ArticlePDF AvailableLiterature Review

From the connectome to brain function

June 2013
Nature Methods 10(6):483-90

DOI:10.1038/nmeth.2451

Source
PubMed

Authors:

Cornelia I Bargmann

Rockefeller University

Eve Marder

Brandeis University

In this Historical Perspective, we ask what information is needed beyond connectivity diagrams to understand the function of nervous systems. Informed by invertebrate circuits whose connectivities are known, we highlight the importance of neuronal dynamics and neuromodulation, and the existence of parallel circuits. The vertebrate retina has these features in common with invertebrate circuits, suggesting that they are general across animals. Comparisons across these systems suggest approaches to study the functional organization of large circuits based on existing knowledge of small circuits.

Content uploaded by Eve Marder

Content may be subject to copyright.

NATURE METHODS

VOL.10 NO.6 | JUNE 2013

483

FOCUS ON MAPPING THE BRAIN HISTORICAL PERSPECTIVE

In this Historical Perspective, we ask what

information is needed beyond connectivity

diagrams to understand the function of nervous

systems. Informed by invertebrate circuits

whose connectivities are known, we highlight

the importance of neuronal dynamics and

neuromodulation, and the existence of parallel

circuits. The vertebrate retina has these features in

common with invertebrate circuits, suggesting that

they are general across animals. Comparisons across

these systems suggest approaches to study the

functional organization of large circuits based on

existing knowledge of small circuits.

An animal’s behavior arises from the coordinated activ-

ity of many interconnected neurons—“many” meaning

302 for Caenorhabditis elegans, 20,000 for a mollusc,

several hundred thousand for an insect or billions for

humans. Determining the connectivity of these neu-

rons, via combined anatomical and electrophysiologi-

cal methods, has always been a part of neuroscience. As

we were writing this, these ideas were being revisited

from the perspective of massively parallel methods for

dense reconstruction, or ‘connectomics’. One thread of

this analysis involves the detailed, high-density map-

ping of point-to-point connections between neurons

at synapses

1–4

. The specialized membrane structures

and synaptic vesicles of synapses can be visualized

with an electron microscope, and consequently dense

reconstructions of nervous-system connectomes rely

on electron microscopy of serial brain sections. In a

complementary approach, detailed electrophysiologi-

cal analysis shows how synapses and circuits function

at high resolution, and is increasingly being applied to

large numbers of interconnected neurons.

The first approaches used to map complete circuits

came from studies of the smaller nervous systems

of invertebrates. In the 1960s and 1970s, systematic

electrophysiological recordings from neurons in

discrete ganglia enabled the identification of neu-

ronal components of circuits that generate specific

behaviors

5–7

. In association with the recordings of

these individually recognizable, identified neurons, the

cells were filled with dye to visualize their structures

and projection patterns via light microscopy

8–10

. In

some cases, electron microscopy was used to observe

the anatomical synapses in these small circuits

11–13

But until the publication of the heroic electron micros-

copy reconstruction of the full nervous system of

C. elegans

in the mid-1980s, it was unimaginable

that the electron microscope could be used to deter-

mine circuit connectivity rather than providing

ultrastructural detail to connectivity determined

either with physiological or light microscopy–based

anatomical methods.

Recent advances in electron microscopy and image

analysis have made it possible to scale up this ultrastruc-

tural approach: to serially section and reconstruct pieces

of both vertebrate and invertebrate nervous systems,

with the stated purpose of using detailed connectomes

to reveal how these circuits work

4,15–18

. Such large-

scale projects will provide new anatomical data that will

offer invaluable insights into the functional organiza-

tion of the structures studied. An unbiased approach

to data acquisition always reveals surprises and new

insights. Moreover, because of the scope and size of

these projects, such efforts will generate unprecedented

amounts of data to be analyzed and understood.

Here we ask what additional information is needed

beyond connectivity diagrams to understand circuit

function, informed by the invertebrate circuits whose

connectivity is known. For the prototypical case, the

complete C. elegans nervous system, the anatomical

connectome was largely established over 25 years

ago

. In a variety of other invertebrate preparations,

connectivity was established using combinations of

electrophysiological recordings and neuronal tracing 30–

40 years ago, which enabled researchers to generate

a wiring map that incorporates activity information.

Despite their different starting points from anatomy and

electrophysiology, these two approaches have uncovered

From the connectome to brain function

Cornelia I Bargmann

& Eve Marder

2,3

Howard Hughes Medical Institute, The Rockefeller University, New York, New York, USA.

Volen Center, Brandeis University, Waltham,

Massachusetts, USA.

Department of Biology, Brandeis University, Waltham, Massachusetts, USA. Correspondence should be addressed to

C.I.B. (cori@rockefeller.edu).

Received 27 FebRuaRy; accepted 5 apRil; published online 30 may 2013; doi:10.1038/nmeth.2451

npg

HISTORICAL PERSPECTIVE FOCUS ON MAPPING THE BRAIN

484

VOL.10 NO.6 | JUNE 2013

NATURE METHODS

similar principles and similar puzzles as to how circuit function

arises from the component neurons and their interactions.

What do functional and anatomical maps reveal?

We begin with the connectivity diagram of the stomatogas-

tric ganglion (STG) of the crab, Cancer borealis (Fig. 1a) and a

graph of the connectome of C. elegans (Fig. 1b). In each case, the

number of neurons is small, ~27 neurons or 302 neurons, respec-

tively, but the number of synaptic connections is much larger;

the neurons are extensively interconnected. The basic function of

each circuit is known: to generate rhythmic stomach movements

for the crab STG and to control locomotion behavior in response

to sensory inputs for C. elegans. The intellectual strength of the

STG system is the ability to relate neuronal connectivity to

neuronal activity patterns; the complementary strength of

C. elegans is the ability to relate neuronal connectivity to whole-

animal behavior.

The STG contains motor neurons and interneurons that gen-

erate two rhythmic motor patterns

. The pyloric rhythm is an

oscillating, triphasic motor pattern that is continuously active and

depends on a set of electrically coupled pacemaker neurons. The

gastric mill rhythm is episodically active and depends on descend-

ing modulatory inputs activated by sensory neurons for its gen-

eration

19,20

. Although these rhythms are easily studied separately,

a close look at the STG connectivity diagram reveals that the neu-

rons that conventionally are thought to be part of the pyloric

circuit (neurons AB, PD, LP, PY, VD and IC) are highly intercon-

nected with those conventionally thought part of the gastric mill

circuit (neurons DG, GM, LPG, MG, LG and Int1) (AM is part

of a third circuit that we will not discuss here). Indeed, many

STG neurons switch their activity between the two rhythms

and the separation of the STG’s connectivity into two discrete

circuits, although convenient for those who study the network,

does not really capture the highly interconnected reality of the

ganglion’s architecture.

Like all nervous systems, the circuit has many chemical synapses,

in which a presynaptic neuron releases a chemical neurotransmit-

ter to activate receptors on the postsynaptic neuron. Chemical

synapses can be inhibitory or excitatory depending on the nature

of the receptor and associated ion channels; the chemical synapses

among STG neurons are inhibitory. Additional connections are

created by the widespread electrical synapses, mediated by direct

cytoplasmic communication through gap junctions, through

which current flows depending on the voltages of the coupled

neurons. In the STG circuit, there are many instances of neurons

that are connected by electrical synapses as well as by chemical

inhibitory synapses (Fig. 1a). There are also many instances of

neurons connected by reciprocal inhibition. These wiring motifs

contribute to circuit properties that are not easily predictable.

In addition, there are many ‘parallel pathways’ in which two neu-

rons are connected via two or more synaptic routes, one direct

route and additional indirect routes (Fig. 1a). The complexity of

this connection map poses the essential question: are all synapses

important, or are some only important under certain conditions

(as appears to be the case)

? How do we understand the impor-

tance of synaptic connectivity patterns that seem to oppose each

other, such as the common motif of electrical coupling between

neurons that also inhibit each other?

The C. elegans wiring diagram was assembled in the near-

complete absence of prior functional information. It allowed an

DD03

DD02

VB04

DD04

VD07

VB06

VD05

VD06

VD04

DB03

VC01

VC02

VA07

VD08

DB02

AS04

VA04

VD03

VA06

VC03

DB04

DD05

VA05

DA04

DA03

VD09

AS03

AS05

AS06

VB08

DD01

VD02

DA05

VA09

VA08

AS02

VA03

VB02

DB01

PDB

VA02

VB09

VD10

PDA

RID

DA02

VB07

DA06

AS11

VD13

VD01

VA12

DD06

VC04

DA09

VD12

DVB

PVDL

AS09

VA11

DA08

DA07

VB11

AS08

PVDR

VA10

PHCL

PHBL

VA01

AS01

VB10

AS07

DB07

PVCR

PHBR

AS10

PVCL

AVAL

DA01

LUAL

DB05

DB06

AVAR

PVWR

PLML

VD11

PHCR

SABD

LUAR

PQR

PVWL

AVDL

PVNR

AVL

AVDR

AVBR

PHAL

AVBL

PLMR

PHAR

FLPR

AVG

PVNL

FLPL

SABVR

AVM

SABVL

AVJL

PVPR

DVC

VC05

PVPL

VB01

AVJR

BDUR

HSNR

PDER

PDEL

PVM

DVA

AVFR

RIFL

AVFL

RIFR

AQR

AVHR

AVHL

PVR

SDQL

ALMR

BDUL

AVKL

ALA

PVQR

PVT

ALML

ADAL

HSNL

SDQR

AVEL

ASJR

ADER

SAADL

RMFR

ADAR

AVER

ADLR

SAAVR

SAAVL

ADLL

AIML

AVKR

ASHR

RIMR

AIMR

RIML

ASHL

PVQL

SIBVL

ASJL

RMFL

ASKR

SAADR

RIGL

AIBR

RICL

RIS

SMBVR

AIBL

PLNL

ADEL

ASKL

PLNR

RIGR

SMBDL

ALNL

AIAR

RICR

SMBDR

AIAL

SMBVL

RMGR

ALNR

AWBR

RIR

BAGR

BAGL

RMGL

ASGL

AUAL

URXL

ASGR

AIZR

AWBL

SMDDR

RIBL

AIZL

ASIR

URYVL

SIBVR

URBL

URYVR

ADFR

RMHL

RIBR

URYDR

ASER

AWAR

ASIL

AINR

RIVL

OLLL

AUAR

ADFL

AWCR

SMDDL

RMHR

RIVR

SIBDL

AWCL

SIADL

ASEL

AINL

AWAL

URYDL

CEPDL

CEPVL

OLLR

RMDR

URXR

SIAVR

CEPDR

SIADR

AIYL

AIYR

SIAVL

SIBDR

AFDR

AFDL

SMDVR

SMDVL

URBR

RMDL

CEPVR

RIAL

RIAR

OLQDR

OLQVL

OLQDL

RMDVR

IL2L

RMDDL

RMED

RMEV

RIH

RMDVL

RMDDR

IL1L

IL2R

IL1R

URADL

OLQVR

IL1DL

IL1VL

RIPL

IL1DR

URADR

RMEL

IL1VR

RIPR

IL2DL

URAVL

IL2VL

RMER

IL2DR

IL2VR

URAVR

AB PD LPG

Electrical synapse

Chemical inhibitory synapses

LP IC LG MG GM

PY VD DG AM

Int1

Figure 1

Connectivity of two

well-studied invertebrate circuits.

(a) Connectivity diagram of the crab

STG based on electrophysiological

recordings. Red and blue background

shading indicates neurons that are

primarily part of the pyloric and

gastric circuits, respectively. Purple

shading indicates that some neurons

switch between firing in pyloric

and gastric time, and that there

is no fixed boundary between the

pyloric and gastric circuits. Yellow

highlights two neurons that are both

electrically coupled and reciprocally

inhibitory. Green highlights one

of many examples of neurons that

are coupled both monosynaptically

and polysynaptically. (b) The

connectome of C. elegans,

showing all 302 neurons and their

chemical synapses but not their

gap junctions. Each neuron has a

three-letter name, often followed

by a spatial designator. This top-

to-bottom arrangement (signal

flow view) is arranged to reflect

dominant information flow,

which goes from sensory neurons

(red) to interneurons (blue) to

motor neurons (green). Reprinted

from ref. 50.

npg

FOCUS ON MAPPING THE BRAIN HISTORICAL PERSPECTIVE

NATURE METHODS

VOL.10 NO.6 | JUNE 2013

485

immediate classification of neurons into large classes: sensory

neurons (with distinctive sensory dendrites and cilia), motor

neurons (with neuromuscular junctions) and interneurons

(a term that is used in C. elegans to describe any neuron that is

not evidently sensory or motor, encompassing projection neurons

and local neurons)

. In each group, neurons were subdivided into

unique types with similar morphologies and connections, collaps-

ing the wiring diagram from 302 neurons to 119 neuronal types.

The flow of information through chemical synapses is predomi-

nantly from sensory to interneuron to motor neuron, with many

parallel pathways linking neurons both directly and indirectly

(as in the STG), as well as gap junctions that may form electrical

synapses (~10% of all synapses). Most neurons are separated from

each other by no more than two or three synaptic connections.

The C. elegans map was immediately used to define neurons

required for the touch-avoidance response, which is still the most

completely characterized of the animal’s behaviors

. Light touch

to the head elicits a reversal, and light touch to the tail elicits

a forward acceleration. The neurons required for the touch-

avoidance response were identified by killing cells with a laser

microbeam and assessing the behavioral repertoire of the worms.

Guided by the wiring diagram, this analysis revealed essential

mechanosensory neurons in the head and tail, key interneurons

required to propagate information, and motor neurons required

for forward and backward movement (Fig. 2). The success of

this approach inspired similar analyses of chemosensory behav-

iors, foraging, egg-laying, feeding and more. At this point, over

60% of C. elegans neuron types have defined functions in one or

more behaviors.

This notable success, however, hides a surprising failure. For

C. elegans, although we know what most of the neurons do, we

do not know what most of the connections do, we do not know

which chemical connections are excitatory or inhibitory, and we

cannot easily predict which connections will be important from

the wiring diagram. The problem is illustrated most simply by

the classical touch-avoidance circuit

(Fig. 2). The PLM sensory

neurons in the tail are solely responsible for tail touch avoid-

ance. PLM forms 31 synapses with 11 classes of neurons, but only

one of those targets is essential for the behavior—an interneuron

called PVC that is connected to PLM by just two gap junctions

and two chemical synapses. An even greater mismatch between

the number of synapses and their importance in behavior is seen

in the avoidance of head touch, where just two of 58 synapses

(again representing gap junctions) are the key link between the

sensory neurons (ALM and AVM) and the essential interneuron

(AVD). This general mismatch between the number of syn-

apses and apparent functional importance has applied wherever

C. elegans circuits have been defined. As a result, early guesses

about how information might flow through the wiring diagram

were largely incorrect.

Clearly, the wiring diagram could generate hypotheses to test,

but solving a circuit by anatomical inspection alone was not suc-

cessful. We believe that anatomical inspection fails because each

wiring diagram encodes many possible circuit outcomes.

Parallel and antagonistic pathways complicate circuits

Both of the wiring diagrams shown in Figure 1 are richly con-

nected. In the STG, a large fraction of the synapses are electrical

synapses. In some cases, the electrical synapses connect multiple

copies of the same neuron, such as the two PD neurons in the

STG. Notably, many electrical synapses connect neurons with dif-

ferent functions. Almost invariably, the combination of electrical

and chemical synapses create ‘parallel pathways’, that is to say,

multiple pathways by which neuron 1 can influence neuron 2

(Fig. 1a). For example, in the STG, the PD neuron inhibits the IC

neuron through chemical synapses but also can influence the IC

neuron via the electrical synapse from LP to IC. Parallel pathways

such as those in the STG can be viewed as degenerate, as they

create multiple mechanisms by which the network output can be

switched between states

(Fig. 3). A simulation study

shows a

simplified five-cell network of oscillating neurons coupled with

electrical synapses and chemical inhibitory synapses. The f1 and

f2 neurons are connected reciprocally by chemical inhibitory

synapses, as are the s1 and s2 neurons. This type of wiring con-

figuration, called a half-center oscillator, often but not universally

causes the neurons to be rhythmically active in alternation

. In

this example, two different oscillating rhythms are generated, one

fast and one slow. The hub neuron at the center of the network

can be switched between firing in time with the fast f1 and f2

neurons to firing in time with the slow s1 and s2 neurons by three

entirely different circuit mechanisms: changing the strength of the

electrical synapses, changing the strength of the synapses between

f1 and s1 onto the hub neuron, and changing the strength of the

reciprocal inhibitory synapses linking f1 to f2 and s1 to s2 in the

half-center oscillators.

An example from the C. elegans connectome illustrates another

twist of circuit logic: divergent circuits that start at a common

point but result in different outcomes. In this example, gap junc-

tions and chemical synapses from ADL sensory neurons generate

opposite behavioral responses to a C. elegans pheromone (Fig. 4a).

The chemical synapses drive avoidance of the pheromone, whereas

the gap junctions stimulate a pheromone-regulated aggregation

behavior

. Differing use of the chemical synapse subcircuit ver-

sus the gap junction subcircuit allows ADL to switch between

these two opposing behaviors in different contexts. ADL illus-

trates the point that is not possible to ‘read’ a connectome if it is

intrinsically ambiguous, encoding two different behaviors.

Parallel and divergent systems of synapses are widespread fea-

tures of invertebrate and vertebrate networks alike, and can be

composed of sets of chemical synapses as well as sets of chemical

AVM PLM

AVD

Gap junction

Chemical synapse

Sensory neuron

Interneuron

Motor neuron

ALM

Anterior touch Posterior touch

PVC

9 classes

of neurons

9 classes

of neurons

6 classes

of neurons

AVB

Reversal Acceleration

Figure 2

C. elegans neurons essential for avoidance of light touch.

Inferred connections necessary for anterior and posterior touch avoidance

are in purple and orange, respectively; other synapses are in black. The

‘essential’ synapses here shown in orange and purple comprise less than

10% of the output synapses of the mechanosensory neurons. Image based

on ablation data from ref. 22.

npg

HISTORICAL PERSPECTIVE FOCUS ON MAPPING THE BRAIN

486

VOL.10 NO.6 | JUNE 2013

NATURE METHODS

and electrical synapses. To understand information flow, there

will be no substitute for recording activity. The methods for moni-

toring neuronal activity have improved dramatically in recent

years, with development of new multi-electrode recording tech-

niques and a suite of genetically encoded indicators that can be

used to measure calcium, voltage and synaptic release at cellular

and subcellular levels. However, improved methods are needed

to detect electrical synapses, which can also be difficult to see

in electron micrographs. The regulation of electrical synapses

by voltage, neuromodulation, phosphorylation and small mole-

cules is understudied

26,27

. A chemical method for measuring gap

junctions, local activation of molecular fluorescent probes, is a

promising new direction that should spawn innovation

Neuromodulation reconfigures circuit properties

Superimposed on the fast chemical synapses and electrical syn-

apses in the wiring diagram are the neuromodulators—biogenic

amines (serotonin, dopamine, norepinephrine and histamine)

and neuropeptides (dozens to hundreds, depending on species)

These molecules are often released together with a fast chemical

transmitter near a synapse, but they can diffuse over a greater dis-

tance. Modulators also can be released from neuroendocrine cells

that do not make defined synaptic contacts or can be delivered as

hormones through the circulation. As a result, the targets of neuro-

modulation are invisible to the electron microscope. Signaling pri-

marily through G protein–regulated biochemical processes rather

than through ionotropic receptors, neuromodulators change neu-

ronal functions over seconds to minutes, or even hours.

Many years of work on the effects of neuromodulators on the

STG have revealed that the functional connections that give rise

to a specific circuit output are specified, or in fact ‘configured’, by

the neuromodulatory environment

. Every synapse and every

neuron in the STG is subject to modulation; the connectivity dia-

gram by itself only establishes potential circuit configurations,

whose availability and properties depend critically on which of

many neuromodulators are present at a given moment

. Under

some modulatory conditions, anatomically ‘present’ synaptic

connections may be functionally silent, only to be strengthened

under other modulatory conditions. Likewise, modulators can

qualitatively alter the neurons’ intrinsic properties, transforming

neurons from tonic spiking to those generating plateau poten-

tials or bursts

. These effects of neuromodulators can activate

or silence an entire circuit, change its frequency and/or the phase

relationships of the motor patterns generated.

C. elegans has over 100 different neuropeptides as well as bio-

genic amine neuromodulators. The integration of neuromodula-

tion into its fast circuits appears to selectively enhance the use of

particular connections at the expense of others. For example, a

‘hub-and-spoke’ circuit drives aggregation of C. elegans by cou-

pling multiple sensory inputs through gap junctions with a com-

mon target neuron, RMG (Fig. 4b). Neuromodulation of RMG

by the neuropeptide receptor NPR-1 effectively silences this gap-

junction circuit, while sparing other functions of the input sensory

neurons that are mediated through chemical synapses

Neuromodulators are prominent in all nervous systems, and

act as key mediators of motivational and emotional states such as

100 mV

1 s

synA

synB

a b c d

synA

synB

Figure 3

Similar changes in

circuit dynamics can arise from

three entirely different circuit

mechanisms. (a) Circuit diagram

(top) shows the ‘control’ condition

in which f1 and f2 are firing

in a fast rhythm (indicated by

red shading) and the remaining

neurons are firing in a slow

rhythm (shaded in blue). hn is the

hub neuron. In these diagrams,

electrical synapses are shown as

resistor symbols and chemical

inhibitory synapses with filled

circuits. Traces (bottom) show the

voltage waveforms of the five neurons. (b–d) Responses when the strength of the chemical synapses to the hub neuron (g

synA

) was decreased (b), when

the strength of the electrical synapses (g

) was decreased (c) and when the strength of the chemical synapses between f1 and f2 and between s1 and s2

synB

) was decreased (d). Image modified from ref. 23.

Figure 4

Two views of a multifunctional

C. elegans circuit. (a) Ambiguous circuitry

of the ADL sensory neurons, which drive

avoidance of the ascaroside pheromone

C9 through chemical synapses onto multiple

interneurons (right) but can also promote

aggregation (attraction toward pheromones)

through gap junctions with RMG (left). Image

modified from ref. 25. (b) Neuromodulation

separates overlapping circuits. Multiple sensory

neurons form gap junctions with the RMG hub

neurons and promote aggregation through this

circuit, but each sensory neuron also has chemical synapses that can drive RMG-independent behaviors. The neuropeptide receptor NPR-1 inhibits

RMG to suppress aggregation. Image modified from ref. 30.

Gap junction

Chemical synapse Sensory neuron Interneuron or motor neuron

AVA

AVD

AIA

AIB

C9 pheromone

Pheromone

attraction

ASK URX

IL2

RMG

AWB

ADL

Nociceptive

avoidance

Oxygen

avoidance

Pheromone

avoidance

NPR-1

a b

RMG

ASH

Aggregation

Pheromone

avoidance Aggregation

ADL

npg

FOCUS ON MAPPING THE BRAIN HISTORICAL PERSPECTIVE

NATURE METHODS

VOL.10 NO.6 | JUNE 2013

487

sleep, arousal, stress, mood and pain. To understand their release

and their effects on circuits, new methods are needed to monitor

neuromodulation in vivo. Electrophysiology remains the best tool

for characterizing the functional effects of neuromodulators but

is low-throughput. Biochemical methods can be used to reveal

the presence of neuromodulators in tissue or in bulk extracellular

fluid but are less effective for detecting them near a particular

synapse or release site. A new genetic method can be used to read

out the neuromodulatory state directly by monitoring receptor

activation but with a timeframe of hours, whereas endogenous

modulation can change within minutes or seconds

. Progress

is needed in all of these domains and beyond: there is a need to

move from individual neurons and modulators to physiologically

relevant modulatory states, which are likely to include multiple

neuromodulators acting at many sites.

Neuronal dynamics shape the activity of circuits

The existence of parallel circuits and neuromodulation means

that connectivity alone does not provide adequate information

to predict the physiological output of circuits. Even without these

factors, the behavior of neurons over time is unpredictable from

anatomy because neuronal behavior is sensitive to intrinsic chan-

nels and electrical properties that vary within and between cell

types. Channels, synapses and biochemical processes interact

to generate explicitly time-delimited features, or dynamics, in

neurons and circuits.

The importance of neuronal dynamics in circuit function can

be seen most simply in a two-cell circuit (Fig. 5). Two isolated

neurons from the STG that are not normally synaptically coupled

were connected using the dynamic clamp, a computer-neuronal

interface that allows a user to manipulate biological neurons

with conductances that imitate ion channels and synaptic con-

nections

. The neurons are connected reciprocally by dynamic

clamp-created inhibitory synapses so that the neurons rhythmi-

cally alternate their activity

. The dynamic clamp allows the

investigator to change the strength of the synapses as well as the

amount of one of the membrane currents, hyperpolarization-

activated inward current (I

)—either of which dramatically

alters the period of the circuit oscillation (Fig. 5). Thus, a given

wiring diagram can produce widely different dynamics with

different sets of circuit parameters, and conversely, different

circuit mechanisms can give rise to similar oscillation dynamics.

Without knowing the strength and time course of the synaptic

connections as well as the numbers and kinds of membrane

currents in each of the neurons, it would not be possible to

simply go from the wiring diagram to the dynamics of even

two neurons. Synaptic connectivity alone does not sufficiently

constrain a system.

Understanding neuron-specific and circuit-specific dynamics

will be essential to understanding mammalian circuits as well as

invertebrate circuits. In some cases, unique dynamic properties

are characteristics of particular cell types—for example, different

classes of inhibitory cortical interneurons are distinguished as

much by their dynamics as by their connectivity

. In other cases,

neuronal dynamics are variable among similar cells or even within

one cell type. For example, pyramidal neurons in specific areas

of the cortex exhibit persistent activity associated with working

memory

, and neurons in brainstem modulatory systems switch

their properties between tonic and phasic firing modes depending

on behavioral states

. Finally, synaptic plasticity can occur on

rapid timescales to strengthen and weaken synapses based on use,

adding complexity to circuit-level dynamics

Analyzing neuronal dynamics often requires the circuit to be

simultaneously monitored and manipulated, as shown in the

example of the dynamic clamp. Emerging techniques of optoge-

netics and pharmacogenetics can be combined with recording

as well, but a limitation of all of these methods is that they act at

the level of neurons or groups of neurons. To understand func-

tional connectivity, it will be useful to develop methods to silence

or activate specific channels and specific synaptic connections

between two specified neurons, without affecting all other func-

tions of the same cells.

Vertebrate retina also has complex circuit properties

What lessons will emerge as connectomes are scaled up from

small-scale to large-scale circuits? Many features will be com-

mon to small and large circuits. Vertebrate circuits, like inverte-

brate circuits, have multiple cell types with nonuniform intrinsic

properties, extensive and massively parallel synaptic connectivity,

and neuromodulation. The balance of these components varies

between animals and brain regions (the STG has more electri-

cal synapses than most vertebrate brain regions; C. elegans uses

mostly graded potentials instead of all-or-none action potentials),

but in reality, the diversity of circuits in the vertebrate brain is

at least as great as the difference between any one vertebrate

region and any invertebrate circuit. The essential distinction we

see in vertebrate brains is not a particular microcircuit property

but their repeating structure (for example, the many cortical

columns) and their enormous scale compared to the worm brain

and the STG.

2 2

20 40 60 80 100

Period (s)

Conductance (nS)

Synaptic

Synaptic conductance

50 nS 60 nS 80 nS

2 s

conductance

80 nS 50 nS 30 nS

10 mV

Figure 5

Changing either intrinsic neuronal properties or synaptic

properties can alter network function. The dynamic clamp, a computer-

neuron interface (top) was used to vary either the strength of synaptic

connections between two neurons (synaptic) or the amount of an

intrinsic hyperpolarization-activated inward current (I

), in one neuron

as graphed (right). Traces (bottom) show the action potentials generated

by the alternating, oscillating neuron pair as those properties were

varied. Image modified from ref. 24.

npg

HISTORICAL PERSPECTIVE FOCUS ON MAPPING THE BRAIN

488

VOL.10 NO.6 | JUNE 2013

NATURE METHODS

Examination of the vertebrate retina has begun to reveal the

relationships between performance of a large circuit and prop-

erties of a small circuit. The special power of studying the iso-

lated retina is the ability to experimentally control visual input

while simultaneously recording output—the spikes from retinal

ganglion cells that project to the brain. The retina represents an

intermediate degree of complexity with features of both a small

circuit and a large circuit, and has been subject to the most com-

plete anatomical and electrophysiological characterizations of

any vertebrate brain region. Current connectomics studies of

the retina, for example, include dense reconstruction of serial-

section electron micrographs accompanied by analysis of the

neurotransmitter phenotype and activity patterns of the recon-

structed neurons

3,17

. The combination of structure and function,

and a rich history of elegant experiments, make this the ideal

system for understanding neural computations in detail.

The retina contains millions of neurons that fall into five major

neuronal classes (photoreceptors, horizontal cells, bipolar cells,

amacrine cells and retinal ganglion cells), which are subdivided

into about 60 discrete cell types

. Each of the 60 cell types is

arrayed in a near-crystalline two-dimensional array, so that any

pixel viewed by the retina is covered by at least one neuron of each

cell type. Ultimately, information leaves the retina through the

20 classes of retinal ganglion cells, each of which is considered to

be a parallel but partially overlapping processing stream.

The first views of the retinal connectome show all of the prop-

erties that we highlight in small circuits: cellular complexity,

extensive interconnectivity, parallel circuits with chemical and

electrical synapses, and neuromodulation. The heterogeneity of

the 60 retinal cell types is substantial, echoing the heterogeneity

of individual neurons in the STG or C. elegans. Anatomically,

some dendritic arbors cover only a tiny area of the visual field, but

others arborize much more broadly. Their intrinsic physiological

properties are also extremely diverse, with some neurons that

spike (such as retinal ganglion cells), and many neurons that do

not spike (such as photoreceptors and bipolar cells)

. There are

even amacrine neurons that perform independent computations

in different parts of their complex arbors

Synaptic connections in the retina are extensive and diverse,

and electron microscopy reconstructions have revealed many

classes of synaptic connections that had not been observed in

physiological studies

3,17

. There is a great variety of excitatory

and inhibitory chemical synapses, and there are many electri-

cal synapses, that all vary in their strength and their modifica-

tion by experience. Both anatomical and physiological studies

demonstrate that the retina, like small circuits, consists of many

partly parallel circuits with overlapping elements. In particu-

lar, the retina operates over many orders of magnitude of light

intensity, and the properties of its circuits change with its visual

inputs. Within a few seconds in a new visual environment, retinal

ganglion cells shift their properties to encode relevant features

of light intensity, contrast and motion, drawing on different

features of the network

. Subsets of retinal ganglion cells change

their weighting of center and surround inputs in a switch-like

fashion as light levels change

. A brief period of visual stimula-

tion can even reverse the apparent direction-selectivity of retinal

ganglion cells

Finally, neuromodulation has a role in retinal processing that

reshapes visual circuits. Dopamine is released from a subset of

amacrine cells around dawn, under the control of acute light

stimuli and circadian rhythm

, and acts on cells throughout the

retina to switch them from properties appropriate to night vision

to day vision. The photoreceptors themselves are modulated, and

their coupling through gap junctions decreases to increase their

resolution but reduce their sensitivity. Downstream of the rod

photoreceptors, which dominate night vision, dopamine closes

the gap junctions between rod bipolar cells and AII amacrine

cells, effectively diminishing rod input to the retinal ganglion cell

output of the retina.

What differences are there between small and large circuits? The

sheer size of the retina shows a sharp transition compared to the

size of the STG and the worm brain, and the level of analysis moves

from single cells to cell classes. Understanding a single pixel is

not sufficient to understand the retina, and here the properties of

simple and complex circuits diverge. For example, long-range com-

munication allows groups of retinal cells to perform computations

that a single cell cannot. Wide-field cells such as starburst amacrine

cells can make judgments about motion that no single-pixel neuron

could make but can then feed that information into narrow-field

single-pixel neurons to bias their properties. The scaling from fine

resolution to broad resolution and back again emerges from the

diversity of spatial scales across the structure of the retina.

Circuits interact to generate behavior

The entire nervous system is connected, but reductionist neu-

roscientists invariably focus on pieces of nervous systems. The

value of these simplified systems should not let us forget that

behavior emerges from the nervous system as a whole. At the

moment, obtaining the connectomes of even small parts of

the vertebrate nervous system is a heroic task. However, estab-

lishing the detailed pattern of connectivity for a small part of the

nervous system may not be sufficient to understand how that

piece functions in its full context. By parceling out small regions,

one invariably loses information about the long-range connec-

tions to and from that area.

The extent to which long-range connectivity clouds our under-

standing of connectomes will vary. For example, the vertebrate

retina is anatomically isolated, functionally coherent and lacks

recurrent feedback synapses from other brain areas that are

prominent in most other parts of the central nervous system. We

might imagine the retina as a two-dimensional circuit, whereas

most vertebrate circuits are three-dimensional; new principles

will certainly arise from connectomes that include recurrent

inputs. In the amygdala, for example, the intermixing of multi-

ple cell types with different long-range inputs and outputs would

preclude a meaningful understanding based on local anatomy

alone

. Choosing well among brain regions, and combining con-

nectomes with molecular and functional information about the

same cells, as is being done in the vertebrate retina

3,17

, will lead

to the most informative results.

How can we ‘solve’ the brain?

As we look to ways that other neural systems may be charac-

terized with similar power to the three described here, we can

draw certain lessons. One is that precise circuit mapping and spe-

cific neuron identification have had great importance for unify-

ing structural and functional data from different laboratories.

Extending this idea, other systems may not have individually

npg

FOCUS ON MAPPING THE BRAIN HISTORICAL PERSPECTIVE

NATURE METHODS

VOL.10 NO.6 | JUNE 2013

489

named cells, but all nervous systems have cell types distinguished

by anatomy, connectivity and molecular profile that can serve as

the basis of a common vocabulary. Improvements in molecular

staining methods will only increase the power of a connectome

anchored in cellular identity.

We can also see that connectomic endeavors will need to be

supplemented by experiments that monitor, manipulate and

model circuit activity. Monitoring and manipulation of circuit

function have been considered above. To complement and inform

these experimental approaches, the third step is to develop mod-

els that describe how a system’s output results from the interac-

tions of its components. There is a tension between the desire to

study abstract models that are amenable to precise mathematical

analysis and the desire to study models with sufficient biologi-

cal realism to represent the system’s underlying structures and

functions. In small circuits, it is now possible to construct

models and families of models that can be quite instructive

In C. elegans, a few testable models emerged directly from analy-

ses of anatomy. One was the concept of a motif, a set of connection

patterns between three or four neurons that are over-represented

in the wiring diagram compared to the statistical expectation

based on individual connections

. Perhaps these motifs per-

form a canonical computation, or a few canonical computations,

so that solving a few of them effectively solves a larger piece

of the diagram.

But how should we approach building models of large networks

without generating models that are as difficult to understand as

the biological systems that motivated and inspired them? The

beauty of the connectome is its precision and specificity, but it

is hard to imagine useful network models that implement all of

the details of cell-to-cell connectivity obtained with the electron

microscope, when building such models would require enormous

numbers of assumptions about other circuit parameters, and these

parameters are likely to change in different modulatory states. So

we face a conundrum: the new anatomical data will be instructive,

but it is not yet obvious what kinds of models will best reveal the

implications of these data for how circuits actually work.

We are in the midst of a fascinating international debate about

whether it is the right time to embark on a ‘big science’ project to

monitor and model large brain regions. There are those who argue

that we are now at the point at which investments in large-scale

projects will considerably advance the field in ways not possible

by a distributed small-lab approach

46–48

. Big science works best

when the goals of a project are well-defined and when the out-

comes can be easily recognized. Both were true about the human

genome project, but neither is true, yet, about large-scale attempts

to understand the brain. Moreover, this is well-recognized, and all

of the proponents of large-scale initiatives are acutely aware of the

necessity to develop new technology

and of the extraordinary

complexity of biological systems

. That said, the largest challenge

we face in future attempts to understand the dynamics of large

circuits is not in collecting the data: what is most needed are new

methods that allow our human brains to understand what we find.

Humans are notoriously bad at understanding multiple nonlinear

processes, although we excel at pattern recognition. Somehow, we

have to turn the enormous data sets that are already starting to be

generated into a form we can analyze and think about. Otherwise,

we will be doomed to creating a machine that will understand the

human brain better than we can!

ACKNOWLEDGMENTS

We thank M. Meister for sharing his knowledge of the retina. C.I.B. is funded by

the Howard Hughes Medical Institute. Research in the Marder laboratory relevant

to this piece is funded by the US National Institutes of Health (NS17813, NS

81013 and MH46742).

COMPETING FINANCIAL INTERESTS

The authors declare no competing financial interests.

Reprints and permissions information is available online at http://www.nature.

com/reprints/index.html.

1. Bock, D.D. et al. Network anatomy and in vivo physiology of visual

cortical neurons. Nature 471, 177–182 (2011).

2. Briggman, K.L. & Bock, D.D. Volume electron microscopy for neuronal

circuit reconstruction. Curr. Opin. Neurobiol. 22, 154–161 (2012).

3. Briggman, K.L., Helmstaedter, M. & Denk, W. Wiring speciﬁcity in the

direction-selectivity circuit of the retina. Nature 471, 183–188 (2011).

4. Kleinfeld, D. et al. Large-scale automated histology in the pursuit of

connectomes. J. Neurosci. 31, 16125–16138 (2011).

5. Burrows, M. Monosynaptic connexions between wing stretch receptors and

ﬂight motoneurones of the locust. J. Exp. Biol. 62, 189–219 (1975).

6. Fentress, J.C. Simpler Networks and Behavior (Sinauer Associates, 1976).

7. Getting, P.A., Lennard, P.R. & Hume, R.I. Central pattern generator

mediating swimming in Tritonia. I. Identiﬁcation and synaptic

interactions. J. Neurophysiol. 44, 151–164 (1980).

8. Stretton, A.O. & Kravitz, E.A. Neuronal geometry: determination with a

technique of intracellular dye injection. Science 162, 132–134 (1968).

9. Remler, M., Selverston, A. & Kennedy, D. Lateral giant ﬁbers of cray ﬁsh:

location of somata by dye injection. Science 162, 281–283 (1968).

10. Goodman, C.S. Isogenic grasshoppers: genetic variability in the morphology

of identiﬁed neurons. J. Comp. Neurol. 182, 681–705 (1978).

11. Maynard, E.A. Electron microscopy of the stomatogastric ganglion in the

lobster, Homarus americanus. Tissue Cell 3, 137–160 (1971).

12. King, D.G. Organization of crustacean neuropil. I. Patterns of synaptic

connections in lobster stomatogastric ganglion. J. Neurocytol. 5, 207–237

(1976).

13. King, D.G. Organization of crustacean neuropil. II. Distribution of synaptic

contacts on identiﬁed motor neurons in lobster stomatogastric ganglion.

J. Neurocytol. 5, 239–266 (1976).

14. White, J.G., Southgate, E., Thomson, J.N. & Brenner, S. The structure of

the nervous system of the nematode Caenorhabditis elegans. Phil. Trans.

R. Soc. Lond. B 314, 1–340 (1986).

15. Seung, H.S. Reading the book of memory: sparse sampling versus dense

mapping of connectomes. Neuron 62, 17–29 (2009).

16. Meinertzhagen, I.A. & Lee, C.H. The genetic analysis of functional

connectomics in Drosophila. Adv. Genet. 80, 99–151 (2012).

17. Anderson, J.R. et al. Exploring the retinal connectome. Mol. Vis. 17,

355–379 (2011).

18. Lu, J., Tapia, J.C., White, O.L. & Lichtman, J.W. The interscutularis muscle

connectome. PLoS Biol. 7, e32 (2009).

19. Marder, E. & Bucher, D. Understanding circuit dynamics using the

stomatogastric nervous system of lobsters and crabs. Annu. Rev.

Physiol. 69, 291–316 (2007).

20. Nusbaum, M.P. & Beenhakker, M.P. A small-systems approach to motor

pattern generation. Nature 417, 343–350 (2002).

21. Thirumalai, V., Prinz, A.A., Johnson, C.D. & Marder, E. Red pigment

concentrating hormone strongly enhances the strength of the feedback

to the pyloric rhythm oscillator but has little effect on pyloric rhythm

period. J. Neurophysiol. 95, 1762–1770 (2006).

22. Chalﬁe, M. et al. The neural circuit for touch sensitivity in Caenorhabditis

elegans. J. Neurosci. 5, 956–964 (1985).

23. Gutierrez, G.J., O’Leary, T. & Marder, E. Multiple mechanisms switch an

electrically coupled, synaptically inhibited neuron between competing

rhythmic oscillators. Neuron 77, 845–858 (2013).

24. Sharp, A.A., Skinner, F.K. & Marder, E. Mechanisms of oscillation in

dynamic clamp constructed two-cell half-center circuits. J. Neurophysiol. 76,

867–883 (1996).

25. Jang, H. et al. Neuromodulatory state and sex specify alternative

behaviors through antagonistic synaptic pathways in C. elegans.

Neuron 75, 585–592 (2012).

26. Pereda, A.E. et al. Gap junction-mediated electrical transmission: regulatory

mechanisms and plasticity. Biochim. Biophys. Acta 1828, 134–146 (2013).

27. Neyton, J. & Trautmann, A. Physiological modulation of gap junction

permeability. J. Exp. Biol. 124, 93–114 (1986).

npg

HISTORICAL PERSPECTIVE FOCUS ON MAPPING THE BRAIN

490

VOL.10 NO.6 | JUNE 2013

NATURE METHODS

28. Guo, Y.M. et al. Imaging dynamic cell-cell junctional coupling in vivo

using Trojan-LAMP. Nat. Methods 5, 835–841 (2008).

29. Marder, E. Neuromodulation of neuronal circuits: back to the future.

Neuron 76, 1–11 (2012).

30. Macosko, E.Z. et al. A hub-and-spoke circuit drives pheromone attraction

and social behaviour in C. elegans. Nature 458, 1171–1175 (2009).

31. Inagaki, H.K. et al. Visualizing neuromodulation in vivo: TANGO-mapping

of dopamine signaling reveals appetite control of sugar sensing. Cell 148,

583–595 (2012).

32. Sharp, A.A., O’Neil, M.B., Abbott, L.F. & Marder, E. Dynamic clamp: computer-

generated conductances in real neurons. J. Neurophysiol. 69, 992–995 (1993).

33. Kawaguchi, Y. & Kondo, S. Parvalbumin, somatostatin and cholecystokinin

as chemical markers for speciﬁc GABAergic interneuron types in the rat

frontal cortex. J. Neurocytol. 31, 277–287 (2002).

34. Egorov, A.V., Hamam, B.N., Fransen, E., Hasselmo, M.E. & Alonso, A.A.

Graded persistent activity in entorhinal cortex neurons. Nature 420,

173–178 (2002).

35. Aston-Jones, G. & Cohen, J.D. An integrative theory of locus coeruleus-

norepinephrine function: adaptive gain and optimal performance. Annu.

Rev. Neurosci. 28, 403–450 (2005).

36. Zucker, R.S. & Regehr, W.G. Short-term synaptic plasticity. Annu. Rev.

Physiol. 64, 355–405 (2002).

37. Masland, R.H. The neuronal organization of the retina. Neuron 76,

266–280 (2012).

38. Grimes, W.N., Zhang, J., Graydon, C.W., Kachar, B. & Diamond, J.S. Retinal

parallel processors: more than 100 independent microcircuits operate

within a single interneuron. Neuron 65, 873–885 (2010).

39. Hosoya, T., Baccus, S.A. & Meister, M. Dynamic predictive coding by the

retina. Nature 436, 71–77 (2005).

40. Farrow, K. et al. Ambient illumination toggles a neuronal circuit

switch in the retina and visual perception at cone threshold. Neuron 78,

325–338 (2013).

41. Rivlin-Etzion, M., Wei, W. & Feller, M.B. Visual stimulation reverses

the directional preference of direction-selective retinal ganglion cells.

Neuron 76, 518–525 (2012).

42. Witkovsky, P. Dopamine and retinal function. Doc. Ophthalmol. 108, 17–40

(2004).

43. Ciocchi, S. et al. Encoding of conditioned fear in central amygdala

inhibitory circuits. Nature 468, 277–282 (2010).

44. Marder, E. & Taylor, A.L. Multiple models to capture the variability in

biological neurons and networks. Nat. Neurosci. 14, 133–138 (2011).

45. Milo, R. et al. Network motifs: simple building blocks of complex

networks. Science 298, 824–827 (2002).

46. Markram, H. The blue brain project. Nat. Rev. Neurosci. 7, 153–160

(2006).

47. Koch, C. & Reid, R.C. Neuroscience: Observatories of the mind.

Nature 483, 397–398 (2012).

48. Alivisatos, A.P. et al. The brain activity map project and the challenge

of functional connectomics. Neuron 74, 970–974 (2012).

49. Koch, C. Systems biology. Modular biological complexity. Science 337,

531–532 (2012).

50. Varshney, L.R., Chen, B.L., Paniagua, E., Hall, D.H. & Chklovskii, D.B.

Structural properties of the Caenorhabditis elegans neuronal network.

PLoS Comput. Biol. 7, e1001066 (2011).

npg

A Perspective on Brain Symbolization Blueprint to Understand Abstraction of the Mind

Preprint

Full-text available

Dec 2024

Yeongseok Chae

To understand how a brain works, neuroscience have emerged and progressed tremendous studies on neurons. Among them, there are many important achievements including the discovery of neural structures and functions. It was found that brain information is transported via neural synapses, and the synapse formation generates memory. Furthermore, artificial intelligence using the neural network model has reached to imitate high cognitive functions of humans. However, understanding fundamental principles of neural computation and plasticity still seem overwhelming and ambitious. We still do not know how the brain deals with information and how neural circuits reflect external environments, which indicates that general research strategy applied to other biological organs is not effective enough. To ameliorate this contradictory situation, here, I discuss that the difference between the brain and other organs stems from the high level of functional abstractions, especially the symbolizations that make a discrepancy between physical and functional boundaries in neural networks. And I review crucial findings in neuroscience in terms of how we can reduce the discrepancy by building a brain symbolization blueprint.

Capillary connections between sensory circumventricular organs and adjacent parenchyma enable local volume transmission

Article

Full-text available

Jan 2025
J NEUROENDOCRINOL

Among contributors to diffusible signaling are portal systems which join two capillary beds through connecting veins. Portal systems allow diffusible signals to be transported in high concentrations directly from one capillary bed to the other without dilution in the systemic circulation. Two portal systems have been identified in the brain. The first was discovered almost a century ago and connects the median eminence to the anterior pituitary gland. The second was discovered a few years ago and links the suprachiasmatic nucleus to the organum vasculosum of the lamina terminalis, a sensory circumventricular organ (CVO). Sensory CVOs bear neuronal receptors for sensing signals in the fluid milieu. They line the surface of brain ventricles and bear fenestrated capillaries thereby lacking blood–brain barriers. It is not known whether the other sensory CVOs, namely the subfornical organ (SFO), and area postrema (AP) form portal neurovascular connections with nearby parenchymal tissue. To preserve the integrity of the vasculature of CVOs and their adjacent neuropil, we combined iDISCO clearing and light‐sheet microscopy to acquire volumetric images of blood vessels and traced the vasculature in two experiments. In the first, the whole brain vasculature was registered to the Allen Brain Atlas in order to identify the nuclei to which the SFO and AP are attached. In the second study, regionally specified immunolabeling was used to identify the attachment sites and vascular connections between the AP, and the SFO to their respective parenchymal attachment sites. There are venous portal pathways linking the capillary vessels of the SFO and the posterior septal nuclei, namely the septofimbrial nucleus and the triangular nucleus of the septum. Unlike the arrangement of portal vessels, the AP and the nucleus of the solitary tract share a common capillary bed. Taken together, the results reveal that all three sensory CVOs bear direct capillary connections to adjacent neuropil, providing a direct route for diffusible signals to travel from their source to their targets.

An integrative data-driven model simulating C. elegans brain, body and environment interactions

Article

Full-text available

Dec 2024

The behavior of an organism is influenced by the complex interplay between its brain, body and environment. Existing data-driven models focus on either the brain or the body–environment. Here we present BAAIWorm, an integrative data-driven model of Caenorhabditiselegans, which consists of two submodels: the brain model and the body–environment model. The brain model was built by multicompartment models with realistic morphology, connectome and neural population dynamics based on experimental data. Simultaneously, the body–environment model used a lifelike body and a three-dimensional physical environment. Through the closed-loop interaction between the two submodels, BAAIWorm reproduced the realistic zigzag movement toward attractors observed in C. elegans. Leveraging this model, we investigated the impact of neural system structure on both neural activities and behaviors. Consequently, BAAIWorm can enhance our understanding of how the brain controls the body to interact with its surrounding environment.

Estimating the impact of structural directionality: How reliable are undirected connectomes?

Preprint

Jan 2018

Directionality is a fundamental feature of network connections. Most structural brain networks are intrinsically directed because of the nature of chemical synapses, which comprise most neuronal connections. Due to limitations of non-invasive imaging techniques, the directionality of connections between structurally connected regions of the human brain cannot be confirmed. Hence, connections are represented as undirected, and it is still unknown how this lack of directionality affects brain network topology. Using six directed brain networks from different species and parcellations (cat, mouse, C. elegans, and three macaque networks), we estimate the inaccuracies in network measures (degree, betweenness, clustering coefficient, path length, global efficiency, participation index, and small worldness) associated with the removal of the directionality of connections. We employ three different methods to render directed brain networks undirected: (i) remove uni-directional connections, (ii) add reciprocal connections, and (iii) combine equal numbers of removed and added uni-directional connections. We quantify the extent of inaccuracy in network measures introduced through neglecting connection directionality for individual nodes and across the network. We find that the coarse division between core and peripheral nodes remains accurate for undirected networks. However, hub nodes differ considerably when directionality is neglected. Comparing the different methods to generate undirected networks from directed ones, we generally find that the addition of reciprocal connections (false positives) causes larger errors in graph-theoretic measures than the removal of the same number of directed connections (false negatives). These findings suggest that directionality plays an essential role in shaping brain networks and highlight some limitations of undirected connectomes.

Parallel and converging multisensory cascades in the Drosophila connectome

Preprint

Full-text available

Dec 2024

Connectomes are network maps of synaptic connectivity. A key functional role of any connectome is to constrain inter-neuronal signaling and to sculpt the flow of activity across the nervous system. Connectomes therefore play a central role in the rapid tranmission of information about an organism’s environment from sensory neurons to higher-order neurons for action planning and ultimately to effectors. Here, we use a parsimonious model of activity spread to investigate the connectome’s role in shaping putative sensory cascades. Our model allows us to simulate the signaling pathways from sensors to the rest of brain, mapping the similarity of these pathways between different sensory modalities and identifying convergence zones–neurons that are activated simultaneously by different sensory modalities. Further, we considered two multisensory integration scenarios – a cooperative case where different sensory modalities interacted to “speed up” (reduce) neurons’ activation times and a competitive “winner take all” case, where different sensory streams vied for the same neural territory. Finally, we use a data-driven algorithm to partition neurons into different classes based on their behavior during the cascade simulations. Our work helps underscore the role of “simple” models for enriching connectome data, while offering a data-driven classification of neurons based on their joint connectional/dynamical properties.

Neuromodulation and homeostasis: complementary mechanisms for robust neural function

Preprint

Full-text available

Dec 2024

Neurons depend on two interdependent mechanisms-homeostasis and neuromodulation-to maintain robust and adaptable functionality. Homeostasis stabilizes neuronal activity by adjusting ionic conductances, whereas neuromodulation dynamically modifies ionic properties in response to external signals. Combining these mechanisms in conductance-based models often produces unreliable outcomes, particularly when sharp neuromodulation interferes with homeostatic tuning. This study explores how a biologically inspired neuromodulation controller can harmonize with homeostasis to ensure reliable neuronal function. Using computational models of stomatogastric ganglion and dopaminergic neurons, we demonstrate that controlled neuromodulation preserves neuronal firing patterns while maintaining intracellular calcium levels. Unlike sharp neuromodulation, the neuromodulation controller integrates activity-dependent feedback through mechanisms mimicking G-protein-coupled receptor cascades. The interaction between these controllers critically depends on the existence of an intersection in conductance space, representing a balance between target calcium levels and neuromodulated firing patterns. Maximizing neuronal degeneracy enhances the likelihood of such intersections, enabling robust modulation and compensation for channel blockades. We further show that this controller pairing extends to network-level activity, reliably modulating central pattern generators in crustaceans. These findings suggest that targeting neuromodulation pathways-rather than ion channels directly-may offer safer pharmacological strategies to manage neuronal dysfunctions. This study highlights the complementary roles of homeostasis and neuromodulation, proposing a unified control framework for maintaining robust and adaptive neural activity under physiological and pathological conditions.

The effects of overnight sleep deprivation or total sleep deprivation on the body

Article

Full-text available

Jun 2024

Nataliya Volotovska

This analytical and scientific work explores the multifaceted effects of overnight and total sleep deprivation on the human body. Sleep deprivation, increasingly prevalent due to the modern pace of life, significantly impacts various physiological systems including the cardiovascular, nervous, reproductive, digestive, and immune systems. It also leads to hormonal imbalances and cognitive impairments, manifesting in serious somatic and mental illnesses. Despite extensive research, the mechanisms through which sleep influences these body systems remain incompletely understood. This review aims to synthesize current knowledge on the subject, emphasizing the importance of early identification and treatment of sleep disorders to prevent a range of health issues and enhance overall well-being.

The recurrent temporal restricted Boltzmann machine captures neural assembly dynamics in whole-brain activity

Article

Full-text available

Nov 2024
eLife

Animal behaviour alternates between stochastic exploration and goal-directed actions, which are generated by the underlying neural dynamics. Previously, we demonstrated that the compositional Restricted Boltzmann Machine (cRBM) can decompose whole-brain activity of larval zebrafish data at the neural level into a small number (∼100-200) of assemblies that can account for the stochasticity of the neural activity (van der Plas et al., eLife, 2023). Here, we advance this representation by extending to a combined stochastic-dynamical representation to account for both aspects using the recurrent temporal RBM (RTRBM) and transfer-learning based on the cRBM estimate. We demonstrate that the functional advantage of the RTRBM is captured in the temporal weights on the hidden units, representing neural assemblies, for both simulated and experimental data. Our results show that the temporal expansion outperforms the stochastic-only cRBM in terms of generalization error and achieves a more accurate representation of the moments in time. Lastly, we demonstrate that we can identify the original time-scale of assembly dynamics by estimating multiple RTRBMs at different temporal resolutions. Together, we propose that RTRBMs are a valuable tool for capturing the combined stochastic and time-predictive dynamics of large-scale data sets.

A brief history of insect neuropeptide and peptide hormone research

Article

Full-text available

Dec 2024
CELL TISSUE RES

Dick Nässel

This review briefly summarizes 50 years of research on insect neuropeptide and peptide hormone (collectively abbreviated NPH) signaling, starting with the sequencing of proctolin in 1975. The first 25 years, before the sequencing of the Drosophila genome, were characterized by efforts to identify novel NPHs by biochemical means, mapping of their distribution in neurons, neurosecretory cells, and endocrine cells of the intestine. Functional studies of NPHs were predominantly dealing with hormonal aspects of peptides and many employed ex vivo assays. With the annotation of the Drosophila genome, and more specifically of the NPHs and their receptors in Drosophila and other insects, a new era followed. This started with matching of NPH ligands to orphan receptors, and studies to localize NPHs with improved detection methods. Important advances were made with introduction of a rich repertoire of innovative molecular genetic approaches to localize and interfere with expression or function of NPHs and their receptors. These methods enabled cell- or circuit-specific interference with NPH signaling for in vivo assays to determine roles in behavior and physiology, imaging of neuronal activity, and analysis of connectivity in peptidergic circuits. Recent years have seen a dramatic increase in reports on the multiple functions of NPHs in development, physiology and behavior. Importantly, we can now appreciate the pleiotropic functions of NPHs, as well as the functional peptidergic “networks” where state dependent NPH signaling ensures behavioral plasticity and systemic homeostasis.

Predicting modular functions and neural coding of behavior from a synaptic wiring diagram

Article

Full-text available

Nov 2024
NAT NEUROSCI

A long-standing goal in neuroscience is to understand how a circuit’s form influences its function. Here, we reconstruct and analyze a synaptic wiring diagram of the larval zebrafish brainstem to predict key functional properties and validate them through comparison with physiological data. We identify modules of strongly connected neurons that turn out to be specialized for different behavioral functions, the control of eye and body movements. The eye movement module is further organized into two three-block cycles that support the positive feedback long hypothesized to underlie low-dimensional attractor dynamics in oculomotor control. We construct a neural network model based directly on the reconstructed wiring diagram that makes predictions for the cellular-resolution coding of eye position and neural dynamics. These predictions are verified statistically with calcium imaging-based neural activity recordings. This work demonstrates how connectome-based brain modeling can reveal previously unknown anatomical structure in a neural circuit and provide insights linking network form to function.

Multiple Mechanisms Switch an Electrically Coupled, Synaptically Inhibited Neuron between Competing Rhythmic Oscillators

Article

Full-text available

Mar 2013

Rhythmic oscillations are common features of nervous systems. One of the fundamental questions posed by these rhythms is how individual neurons or groups of neurons are recruited into different network oscillations. We modeled competing fast and slow oscillators connected to a hub neuron with electrical and inhibitory synapses. We explore the patterns of coordination shown in the network as a function of the electrical coupling and inhibitory synapse strengths with the help of a novel visualization method that we call the "parameterscape." The hub neuron can be switched between the fast and slow oscillators by multiple network mechanisms, indicating that a given change in network state can be achieved by degenerate cellular mechanisms. These results have importance for interpreting experiments employing optogenetic, genetic, and pharmacological manipulations to understand circuit dynamics.

The Structure of the Nervous System of the Nematode Caenorhabditis elegans

Article

Jan 1986

Dynamic Clamp - Computer-Generated Conductances in Real Neurons

Article

Apr 1993

1. We describe a new method, the dynamic clamp, that uses a computer as an interactive tool to introduce simulated voltage and ligand mediated conductances into real neurons. 2. We simulate a gamma-aminobutyric acid (GABA) response of a cultured stomatogastric ganglion neuron to illustrate that the dynamic clamp effectively introduces a conductance into the target neuron. 3. To demonstrate an artificial voltage-dependent conductance, we simulate the action of a voltage-dependent proctolin response on a neuron in the intact stomatogastric ganglion. We show that shifts in the activation curve and the maximal conductance of the response produce different effects on the target neuron. 4. The dynamic clamp is used to construct reciprocal inhibitory synapses between two stomatogastric ganglion neurons that are not coupled naturally, illustrating that this method can be used to form new networks at will.

S HORT -T ERM S YNAPTIC P LASTICITY

Article

Mar 2002

Abstract Synaptic transmission is a dynamic process. Postsynaptic responses wax and wane as presynaptic activity evolves. This prominent characteristic of chemical synaptic transmission is a crucial determinant of the response properties of synapses and, in turn, of the stimulus properties selected by neural networks and of the patterns of activity generated by those networks. This review focuses on synaptic changes that result from prior activity in the synapse under study, and is restricted to short-term effects that last for at most a few minutes. Forms of synaptic enhancement, such as facilitation, augmentation, and post-tetanic potentiation, are usually attributed to effects of a residual elevation in presynaptic [Ca2+]i, acting on one or more molecular targets that appear to be distinct from the secretory trigger responsible for fast exocytosis and phasic release of transmitter to single action potentials. We discuss the evidence for this hypothesis, and the origins of the different kinetic phases of synaptic enhancement, as well as the interpretation of statistical changes in transmitter release and roles played by other factors such as alterations in presynaptic Ca2+ influx or postsynaptic levels of [Ca2+]i. Synaptic depression dominates enhancement at many synapses. Depression is usually attributed to depletion of some pool of readily releasable vesicles, and various forms of the depletion model are discussed. Depression can also arise from feedback activation of presynaptic receptors and from postsynaptic processes such as receptor desensitization. In addition, glial-neuronal interactions can contribute to short-term synaptic plasticity. Finally, we summarize the recent literature on putative molecular players in synaptic plasticity and the effects of genetic manipulations and other modulatory influences.

Modular Biological Complexity

Article

Aug 2012

Christof Koch

It has been argued that the technological capability to fully simulate the human brain on digital computers will exist within a decade. This is taken to imply that we will comprehend its functioning, eliminate all diseases, and “upload” ourselves to computers (1). Although such predictions excite the imagination, they are not based on a sound assessment of the complexity of living systems. Such systems are characterized by large numbers of highly heterogeneous components, be they genes, proteins, or cells. These components interact causally in myriad ways across a very large spectrum of space-time, from nanometers to meters and from microseconds to years. A complete understanding of these systems demands that a large fraction of these interactions be experimentally or computationally probed. This is very difficult.

Ambient Illumination Toggles a Neuronal Circuit Switch in the Retina and Visual Perception at Cone Threshold

Article

Mar 2013

Video abstract: Gradual changes in the sensory environment can lead to abrupt changes in brain computations and perception. However, mechanistic understanding of the mediating microcircuits is missing. By sliding through light levels from starlight to daylight, we identify retinal ganglion cell types in the mouse that abruptly and reversibly switch the weighting of center and surround interactions in their receptive field around cone threshold. Two-photon-targeted recordings and genetic and viral tracing experiments revealed that the circuit element responsible for the switch is a large inhibitory neuron that provides direct inhibition to ganglion cells. Our experiments suggest that weak excitatory input via electrical synapses together with the spiking threshold in inhibitory cells act as a switch. We also reveal a switch-like component in the spatial integration properties of human vision at cone threshold. This work demonstrates that circuits in the retina can quickly and reversibly switch between two distinct states, implementing distinct perceptual regimes at different light levels.

Simpler Networks and Behavior

Book

Jan 1976

John C. Fentress

Visual Stimulation Reverses the Directional Preference of Direction-Selective Retinal Ganglion Cells

Article

Nov 2012

Direction selectivity in the retina is mediated by direction-selective ganglion cells. These cells are part of a circuit in which they are asymmetrically wired to inhibitory neurons. Thus, they respond strongly to an image moving in the preferred direction and weakly to an image moving in the opposite (null) direction. Here, we demonstrate that adaptation with short visual stimulation of a direction-selective ganglion cell using drifting gratings can reverse this cell’s directional preference by 180°. This reversal is robust, long lasting, and independent of the animal’s age. Our findings indicate that, even within circuits that are hardwired, the computation of direction can be altered by dynamic circuit mechanisms that are guided by visual stimulation. Video Abstract eyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiIzNzE5M2RkNzUzZDc0ZjEwNzdhNmVlNTFkZTc0YzA5OSIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjAwNTI3NDAzfQ.OUv-PiwLDyedPvuvjsHdwijXWkUIuG-2CmXKW4MYhiI0btc0kfAyy_sm1DStSo1oGZhljDF0gk4cLg8ucUxbMqVjGtRL942AlK74Xy0ZPGy7Zb9Zz9vq21xWcmcsM3oh8CrO964httLmxIaILu3osQ3ZesDrY_EXNoYVQbka60FHAq1wv5PH8FmisirXUJEt7mOKAwz3yMBP2EE8_6GUIEQL-scduvBA0duTTOpWZQCxkus-72EdlchN4IFDi4lYcRP8YXvwfYJ57ddp8l17A1198v_oEhBm5QRBNhcW8CW8JttNwhLtkcMnToxFvE5wI81cwePn_w7XUVpjFdGFgQ (mp4, (13.25 MB) Download video

The Genetic Analysis of Functional Connectomics in Drosophila

Article

Oct 2012
Adv Genet

Fly and vertebrate nervous systems share many organizational features, such as layers, columns and glomeruli, and utilize similar synaptic components, such as ion channels and receptors. Both also exhibit similar network features. Recent technological advances, especially in electron microscopy, now allow us to determine synaptic circuits and identify pathways cell-by-cell, as part of the fly's connectome. Genetic tools provide the means to identify synaptic components, as well as to record and manipulate neuronal activity, adding function to the connectome. This review discusses technical advances in these emerging areas of functional connectomics, offering prognoses in each and identifying the challenges in bridging structural connectomics to molecular biology and synaptic physiology, thereby determining fundamental mechanisms of neural computation that underlie behavior.

The Neuronal Organization of the Retina

Article

Oct 2012

Richard H. Masland

The mammalian retina consists of neurons of >60 distinct types, each playing a specific role in processing visual images. They are arranged in three main stages. The first decomposes the outputs of the rod and cone photoreceptors into ∼12 parallel information streams. The second connects these streams to specific types of retinal ganglion cells. The third combines bipolar and amacrine cell activity to create the diverse encodings of the visual world-roughly 20 of them-that the retina transmits to the brain. New transformations of the visual input continue to be found: at least half of the encodings sent to the brain (ganglion cell response selectivities) remain to be discovered. This diversity of the retina's outputs has yet to be incorporated into our understanding of higher visual function.