Download Neuronal networks for induced `40 Hz` rhythms

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Transcript
R
V. Cornilleau-P6r&
E
and C. Gielen – Processing of optic flow
36 Buizza, A. et al. (1980) Exp. Brain Res. 39, 165-176
37 Nakayama, K. (1981) VisionRes. 21, 1475-1482
38 Dijkstra,T.M.H. et al. (1995) VisionRes. 35, 453-462
39 Van den Berg, A.V. and Collewijn, H. (1986) Vision Res. 26,
1209–1222
Acknowledgements 40 Ungerleider,L.G. and Mishkin, M. (1982) in Analysis of Visual
Behavior (Ingle, D.J., Goodale,M.A. and Mansfield,R.J.W., eds),
This work was
pp. 549–586, MIT Press
supportedby Esprit
41 Maunsell, J.H.R. and Van Essen, D.C. (1983) J. Neurosci.3,
ProjectMucorrc2
2563-2586
6615. We thank 42 Koenderink, J.J. and Van Doom, A.J. (1975) Optica Acta 22,
773-791
TjeerdDijkstra,
43 Saito, H. et al. (1986) J. Neurosci. 6, 145-157
[acques Droulez,
44 Lagae, L. et aZ. (1994) /. NeurophysioL71, 1597-1626
FranckBrernmer 45 Duffy, C.J. and Wurtz, R.H. (1991) ]. NeurophysioL 65,
1329-1345
and WernerGraf
for valuable 46 Motter, B.C. and Mountcastle, V.B. (1981) 1. Neccrosci.1, 3–26
47 Orban, G.A.et al. (1992) Proc.NatlAcad. Sci. USA89, 2595–2599
comments on the 48 Allman,J., Miezin,F. and McGuinness,E. (1985) Perception14,
105–126
manusm”pt.
V
I
49 Maunsell, J.H.R. and Van Essen, D.C. (1983) J. Neurophysiol.
49, 1148-1167
50 Roy, J.P. and Wurtz, R.H. (1990) Nature 348, 160-162
51 Bradley, D.C., Qian, N. and Andersen,R.A. (1995) Nature 373,
609-611
52 Andersen, R.A. (1987) in Handbook of Physiology (Section 1,
Vol. 5) (Brookhart, J.M. and Mountcastle, V.B., eds),
PP. 483-518, AmericanPhysiologicalSociety
53 Komatsu, K.and Wurtz, R.H. (1988) ]. NeurophysioL60, 580-603
54 Sakata, H., Shibutani, H. and Kawano, K. (1980)
J. Neurophysiol.43, 1654-1672
55 Kawano, K., Sasaki, M. and Yamashita, M. (1984)
J. Neurophysiol.51, 340-351
56 Thier, P. and Erickson, R.G. (1992) Eur.J. Neurosci. 4, 539-553
57 Berman, N., Blakemore, C. and Cynader, M. (1975) J. PhysioL
246, 595-615
58 Hoffmann, K.P. (1982) in Functional Basis of Ocular Motility
Disorder (Lennerstrand, G., Zee, D.S. and Keller, E.L., eds),
pp. 303–311, Pergamon
59 Grasse,K.L. (1994) VisionRes. 34, 1673-1689
Neuronal networks for induced ’40 Hz’
rhythms
John G.R. Jefferys, Roger D. Traub and Miles A. Whittington
A fast,coherentEEG rhythm,calleda gamma or a ’40Hz’ rhythm,hasbeen implicatedboth in
higherbrainfunctions,suchasthe‘binding’of featuresthat are detectedby sensorycorticesinto
perceivedobjects,and in lower levelprocesses,
suchas the phasecodingof neuronalactivity.
Computer simulationsof severalparts of the brain suggestthat gamma rhythms can be
generatedby poolsof excitatoryneurones,networksof inhibitoryneurones,or networksof both
excitatoryand inhibitoryneurones.The strongestexperimentalevidencefor rhythm generators
J.G.R.Jeff2rysis hasbeen shownfor: (1) neocorticaland thalamicneuronesthat are intrinsic’40Hz’ oscillators,
at the Dept of although synchronystill requiresnetwork mechanisms;
and (2) hippocampaland neocortical
Physiology,The
networksof mutually inhibitory interneuronesthat generatecollective40Hz rhythms when
MedicalSchool,
[email protected] excitedtonically.
Birmingham,
Trends Neurosci. (1996) 19, 202-208
Birmingham,UX
B1527T.
In humans the auditory response includes brief
AST, GAMMA RHYTHMS have been implicated in
MilesA.
Whittbcgtonis at
higher cognitive function. They are also known as ‘40Hz transientresponses’’9’20,which increasewhen the
the Neuronal ‘40 Hz’rhythms, but actually range from 30 to 100Hz subjectpaysattention, andwhich disappearwith lossof
Repetitiveauditory
NetworksGroup, and might vary in frequency during a response. The consciousnessduringanaesthesia21.
DeptofPhysiology 20-100 Hzrangeweconsiderhereoverlapswith the beta stimulation at -40 Hz generates a large ’40 Hz steadyand Biophysics, band (15–30Hz)of the EEG,but wewillignorethe finer state response’zz.Recordingsof brain magnetic activity
StMary’sHospital points of EEG classification. The natural history and (magnetoencephalogramsor MEGs)in humans suggest
during
MedicalSchool, functionalrolesof synchronousgammaoscillationshave that gammarhythms can be verywidespread23,
Imperial College, been reviewedrecentlyl-3,and sowillbe consideredonly both waking and dream states. Other MEG measurements in humans suggestthat gamma rhythms might
London, briefly.
Gamma rhythms occur in humans and other mam- be organizedto sweepacrossthe whole brain, perhaps
UK W2 lPG.
RogerD. Traub is at mals following sensory stimuli, often in brief runs. providing ‘temporal binding...into a single cognitive
the IBMResearch ‘Inducedrhythms’at 50-60 Hzwerefirstdescribedin the experience’24.
Division,T.!. olfactorybulbby Adrian4,andhavesincebeenidentified
Neuronalfiring
auditorycorWatson Research in the olfactory cortexs, visualcortex3’&9,
somatosensory cortexlz and motor cortex13-15. Single-unit recordings in vivo have revealed much
Center,Yorktown tex10’11,
Heights,NY 10598, Gammaoscillationsalsooccur in the hippocampuslc’17, aboutthe eventsor featuresto which neuronesrespond.
USA,and the Dept wherethe linkwithexternalsensorystimuliis lessdirect, Individualneuronesdonot detecttheirpreferredsensory
of Neurology, but might stillexistin the formof multimodalinputsre- featuresin isolation,but formpartof neuronalnetworks
Columbia ceivedfrom higher-ordersensorycortices.Hippocampal whoseemergentpropertiesdefinethe feature-detection
Universi~,New gamma rhythms tend to occur during the theta band propertiesof the corticalcolumn. In the visualsystem,it
York,NY 10032, (4-12 Hz)of the EEG,whichis a prominentfeatureof the usedto be thought that successivehierarchiesof neurespeciallyduringexploration. ones encoded progressivelymore-complex features of
USA. hippocampusin vivo1618,
F
202
TINS Vol. 19, No. 5, 1996
Copyright @ 1996, Elsevier Science Ltd. All rights reserved. 0166-
2236/96/S15.00
PII: S0166-2236(96)1
OO23-O
REVIEW
J.G.R. Jefferys et cd. – Neuronal networks for induced ‘40 Hz’ rhythms
A
c
B
Recurrent
inhibition
Mutual
excitation
D
Intrinsic
oscillator
in network
Mutual
inhibition
Fig. 2. Simplified representationsto illustrate the essentialfeaturesof severalmechanismsproposed ta be involved
in the generation of gamma oscillations. In each case, E (excitatory) and / (inhibitory) representnetworks of
neuronesthat are mutually connected,the continuouslines indicate the key connectionsfor their respectivemechanisms,and the dot-dash arrow indicatesthe ffow of specificinformation through the network.(A) i//ustratesthe recurrent inhibitory loop model proposedby Freemanet al.s Computertheoreticalanalysishas subsequentlyshown that
mutual excitationisrequired,but that mutua/ inhibition isnot. (B) showsa simi/armode/in whichthe time de/aysa/ong
the axonscouplinggroupsof excitatoryneuronesplay a keyro/e33,and whichalsoreceivescontributionsfrom recurrent
inhibition. (C) proposesthat neuroneswith intrinsic-osci//atorpropertiescan imposetheir own rhythm on the synaptic
networkin which theyare embedded34-37.
(D) representsour own mode/of gamma oscillationsin the hippocampus,in
which interneuronesare tonica//yexcited(thin arrows)so that they will fire at a rate >40 Hz. Thedivergentinhibitory
connectionsbetween theseneuronesresult in synchronizedinhibition acrossthe population. When this decays,the
neuraneswill discharaedue
to the tanic excitationthat drivesthe rhvthm38imDosinoo
rhvthm of about 40 Hz. Notice
.
,./,
that in its simplestform (D) separatesthe role of the oscillator(or clock)and the processor.
sites, or to technical differences(gammarhythms can
occur in brief bursts with a considerable jitter in the
frequency”, so any correlation could conceivably be
smudgedout when measurementsare averagedduring
0.5s runs of an EEG, or 20 cycles at 40 Hz; Ref. 28).
However,the reasons for these discrepanciesremain
unresolved.
Coherent rhythms might have other functions. One
idea is that they providea timing referencefor a neural
code that depends on the phase relationship of individual neurones with the reference oscillation. The
stronger the excitation to an individualneurone, the
earlierin the cycle it willfire.Thus neuronesthat fire at
similarphasesin the rhythm will have receivedsimilar
intensifies of input which might be used,for example,
to lock their outputstogether for a more effectivesummation.Thishypothesiswasproposedfortheta rhythms
in the hippocampus30,and also, more recently, for
gamma rhythms31.
The role of gammarhythmsis unknown.They might
be central to our cognitivefunction, be fundamentalto
the neural code, have some entirely different role, or
simplybean epiphenomenon32with no deepmeaning.
We believethat one key step to resolvingthese issuesis
to understandthe cellularandnetworkmechanismsthat
generate gamma rhythms, and to developpharmacologicaltools that willallowus to probetheirrolesin vivo.
What drivesthegammarhythm?
The original models for binding and segmentation
introduced the idea that neurones that oscillated
togetheralsoworkedtogethe~s,reflectingthat synchronization is a more important factor than a narrow
bandwidth327.Although single episodes of neuronal
synchronization might occur by chance, repeatedsynchronizationis much lesslikelyto do so. Atthe cellular
204
TINS VOI. 19, NO. 5,1996
levelrepeatedsynchronizationcould
promote temporal summation at
active synapses.
Severaltheories exist for the generation of gamma oscillations in
variouspartsof the brain, which all
need furtherexperimentaltesting.It
is possible that gamma oscillations
ariseby differentmechanismsin different parts of the brain, and that
severalmechanismscan combine in
individual regions. Figure 2 shows
highly simplifiedrepresentationsof
some of the components of these
differentmechanisms,whichwewill
consider below in roughly chronological order.
Feedback loops between excitatory and
inhibitory neurones
(The main regionsimplicatedinclude: the olfactory bulb, the piriform cortex, the entorhinal cortex
and the primary visual cortex.)
Freemanand colleaguesdevelopeda
model for induced rhythms in several olfactorystructures,which proposed that synchronous oscillation
is generated by a feedback loop
between excitatory and inhibitory
neuroness.Theyproposedthat some
mutual connectivity was also requiredwithinthe poolsof both excitato~ andinhibitory
neuronesto stabiiizethe oscillations. Ermentrout39has
shown that mutual excitation amongst the excitatory
neurones is necessaryfor stable oscillations to be generated by a recurrent inhibitory loop. However, our
recent simulationssuggestthat other conditions might
sufficefor stable oscillations (R.D.Traub, unpublished
observations).
Freemanet al.s predictedthat inhibitory cells should
lag behind the excitatory cells by a quarter of a cycle
(6.5 ms at 40Hz). Experimental support came from
single-unitand EEGrecordingsin vivofromthe olfactory
bulb, anteriorolfactorynucleus,prepiriformcortex and
entorhinalcortexs.The signalsfell into two groups:one
set firedin phasewith the gamma EEG,and one either
led or laggedthe gamma EEGby a quarter of a cycle.
Unfortunatelythese measurementscannot identifythe
types of neurones in each group. In contrast, hippocampalinterneuronesrecordedduringthe gammaEEG
fire in phasewith pyramidalcells”. This is predictedby
our inhibitory networkmodel (see below), both when
isolatedfrom the excitatory network38,and when connected with pyramidalcells (R.D. Traub et aZ.,unpublished observations). Why the hippocampal and the
(superficiallysimilar)olfactorycortical circuitryshould
differ remains unclear5.
Wilson and Bowermade similar models of the piriform cortex33and the primaryvisualcortex32.The geometric structure of these models differed, but the
essential idea in both was that the amplitudeand the
frequency of coherent 30-60Hz oscillations, elicited
by afferent volleys, were determined (or ‘tuned’) by a
fast-feedbackinhibitory loop (Fig. 2A). Essentially, if
the stimulus is appropriate (not too strong), enough
activityin the recurrentexcitatoryconnectionsbetween
pyramidalcells persists after the recurrent inhibition
J.G.R. Jefferys et al.–
wanes in order to re-excite the pyramidal-cellpopulation. In the case of the piriform-cortex model, they
showed that the time constant of inhibition ‘tuned’
the frequency of the gamma rhythm, so that longer
time spent open for the chloride channels resultedin
slowerrhythms (and also a loss of power).
In their model of the primaryvisual cortex Wilson
and Bower32note, in passing, that local mutual inhibition betweenthe interneurones‘improvedfrequency
locking and produced auto- and cross-correlations
with more pronouncedoscillatorycharacteristics’.This
differs from the central role of similar connections in
the generationof gammarhythmsin the hippocampus,
where they were both necessaryand sufficient38,40.
In the visual-cortex model, horizontal pyramidal
cell axons wereessential for long-range(>1mm) crosscorrelations32.These had zero phase lag as long as the
excitatory postsynaptic potentials (EPSPS)that they
generatedwerenot too strong. StrongerEPSPSresultin
phaselagsconsistentwith delaysin axonal conduction,
while weakerEPSPSwere reminiscent of other kinds of
looselycoupledoscillators.In both the visual-cortexand
the piriform-cortex versions of this model, gamma
rhythms arose from interactions between networksof
excitatory neurones, could dependon the conduction
velocitiesof intrinsic cortical connections (Fig.2B), and
were tuned by the time constants of excitatory and
inhibitory synapses.We are not awareof any attempts
to dissectthese complexinteractionsexperimentally;in
particular,an investigationof the effectsof conduction
delayson cortical oscillations wouldbe instructive.
Intrinsic oscillations in individual neurones
(The main regions implicated include the thalamus
and the neocortex; seeFig.2C.) Neuronesin many parts
of the brain have the intrinsic capacity to oscillate at
about 40 Hz. Space does not permit an exhaustive reviewof intrinsic oscillators;here we outline one or two
relevant cases. For example, severaltypes of neurones
in the thaiamocortical system such as the reticular34
and intralaminar35neurones do so. In the neocortex
itself, examplesof intrinsic cellularoscillators include:
sparselyspiny, layer-4 neurones3b,about 20% of longaxon projection neurones in layers 5 and 6 (Ref. 37),
and ‘chattering cells’ (cells that fire brief trains of
action potentials at 200 Hz about 40times a second),
which were recently reported in vivo41.
Slice studies revealed that oscillations of 40Hz in
sparselyspiny neurones in the frontal cortex are generatedby persistent,voltage-dependentNa+currentsand
delayed voltage-dependentrectifier currents3b.Other
frontal-cortexneuronesuse fast persistentNa+currents,
leak and slow non-inactivating K+currentsto generate
oscillations of 4–20 Hz (Ref. 42). Various models suggest that similarmechanisms can generateoscillations
of 40 Hz (Ref.43). At least some cortical neuroneswith
intrinsic oscillator mechanismsproject to contralateral
areas, and to the thalamus, providingroutes for longrangesynchronizationof these oscillations37.The existence of cells with intrinsic oscillations at -40 Hz does
not in itself explain the synchronizationof local populations of neurones, but it is likely to pace population
rhythms when the neurones are suitably coupled by
chemical or electrical synapsesor both44.
A
REVEW
Neuronal networks for induced ‘40 Hz’ rhythms
Isolated interneurone
/
B
Interneurone network
50 mV
50 ms
Fig. 3. Inhibitory neuronal networks generate gamma oscillations.
(A) Computersimulationof the briskexcitationof an iso/atedinhibitory
interneurone by an injection of current to mimic the activation of
metabotropic glutamate (mG/u) receptors.(B) The same inhibitory
interneuroneas part of a network of inhibitory neuronescoupledby
fast, GABA,-mediatedinhibitory postsynapticpatentiak (/f2Ps), Its
responseto mCJureceptoractivationis now sculptedinto an osci//ation
of 33 Hz by synchronized/P5Psgeneratedby the inhibitory netwark.
based on experiments and computer simulations on
the hippocampal slice (Fig. lB,C). Essentially, when
networksof inhibitory neurones are tonically excited,
they tend to entrain each other into rhythmic firing
throughtheir mutualinhibitoryconnections. Figure3A
showsa computersimulationof the effect of a depolarizing current in an isolated interneurone. The rapid
dischargebecomes organizedinto a rhythmic pattern
of -40Hz whenthe interneuroneis synapticallycoupled
to a networkof similarlyactivatedinterneurones (Fig.
3B). In effect the rhythm is sculpted from the tonic
discharge by synchronous inhibitory postsynaptic
potentials (IPSPS).The experimental evidence for this
model is that synchronous IPSPSat frequenciesin the
gamma band occur in hippocampal and neocortical
slices where all monotropicglutamate-receptor containing synapses are blocked. In these experiments,
interneuronesare excited by the activation of metabotropic glutamate (mGlu) receptors. The experimental
blockadeof fast EPSPSexcludesmodelsthat dependon
these (Fig. 2A–C). We are not aware of hippocampal
neuronesthat oscillate preferentiallyat 40 Hz,and our
computer simulations show that such intrinsic oscillators are not necessary for gamma oscillations in a
neuronal network. In this model, what is necessaryis
tonic excitationof the interneurones(forexample,from
Networks of inhibitory neurones
(The main regions implicated include the hippo- metabotropic or NMDAreceptors). Fast EPSPSmight
campus and the parietal neocortex; see Fig. 2D.) We be superimposedon the tonic excitation, but they are
haverecentlyproposeda newmodelof gammarhythms not required,as shown by the original experiments38.
TZNSVOL 19, No. 5,1996
205
REVIEW
J.G.R. Jefferys et al. - Ne.renal
networks for induced ‘40 Hz’ rhythms
tonic excitation of the interneurones (which is the case experimen1.0
200 PA
200 PA
c
tally in the hippocampus and at
o
least part of the neocortex). The
100 ms
100 ms
~
0.5
recurrentinhibition model appears
k
o
to
be stabilizedby mutual connecso
tions
within the population of
2
1
I
1
1
1
I
inhibitory neurones32(and also by
-0.5
20
40
60
20
40
60
mutuallyexcitatoryneurones39),but
t(ms)
t(ins)
this is very different from the central role that such connections play
in the inhibitory network model.
Thisnewmodelpredictsthat both
B
50
excitatory and inhibitory neurones
fire in phase with the gamma
a
rhythm, becauseboth typesof neurone are clocked by the same popu-7
&40
lation IPSPS.This is the case in the
>
:
hippocampus in vivo16,but appar0
ently not in the olfactory bulb and
related areaswhere a phase lag was
predictedas a resultof the reciprocal
.-:30
activity in the recurrentexcitatory–
inhibitory loops. Interestingly, the
b
olfactorybulb and anteriorolfactory
nucleus have a peak in power at
20
oscillation frequencies of around
1,~,
75 Hz, comparedwith 40-50Hz for
the hippocampus,and couldusedif0
30
35
10
15
20
25
5
0
ferent mechanisms. Lateral entoIPSC decay constant, TD(ins)
rhinal cortex and prepiriformcortex
have
peaks in power at both these
Fig. 4. The frequency of oscillation in the inhibitory neuronal network is a function of the decay constant of the
frequencies.
inhibitory postsynaptic current (IPSC). (A) shows autocorrelationsof voltage-clamprecordingsfrom inhibitory
Inhibitory network mechanisms
interneuronesin stratum oriensmade during an application of glutamate in the presenceof drugs to blockianotropic
glutamate receptors.(a) Priorto additian of 20~~ pentobarbital, the netwarkoscillatedat 22.7ms [44 Hz; IPSCdecay might also function in the neoconstant (TJ was 9.1 i 0.4ms], which is faster than pyramidal cells which have a T. of 22.4 *0.8ms. (b) After cortex. Metabotropic glutamate
equilibration with pentobarbital the period slowedto 44.5 ms (22 Hz; TDreached>30ms). (B) Measurementsmadeof agonistselicitedgammaoscillations
both networkfrequencyand T. (opencircles)during the wash-inof 2pMpentobarbital reveala closerelationship,which when the monotropic glutamate
matchesthat predictedby computersimulations(filled diamonds).More recentcomputersimulationsmatch the non- receptors were blocked pharmacolinearity found at lower frequenciesand the upper and lower limits to the synchronousnetworkoscillations,fo//owing
logically, much as they did in the
an increasein the connectivityof the simulatednetwork40. Figure adapted, with permission,from Ref.38.
hippocampus38.
This meansthat the
cortical inhibitory networkcan susThe frequency of the oscillation is controlled, in part, tain gamma oscillations,but we cannot exclude other
by the time constant of the fast GABAA-mediated
IPSP. parallel mechanisms. As mentioned above, the neoComputersimulationshavepredicted,and experiments cortex contains neurones that are intrinsic oscillators
have confirmed, that drugsthat slow the decay of the at -40 Hz. We predict that inhibitory neurones that
IPSP(forexample,barbiturates),alsoslowthe frequency oscillatewilltend to stabilizethe gammarhythm of the
of the oscillation(Fig.4). (In 1950, Adrianmadea similar inhibitory network. At least some intrinsic oscillator
observationon the olfactorybulb when he noted that neurones are inhibitory.
Golomb, Wang and Rinze14’made simulations that
frequenciesof inducedoscillationsin the olfactorybulb
differed between urethane-induced and barbiturate- showthat mutualinhibition can entraina network,proinduced anesthesia, at around 50 and 15 Hz respec- vided that the individualneurones, when uncoupled,
tively.) In a more extensiveexploration of the control oscillate with a short period relative to the inhibitory
of these synchronous, inhibitory gamma oscillations, timecourse. In the case of gamma oscillations in the
we findthat they can existovera rangeof 20-70 Hz,and hippocampus,these conditions are met when TGABAis
that they desynchronize outside of this range40.The in the range 8–13 ms, and there is sufficient tonic
oscillations speed up, in both experiment and simu- ‘drive’to the interneurones (Figs3 and 4). This model
lation, with an increasedexcitatory drive,a shortened was developedfor the reticular nucleus of the thalainhibitory postsynapticcurrent (IPSC)decay constant mus, which participatesin the generation of synchro(.~A,A),or a decreasedIPSC amplitude.The recordings nous 7–12Hz ‘spindle’ discharges and 3 Hz absenceIn both cases,the low threshold
also show that at least two classesof interneuronepar- seizuredischarges4s’4G.
ticipate in this activity:fast spikingcellsin strataoriens (T), voltage-dependentCa2+currentplaysa crucialrole,
generating rebound excitation when the IPSPSdecay.
and pyramidale40.
This inhibitory networkmodel resemblesthe recur- Computer models showed that full synchronization
rent inhibitory loop mechanism (above), in that both dependson a sufficientlyslowinhibitorypotentialcomare sensitive to IPSPdecay constants. It differsin that paredwiththe excitationcomponent(the low-threshold
it doesnot requireintact fast EPSPS,and it doesrequire Ca2+spike)4547.In the case of spindledischargesthis is
Aa
b
o
0
5
2
0al
4=
z
=
8
0
00
~o
0
206
TINSVOL 19, NO. 5,1996
REVIEW
J.G.R. Jefferys et al. - Neuronal networks for induced ‘40 Hz’ rhythms
providedby fast GABAA-mediated
IPSPS,and in the case coherence providedby this kind of mechanism is too
of 3 Hzabsence-likeseizures,by slow GABA~-mediated weakto providefor reliablebinding, in that it appears
IPSPS.In practice,in the diencephalicslice,the thalamic in simulations only when multiple trials are used32.
reticular nucleus alone does not produce spindle dis- Scaling to largernetwprksor the presence of intrinsic
charges. Coherent oscillations also require excitatory oscillators, or both, might circumvent this problem.
synaptic input from thalamocortical-projection neur- Alternatively(or additionally)oscillations in different
one545,48<
Without
the
projection
neuronesthe network partsof the cortex couldresonatethroughthe thalamotends to generate ‘clusters’, in which only part of the cortical loop3s(but see Ref. 1). We have a long way to
population participatesin the oscillation, which then go before we can identify how these long-rangelinks
has a frequencythat is an integer multiple (usuallyx 2 are made and broken, and understandingthe mechaor x 3) of the mean firing rate of individual neur- nisms of local networkoscillations is a key step in this
ones45.This is perhapseasiestto imagine in the case of direction.
We startedthis reviewdescribingthe association of
two mutually inhibitory neurones with rebound excitation; as long as the IPSPSlast long enough to de- fast rhythms with higher cortical functions. Evidence
inactivatethe T current,then stimulatingone will lead from MEG recordings suggests that impairments of
to a persistentsequenceof alternatingdischargesin the these rhythms could be involvedin brain diseasesuch
two neurones.Asyet there is no experimentalevidence as Alzheimerlsdisease23.The relationship of gamma
rhythms with selective attention has prompted ideas
as to whether clusteringexists in thalamic tissue.
Inhibitory networksynchronization differsbetween that disruptionof its mechanism could play a role in
“ lg. Understandingthe cellular and netthalamus and hippocampus.In gamma oscillations in schizophrema
the hippocampus,the ‘rebound excitation’ is not due work mechanisms that generategamma rhythms proto a Ca2+current, but rather to the sustained inward vides a starting point for thinking how they could be
current that is turned on by activating mGlu recep- manipulatedtherapeutically.
torsqg.This causes inhibitory neurones to dischargeas
soon as the IPSPhas decayed sufficiently. More work Selectedreferences
is needed to find out whether this mechanism applies 1 Gray, C.M. (1994) J. CompuL Neurosci. 1, 11-38
218-226
2 Engel, A.K. et aL (1992) TrendsNeurosci. 1S,
in other regions that can generate gamma rhythms,
3 Singer, W. and Gray, C.M. (1995) Annu. Rev. Neurosci. 18,
such as olfactory structures and neocortex. Some of
555-586
the mechanisms outlined above could well coexist in
4 Adrian, E.D. (1950) Electroencephalogr. Clin. Neurophysiol.2,
377-388
particular parts of the brain: inhibitory network and
5 Eeckman, F.H.and Freeman,W.J. (1990) BrainRes.528,238-244
intrinsic mechanismscould combine in the neocortex;
6 Gray, C.M. et al. (1989) Nature 338, 334-337
and excitatory networks might be essential for the re7 Engel, A.K. et al. (1991) Science 252, 1177-1179
current inhibition loop to work39.Ourmuch improved 8 Engel, A.K., Konig, P. and Singer, W. (1991) Proc.Natl Acad.
Sci. USA88, 91369140
understandingof networkoscillatorshas startedto make
9 Freeman,W.J. and van Dijk,B.W. (1987) BrainRes.422,267-276
their mechanisms amenable to direct experimental 10 Madler, C. et aL (1991) Br. 1.Anaestlr.66, 81-87
testing.
11 Keller,L et aL (1990) Clin.Hectroencephakgr. 21, 88-92
Functionalconsequences
of rhythmicinhibition
Both theorysoand experiment38r51-53
show that inhibitory neurones are very effective at determining
when a pyramidalcell will fire. Ourproposalis that the
inhibitory networkreceivesa steadyor slowexcitatory
drivewhich makesit oscillate,providinga clock which
determines when pyramidalcells can fire, if they receive suprathreshold, excitatory afferent inputs. The
fact that the inhibitory network can, by itself, sustain
a rhythm in the gamma frequency range without a
requirement for fast EPSPS,separatesthe synchronizing control or ‘clock’ from the specific neuronal processingof information ‘the Central ProcessingUnit’. A
possibleimplication is this: if firing of pyramidalneurones is constrainedto occur at particulartimes imposed
by a 40Hz interneuronal network clock, then brain
regions expressing40 Hz might not use rate-encoding
of information, but rather might encode information
through selection of which pyramidalcells fire at all.
Furtherwork also is requiredto find out whether this
provides the means for associating or binding attributes of specificobjects, for controlling summationand
potentiation, or for a phase code for neural signaling.
If the role of gamma rhythms is indeed to mediate
binding, then mechanisms must exist for the selective
coupling of areas involved in processingcommon entities. One ideais that reciprocalexcitatoryconnections
can do this, although there seemsto be a constraint in
that the conduction delaymust be less than one third
of the period of the rhythm54.Others argue that the
12 Bouyer, J.J. et aZ. (1987) Neuroscience22, 863-869
13 Ptimtscheller, G., Flotzinger, D. and Neuper, C. (1994)
Electroencephalogr.Clin. Neurophysiol.90, 456-460
14 Murthy, V.N. and Fetz, E.E. (1992) Proc. Natl Acad.Sci.USA89,
5670-5674
15 Sanes,J.N. and Donoghue, J.P. (1993) Proc. Natl Acad.Sci. USA
90, 4470-4474
16 Bragin, A. et al. (1995)]. Neurosci.15, 47-60
17 Stumpf, C. (1965) Electroencephalogr. CZin. NeurophysioL 18,
477-486
18 Soltesz,L and Deschi2nes,M. (1993) J NeurophysioL70, 97-116
19 Tiitinen, H. et al. (1993) Nature 364, 59-60
20 Pantev, C. et al. (1991) Proc.Natl Acad. Sci. USA88, 8996-9000
21 Kulli,J. and Koch, C. (1991) TrendsNeurosci. 14, 6-10
22 Galambos, R., Makeig, S. and Talmachoff, P.J. (1981) Proc.
Natl Acad. Sci. USA78, 2643-2647
23 Ribary,U. et aL (1991) Proc.NatZAcad.Sci. USA88,11037-11041
24 LlinAs,R. and Ribary, U. (1993) Proc. NatZAcad. Sci. USA90,
2078-2081
25 von der Malsburg, C. and Schneider, W. (1986) BioZ.Cybem.
54, 29-40
26 Gray, C.M. et al. (1992) Visual Neurosci. 8, 337-347
27 Engel, A.K., Konig, P. and Schillen, T.B. (1992) Cur’r.BioL2,
33%334
28 Young, M.P., Tanaka, K. and Yamane, S. (1992)
f. Neurophysiol.
67, 1464-1474
29 Tovee, M.J. and Rolls, E.T. (1992) NeuroReport3, 369-372
30 O’Keefe,J. and Recce, M.L. (1993) Hippocampus 3, 317-330
31 Hopfield,J.J. (1995) Nature 376, 33-36
32 Wilson, M.A.and Bower,J.M. (1991) NeuralComput.3,498-509
33 Wilson, M. and Bower,J.M. (1992) J. NeurophysioL67, 981-995
34 Pinault, D. and [email protected],M. (1992) Neuroscience51,245-258
35 Steriade, M., Curro Dossi, R. and Contreras, D. (1993)
Neuroscience
56, 1-9
36 Llinfis,R.R., Grace, A.A.and Yarom, Y. (1991) Proc. Natl Acad.
Sci. USA88, 897–901
37 Nuilez,A., Amzica, F. and Steriade,M. (1992) Neuroscience
51,
7-1o
38 Whittington, M.A., Traub, R.D. and Jefferys, J.G.R. (1995)
Nature 373, 612-615
TINS Vol. 19, ~0, S, 1996
Acknowledgements
This workwas
supportedby the
Wellcome Trustand
IBM. We thank
GyorgyBuzsaki,
Bard Errrrentrout,
Charles Grayand
John Rinzelfor
helpful discussions
duringthe
preparationof this
manuscript.
207
REVIEW
J.G.R. Jefferys et al.–
Neuronal networks for induced ‘40 Hz’ rhythms
39 Ermentrout, G.B. (1995) in The Handbook of Brain Theory and
NeuraZNetworks (Arbib,M.A., cd.), pp. 732–738, MIT Press
40 Traub, R.D. et aL 1. I%ysiol. (in press)
41 McCormick, D.A., Gray, C.M. and Wang, Z. (1993) Soc.
Neurosci.Abstr. 19, 359
42 Gutfreund, Y., Yarom, Y. and Segev, I. (1995) ~.Physiol. 483,
621-640
43 Wang, X-J. (1993) NeuroReport5, 221-224
44 Llinfis,R.R. (1988) Science 242, 1654-1664
45 Golomb, D., Wang, X-J. and Rinzel,J. (1994) J. NeurophysioL
72, 1109-1126
46 Wang, X-J. and Rinzel,J. (1993) Neuroscience53, 899-904
47 Destexhe, A. et d. (1994) J. NeurophysioL72, 803-818
48 Von Krosigk,M., Bal, T. and McCormick, D.A. (1993) Science
261, 361-364
49 McBain, C.J., DiChiara, T.J. and Kauer,J.A. (1994) J. Neurosci.
14, 4433-4445
50 Lytton, W.W. and Sejnowski, T.J. (1991) J. [email protected],
1059-1078
51 Buhl, E.H., Halasy, K. and Somogyi, P. (1994) Nature 368,
823-828
52 Lacaille,J.C. et aL (1987) J. Neurosci. 7, 1979-1993
53 Knowles, W.D. and Schwartzkroin, P.A. (1981) J. Neurosci. 1,
318-322
54 Konig, P., Engel, A.K. and Singer, W. (1995) Proc. Natl Acad.
Sci. USA92, 29&294
BOOK REVIEW
Methods in Enzymology.Vol. 255- Small GTPases
and Their Regulators,Part A: Ras Family;
Vol. 256- Small GTPasesand Their Regulators,
Part B: Rho Family
edited by W.E. Balch,CJ Der andA. Hall,AcademicPress,1995.$99.00 (xxxi + 548 pages)
ISBFJO /2 /82/56 O (Vol.255); $80.00 (xxix+ 40/ pages)ISBNO 12 182/57 9
(Vo/.256)
The Rasand Rho proteinsare membersof
a large superfamilyof small GTPasesthat
are activateduponGTP bindingand return
to an ‘off’ state when GTP is cleavedto
GDP + Pi as a result of their intrinsic
GTPase property. The Ras proteins encoded by viral Rasgenesdifferfrom those
encodedby cellulargenesby a few amino
acids.These point mutationsimpair their
GTPaseactivityandthereforeinterferewith’
their normal shut-off mechanism,making
them constitutivelyactive.
The different interconversionstatesof
the GTPases(whichhasmadethem popularly known as ‘molecularswitches’)regulate many intracellularsignalingpathways.
Although both the Rasand Rho family of
GTPasesbelongto a classof proteinsthat
end with the sequenceCXXXX, they are
functionallydistinct.Rasproteins regulate
cell growth and differentiationby providinga linkbetweengrowth factor receptors
and gene expression’,whereas Rho proteins regulate the assemblyof focal adhesionsand cell movement.AlthoughRho
proteins belong to the GTPase superfamily,they are mainlyinvolvedin the regulation of actin cytoskeletalorganization2’3.
Volumes 255 and 256 of Methods in
Enzymologyare dedicatedto the biology
and biochemistryof Ras-and Rho-related
proteins, respectively, and are divided
into four sections.Both volumesare categorized in a similar fashion. The first
sectiondescribesthe methodsusedfor the
cloning and purification of recombinant
Ras or Rho proteins from bacteria,yeast
and the baculovirus–insect-cell
system.In
this section, emphasisis placed on the
post-translationalmodificationsof the Ras
proteins,which makesthem hydrophobic
208
TINSVoL 19, No. S, 1996
mightbe. The lackof detailedintroduction
in some of the chapters might make it
diftlcult to follow for students or those
enteringthe area of signaltransductionfor
the first time, and a few articleshave not
been appropriately assignedto the relevant chapters. In some articles the
authors have failed to emphasize the
importance of post-translational differences (such as glycosylation,phosphorylation andfarnesylation)for the functionof
Ras-related proteins when expressing
them in different systems (for example,
E. coli,yeast and insectcells).This distinction isimportantandcouldhavebeenindicated. The importance of selecting the
correct system or cell line has been
ignored in the chapters that describe
protein–protein interactions,which could
lead to bona fide protein-protein interactions being missed due to weak expressionor lackof interactingproteins.In
addition, misfolding of proteins (for
example, in E. coli, yeast, Sf9 cells or
mammaliancell lines)or failure of certain
post-translationaleventsin the target proteins might lead to incorrect conclusions.
The antisenseapproachthat hasbeenused
to inhibit Ras function is convincing,but
control experiments are missingor have
not been describedhere4.
Overall, these books are a most useful
and valuable resource to everyone
involved in the field of protein research.
They will certainlyserveasguidancebooks
and many of the techniques described
might remain central to the field of signal
transductionin the future.
and allows them to become attached to
the plasmamembrane.The importanceof
this modificationiswell addressed.
The second section deals mainly with
the cyclic processof interconversionbetween the GTP (’on’) state and the GDP
(’off’) state of Rasproteins.The technical
details required to monitor the GTPbindingproperty both at an in vivo(in situ)
level and at an in vitro level have been adequatelycovered.
The third section describes the various
approachestaken to identifythe proteinprotein interactionsbetween components
of the Ras-relatedsignaltransductionpathway, for which a rangeof techniqueshas
beenwidelyused,from classicaltechniques
(such as metabolic labelingand immunoprecipitation)to more recentmolecularapproaches(suchasthe two-hybrid system).
The final section describesthe various
fascinatingapproachesthat havebeenused
to monitor the biologicalactivity of Ras
genes, which include oocyte and mammalian microinjection assays, fibroblast
complementationassaysandthe screening
of phagepeptidelibrariesfor SH3 Iigands.
Both of these volumes have several
strengths:the readableand concisecollecTilat A.RJzvi
tion of chaptersare written by some of
Llept of Cc//Biology,Neurobiologyand
the leadersin the field of signaltransducAnatomy, Universityof Cincinnati Medical
tion, the logical organization of inforCenter,231 8ethesdaAvenue,Cincinnati,
mation makes it quick to access inforOH 45267-0521, USA.
mation related to specific Ras or Rho
genes,and many of the figuresare generReferences
ally convincingand well done, particularly
1 Barbacid, M. (1987) Annu. Rev.Biochern.
the photomicrographsin the section ‘the
56, 779-827
2 Burridge, K. et aL (1988) Annu. Rev. Cell
role of Rho proteins in cellularfunction’,
BioL 4, 487-525
which are of excellentquality.
3 Vincent, S., Jeanteur, P. and Fort, P.
However, there are a few areaswhere
(1992) Mol.CeKBioZ.12, 3138-3148
4 Gura, T. (1995) Science 270, 575-577
these books are not as strong as they
Copyright 01996,
Elsevier Science Ltd. All rights reserved. 0166-
2236/96/$15.00
PII: S0166-2236(96)60012-5