclei excitatory neurons regulate developmental scaling
of presynaptic Purkinje cell number and organ growth. Elife 8, e50617 (2019).
128. Tierney, A. J. Evolutionary implications of neural circuit structure and function. Behav.
Process. 35, 173–182 (1995).
129. Cande, J., Prud’homme, B. & Gompel, N. Smells like evolution: the role of chemoreceptor
evolution in behavioral change. Curr. Opin. Neurobiol. 23, 152–158 (2013).
130. Jacobs, G. H., Williams, G. A., Cahill, H. & Nathans, J. Emergence of novel color vision
in mice engineered to express a human cone photopigment. Science 315, 1723–1725
(2007).
131. Frangeul, L. et al. A cross-modal genetic framework for the development and plasticity
of sensory pathways. Nature 538, 96–98 (2016).
132. Ramdya, P. & Engert, F. Emergence of binocular functional properties in a monocular
neural circuit. Nat. Neurosci. 11, 1083–1090 (2008).
133. Ramaekers, A. et al. Altering the temporal regulation of one transcription factor drives
evolutionary trade-os between head sensory organs. Dev. Cell 50, 780–792 (2019).
134. Jourjine, N. & Hoekstra, H. E. Expanding evolutionary neuroscience: insights from
comparing variation in behavior. Neuron 109, 1084–1099 (2021).
Article that introduced the concept of ‘model clades’ for evolutionary neuroscience
studies.
135. Herman, A . et al. The role of gene low in rapid and repeated evolution of cave-related
traits in Mexican tetra, Astyanax mexicanus. Mol. Ecol. 27, 4397–4416 (2018).
136. Torres-Méndez, A. et al. A novel protein domain in an ancestral splicing factor drove the
evolution of neural microexons. Nat. Ecol. Evol. 3, 691–701 (2019).
137. Lillvis, J. L. & Katz, P. S. Parallel evolution of serotonergic neuromodulation underlies
independent evolution of rhythmic motor behavior. J. Neurosci. 33, 2709–2717 (2013).
138. Euler, T. & Baden, T. Computational neuroscience: species-speciic motion detectors.
Nature 535, 45–46 (2016).
139. Loomis, C. et al. An adult brain atlas reveals broad neuroanatomical changes in
independently evolved populations of Mexican caveish. Front. Neuroanat. 13, 88
(2019).
140. Zakon, H. H., Lu, Y., Zwickl, D. J. & Hillis, D. M. Sodium channel genes and the evolution
of diversity in communication signals of electric ishes: convergent molecular evolution.
Proc. Natl Acad. Sci. USA 103, 3675–3680 (2006).
141. Lui, J. H., Hansen, D. V. & Kriegstein, A. R. Development and evolution of the human
neocortex. Cell 46, 18–36 (2011).
Nature Reviews Neuroscience
Review article
142. Haalck, L., Mangan, M., Webb, B. & Risse, B. Towards image-based animal tracking in natural
environments using a freely moving camera. J. Neurosci. Methods 330, 108455 (2020).
143. Berck, M. E. et al. The wiring diagram of a glomerular olfactory system. Elife 5, e14859
(2016).
144. Eichler, K. et al. The complete connectome of a learning and memory centre in an insect
brain. Nature 548, 175–182 (2017).
145. Mackay, T. F. C. C. et al. The Drosophila melanogaster genetic reference panel. Nature
482, 173–178 (2012).
146. Brown, E. B., Layne, J. E., Zhu, C., Jegga, A. G. & Rollmann, S. M. Genome-wide
association mapping of natural variation in odour-guided behaviour in Drosophila. Genes
Brain Behav. 12, 503–515 (2013).
147. Harbison, S. T., McCoy, L. J. & Mackay, T. F. C. Genome-wide association study of sleep in
Drosophila melanogaster. BMC Genomics 14, 281 (2013).
148. Andersson, L. L . S. et al. Mutations in DMRT3 aect locomotion in horses and spinal
circuit function in mice. Nature 488, 642–646 (2012).
149. Chen, S. et al. Frequent recent origination of brain genes shaped the evolution of
foraging behavior in Drosophila. Cell Rep. 1, 118–132 (2012).
150. Capra, J. A., Erwin, G. D., Mckinsey, G., Rubenstein, J. L. R. & Pollard, K. S. Many human
accelerated regions are developmental enhancers. Philos. Trans. R. Soc. B Biol. Sci. 368,
20130025 (2013).
151. Pollard, K. S. et al. Forces shaping the fastest evolving regions in the human genome.
PLoS Genet. 2, e168 (2006).
152. Dennis, M. Y. et al. Evolution of human-speciic neural SRGAP2 genes by incomplete
segmental duplication. Cell 149, 912–922 (2012).
153. Boyd, J. L. et al. Human-chimpanzee dierences in a FZD8 enhancer alter cell-cycle
dynamics in the developing neocortex. Curr. Biol. 16, 772–779 (2015).
154. Kawecki, T. J. et al. Experimental evolution. Trends Ecol. Evol. 27, 547–560 (2012).
155. Mery, F. & Kawecki, T. J. Experimental evolution of learning ability in fruit lies. Proc. Natl
Acad. Sci. USA 99, 14274–14279 (2002).
156. Mackay, T. F. C. et al. Genetics and genomics of Drosophila mating behavior. Proc. Natl
Acad. Sci. USA 102, 6622–6629 (2005).
157. Jaksic, A. M. et al. Neuronal function and dopamine signaling evolve at high temperature
in Drosophila. Mol. Biol. Evol. 37, 2630–2640 (2020).
158. Turner, T. L., Stewart, A. D., Fields, A. T., Rice, W. R. & Tarone, A. M. Population-based
resequencing of experimentally evolved populations reveals the genetic basis of body
size variation in Drosophila melanogaster. PLoS Genet. 7, e1001336 (2011).
159. Versace, E. Experimental evolution, behavior and genetics: associative learning as a case
study. Curr. Zool. 61, 226–241 (2015).
160. Schlötterer, C., Koler, R., Versace, E., Tobler, R. & Franssen, S. U. Combining experimental
evolution with next-generation sequencing: a powerful tool to study adaptation from
standing genetic variation. Heredity 114, 431–440 (2015).
Acknowledgements
The authors thank T. Baden from the University of Sussex, R. Arguello from the University of
Lausanne and J. L. Ramos from the L.L.P.-G. laboratory for discussions and comments on
the manuscript. They thank J. Brock at the Francis Crick Institute for help with illustrations.
R.J.V.R. is supported by a Boehringer Ingelheim Fonds Ph.D. fellowship. L.L.P.-G.’s laboratory
is supported by a European Research Council Starting Investigator Grant (802531), an Allen
Distinguished Investigator Award, a Human Frontiers Science Grant (RGY0052/2022) and the
Francis Crick Institute, which receives its core funding from Cancer Research UK (FC001594),
the UK Medical Research Council (FC001594) and the Wellcome Trust (FC001594).
Author contributions
The authors contributed equally to all aspects of the article.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains supplementary material available at
https://doi.org/10.1038/s41583-022-00644-y.
Correspondence should be addressed to Lucia L. Prieto-Godino.
Peer review information Nature Reviews Neuroscience thanks P. Katz and the other,
anonymous, reviewer(s) for their contribution to the peer review of this work.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional ailiations.
Related links
Retinal Functomics: http://retinal-functomics.net/
© Springer Nature Limited 2022
Springer Nature or its licensor holds exclusive rights to this article under a publishing
agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted
manuscript version of this article is solely governed by the terms of such publishing
agreement and applicable law.
