Special Issue: Neuropsychiatric Disorders

Cortical parvalbumin interneurons andcognitive dysfunction in schizophreniaDavid A. Lewis, Allison A. Curley, Jill R. Glausier and David W. Volk

Translational Neuroscience Program, Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA 15213, USA

Review

Deficits in cognitive control, a core disturbance of schizo-phrenia, appear to emerge from impaired prefrontalgamma oscillations. Cortical gamma oscillations requirestrong inhibitory inputs to pyramidal neurons fromthe parvalbumin basket cell (PVBC) class of GABAergicneurons. Recent findings indicate that schizophrenia isassociated with multiple pre- and postsynaptic abnor-malities in PVBCs, each of which weakens their inhibi-tory control of pyramidal cells. These findings suggest anew model of cortical dysfunction in schizophrenia inwhich PVBC inhibition is decreased to compensate foran upstream deficit in pyramidal cell excitation. Thiscompensation is thought to rebalance cortical excitationand inhibition, but at a level insufficient to generate thegamma oscillation power required for high levels ofcognitive control.

IntroductionPsychosis (e.g. hallucinations, delusions and disorganizedbehavior) is the most striking clinical feature of schizophre-nia, but impairments in cognition are now recognized as thecore domain of dysfunction in the illness [1]. Cognitivedeficits are present and progressive years before the onsetof psychosis [2] and thedegree of cognitive impairment is thebestpredictor of long-termfunctional outcome[3].The rangeof cognitive deficits in schizophrenia suggests an overarch-ing alteration in cognitive control; that is, the ability toadjust thoughts or behaviors to achieve goals [4]. Cognitivecontrol depends on the coordinated activity of several brainregions, including the dorsolateral prefrontal cortex(DLPFC) [5], and gamma frequency (30–80 Hz) oscillationsin DLPFC neural networks are thought to be a key neuralsubstrate for cognition [6]. Consistent with these observa-tions, when performing tasks that require cognitive control,individuals with schizophrenia exhibit altered activation ofthe DLPFC [7] and lower power of frontal lobe gammaoscillations [8,9].

Because cortical gamma oscillations require the strongand synchronous inhibition of networks of pyramidal neu-rons (reviewed in [10]), deficient GABA neurotransmissionin the DLPFC has been hypothesized to contribute toaltered gamma oscillations and impaired cognition inschizophrenia [11]. Consistent with this interpretation,manipulations in animal models that reduce GABA-mediated inhibition diminished gamma oscillations [12]and impaired cognitive function [13–16]. In addition, in

Corresponding author: Lewis, D.A. (lewisda@upmc.edu).

0166-2236/$ – see front matter � 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tins.2011.

individuals with schizophrenia, negative modulation ofGABAergic neurotransmission exacerbated symptoms[17], whereas positive modulation was associated withincreased frontal lobe gamma oscillations during a cogni-tive control task [18].

However, surprising recent findings regarding the func-tional properties of certain subtypes of cortical interneur-ons and new observations regarding cell type-specificalterations in markers of GABAergic neurotransmissionin schizophrenia require a new conceptualization of therole of altered cortical GABAergic signaling in the cogni-tive deficits of schizophrenia. Consequently, here we (i)review recent findings both from cellular physiologyexperiments and postmortem studies of schizophrenia thatdemonstrate the limitations of existing circuitry models ofcognitive dysfunction in schizophrenia based on earlierdata; (ii) propose a new pathophysiological model of therole of altered GABA neurotransmission in cortical circuit-ry dysfunction in schizophrenia; and (iii) discuss the keyresearch questions raised by the new data and model.

Deficient cortical GABA synthesis is a conserved featureof schizophreniaGABAergic signaling is regulated, in part, by the enzymat-ic activity of two isoforms of glutamic acid decarboxylase(GAD), which differentially contribute to GABA synthesis.Inmice, deletion of the gene encoding the 67-kDa isoform ofGAD (GAD67) results in a 90% reduction in brain GABAlevels and is embryonically lethal [19], whereas deletion oftheGAD65 gene is associated with only a 20% reduction intotal brain GABA [20] and normal survival. In multiplestudies using a variety of techniques, levels of GAD67mRNA [21] and protein [22,23] have been found consis-tently to be lower in the DLPFC of subjects with schizo-phrenia. Similar deficits in GAD67 mRNA are also presentin other cortical regions, including sensory, motor andlimbic regions [24–27]. By contrast, cortical expression ofGAD65 appears to be normal or only slightly altered inschizophrenia [22,28], and the density of GAD65-labeledaxon terminals in the DLPFC is unchanged [29].

The magnitude of the GAD67 deficit in schizophreniadiffers substantially across individuals, raising the ques-tion of the extent to which the deficit reflects the diseaseprocess or comorbid factors. Because GAD67 expression isactivity regulated [30], lower GAD67 expression in schizo-phrenia could reflect reduced cortical activity secondary toother factors that accompany a chronic psychiatric illness.However, the variability in GAD67 mRNA levels across

10.004 Trends in Neurosciences, January 2012, Vol. 35, No. 1 57

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

subjects with schizophrenia is not attributable to potentialconfounds, such as substance abuse or antipsychotic med-ications, predictors (e.g. male sex, a family history ofschizophrenia and early age of onset), or measures ofdisease severity (e.g. suicide, lower socioeconomic status,not living independently and no history of marriage), orduration of illness [23,28]. Thus, lower cortical GAD67mRNA levels appear to be a conserved feature that is acore common component, and not a consequence, of thedisease process of schizophrenia.

However, less GAD67 mRNA and protein does notnecessarily support the conclusion that cortical GABAlevels are lower in schizophrenia. For example, GAD67expression could be downregulated in response to reducedGABA metabolism; indeed, pharmacological inhibition ofGABA degradation results in elevated cortical GABA andless GAD67 protein [31]. Unfortunately, current attemptsto measure cortical GABA levels in vivo with magneticresonance spectroscopy (MRS) have produced mixedresults in subjects with schizophrenia [32–34]. However,lower GABA levels in the visual cortex in subjects withschizophrenia were correlated with reductions in a behav-ioral measure of visual inhibition that depends on GABA[(Figure_1)TD$FIG]

PVBC

CCKBC

μOR

CCK2

GABAA α1Key:GABAA α2

GAD67

GABAA α1

µORmRNAs GAD67

CCKmRNA

Gα1

1

2

3

4

Cor

tical

laye

rs

Figure 1. Schematic summary of alterations in neuronal circuitry in layer 3 of the dors

inhibition of pyramidal neurons (gray neurons) by parvalbumin-positive basket cells (P

expression [39] and lower GAD67 protein [23] and, hence, less GABA synthesis; (ii) h

activity and suppresses GABA release [77]; (iii) reduced expression of cholecystokinin (C

[79–81]; and (iv) less mRNA for, and presumably fewer, postsynaptic GABAA a1 recept

neuron in deep layer 3, but are probably present throughout layer 3. Chandelier neuron

axon terminals [47,48] and increased postsynaptic GABAA a2 receptors in pyramidal neu

excitation of pyramidal neurons if these inputs are depolarizing [51,53]. Because relative

the adult primate DLPFC [100], perhaps reflecting the postnatal developmental decline in

postsynaptic GABAA a2 receptors are increased in deep layer 3 pyramidal neurons in sch

not known. Abbreviation: CCKBC, CCK basket cell.

58

neurotransmission [34], and frontal lobe GABA levelstended to be correlated with working memory performancein subjects with early-stage schizophrenia [35]; both ofthese findings support the idea that lower GABA synthesisin schizophrenia results in cognitive impairments. Howev-er, becauseMRS assesses total tissue GABA levels, and notGABA levels in synaptic vesicles or the extracellular space,the relevance of MRSmeasures to cortical GABA synthesisand transmission remains uncertain. Alternative strate-gies to measure shifts in the levels of extracellular GABAare emerging and preliminary findings support a positiverelationship between the capacity to increase extracellularGABA and physiological correlates (i.e. gamma oscilla-tions) of cognitive control [36]. However, as indicated inthe following sections, in vivo methods that can assess thesynthesis and release of GABA from particular populationsof interneurons may be required.

GAD67 deficit is prominent in parvalbumin-positiveinterneuronsUnderstanding the functional significance of lower corticalGAD67 levels requires knowledge of the affected class ofinterneurons. In schizophrenia, GAD67 mRNA levels are

PVChCGAT1

PVChC

GABAA α2GAT1

ABAA mRNA

TRENDS in Neurosciences

olateral prefrontal cortex (DLPFC) in subjects with schizophrenia. The perisomatic

VBCs) is reduced owing to (i) lower glutamic acid decarboxylase (GAD)-67 mRNA

igher levels of m opioid receptor (mOR) expression in PVBCs, which reduces their

CK) mRNA [28,41], which stimulates the activity of, and GABA release from, PVBCs

ors in pyramidal neurons [63]. These alterations are shown only for the pyramidal

s (PVChCs) have decreased GABA membrane transporter 1 (GAT1) protein in their

ron axon initial segments [49], suggesting enhanced GABA signaling and increased

ly few GABAA a2-labeled axon initial segments are detectable in layers deep 3–4 of

mRNA expression of this subunit in the primate DLFPC [101], it is unclear whether

izophrenia. Levels of GAD67 protein in PVChC axon terminals in schizophrenia are

Table 1. Comparison of the properties of PVChCs and PVBCs in schizophrenia

Properties PVChCs PVBCs Refs

Location of synapses on

pyramidal neuron

AIS Soma and proximal dendrites and spines

Predominant GABAA

receptor a subunit

a2 a1 [57]

Alterations in the DLPFC

in schizophrenia

GAD67 mRNA and protein levels unknown # GAD67 protein in axon terminals in layer 3 [23]

# Density of GAT1-positive axon

cartridges in layers 2–4

$ Density of GAT1-positive axon terminals in layer 3 [48]

Density of PV-positive axon cartridges unknown # Density of PV-positive axon terminals in layer 3 [65]

" GABAA a2 mRNA in layer 2 # GABAA a1 mRNA in layer 3 [50]

" Density of GABAA a2-positive

AIS in layers 2 and superficial 3

# GABAA a1 mRNA selectively in layer 3 pyramidal cells [49,63]

" m opioid receptor mRNA [77]

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

markedly lower only in 25–35% of DLPFC interneurons[37,38], and GAD67 mRNA is not detectable in approxi-mately 50% of the subset of interneurons that express thecalcium-binding protein parvalbumin (PV) [39]. Levels ofPV mRNA are also lower in schizophrenia [24,40,41], butthe densities of neurons labeled for either PVmRNA [39] orprotein [42–44] in the DLPFC do not differ from compari-son subjects. Together, these findings suggest that thenumber of cortical PV neurons is not altered in schizophre-nia, but that GAD67 is markedly reduced in a subset ofthese neurons. By contrast, calretinin-containing inter-neurons, which comprise approximately 45% of GABAneurons in the primate DLPFC, do not appear to beaffected in the illness [11].

Recent in vivo studies using optogenetic techniqueshave clearly established that activity in PV-positive inter-neurons is essential for driving cortical gamma oscillationsin mice [45,46], although the particular subclass(es) of PV-positive interneurons responsible could not be determinedusing this approach. Cortical PV-positive interneuronsconsist of two main types: chandelier and basket cells(Figure 1). The axon terminals of PV-positive chandelier(aka ‘axo-axonic’) cells (PVChCs) form distinctive verticalarrays (termed ‘cartridges’) that exclusively innervate theaxon initial segment (AIS) of pyramidal neurons just prox-imal to the site of action potential generation. By contrast,PV-positive basket cells (PVBCs) innervate the cell bodyand proximal dendrites of pyramidal neurons. In the fol-lowing sections, we review recent findings indicating thatthe nature of the alterations in, and the functional con-sequences of, schizophrenia-related molecular alterationsare markedly different for PVChCs and PVBCs. It isimportant to note that several of the recent findings citedon the biological properties of PVChCs and PVBCs arefrom studies of the rodent neocortex or hippocampus and,thus, the extent to which they are generalizable to theprimate DLPFC remains to be determined.

Alterations in PVChCs in schizophrenia: increasingpyramidal cell excitation?In schizophrenia, the density of PVChC axon cartridgesthat are immunoreactive for the GABA membrane trans-porter 1 (GAT1) is reduced in DLPFC layers 2–4 [47,48](Figure 1). In layers 2 and superficial 3 of subjects withschizophrenia, the lower density of GAT1-labeled car-tridges is inversely correlated with an increase in the

59

density of AISs that are immunoreactive for the a2 subunitof the GABAA receptor [49] (Table 1). Consistent with thesefindings, GABAA a2 subunit mRNA levels are elevated inthe same laminar location in schizophrenia [50]. The de-crease in presynaptic GAT1 protein and increase in post-synaptic GABAA a2 receptor protein levels wereinterpreted as coordinated compensations in response toa PVChC deficit in GABA synthesis in schizophrenia that,by reducing GABA reuptake and increasing the probabilityof receptor binding, respectively, would strengthen GABAsignaling [11]. However, although GAD67 mRNA expres-sion has been shown to be lower in PV neurons [39], GAD67mRNA and protein levels have not been assessed specifi-cally in PVChCs or their axon cartridges in schizophrenia.

Given the presumed potent inhibitory regulation of py-ramidal neurons by PVChCs, deficient GABA neurotrans-mission from PVChCs was previously postulated to be theneural substrate for the lower frontal lobe gamma bandpowerobservedduringcognitive control tasks in schizophre-nia [11]. However, recent findings raise questions about thisinterpretation. For example, rather than providing thestrong hyperpolarization of pyramidal neuron networksrequired for gamma oscillations, recent findings indicatethat the synaptic inputs from neocortical chandelier cellsare actually depolarizing in some cases. For example, stim-ulation of neocortical chandelier cells can initiate spikes inpostsynaptic pyramidal cells [51]. This excitatory effectmaybe the result of the much lower levels of the K+-Cl�co-transporter 2 (KCC2), which extrudes chloride, at the AISrelative to the soma or dendrites [51,52]. The resultinghigher intracellular levels of chloride in the AIS would leadto the flow of chloride out of, rather than into, the AIS whenGABAA receptors are activated, depolarizing the mem-brane. Indeed, under experimental conditions that preservethe physiological intracellular chloride concentration,PVChC inputs depolarize postsynaptic pyramidal neurons,whereas inputs from PVBCs are, as expected, hyperpolariz-ing [51,53].

Other recent findings also raise questions about the roleof PVChCs in the generation of gamma oscillations. Forexample, gamma oscillations require a fast decay of theinhibitory postsynaptic current in pyramidal cells and theduration of this postsynaptic inhibitory current depends onthe subunit composition of the GABAA receptors thatmediate it [54]; the kinetics of the a2 subunit-containingGABAA receptors that are postsynaptic to PVChCs appear

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

to be too slow to drive a circuit at gamma frequency [10]. Inaddition, the firing of PVChCs in the rodent hippocampusis not strongly coupled to the gamma oscillation cycle [55],but to the much slower cycle of theta oscillations (reviewedin [56]).

Alterations in PVBCs in schizophrenia: decreasingpyramidal cell inhibition?The PV neurons with reduced GAD67mRNA expression inschizophrenia do include PVBCs, as lower GAD67 proteinlevels have been found in PVBC axon terminals (identifiedby excluding PVChC axon cartridges) in DLPFC layers 3–4in subjects with schizophrenia [23] (Figure 2). The cell-typespecificity of this finding was supported by the observationthat the GAD67 protein deficit in these terminals was

[(Figure_2)TD$FIG]

(a)

(b)

PVBC

P

Figure 2. Summary of regulators of GABA neurotransmission from a parvalbumin-posit

dorsolateral prefrontal cortex (DLPFC) from a healthy subject (a) and a subject with schiz

the magnitude and direction of chloride (Cl–) ion flow mediated by the Cl– transporte

channels in a1-containing GABAA receptors. In healthy adult neurons (a), intracellular C

binding of GABA (red dots) to a1-containing GABAA receptors triggers Cl– entry and

decarboxylase (GAD)-67 (light-blue shading) in PVBC axon terminals [23] lead to less

through upregulated m opioid receptors (mOR) [77], and the effects of GABA are reduced

two kinases [oxidative stress response kinase (OXSR1) and with no K (lysine) protein kin

Cl– transporters and, consequently, increased NKCC1 activity and decreased KCC2 activ

a1-containing GABAA receptors, Cl– influx is reduced and GABA neurotransmission is

60

approximately ten times greater than in the total DLPFCgray matter from the same subjects [23]. Thus, althoughnot conclusive, the existing data suggest that the GAD67deficit in schizophrenia is specific to, or at least particularlypronounced in, PVBCs. By contrast, GAT1 protein levels donot appear to be altered in the axon terminals of PVBCs inschizophrenia [47], suggesting that GABA uptake is notreduced in these terminals as it is in PVChC axoncartridges.

In addition to the evidence for lower GABA synthesis inPVBCs, postsynaptic findings in pyramidal cells also sup-port a role for weaker inhibitory inputs from PVBCs in thepathophysiology of DLPFC dysfunction in schizophrenia.These inputs are principally mediated by a1-containingGABAA receptors in hippocampal pyramidal cells [57,58].

[Cl-]

[Cl-]

μOR

GABA α1

Key:

NK

CC

1

KC

C2

P

P

P

P

NK

CC

1

KC

C2

P

P

PP

PP

PP

P

P

PP

OX

SR

1W

NK

3

OX

SR

1

WN

K3

TRENDS in Neurosciences

ive basket cell (PVBC) terminal to a layer 3 pyramidal cell (gray neuron, insert) in the

ophrenia (b). In both panels, the size and orientation of the yellow arrows indicates

rs N+-K+-Cl�-co-transporter 1 (NKCC1) and K+-Cl–-co-transporter 2 (KCC2) and Cl–

l– concentration is low owing to low levels of NKCC1 and high levels of KCC2. The

membrane hyperpolarization. In schizophrenia (b), lower levels of glutamic acid

GABA synthesis. The release of GABA is further suppressed by greater signaling

owing to fewer postsynaptic a1-containing GABAA receptors [63]. Higher levels of

ase (WNK3)] [69] are thought to lead to increased phosphorylation (green P) of both

ity, producing a greater intracellular Cl– concentration. Thus, upon activation of the

less hyperpolarizing.

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

Most [24,28,50,59], but not all [60], well-controlled studieshave reported lower levels of GABAA a1 subunit mRNA inthe DLPFC of schizophrenia subjects, with this decreasemost prominent in layers 3 and 4 [50]. Expression of theGABAA receptor b2 subunit, which preferentially assem-bles with a1 subunits [61], was also lower selectively inlayers 3–4 [50]. By contrast, other GABAA receptor sub-units were unchanged in layer 3 (or showed a differentlaminar pattern of altered expression) [50,59,62]. WithinDLPFC layer 3, cellular-level analyses revealed that a1subunit mRNA expression was markedly lower in pyrami-dal cells (Figure 2), but was unaltered in interneurons [63].However, whether these expression changes are associatedwith alterations in the number of membrane-boundGABAA receptors [64] containing a1 and b2 subunitsremains to be determined.

Lower pyramidal cell a1 subunit expression could indi-cate a lower number of inputs from PVBC axons in schizo-phrenia, an interpretation consistent with the lowerdensity of PV-immunoreactive axon terminals in the samelaminar location [65]. Whether this deficit reflects fewerPVBC terminals or fewer PV-containing axonal projectionsfrom the thalamus remains unresolved [65]. Interestingly,genetic reductions of GAD67 in PV interneurons duringdevelopment result in less PVBC axonal arborization andsynapse formation [66]. Thus, the undetectable levels ofGAD67 mRNA in approximately 50% of DLPFC PV inter-neurons in schizophrenia [39], if present early in life, couldlead to fewer PVBC axon terminals and, thus, fewer syn-apses needing GABAA a1 receptors in postsynaptic cells.However, diminished PVBC axonal arbors would beexpected to affect all postsynaptic cells, including otherinterneurons [67]. Given that a1 subunit mRNA levels ininterneurons were not altered in schizophrenia [63], thereductions in both GAD67 and GABAA a1 levels are likelyto be attributable to some other common factor.

In addition to lower levels of these pre- and postsynapticmediators of PVBC inputs to DLPFC layer 3 pyramidalneurons, several other findings suggest that inhibitoryinputs from PVBCs are weaker in schizophrenia. First,the extent to which DLPFC pyramidal neurons can behyperpolarized may be lower in schizophrenia. Thestrength of the postsynaptic response to GABA dependson the driving force for the influx of chloride when GABAA

receptors are activated. Intracellular chloride levels aredetermined by the balance of activity between the chloridetransporters N+-K+-Cl–-co-transporter 1 (NKCC1) andKCC2, which mediate chloride uptake and extrusion, re-spectively [68]. The expression of these transporters is notaltered in the DLPFC of subjects with schizophrenia[69,70], but two kinases, oxidative stress response kinase(OXSR1) and with no K (lysine) protein kinase (WNK3),which phosphorylate both chloride transporters, aremarkedly overexpressed in schizophrenia [69]. Interest-ingly, protein levels of OXSR1 are most prominent in layer3 pyramidal neurons in the DLPFC [69]. Assuming thatthe elevated levels of OXSR1 and WNK3 expression rep-resent greater kinase activity, then increased phosphory-lation of the chloride transporters would decrease theactivity of KCC2 and increase the activity of NKCC1,resulting in elevated chloride levels in layer 3 pyramidal

61

neurons in schizophrenia (Figure 2). As a result, the gra-dient for chloride entry associated with activation ofGABAA receptors would be reduced, resulting in lesshyperpolarization of layer 3 pyramidal neurons whenGABA is released from PVBCs.

Second, other molecular alterations in schizophreniapoint to both reduced activity of, and a suppression ofGABA release from, PVBCs. For example, m opioid recep-tors are present on the perisomatic region and axon term-inals of hippocampal PV neurons [71,72]. Stimulation ofthe perisomatic m opioid receptors activates G protein-coupled inwardly rectifying potassium channels, hyperpo-larizing the cell body [73,74]; stimulation of m opioidreceptors on axon terminals suppresses vesicular GABArelease [75,76]. Consistent with the role of PVBC firing andGABA neurotransmission in generating gamma oscilla-tions, stimulation of m opioid receptors disrupts neuralnetwork activity at gamma frequencies in the hippocam-pus [55]. In the DLPFC of subjects with schizophrenia, m

opioid receptor transcript levels are higher than in healthycomparison subjects (Figure 1), a difference not attribut-able to psychotropic medications or other factors comorbidwith schizophrenia [77]. By contrast, other markers ofopioid signaling (e.g. d and k opioid receptors, proenkepha-lin and prodynorphin) are not altered in the illness [77,78].Thus, by both reducing the activity of PVBCs and suppres-sing GABA release from their axon terminals, elevatedlevels of m opioid receptors in schizophrenia could serve asyet another means of weakening PVBC inhibition of pyra-midal neurons (Figure 2).

Third, the activity of PVBCs is also regulatedby inputs from cholecystokinin-positive basket cells(CCKBCs) via CCK2 receptors. CCK stimulates the activi-ty of, and GABA release from, PV cells in the hippocampus[79–81]. Thus, the observed downregulation of CCKmRNAexpression in schizophrenia [28,41] might also contributeto weaker PVBC inputs to pyramidal neurons (Figure 1).Interestingly, recent studies in monkey DLPFC indicatethat the terminals of CCKBCs (at least the subclass thatalso expresses the cannabinoid 1 receptor) principallycontain GAD65, and not GAD67, protein [82]. Thus, be-cause cortical expression of GAD65 appears to be normal oronly slightly altered in schizophrenia [22,28], GABA neu-rotransmission may not be altered in CCK neurons; thepositive correlation between the deficits in GAD67 andCCK mRNAs in schizophrenia [22,28] may reflect theirshared downregulation in different cell types in response toan upstream reduction in cortical excitation (Box 1).

Role of PVBCs in lower DLPFC gamma oscillation powerin schizophreniaThe findings summarized above suggest that several mo-lecular alterations in different cell types converge to weak-en PVBC inputs to pyramidal neurons in DLPFC layer 3 inschizophrenia. Interestingly, PVBCs play a critical role inthe generation of cortical gamma band oscillations[10,45,46]. For example, the a1-containing GABAA recep-tors that are postsynaptic to PVBC inputs in hippocampalpyramidal cells produce currents with a decay periodappropriate for gamma oscillations [10]. In addition,electrophysiological findings in the hippocampus indicate

Box 1. Balancing cortical excitation and inhibition in schizophrenia

The dynamic balance between excitation and inhibition (E/I balance)

allows activity to propagate through local cortical networks without

either dying out or increasing uncontrollably [94]. E/I balance is

maintained in the face of perturbations in circuit activity [102] by, at

least in part, reciprocal, sustained adjustments in the levels of

excitatory and inhibitory synaptic transmission through a process

termed ‘synaptic scaling’ or ‘homeostatic synaptic plasticity’

[94,103,104]; that is, the amplitudes of excitatory and inhibitory

postsynaptic currents are independently adjusted via scaled changes

in the neurotransmitter content of synaptic vesicles and in the density

of postsynaptic receptors (e.g. [104]; reviewed in [104,105].

In schizophrenia, current findings suggest the hypothesis that the

pre- and postsynaptic strength of PVBC inputs to pyramidal neurons

is a downstream response to maintain the E/I balance in the face of

persistently lower DLPFC network excitatory activity. The idea of an

upstream deficit in cortical excitatory activity is supported by

evidence that schizophrenia is associated with an intrinsic deficit in

excitatory drive to layer 3 pyramidal neurons. The density of dendritic

spines, which reflect the number of glutamatergic synapses to

pyramidal cells, is significantly decreased in schizophrenia, and this

deficit is most pronounced in DLPFC deep layer 3 and present to a

milder degree in superficial layer 3 [106,107]. In addition, the somal

size [108,109] and dendritic tree [106] of layer 3 pyramidal neurons are

smaller in schizophrenia, suggestive of a developmental disturbance

in these neurons. Consistent with this interpretation, in cortical

regions with layer 3 pyramidal neurons that are smaller and have

lower spine density in schizophrenia (e.g. the superior temporal

gyrus) [109,110], gray matter volume is lower at baseline in high-risk

youth [111] and progressively declines in those who subsequently

convert to psychosis [112,113].

The idea that an intrinsic deficit in excitatory drive to layer 3

pyramidal cells is an upstream event is supported by the findings

that the expression of certain gene products [e.g. Duo, cell division

cycle 42 (Cdc42)] that regulate dendritic spine formation and

maintenance are altered in schizophrenia, and these alterations are

strongly correlated with DLPFC layer 3 spine deficits [114]. The

particular prominence of the deficit in dendritic spines on layer 3

pyramidal neurons may reflect the layer specificity of certain

molecular abnormalities. For example, Cdc42 effector protein 3

(Cdc42EP3) mRNA, which is preferentially expressed in layer 3 of

human DLPFC [115], is upregulated in schizophrenia [116]. The

activation of Cdc42 by glutamate stimulation is thought to inhibit

Cdc42EP3 activity, which in turn dissociates the complex of septin

filaments in spine necks, enabling the movement from the parent

dendrite of molecules required for synaptic potentiation [116]. Thus,

lower levels of Cdc42 and higher levels of Cdc42EP3 might lead to a

reduced capacity for glutamatergic stimuli to open the septin

filament barrier in the spine neck, resulting in impaired synaptic

plasticity and spine loss.

Thus, intrinsic deficits in dendritic spines resulting in lower activity

of layer 3 pyramidal neurons might represent an upstream pathology

that induces several homeostatic responses (Figure 1, main text) to

reduce PVBC inhibition of pyramidal neurons to rebalance network

levels of excitation and inhibition.

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

that PVBC firing is more strongly coupled to the gammaoscillation cycle than is the firing of PVChCs or CCKBCs[55]. Furthermore, gamma oscillations are significantlyreduced by stimulation of the presynaptic m opioid recep-tors that suppress GABA release fromPVBCs, but not fromPVChCs or CCKBCs [55].

Thus, the multiple alterations that weaken PVBC inhi-bition of pyramidal neurons could provide the neural sub-strate for the lower power of frontal lobe gamma bandoscillations during cognitive control tasks [8,9]. The pres-ence of such abnormalities in both the first episode [9] andchronic [8] phases of the illness is consistent with the ideathat the alterations in GABA markers observed in post-mortem studies are not a consequence of the treatment orchronicity of the illness [23]. Furthermore, the tendency forthese alterations to be most prominent in layer 3 is consis-tent with evidence that circuitry in this laminar location ofthe primate neocortex is critical for both gamma oscilla-tions [83] and delay-dependent cognitive control tasks [84].However, these findings raise the question of whether themultiple pre- and postsynaptic factors that lower PVBCinhibition could each represent a different type of primarypathology, any of which could lead to impaired gammaoscillations and the resulting cognitive control deficits inschizophrenia. However, given the frequency of these find-ings in individuals with schizophrenia [23,63,69,77], andtheir co-occurrence in the same individuals, it seems moreprobable that together they represent convergent conse-quences or compensations to some common factor that isupstream in the disease process.

PV interneurons and cortical excitatory-inhibitorybalance in schizophreniaOne possible upstream factor might be the dendritic spinedeficit on layer 3 pyramidal neurons [85], which could lead

62

to a net reduction in local DLPFC excitatory activity andimpaired gamma oscillations. According to the pyramidal-interneuron network gamma (PING) model of gammaoscillations, PVBCs are recruited by phasic, glutamatergicinputs from pyramidal neurons, and PVBCs provide strongand fast (i.e. GABAA a1-receptor-mediated) feedback inhi-bition to pyramidal neurons [86]. The divergent connec-tions of neocortical PVBCs [87] result in the simultaneoushyperpolarization of a distributed group of pyramidal neu-rons, and the fast and synchronous decay of this inhibitionpermits the simultaneous firing of the pyramidal cells atgamma frequency (reviewed in [10]). The strength of theexcitatory inputs to PVBCs from neighboring pyramidalcells is likely to be lower in schizophrenia because thesepyramidal cells, owing to a deficit in dendritic spines andpresumably fewer excitatory inputs, are thought to be lessactive (Box 1). Lower expression of the NR2A subunit ofglutamatergic NMDA receptors in DLPFC layer 3 PVneurons in schizophrenia [88] could also contribute tothe reduced strength of excitatory inputs to PVBCs; inter-estingly, computational modeling suggests that loweringthe slow excitatory current fromNMDA receptors (relativeto the fast excitatory current provided by AMPA receptors)increases gamma band power [89], suggesting that down-regulation of NMDA receptors represents a compensatoryresponse in PVBCs.

The net reduction in network excitatory activity owingto layer 3 pyramidal neuron spine deficits (Box 1) mightthen evoke homeostatic mechanisms to reduce the inhibi-tion of these pyramidal cells. From this perspective, all ofthe molecular alterations described above (Figures 1 and2) that weaken PVBC inhibition of pyramidal neuronscould be understood as compensatory responses to lowerpyramidal cell inhibition and to restore the excitatory–

inhibitory balance (E/I balance) in DLPFC circuitry

Box 2. Outstanding questions

� Are protein levels of GAD67 altered in the axon terminals of

PVChCs? The idea that PVBC inhibition is reduced to compensate

for an intrinsic deficit in pyramidal cell excitability (Box 1), and that

PVChCs are excitatory, makes the prediction that GAD67 levels and

GABA synthesis are either normal or increased in the axon

terminals of PVChCs; that is, PVChCs are predicted to be among

the approximately 50% of PV neurons that do not lack detectable

GAD67 mRNA in schizophrenia.

� Are protein levels of GAT1 altered in the axon terminals of

PVBCs? Although changes in GAT1 levels have not been reported

in experimental models of homeostatic synaptic plasticity, the

idea that changes in PVBCs and PVChCs are compensatory to

reduce inhibition and increase excitation of pyramidal cells,

respectively, predicts that the GAD67:GAT1 ratio is increased in

PVChC axon cartridges and decreased in PVBC terminals in

schizophrenia, a prediction that can now be tested with new

methods of quantifying protein levels in specific populations of

axon terminals [23,82].

� Is the deficit in PV mRNA expression in schizophrenia [24,40,41]

present in PVBCs and/or PVChCs? Because the absence of PV

facilitates repetitive GABA release and augments gamma oscilla-

tions [117], the current model predicts that PV protein levels would

be lower in the axon cartridges of PVChCs and unchanged in PVBC

terminals in schizophrenia.

� Are markers of excitatory inputs to PVBCs lower in schizophrenia

as an additional means of reducing their inhibitory output?Unfortunately, because molecular markers that differentiate PVBCs

from PVChCs are still unknown, it is not yet possible to determine

whether AMPA and NMDA receptor subunits are specifically

downregulated in PVBCs in schizophrenia.

� Are levels of the vesicular GABA transporter (vGAT), which

regulates the loading of GABA into synaptic vesicles, altered in

PVBCs? Because changes in vGAT contribute to homeostatic

synaptic plasticity in experimental systems [118], the current model

predicts that vGAT would be lower in the axon terminals of PVBCs.

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

(Box 1). The idea that altered excitation is upstream ofaltered inhibition is supported by recent findings thatgenes related to glutamate signaling (and not those relat-ed to GABA signaling) appear to mediate schizophreniasusceptibility [90].

If neocortical chandelier cells are, in fact, a potentsource of a slow depolarizing current in cortical pyramidalcells, then each of the protein alterations at PVChC inputsto pyramidal cells (e.g. lower presynaptic GAT1 and higherpostsynaptic GABAA a2 receptors; Figure 1) would prolongthe duration and increase the strength of the excitatorypostsynaptic current in the AIS, respectively. Thus, incontrast to the previous interpretation that these changesin PVChC inputs are compensations to augment pyramidalcell inhibition [11], the idea that PVChCs are excitatorysuggests that the pre- and postsynaptic changes in theirinputs to pyramidal neurons provide a means to increasethe type of slow, NMDA-like depolarization of pyramidalcells that is thought to be essential for the DLPFC neuralnetwork activity associated with cognitive control tasks[91]. Thus, the well-documented alterations in PVChCsynapses in schizophrenia might provide another meansto increase excitation and restore the E/I balance. Such amechanism might explain the increase in frontal gammaoscillations during a cognitive control task seen in patientswith schizophrenia following treatment with a GABAA a2receptor agonist [18].

This hypothesis, that lower excitatory activity in layer 3pyramidal neurons leads to multiple compensatoryresponses to reduce inhibition from PVBCs and increaseexcitation from PVChCs, requires answers to additionalquestions regarding the nature of the alterations in PVBCsand PVChCs in schizophrenia (Box 2). It also raises inter-esting interpretations and predictions for clinical observa-tions in schizophrenia. First, although the compensatorychanges in PVBC activity are predicted to rebalance excita-tion and inhibition, the new level of E/I balance in theDLPFC of subjects with schizophrenia would lack thestrength of both excitation and inhibition required for gen-erating sufficient levels of gamma band power to supportcognitive control (Figure 3). For example, cell type-specificexperimental manipulations in mice have demonstrated

63

that lowering either AMPA receptor-mediated excitatoryinputs to PV-positive neurons [92] or inhibitory output fromPVneurons [45,46] (butnot inhibitory inputs toPVBCs [93])reduces gammabandpower. The deleterious effects of lowerlevels of both excitation and inhibition in the PVBC-pyra-midal neuron circuit on generating gamma oscillationsmight be expected to be most evident under task conditionsdemanding high levels of cognitive control, as demonstratedin experimental studies of subjects with schizophrenia [8,9].

Second, the reset E/I balance in individuals with schizo-phrenia, at lower levels of both excitation and inhibition,would have less dynamic range for adjusting excitation andinhibition in the face of new forces that alter one or theother. That is, homeostatic synaptic plasticity (the capacityto scale the strength of all excitatory and inhibitory inputsto a neuron in response to large scale changes in networkactivity [94]) would be limited. This reduced capacity of thecircuitry to respond to challenges altering either excitationor inhibition might explain: (i) the tendency for thesymptoms of schizophrenia to worsen in the face ofstress-induced changes in prefrontal activity [95]; (ii) theworsening of cognitive deficits [2] that is temporally asso-ciated with the normal adolescence-related pruning ofexcitatory synapses, which is most prominent in layer 3of the DLPFC [96]; and (iii) the increased liability to, andseverity of, schizophrenia associated with marijuana use[97], which can suppress cortical GABA release [98].

It is important to note that an alternativemodel [that anupstream GAD67 deficit in PV neurons (owing to NMDAreceptor hypofunction in these neurons) leads to a disinhi-bition of pyramidal cells] has support from pharmacologi-cal animal models of schizophrenia [99]. However, theresulting increase in cortical network activity would beexpected to evoke compensatory responses to enhanceinhibition, but as summarized above, schizophrenia isaccompanied by multiple molecular changes, each of whichwould decrease PVBC-mediated inhibition in DLPFCcircuitry.

The current findings and suggested model for inter-preting these findings focus on only a limited portion ofcortical circuitry; a full accounting of the pathophysiologyunderlying cognitive control deficits in schizophrenia

[(Figure_3)TD$FIG]

PrefrontalGamma Band Power

PVBC

PVBC

RecurrentExcitation

GABAA α1 ReceptorGABAA α2 Receptor Glutamate Receptor

FeedbackInhibition

↓ RecurrentExcitation

↓FeedbackInhibition

Healthy: Normal E/I Balance

Schizophrenia: “Re-Set” E/I Balance

Key:

20

30

40

50

60

70

80

Time

Fre

quen

cy (

Hz)

PrefrontalGamma Band Power

20

30

40

50

60

70

80

Time

Fre

quen

cy (

Hz)

PVChC

Lower recurrent excitationbetween P neurons

• Smaller dendritic arbor • Fewer dendritic spines

Lower feedback inhibitionfrom PVBCs

• Reduced GABA synthesis • Increased suppression of GABA release • Fewer GABAA α1 receptors • Reduced chloride influx

Greater P neuron depolarizationfrom PVChCs

• Reduced GABA re-uptake • More GABAA α2 receptors

PVChC

TRENDS in Neurosciences

Figure 3. Connectivity between pyramidal neurons (gray neurons) and parvalbumin-positive basket (PVBC) and chandelier (PVChC) cells in dorsolateral prefrontal cortex

(DLPFC) layer 3. Reciprocal connections formed by the local axon collaterals of pyramidal neurons provide recurrent excitation, whereas the excitatory inputs from pyramidal

neurons to PVBCs furnish feedback inhibition. These connections are critical for generating gamma band oscillations, and the strengths of these connections are adjusted to

maintain a normal excitation–inhibition (E/I) balance in the healthy brain (a). In schizophrenia (b), lower spine density in layer 3 pyramidal neurons is hypothesized to result in

lower network excitation, evoking a compensatory reduction in feedback inhibition of pyramidal neurons from PVBCs [less presynaptic glutamic acid decarboxylase (GAD)-67;

fewer postsynaptic GABAA a1 receptors) and increased depolarization of pyramidal neurons by PVChCs (less presynaptic GABA membrane transporter 1; more postsynaptic

GABAA a2 receptors). The resulting ‘reset’ of the E/I balance at a lower level of both excitation and inhibition renders the circuit less able to generate normal levels of gamma

band power, resulting in impaired cognition. Heat maps are reproduced, with permission, from [8]. � (2006) Proceedings of the National Academy of Sciences.

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

requires both better knowledge of the patterns ofconnectivity within the DLPFC and more sensitive meth-ods for assessing the functional integrity and compensa-tions of these connections in the illness. In addition, theuse of appropriate animal models as proof-of-concept testsof the hypothesized upstream versus downstreamrelationships between cellular level alterations in schizo-phrenia are essential. Such future advances will provide a

64

more informed substrate for designing rational interven-tions to enhance cognitive function in people withschizophrenia.

Disclosure statementD.A.L. currently receives investigator-initiated researchsupport from Bristol-Myers Squibb (BMS), Curridium andPfizer and, in 2009–2011, served as a consultant in the areas

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

of target identification and validation and new compounddevelopment to BioLine RX, BMS, Merck and SK LifeScience.

AcknowledgmentsThe authors thank Mary Brady for technical assistance. The workconducted by the authors and cited in this review was supported byNational Institutes of Health (NIH) grants MH084053 and MH043784.The content is solely the responsibility of the authors and does notnecessarily represent the official views of the NIH or the NationalInstitute of Mental Health.

References1 Keefe, R.S. and Fenton, W.S. (2007) How should DSM-V criteria for

schizophrenia include cognitive impairment? Schizophr. Bull. 33,912–920

2 Reichenberg, A. et al. (2010) Static and dynamic cognitive deficits inchildhood preceding adult schizophrenia: a 30-year study. Am. J.Psychiatry 167, 160–169

3 Green, M.F. (2006) Cognitive impairment and functional outcome inschizophrenia and bipolar disorder. J. Clin. Psychiatry 67 (Suppl. 9),3–8

4 Lesh, T.A. et al. (2011) Cognitive control deficits in schizophrenia:mechanisms and meaning. Neuropsychopharmacology 36, 316–338

5 Miller, E.K. and Cohen, J.D. (2001) An integrative theory of prefrontalcortex function. Annu. Rev. Neurosci. 24, 167–202

6 Howard,M.W. et al. (2003) Gamma oscillations correlate withworkingmemory load in humans. Cereb. Cortex 13, 1369–1374

7 Minzenberg, M.J. et al. (2009) Meta-analysis of 41 functionalneuroimaging studies of executive function in schizophrenia. Arch.Gen. Psychiatry 66, 811–822

8 Cho, R.Y. et al. (2006) Impairments in frontal cortical gammasynchrony and cognitive control in schizophrenia. Proc. Natl. Acad.Sci. U.S.A. 103, 19878–19883

9 Minzenberg, M.J. et al. (2010) Gamma oscillatory power is impairedduring cognitive control independent of medication status in first-episode schizophrenia. Neuropsychopharmacology 35, 2590–2599

10 Gonzalez-Burgos, G. and Lewis, D.A. (2008) GABA neurons and themechanisms of network oscillations: implications for understandingcortical dysfunction in schizophrenia. Schizophr. Bull. 34, 944–961

11 Lewis, D.A. et al. (2005) Cortical inhibitory neurons andschizophrenia. Nat. Rev. Neurosci. 6, 312–324

12 Lodge, D.J. et al. (2009) A loss of parvalbumin-containinginterneurons is associated with diminished oscillatory activity inan animal model of schizophrenia. J. Neurosci. 29, 2344–2354

13 Enomoto, T. et al. (2011) Reducing prefrontal gamma-aminobutyricacid activity induces cognitive, behavioral, and dopaminergicabnormalities that resemble schizophrenia. Biol. Psychiatry 69,432–441

14 Gruber, A.J. et al. (2010) More is less: a disinhibited prefrontal corteximpairs cognitive flexibility. J. Neurosci. 30, 17102–17110

15 Paine, T.A. et al. (2011) Schizophrenia-like attentional deficitsfollowing blockade of prefrontal cortex GABA(A) receptors.Neuropsychopharmacology 36, 1703–1713

16 Sawaguchi, T. et al. (1989) Delayed response deficits produced by localinjection of bicuculline into the dorsolateral prefrontal cortex inJapanese macaque monkeys. Exp. Brain Res. 75, 457–469

17 Ahn, K. et al. (2011) Probing GABA receptor function in schizophreniawith iomazenil. Neuropsychopharmacology 36, 677–683

18 Lewis, D.A. et al. (2008) Subunit-selective modulation of GABA type Areceptor neurotransmission and cognition in schizophrenia. Am. J.Psychiatry 165, 1585–1593

19 Asada, H. et al. (1997) Cleft palate and decreased brain gamma-aminobutyric acid in mice lacking the 67-kDa isoform of glutamicacid decarboxylase. Proc. Natl. Acad. Sci. U.S.A. 94, 6496–6499

20 Asada, H. et al. (1996) Mice lacking the 65 kDa isoform of glutamicacid decarboxylase (GAD65) maintain normal levels of GAD67 andGABA in their brains but are susceptible to seizures. Biochem.Biophys. Res. Comm. 229, 891–895

21 Gonzalez-Burgos, G. et al. (2010) Alterations of cortical GABAneurons and network oscillations in schizophrenia. Curr.Psychiatry Rep. 12, 335–344

65

22 Guidotti, A. et al. (2000) Decrease in reelin and glutamic aciddecarboxylase67 (GAD67) expression in schizophrenia and bipolardisorder. Arch. Gen. Psychiatry 57, 1061–1069

23 Curley, A.A. et al. (2011) Cortical deficits of glutamic aciddecarboxylase 67 expression in schizophrenia: clinical, protein, andcell type-specific features. Am. J. Psychiatry 168, 921–929

24 Hashimoto, T. et al. (2008) Conserved regional patterns of GABA-related transcript expression in the neocortex of subjects withschizophrenia. Am. J. Psychiatry 165, 479–489

25 Thompson,M. et al. (2009) Decreased glutamic acid decarboxylase(67)mRNA expression in multiple brain areas of patients withschizophrenia and mood disorders. J. Psychiatr. Res. 43, 970–977

26 Woo, T-U. et al. (2004) Density of glutamic acid decarboxylase 67messenger RNA-containing neurons that express the N-methyl- D-aspartate receptor subunit NR2A in the anterior cingulate cortex inschizophrenia and bipolar disorder. Arch. Gen. Psychiatry 61, 649–657

27 Impagnatiello, F. et al. (1998) A decrease of reelin expression as aputative vulnerability factor in schizophrenia. Proc. Natl. Acad. Sci.U.S.A. 95, 15718–15723

28 Hashimoto, T. et al. (2008) Alterations in GABA-relatedtranscriptome in the dorsolateral prefrontal cortex of subjects withschizophrenia. Mol. Psychiatry 13, 147–161

29 Benes, F.M. et al. (2000) Glutamate decarboxylase(65)-immunoreactive terminals in cingulate and prefrontal cortices ofschizophrenic and bipolar brain. J. Chem. Neuroanat. 20, 259–269

30 Benson, D.L. et al. (1994) Activity-dependent changes in GAD andpreprotachykinin mRNAs in visual cortex of adult monkeys. Cereb.Cortex. 4, 40–51

31 Mason, G.F. et al. (2001) Decrease in GABA synthesis rate in ratcortex following GABA-transaminase inhibition correlates with thedecrease in GAD(67) protein. Brain Res. 914, 81–91

32 Ongur, D. et al. (2010) Elevated gamma-aminobutyric acid levels inchronic schizophrenia. Biol. Psychiatry 68, 667–670

33 Goto, N. et al. (2009) Reduction of brain gamma-aminobutyric acid(GABA) concentrations in early-stage schizophrenia patients: 3TProton MRS study. Schizophr. Res. 112, 192–193

34 Yoon, J.H. et al. (2010) GABA concentration is reduced in visual cortexin schizophrenia and correlates with orientation-specific surroundsuppression. J. Neurosci. 30, 3777–3781

35 Goto, N. et al. (2009) Associations between plasma levels of3-methoxy-4-hydroxyphenylglycol (MHPG) and negativesymptoms or cognitive impairments in early-stage schizophrenia.Hum. Psychopharmacol. 24, 639–645

36 Frankle, W.G. et al. (2009) Tiagabine increases [11C]flumazenilbinding in cortical brain regions in healthy control subjects.Neuropsychopharmacology 34, 624–633

37 Akbarian, S. et al. (1995) Gene expression for glutamic aciddecarboxylase is reduced without loss of neurons in prefrontalcortex of schizophrenics. Arch. Gen. Psychiatry 52, 258–266

38 Volk, D.W. et al. (2000) Decreased glutamic acid decarboxylase67messenger RNA expression in a subset of prefrontal cortical gamma-aminobutyric acid neurons in subjects with schizophrenia. Arch. Gen.Psychiatry 57, 237–245

39 Hashimoto, T. et al. (2003) Gene expression deficits in a subclass ofGABA neurons in the prefrontal cortex of subjects with schizophrenia.J. Neurosci. 23, 6315–6326

40 Mellios, N. et al. (2009) Molecular determinants of dysregulatedGABAergic gene expression in the prefrontal cortex of subjectswith schizophrenia. Biol. Psychiatry 65, 1006–1014

41 Fung, S.J. et al. (2010) Expression of interneuron markers in thedorsolateral prefrontal cortex of the developing human and inschizophrenia. Am. J. Psychiatry 167, 1479–1488

42 Woo, T-U. et al. (1997) Schizophrenia and the parvalbumin-containing class of cortical local circuit neurons. Am. J.Psychiatry 154, 1013–1015

43 Beasley, C.L. et al. (2002) Selective deficits in prefrontal corticalGABAergic neurons in schizophrenia defined by the presence ofcalcium-binding proteins. Biol. Psychiatry 52, 708–715

44 Tooney, P.A. and Chahl, L.A. (2004) Neurons expressing calcium-binding proteins in the prefrontal cortex in schizophrenia. Prog.Neuropsychopharmacol. Biol. Psychiatry 28, 273–278

45 Cardin, J.A. et al. (2009) Driving fast-spiking cells induces gammarhythm and controls sensory responses. Nature 459, 663–667

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

46 Sohal, V.S. et al. (2009) Parvalbumin neurons and gamma rhythmsenhance cortical circuit performance. Nature 459, 698–702

47 Woo, T-U. et al. (1998) A subclass of prefrontal gamma-aminobutyricacid axon terminals are selectively altered in schizophrenia. Proc.Natl. Acad. Sci. U.S.A. 95, 5341–5346

48 Pierri, J.N. et al. (1999) Alterations in chandelier neuron axonterminals in the prefrontal cortex of schizophrenic subjects. Am. J.Psychiatry 156, 1709–1719

49 Volk, D.W. et al. (2002) Reciprocal alterations in pre- and postsynapticinhibitory markers at chandelier cell inputs to pyramidal neurons inschizophrenia. Cereb. Cortex 12, 1063–1070

50 Beneyto, M. et al. (2011) Lamina-specific alterations in corticalGABAA receptor subunit expression in schizophrenia. Cereb.Cortex 21, 999–1011

51 Szabadics, J. et al. (2006) Excitatory effect of GABAergic axo-axoniccells in cortical microcircuits. Science 311, 233–235

52 Baldi, R. et al. (2010) Differential distribution of KCC2 along the axo-somato-dendritic axis of hippocampal principal cells.Eur. J. Neurosci.32, 1319–1325

53 Woodruff, A.R. et al. (2010) The enigmatic function of chandelier cells.Front Neurosci. 4, 201

54 Capogna, M. and Pearce, R.A. (2011) GABA(A,slow): causes andconsequences. Trends Neurosci. 34, 101–112

55 Gulyas, A.I. et al. (2010) Parvalbumin-containing fast-spikingbasket cells generate the field potential oscillations induced bycholinergic receptor activation in the hippocampus. J. Neurosci. 30,15134–15145

56 Klausberger, T. and Somogyi, P. (2008) Neuronal diversity andtemporal dynamics: the unity of hippocampal circuit operations.Science 321, 53–57

57 Nusser, Z. et al. (1996) Differential synaptic localization of two majorg-aminobutyric acid type A receptor a subunits on hippocampalpyramidal cells. Proc. Natl. Acad. Sci. U.S.A. 93, 11939–11944

58 Doischer, D. et al. (2008) Postnatal differentiation of basket cells fromslow to fast signaling devices. J. Neurosci. 28, 12956–12968

59 Akbarian, S. et al. (1995) GABAA receptor subunit gene expression inhuman prefrontal cortex: comparison of schizophrenics and controls.Cereb. Cortex 5, 550–560

60 Duncan, C.E. et al. (2010) Prefrontal GABA(A) receptor alpha-subunitexpression in normal postnatal human development andschizophrenia. J. Psychiatry Res. 44, 673–681

61 Olsen, R.W. and Sieghart, W. (2009) GABA A receptors: subtypesprovide diversity of function and pharmacology. Neuropharmacology56, 141–148

62 Maldonado-Aviles, J.G. et al. (2009) Altered markers of tonicinhibition in the dorsolateral prefrontal cortex of subjects withschizophrenia. Am. J. Psychiatry 166, 450–459

63 Glausier, J.R. and Lewis, D.A. (2011) Selective pyramidal cellreduction of GABA(A) receptor alpha1 subunit messenger RNAexpression in schizophrenia. Neuropsychopharmacology 36, 2103–

211064 Benes, F.M. et al. (1996) Up-regulation of GABA-A receptor binding on

neurons of the prefrontal cortex in schizophrenic subjects.Neuroscience 75, 1021–1031

65 Lewis, D.A. et al. (2001) Lamina-specific deficits in parvalbumin-immunoreactive varicosities in the prefrontal cortex of subjectswith schizophrenia: evidence for fewer projections from thethalamus. Am. J. Psychiatry 158, 1411–1422

66 Chattopadhyaya, B. et al. (2007) GAD67-mediated GABA synthesisand signaling regulate inhibitory synaptic innervation in the visualcortex. Neuron 54, 889–903

67 Melchitzky, D.S. et al. (1999) Parvalbumin-immunoreactive axonterminals in macaque monkey and human prefrontal cortex:laminar, regional and target specificity of Type I and Type IIsynapses. J. Comp. Neurol. 408, 11–22

68 Farrant, M. and Kaila, K. (2007) The cellular, molecular and ionicbasis of GABA(A) receptor signalling. Prog. Brain Res. 160, 59–87

69 Arion, D. and Lewis, D.A. (2011) Altered expression of regulators ofthe cortical chloride transporters NKCC1 andKCC2 in schizophrenia.Arch. Gen. Psychiatry 68, 21–31

70 Hyde, T.M. et al. (2011) Expression of GABA signaling moleculesKCC2, NKCC1, andGAD1 in cortical development and schizophrenia.J. Neurosci. 31, 11088–11095

66

71 Drake, C.T. and Milner, T.A. (2002) Mu opioid receptors are in discretehippocampal interneuron subpopulations. Hippocampus 12, 119–136

72 Stumm, R.K. et al. (2004) Neuronal types expressing mu- and delta-opioid receptor mRNA in the rat hippocampal formation. J. Comp.Neurol. 469, 107–118

73 Wimpey, T.L. and Chavkin, C. (1991) Opioids activate both an inwardrectifier and a novel voltage-gated potassium conductance in thehippocampal formation. Neuron 6, 281–289

74 Glickfeld, L.L. et al. (2008) Complementary modulation of somaticinhibition by opioids and cannabinoids. J. Neurosci. 28, 1824–1832

75 Capogna, M. et al. (1993) Mechanism of mu-opioid receptor-mediatedpresynaptic inhibition in the rat hippocampus in vitro. J. Physiol. 470,539–558

76 Lupica, C.R. (1995) Delta and mu enkephalins inhibit spontaneousGABA-mediated IPSCs via a cyclic AMP-independent mechanism inthe rat hippocampus. J Neurosci. 15, 737–749

77 Volk, D.W. et al. (2011) Cortical opioid markers in schizophrenia andacross development. Cereb. Cortex DOI: 10.1093/cercor/bhr202

78 Peckys, D. and Hurd, Y.L. (2001) Prodynorphin and kappa opioidreceptor mRNA expression in the cingulate and prefrontal cortices ofsubjects diagnosed with schizophrenia or affective disorders. BrainRes. Bull. 55, 619–624

79 Foldy, C. et al. (2007) Cell type-specific gating of perisomaticinhibition by cholecystokinin. Nat. Neurosci. 10, 1128–1130

80 Karson,M.A. et al. (2008)Cholecystokinin inhibits endocannabinoid-sensitive hippocampal IPSPs and stimulates others.Neuropharmacology 54, 117–128

81 Karson, M.A. et al. (2009) Synaptic cross talk between perisomatic-targeting interneuron classes expressing cholecystokinin andparvalbumin in hippocampus. J. Neurosci. 29, 4140–4154

82 Fish, K.N. et al. (2011) Differential distribution of proteins regulatingGABA synthesis and reuptake in axon boutons of subpopulations ofcortical interneurons. Cereb. Cortex 21, 2450–2460

83 Buffalo, E.A. et al. (2011) Laminar differences in gamma and alphacoherence in the ventral stream. Proc. Natl. Acad. Sci. U.S.A. 108,11262–11267

84 Goldman-Rakic, P.S. (1995) Cellular basis of working memory.Neuron 14, 477–485

85 Lewis, D.A. and Gonzalez-Burgos, G. (2008) Neuroplasticity ofneocortical circuits in schizophrenia. NeuropsychopharmacologyRev. 33, 141–165

86 Hajos, N. and Paulsen, O. (2009) Network mechanisms of gammaoscillations in the CA3 region of the hippocampus. Neural Netw. 22,1113–1119

87 Packer, A.M. and Yuste, R. (2011) Dense, unspecific connectivity ofneocortical parvalbumin-positive interneurons: a canonicalmicrocircuit for inhibition? J. Neurosci. 31, 13260–13271

88 Bitanihirwe, B.K. et al. (2009) Glutamatergic deficits andparvalbumin-containing inhibitory neurons in the prefrontal cortexin schizophrenia. BMC Psychiatry 9, 71

89 Rotaru, D.C. et al. (2011) Glutamate receptor subtypes mediatingsynaptic activation of prefrontal cortex neurons: relevance forschizophrenia. J. Neurosci. 31, 142–156

90 Greenwood, T.A. et al. (2011) Analysis of 94 candidate genes and 12endophenotypes for schizophrenia from the consortium on thegenetics of schizophrenia. Am. J. Psychiatry 168, 930–946

91 Wang, X.J. (2010) Neurophysiological and computational principles ofcortical rhythms in cognition. Physiol. Rev. 90, 1195–1268

92 Fuchs, E.C. et al. (2007) Recruitment of parvalbumin-positiveinterneurons determines hippocampal function and associatedbehavior. Neuron 53, 591–604

93 Wulff, P. et al. (2009) Hippocampal theta rhythm and its coupling withgamma oscillations require fast inhibition onto parvalbumin-positiveinterneurons. Proc. Natl. Acad. Sci. U.S.A. 106, 3561–3566

94 Turrigiano, G.G. and Nelson, S.B. (2004) Homeostatic plasticity in thedeveloping nervous system. Nat. Rev. Neurosci. 5, 97–107

95 Arnsten, A.F. (2009) Stress signalling pathways that impairprefrontal cortex structure and function. Nat. Rev. Neurosci. 10,410–422

96 Bourgeois, J-P. et al. (1994) Synaptogenesis in the prefrontal cortex ofrhesus monkeys. Cereb. Cortex 4, 78–96

97 van Os, J. et al. (2010) The environment and schizophrenia. Nature468, 203–212

Review Trends in Neurosciences January 2012, Vol. 35, No. 1

98 Katona, I. et al. (1999) Presynaptically located CB1 cannabinoidreceptors regulate GABA release from axon terminals of specifichippocampal interneurons. J. Neurosci. 19, 4544–4558

99 Homayoun, H. and Moghaddam, B. (2007) NMDA receptorhypofunction produces opposite effects on prefrontal cortexinterneurons and pyramidal neurons. J Neurosci. 27, 11496–11500

100 Cruz, D.A. et al. (2003) Postnatal development of pre- and post-synaptic GABA markers at chandelier cell inputs to pyramidalneurons in monkey prefrontal cortex. J. Comp. Neurol. 465, 385–400

101 Hashimoto, T. et al. (2009) Protracted developmental trajectories ofGABAA receptor alpha1 and alpha2 subunit expression in primateprefrontal cortex. Biol. Psychiatry 65, 1015–1023

102 Yizhar, O. et al. (2011) Neocortical excitation/inhibition balance ininformation processing and social dysfunction. Nature 477, 171–178

103 Rich, M.M. and Wenner, P. (2007) Sensing and expressinghomeostatic synaptic plasticity. Trends Neurosci. 30, 119–125

104 Pozo, K. and Goda, Y. (2010) Unraveling mechanisms of homeostaticsynaptic plasticity. Neuron 66, 337–351

105 Kilman, V. et al. (2002) Activity deprivation reduces miniature IPSCamplitude by decreasing the number of postsynaptic GABA(A)receptors clustered at neocortical synapses. J. Neurosci. 22, 1328–1337

106 Glantz, L.A. and Lewis, D.A. (2000) Decreased dendritic spine densityon prefrontal cortical pyramidal neurons in schizophrenia. Arch. Gen.Psychiatry 57, 65–73

107 Kolluri, N. et al. (2005) Lamina-specific reductions in dendritic spinedensity in the prefrontal cortex of subjects with schizophrenia. Am. J.Psychiatry 162, 1200–1202

108 Pierri, J.N. et al. (2001) Decreased somal size of deep layer 3pyramidal neurons in the prefrontal cortex of subjects withschizophrenia. Arch. Gen. Psychiatry 58, 466–473

67

109 Sweet, R.A. et al. (2004) Pyramidal cell size reduction inschizophrenia: evidence for involvement of auditory feedforwardcircuits. Biol. Psychiatry 55, 1128–1137

110 Sweet, R.A. et al. (2009) Reduced dendritic spine density in auditorycortex of subjects with schizophrenia. Neuropsychopharmacology 34,374–389

111 Pantelis, C. et al. (2003) Neuroanatomical abnormalities before andafter onset of psychosis: a cross-sectional and longitudinal MRIcomparison. Lancet 361, 281–288

112 Takahashi, T. et al. (2009) Progressive gray matter reduction of thesuperior temporal gyrus during transition to psychosis. Arch. Gen.Psychiatry 66, 366–376

113 McIntosh, A.M. et al. (2011) Longitudinal volume reductions in peopleat high genetic risk of schizophrenia as they develop psychosis. Biol.Psychiatry 69, 953–958

114 Hill, J.J. et al. (2006) Molecular mechanisms contributing to dendriticspine alterations in the prefrontal cortex of subjects withschizophrenia. Mol. Psychiatry 11, 557–566

115 Arion, D. et al. (2007) Molecular markers distinguishingsupragranular and infragranular layers in the human prefrontalcortex. Eur. J. Neurosci. 25, 1843–1854

116 Ide, M. and Lewis, D.A. (2010) Altered cortical CDC42 signalingpathways in schizophrenia: implications for dendritic spine deficits.Biol. Psychiatry 68, 25–32

117 Vreugdenhil, M. et al. (2003) Parvalbumin-deficiency facilitatesrepetitive IPSCs and gamma oscillations in the hippocampus.J. Neurophysiol. 89, 1414–1422

118 De Gois, S. et al. (2005) Homeostatic scaling of vesicular glutamateand GABA transporter expression in rat neocortical circuits.J. Neurosci. 25, 7121–7133