C H A P T E R 1 Neuroimaging Modalities Description, Comparisons, Strengths, and Weaknesses Richard G. Wise Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff University, Park Place, Cardiff CF10 3AT, United Kingdom 1.0. Introduction 2 3.2. Magnetoencephalography 10 1.1. Windows to the Brain 2 4.0. MagneticResonance Techniques 10 1.1.1. BrainStructure 3 4.1. MRS 11 1.1.2. BrainFunction 3 4.2. Structural MRI 12 2.0. RadiotracerTechniques 4 4.3. fMRI 13 2.1. Single-Photon Emission Computed 4.3.1. BOLDfMRI 14 Tomography 4 4.3.2. PharmacologicalfMRI 15 2.2. PositronEmission Tomography 5 4.3.3. ArterialSpinLabeling 2.2.1. CerebralMetabolismand fMRI 16 BloodFlow 6 5.0. Advantages, Disadvantages, and 2.2.2. ReceptorStudies 6 Practical Considerations 16 3.0. Electrophysiological Techniques 7 3.1. Electroencephalography 8 Summary Inthischapter,Ireviewthemostimportantneuroimagingmethodsthataregaininggroundastranslational tools in central nervous system (CNS) drug discovery, drug development and treatment. I consider the informationthatmaybegainedaboutbrainstructureandbrainfunctionfromeachofthetechniques,focusing inparticularonpositronemissiontomography(PET),magneticresonanceimaging(MRI),magneticresonance TranslationalNeuroimaging http://dx.doi.org/10.1016/B978-0-12-386945-6.00001-9 1 (cid:1)2013ElsevierInc.Allrightsreserved. 2 1. NEUROIMAGINGMODALITIES spectroscopy (MRS), electroencephalography (EEG) and magnetoencephalography (MEG). I explore the signalsthattheyprovideandhowtheymaybeused,alongwiththeirrelativeadvantages,disadvantages,and thepracticalitiesofusingthedifferenttechniques.Thisroadmapthroughtherangeofthemostwidelyapplied neuroimagingtechniquesrelevanttotranslationalresearchprovidesanintroductionforthelaterchaptersthat describeapplicationsinspecificdiseaseareas. 1.0. INTRODUCTION This chapter aims to provide an introduction to the neuroimaging modalities that are important to drug discovery, development and treatment. I describe the basis of each tech- nique and compare each one with the others, offering a roadmap through the jungle of different types of neuroimaging, including in particular magnetic resonance (MR) tech- niques, positron emission tomography (PET), and electrophysiological measurements. I will describe thenatureof eachmeasurementorimagingsignal andthe functional readout from the biology, as well as what they areused for and how they areapplied. I discuss the strengths and weaknesses of the different techniques, as one modality rarely has the capa- bility of answering all of the relevant questions. I discuss principally the application of the neuroimagingmodalitiescommontoanimalsandhumans,focusinginparticularonnonin- vasivetechniques. Ibeginthechapterwithanoverview ofthebiologicaldomainsoverseenbythedifferent imaging techniques before describing each modality in turn. I then discuss the advantages and disadvantages ofthe modalities. 1.1. Windows to the Brain There is no single neuroimaging modality that can offer the complete picture of brain structureorfunctionthatisneededfordrugdevelopmentandassessingdiseaseortreatment effects. Each modality has its own sensitivity profile that may offer a piece of information directlylinkedtotheactionofapharmacologicalagent,e.g.receptorbinding,ormorelikely anindirectmarkerofpharmacologicalordiseaseaction,e.g.localhemodynamicactivity.Itis necessarytoappreciatethebasisoftheneuroimagingsignalsandtheirlimitationsinorderto interpret them. This is particularly important in the process of decision making in drug discovery and development. Over- or misinterpretation of neuroimaging data could lead to the expensive failure of new compounds late in the drug development process, which could havebeen anticipated earlier. Underuseofvaluable informationcontainedin neuroi- maging data might leadto promising compounds being abandonedunnecessarily. The neuroimaging modalities applied in humans and animals may be categorized as offering either a structural or a functional readout. While often a useful distinction, the boundary between structure and function may be blurred, in particular at the micro scale andwherefunctionandstructureareintimatelylinked.Anelegantexampleliesinthechar- acterization of the diffusion and distribution of water in the human brain. Diffusion-based MRimagingisregardedasofferingstructuralinformation.However,inthecontextofstroke, where there is a redistribution of intra- and extracellular water, structural information and functional viability of tissue become closely linked. Furthermore, the diffusion of water in the brain has been demonstrated to change over the comparatively short timescales 3 1.0. INTRODUCTION associatedwithstructuralplasticityduringlearning.1Inmoregeneralterms,thefunctionof aneuronmightbesaidtobedefinedbyitsconnections,localandremote,tootherneurons, suchconnectionsbeingidentifiednoninvasivelybydiffusiontractographytechniquesinwhite matter. 1.1.1. Brain Structure Alterationsinbrainstructurewithdiseaseortreatmentrangefromthescaleofgrossstruc- ture, including regional tissue volumes, through to micro and molecular structure, such as receptordistributions.Techniquessensitivetoalterationsinbrainstructuremostcommonly offer a long-term marker of disease progression or modification with treatment. This is because macro-scale alterations in structure, e.g. brain atrophy, normally develop over long periods, typically years in humans. Structuralmagneticresonanceimaging(MRI)techniquesareparticularlysuitedtoexam- ining long-term structural alterations because of their noninvasiveness and therefore their ability to follow longitudinal changes in a research cohort. Traditional MRI structural markers,suchasT1andT2relaxationtimeconstants,icanrevealfatandwaterdistribution (aswellasotherphysicochemicaldifferences),thusdistinguishinggraymatter,whitematter, and cerebrospinal fluid, while developments in MR image contrast, such as magnetization transfer, can offer information on pools of liquid and macromolecular water with potential uses in examining demyelinating conditions.2 Microscopy isinvasiveand so only available post mortem, but is useful for characterizing altered brain microstructure. Microstructural alterations at the scale of molecular receptors can be probed with radiotracers using PET, looking for example, at alterations in spatial distributions of specific receptors. Such molecular-levelalterations areoften interpretedas having directfunctional consequences. 1.1.2. Brain Function Brainfunctionischaracterizedbymanydifferentactivitiesofthebrainaccessibletoneuro- imagingtechniques.Functionalprocessesarenormallythoseoccurringovershorttimescales fromsmallfractionsofasecondtominutes.Theysupportorareassociatedwithinformation processingorthetransmissionandreceptionofsignalsinthebrain.Inanimals,itispossible toplaceelectrodeswithinbraintissueandtorecordindividualcellularpotentialsortheelec- tricalactivityofgroupsofcells.Inhumans,thisisonlypossibleinrarecircumstanceswhen electrodesareimplantedforreasonsofsurgicaldiagnosis,forexampletoidentifyseizurefoci or brain stimulation. More commonly, we rely on indirect or ensemble measures of brain function, which may either be focused on specific portions of brain tissue or distributed acrossthe brain tostudythe brain at the systems level. The brain’s neuronal activity is linked orcoupled to its blood supply,3,4 allowing altered hemodynamics and therefore regional blood oxygenation to be used as a marker of altered function. The growth of functional neuroimaging studies in humans in the last 20years iPleaserefertoSection4.2inthischapterforfurtherdetailsregardingrelaxationtimeconstantsinMRI;and toBrowninChapter2,MagneticResonanceImagingasaToolforModelingDrugTreatmentofCNS Disorders:StrengthsandWeaknesses;andNovakandEinsteininChapter4,StructuralMagneticResonance ImagingasaBiomarkerfortheDiagnosis,Progression,andTreatmentofAlzheimerDisease,inthisvolume foracomprehensivediscussionofthepropertiesofMRI. 4 1. NEUROIMAGINGMODALITIES hasexploitedthesephenomena,firstwithPETandmorerecentlyandonalargerscalewith functionalmagneticresonanceimaging(fMRI).Opticalimagingtechniques,includinginva- sive cortical infrared imaging and noninvasive near-infrared spectroscopy (NIRS) rely on changesincerebralblood oxygenation. Acontinuousenergysupplytobraintissueisessentialformaintainingionconcentration gradients and therefore electrical potentials. Some neuroimaging techniques have been developed that are sensitive to alterations in cerebral metabolism and, in particular, the biochemical species involved in energy supply. PET can be made sensitive to oxygen or glucose metabolism. fMRI techniques are also emerging that are able to quantify cerebral oxygen consumption. Magnetic resonance spectroscopy (MRS) can measure chemical concentrationswithinbraintissueandthereforemonitortheenergystatusoftissuethrough speciessuchashigh-energyphosphates.Withtheappropriateuseoftracersdetectablewith MRS,rateconstantsandchemicalfluxescanbeestimatedtoquantifycerebralmetabolism.5 Specific molecules engaged in signaling or processes associated with synaptic transmis- sioncanbestudiedusingPETandMRS.WiththedevelopmentofappropriatePETligands, specific receptor activity and the distribution of receptors can now be assessed. Receptor activity has the advantage of being directly associable with the action of pharmacological agents in the brain, whereas MRS is able to measure the bulk concentration of the more common neurotransmitters and their modulation with disease and pharmacological inter- vention.Itmustbeborneinmind,however,thattherelationshipbetweenneurotransmitter concentration and brain function may be a complex one depending on the availability or otherwiseof the neurotransmitter. Noninvasive measures of electrophysiological activity can be made from the scalp by recordingelectricalpotentials(i.e.electroencephalography;EEG)orthetinymagneticfields associatedwithneuronalactivity(i.e.magnetoencephalography;MEG).Inordertobedetect- able at a distance of centimeters from their source, these signals necessarily arise from the coordinated activity of populations of neurons. 2.0. RADIOTRACER TECHNIQUES Radiotracer techniques use radionuclides as probes to quantify physiological processes, e.g. cerebral blood flow, or to label biochemical pathways or specific molecules. The probe isspatiallylocalizedbydetectingtheemittedradioactivity,whilevariationsinlocalradioac- tivityovertimecanbeusedtoidentifyrateconstantsandphysiologicalfluxes.Radiotracers are sensitive and can be adapted to different uses, including marking substrates to investi- gate biochemical processes, labeling a drug target, or labeling the drug itself. Molecular tracersaretheonlywaytomeasurereceptorsandtheirfunctionbecauseoftheirlowconcen- trations(nano- andpicomolar range). 2.1. Single-Photon Emission Computed Tomography Single-photonemission computed tomography(SPECT)detectstracer molecules labeled with gamma-emitting radioisotopes. Typically it uses an array of two or three gamma cameras that rotate around the subject and is increasingly combined with a computed 5 2.0. RADIOTRACERTECHNIQUES tomography(CT)systemthatoffersimprovedspatialresolutionandregistrationwithstruc- tural images. SPECT is more widely available than PET, although the range of tracers for SPECTismorelimited. SPECTiscommonlyusedformeasuringcerebralperfusion(regionalcerebralbloodflow; rCBF)using,forexample,99mTc-or123I-labeledradiopharmaceuticals.Receptorimagingwith SPECTislesscommonbutisusedinresearchwithagentsavailablefordopamineandsero- tonin transporter imaging (123I-2b-carbomethoxy-3b-(4-iodophenyl)tropane; 123I-b-CIT) and dopaminergicD receptors(123I-IBZM).6 2 rCBFcanbemeasuredusingfreelydiffusibletracerssuchas133Xe.Thecerebraltransitof thetracerismeasuredand,whencombinedwithanestimateoftheinputfunction,amathe- maticalmodelcanbeusedtoestimatecerebralbloodflow.7Morecommonistheuseofaso- called static tracer such as 99mTc-exametazime. Such tracers pass through the bloodebrain barrier and arethen retained by the brain for several hours.The accumulation of the tracer overtimeisproportionaltotherCBF,althoughabsolutequantitationisdifficultasaccumu- lation also depends on the mechanismof retentionin the brain. 2.2. Positron Emission Tomography PETusesradiotracerslabeledwithpositron-emittingradioisotopes.Ithasahighersignal- to-noise ratio and better spatial resolution than SPECT. PET has an exquisite sensitivity allowing detection of tracers in the nano- to picomolar range. This allows the biochemical system to be probed without significant pharmacological effect, as well as with significant pharmacological effect if advantageous. PET has been used to map receptor systems, measure receptor occupancy by pharmacological agents, assay enzyme activity, measure cerebral oxygen and glucose metabolism and cerebral blood flow, and therefore to map task-specific functions in the human brain, including cognitive, drug-stimulated, motor, and sensory activity. As for SPECT, PET is normally combined in the same machine with CT to allow the collection of combined structural and functional information. In general, a cyclotron is needed to generate a wide range of positron-emitting radionu- clides. These radionuclides are used to label compounds of interest, which are normally introduced intothe subject intravenously. Decay of theradionuclide resultsin theemission ofapositronandaneutrino.Theneutrinoisnotdetected,whilethepositroninteractswith electronsinthetissuefinallybeingannihilatedalongwithoneoftheelectronsandliberating twogammarays,eachwithanenergyof511keV.Thegammaraysareemittedapproximately inoppositedirectionsandaredetectedbyringsofdetectorsarrangedaroundthesubject.The solid-state detectors look for coincident events in which gamma rays are detected almost simultaneouslyonoppositesidesofthehead.Thelinealongwhichthepositronannihilation tookplacecanthereforebereconstructed.Bycombiningdatafromallofthedifferentangles, imagescanbereconstructedthatreflecttheconcentrationofthepositron-emittingradionu- clide. The higher signal-to-noise ratio of PETwith respect to SPECTarises from the lack of a requirement for collimation at the detectors, reduced attenuation of the higher-energy gamma rays, and the use of complete rings of detectors rather than a moving camera. It is possible to construct kinetic models of the PET tracers that, when combined with a time- series of PET images, can reveal the rates of biological reactions in which the tracers are engaged. 6 1. NEUROIMAGINGMODALITIES The wide range of biological measurements that can be made with PETis only possible becausesomeofthemostcommonbiologicalelementscanformpositron-emittingisotopes, including carbon (11C), nitrogen (13N), and oxygen (15O). Furthermore, in many molecules, fluorine-18 (18F) can be used as an analog of hydrogen. Once the appropriate isotope is made, it must be incorporated into an organic molecule that has the required biological activity forinvestigatingthebiologicalprocessofinterest.Inow describespecificexamples of physiological, metabolic, and receptor-based PET measurements. 2.2.1. Cerebral Metabolism and Blood Flow The cerebral rate of metabolic oxygen consumption can be measured using 15O-labeled O 8. However, the most commonly used PETradiotracer is 18F-fluorodeoxyglucose (FDG), 2 which is transported across the bloodebrain barrier. In cells it is converted to FDG-6- phosphate, which is then trapped inside the cell. Knowing the input function of the tracer from arterial blood and with a time-series of FDG measurements, the cerebral metabolic rate for glucose (CMRGl) can be mapped. This has found application in the assessment and identification of brain tumors, which often show increases in aerobic and anaerobic glycolysis.9 FDG-PETcan be used to characterize different types of neurodegeneration that may resultin dementia, including Alzheimerdisease (AD).ii InAD,reducedCMRGlisseenintheparietalandtemporallobes.10Thesepatternsoflow metabolismcanbeusedtohelpdistinguishbetweenADandotherdementiassuchasfron- totemporal dementia. EarlystudiesoffunctionalmappingofthehumanbrainwereperformedusingFDG-PET withsubjectsperformingtasksanddemonstratedincreasedCMRGl.11,12Thislaidthefoun- dationsforthevastlyexpandedfieldofhumanfunctionalbrainmapping.Inthedomainof PET, FDG studies of brain activation have been largely superseded by cerebral blood flow measurements because of the opportunities to perform repeated measurements in the same individual. Cerebralbloodflowismeasuredusingadiffusibletracer,e.g.15O-H O,whichistherefore 2 abletopassthroughthebloodebrainbarrier.Thequantificationofbloodflowreliesonentry ofthetracerintothebrain,whichdependsontheflowratetothetissueandisnotlimitedby thediffusionofthetracerintothetissue.Acompartmentalmodelcanbeapplieddescribing thekineticsof15O-H Oinwhichbloodflowisoneoftheparameters.13Forstudiesaimingto 2 map brain functions, relative cerebral blood flow (CBF) measurements are often made, relying,forshortscans,onthenear-linearrelationshipbetweenbloodflowandthedistribu- tion of 15O-H O.14 2 2.2.2. Receptor Studies Anappropriatebiologicalprobelabeledwithapositron-emittingradioisotopecanbeused to isolate a receptor system of interest. Thereis an increasing array of tracers available. For use in drug discovery and development, it is important to develop tracers for each new iiPleaserefertoSchmidtetal.inChapter5,PositronEmissionTomographyinAlzheimerDisease:Diagnosis andUseasBiomarkerEndpoints,ofthisvolumeforfurtherdetailsregardingtheuseofPETimagingin aneurodegenerativedisorder. 7 3.0. ELECTROPHYSIOLOGICALTECHNIQUES proteindrugtargetalongsidethedevelopmentofthedrugitself.ThisenablesPETtobeused in studying boththe receptor system and theinfluence of the drug on it. UseofaPETtracerinpreclinicalphasesofdrugdevelopmentcanhelpselectleadmole- culesbyidentifyingthosewiththehighesttargetengagement.Targetengagementorbinding to a receptor can be used in preclinical testing to identify whether a drug is likely to have a pharmacological effect. Later on in clinical development, the proof-of-concept is demon- stratedwhen a clinical endpoint can be associated with targetengagement.15 Using aradiolabeled probe,thedensity of receptors andthe binding affinity with which theligandbindstothereceptorofinterestcantypicallybeestablished.16,17Receptordensity andbindingaffinityareoftenexpressedasaratioknownasthebindingpotential.PETcanbe used to identify neurotransmitters or receptors at different sites that play different roles in brain function. These include the presynaptic neuron, the postsynaptic neuron, and those with modulatory functions on the cell membrane or engaged in neuronal metabolism.18 Forexample,forthedopaminergicsystem,D receptorsatthepostsynapticlevelcanbeiden- 2 tified by 11C-raclopride binding. Other examples include 11C-diprenorphine for opioid receptorimaging19and11C-flumazenilforimagingg-aminobutyricacid(GABA) /benzodi- A azepinereceptors. PETcan be particularly useful in evaluating the receptor occupancy of a drug through measuring how the drug inhibits the binding of the radioligand for the receptor system of interest at different doses.20 The drug competes with the radioligand at the same site for binding. Receptoroccupancy can be compared with the pharmacological effect of the drug andthiscanbeparticularlyusefulinestablishingdoserangesorproof-of-conceptofadrug’s action. An example of the usefulness of receptor occupancy studies was provided by the assessment of schizophrenia treatment with antipsychotics and the demonstration of D 2 receptorblockadeusing11C-raclopridebinding,21whichshowedcorrelationswithpharma- cological effectsof the drugs. 3.0. ELECTROPHYSIOLOGICAL TECHNIQUES Noninvasiveelectrophysiologicaltechniquesinvolvethemeasurementofelectricalpoten- tialsatthescalp(i.e.EEG)orweakmagneticfieldsatadistanceofafewmillimetersfromthe scalp(i.e.MEG).Comparedtofunctionalneuroimagingtechniquesbasedonhemodynamic changes, electrophysiological techniques offer a more direct measurement of neuronal activitywithacorrespondinglybettertemporalresolutionoverthetimescaleofmilliseconds. These noninvasive techniques are sensitive to current flow, generating magnetic field changes (MEG) or the differences in potential associated with such currents (EEG). These changes are not generally those associated with the fast transmission of action potentials along the axon but are rather associated with longer-lasting postsynaptic currents. To generate a measurable signal at the scalp, the currents associated with a large number of neurons must combine coherently. This situation is most commonly reached in pyramidal neurons in cortical layers. For EEG, greatest sensitivity arises with pyramidal neurons perpendiculartothescalpwhereasforMEGgreatestsensitivityisachievedwithpyramidal neurons parallel to the scalp. This difference arises because of the orthogonal nature of the electric and magnetic fields associated with current flow. Coherent postsynaptic currents 8 1. NEUROIMAGINGMODALITIES areneededacrossasignificantareaofcorticalsurfacetoobtaindetectablesignals.Itisthere- forelikelythatonlyasmallproportionofthetotalbrainactivityisdetectablewiththesetech- niques. Indeed, intracranial EEG yields a richer signal than scalp EEG, showing that much brain activity isnot visible with conventional EEG or MEG. For both of these techniques, signals become weaker with increasing distance from the source.Signals from deep structuresarethereforeattenuated.However,despitethelimited window provided to the total activity of the brain, these methods can be used to monitor neuraldynamicswithsomesensitivity,forexamplethroughchangesinthespatialdistribu- tionandfrequencycharacteristicsofoscillatoryelectricalactivity.Thisutilitymaysometimes be independent of a precise knowledge of the neural sources and direct functional signifi- canceoftheobservedchanges,althoughagooddealofempiricalevidencehasbeencollected ontheassociationofspecificchangesinEEGsignalsanddiseaseordrugeffectsthathasbeen related to individual brainsystems. Inthefieldofdrugdevelopment,thedemonstrationofcentralelectrophysiologicalalter- ationswithdrugdosingprovidesgoodevidenceofcentralpenetration,i.e.entryofthedrug into thebrainand a resultant effectonneuronalactivity there. 3.1. Electroencephalography EEGbiomarkersshowhighsensitivitytochangesinthestateofthebrainandhavebeen used to assess pharmacological effects in volunteers and patients for a number of years.22 Thereisalargerangeofpotentialmeasurementsthatcanbemade,includingrestingongoing rhythms or oscillations and transient potentials evoked by a specific task (event-related potentials; ERPs). The range of tasks is wide and can be chosen in order to mark a specific disease or drug effect. They range from basic sensory tasks, such as laser sensory evoked potentials for the study of thenociceptive system,23to cognitive tasks for probingattention and workingmemory.24 Scalp EEG signals lie in the range of millivolts. EEG measurements are normally made using at least 21 scalp electrodes, with an increase in the number of recording channels offering improved spatial resolution of voltage changes across the scalp: either 64 or 128 channelsarecommonlyrecorded with modern EEG systems. However, the resolution with which the cortical current distribution or putative dipole source can be located is limited byspatialblurringofthecurrentdistributionimposedbythepoorconductivityoftheskull. However, if one is interested primarily in electrophysiological changes associated with disease or drug administration then the demonstration of characteristic differences in EEG may besufficient without accuratesourcelocalization. Ongoing rhythms, observed without reference to a specific time-locked stimulus, are conventionally considered in frequency bands, although full spectral analysis is always possible. These bands are typically defined as delta (1.5e6Hz), theta (6e8.5Hz), alpha (8.5e12.5Hz), beta (12.5e30Hz), and gamma (30Hz and above), with signals often recorded in the 0.5e70Hz range. In general, there is no specific physiological meaning associated with belonging to a specific band and the boundaries between them are somewhat arbitrary. However, the bands do provide a shorthand to describe frequency ranges. 9 3.0. ELECTROPHYSIOLOGICALTECHNIQUES OccipitalalphawavesaretypicallythestrongestEEGphenomenaandoccurintherelaxed awakesubject.Theyarereducedwhentheeyesareopen.Thalamicnuclei,includingthepul- vinar,influence thegeneration ofalpharhythmsinthe visual cortex. Howeverintracortical connections are responsible for propagation over the cortex.25 Rhythms of a similar frequency range are also observed in the somatosensory cortex; known as rolandic mu rhythms,theyarealsothoughttoarisefromthalamocorticalinteractions.Higher-frequency rhythms such as beta and gamma may involve an influence of the thalamus but evidence from cortical recordings suggests that they may be primarily generated in the cortex and depend on intrinsic oscillatory properties of cortical circuits.25 Commonly, event-related modulation of the amplitude of beta and gamma oscillations is seen, reflecting a change in synchronous activity of underlying neurons known as event-related desynchronization (ERD;adecrease)orsynchronization(ERS;anincrease).EEGsignalsalsoprovideamarker of different levels of arousal or consciousness. In particular, EEG changes during sleep can offeramarkerofdiseaseordrugeffects,therebeingtwomainEEGsleepphenomena:spin- dles or waves between 7 and 14Hz at the start of sleep and delta waves associated with deepersleep stages. EEGmarkershaveprovedtobesensitiveacrossawiderangeofdiseasesandpharmaco- logical treatments, potentially offering useful information in making go/no-go decisions during drug development using preclinical testing and, in the later stages, in smallcohorts of human volunteers.26 In particular, EEG appears to be sensitive in affective disorders, including depression and its pharmacological treatment, especially in the examination of sleepstates.27Inschizophrenia,alterationsinongoingrhythms(EEGfrequencybandpower) and specific sensory evoked responses, e.g. auditory evoked potential (AEP), have been observed leading to suggestions of impaired sensory processing or sensory gating.26,28 This can be revealed in the human P50eN100 component of the AEP, arising from primary and associationauditorycorticesandgeneratedbyunpredictablestimuli.29Theseresponsesoffer an opportunity for translation between animals (rodents) and humans, as models of P50 sensory gating deficits are validated in rats and mice. There is good evidence of translat- ability of AEP endpoints from work with nicotinic acetylcholine receptor agonists and measuresof AEP sensory gating.30 Neurobiological changes associated with AD can go beyond the alteration of cognitive function. Changes can be observed in EEG activity, for example spectral shifts, and alter- ations in evoked responses (ERPs). There are also alterations in the EEG during sleep.31 AgeneralslowingofEEGrhythmsisobservedinAD,includingincreasesindeltaandtheta power and decreases in alpha and beta power.32 There is an association between these changes and the development of cognitive impairment.33 Event-related cognitive markers intheEEG,e.g.P300associatedwithattentionalandperceptualprocesses,arealsopotential markersof AD.34 EEG may beparticularly useful in AD as abnormalities oftenprecede the development ofdefinitive clinical features.26 EEGhasimportantpracticaladvantagesovermostoftheotherneuroimagingtechniques describedinthischapter.Itischeaptoconductanddoesnotrequireexpensiveandperma- nentlyfixedscanningequipment.ItisthereforeeasytocombineEEGwithsleepstudies,inan ambulatorycontext,andattheearlystagesofdrugtestinginhumansubjects.Itisnoninva- sive and extended periods of recording arepossible.It also offers conveniently translatable 10 1. NEUROIMAGINGMODALITIES methods from animals to humans, facilitating the comparison of drug-induced alterations observed in preclinical and clinical testing. However, it should be remembered that the frequency characteristics of specific oscillatory phenomena may differ between small mammals and humans, largely because of frequency differences and latency shifts caused by substantial differences in brain size. 3.2. Magnetoencephalography MEG provides signals of brain function qualitatively similar to EEG as the techniques share a common electrophysiological basis. MEG uses sensitive magnetic field detectors known as superconducting quantum interference devices (SQUIDs). The principal advan- tageofMEGoverEEGliesintheneartransparencyoftheskullandscalptothetinymagnetic field perturbations caused by postsynaptic currents. EEG signals tend to suffer a greater spatialsmearing over the scalp. Todate,MEGhasbeenlesswidelyappliedthanEEGinthestudyofdiseaseandindrug developmentbecauseofthecomparativerarityofMEGrecordingsystemscomparedtoEEG. However,itisprovingavaluabletoolinbasicandclinicalneuroscience.35MEGsystemsare considerably more expensive than EEG both at time of purchase and in ongoing running costs. A MEG installation is fixed, requiring the system to be sited in a heavily shielded room, and regular refills of liquid helium are needed to maintain the SQUIDs at supercon- ductingtemperatures. AtypicalMEGsystemmayhaveapproximately300SQUIDdetectorsand,similartoEEG, canbeusedtodetectongoingrhythmsaswellasevokedfields.36MEGiscomplementaryto EEG, being sensitive to orthogonal components of the postsynaptic current flow, namely those tangential components.37 Using MEG it is possible to reliably detect high-frequency, e.g. gamma band, components of electrophysiological activity that are comparable to those measuredinanimalmodelsusinginvasivetechniques.38Thesehaverecentlyshownanasso- ciationwithcorticalGABAconcentrationinthevisualcortex,demonstratingalinkbetween neurochemistry and neurophysiological activity measured in humans using a noninvasive multimodal imaging approach.39 WhileMEGhasnotbeenusedaswidelyasEEGtoevaluatepharmacologicalaction,there isgrowinginterestinthisarea.40Inparticular,theeffectofmodulatingtheGABAsystemon corticalrhythmshasbeeninvestigated,withtheeffectsofbenzodiazepinesonincreasingthe power and decreasing the frequency of beta rhythms over the sensorimotor cortex being demonstrated.41 4.0. MAGNETIC RESONANCE TECHNIQUES The phenomenon of nuclear magnetic resonance underlies the wide range of MRI and magnetic resonance spectroscopy (MRS) techniques that have been developed. While MRI can offer structural and functional information on the brain at good spatial resolution, the advantageof MRS lies inits ability to probethe brain’s neurochemical environment.