Neuroimaging in Child and Adolescent Psychiatric Disorders

412 Arch Dis Child 2000;82:412–419

CURRENT TOPIC

Neuroimaging in child and adolescent psychiatric disorders

Paramala Janardhanan Santosh

Abstract detailed account of the physical basis of the Neuroimaging in child psychiatry is a various neuroimaging techniques. rapidly developing field and the number Brain imaging may be performed for a vari- of diVerent techniques being used is ety of reasons: to address needs in clinical increasing rapidly. This review describes research, patient management, or diagnosis. In the current status of neuroimaging in clinical practice, structural neuroimaging is childhood psychopathology and discusses indicated when one suspects intracranial limitations of the various studies. As yet, pathology—if the patient exhibits focal neuro- no specific and consistent abnormality logical signs on examination, or when there is a has been detected in childhood psychiat- history of significant head trauma (that is, with ric disorders. Obsessive compulsive dis- extended loss of consciousness, enduring neu- order has shown the most consistent rologic sequelae of any kind, or when there is a findings so far, with orbitofrontal cortex close temporal relation to the onset of psycho- and the caudate nucleus being impli- pathology), refractory epilepsy, childhood cated. Better understanding of the corti- onset psychoses, and when the disorder is costriatal neural networks will shed more refractory to an extensive array of conventional light on the neurodevelopmental disor- treatments. Apart from being indicated in ders of childhood. complicated epilepsy, functional neuroimaging (Arch Dis Child 2000;82:412–419) is currently only a research tool in psychiatry. Functional abnormalities are often picked up Keywords: neuroimaging; psychiatry; psychopathology in subjects with negative structural scans. Table 1 compares the diVerent techniques. Although structures may show group diVer- In the study of the human brain, a picture is ences, they need not show the same degree of worth a thousand words. In the past 20 years, diVerence in, for example, electron density technological advances have provided and (CT), phosphorylation (MRS), or oxygenation refined a variety of neuroimaging methods of blood (fMRI). Thus the fMRI cerebral acti- based on diVerent physical phenomena. There vation may not equate to the PET activation, is now considerable literature on the applica- and both may diVer from the anatomy depicted tion of neuroimaging to children with neurode- by visual inspection or tissue typing. velopmental disorders. At this stage in the Neurophysiological imaging refers to the use of field’s development, replicable findings are suf- cerebral blood flow radiotracers (15O labelled ficient to permit an appraisal of the progress water, 133Xe) or metabolic radiotracers (18F- made so far, while attempting to outline poten- fluorodeoxyglucose) to spatially resolve the tial pitfalls and diYculties faced. haemodynamic and metabolic correlates of Structural imaging measures anatomic neural circuit activity. Neuroreceptor imaging structure (radiographs, computed tomography refers to the use of PET or SPECT radionu- (CT), magnetic resonance (MR) imaging) clides bound to ligands possessing a high and while functional imaging measures the intrinsic selective aYnity for neurotransmitter receptors physical properties of tissues (for example, or transporters. Tracer kinetic models are used metabolism or blood flow) and changes that to convert local positron annihilations into occur in disease (magnetic resonance spectros- estimates of receptor or transporter density, copy (MRS), functional MR imaging (fMRI), distribution, or occupancy. Neurochemical im- positron emission tomography (PET), single aging refers to the use of PET or SPECT radionuclides bound to precursors (for exam- Institute of Psychiatry, photon emission CT (SPECT), magnetic Denmark Hill, London encephalography (MEG)). The nature of the ple, tryptophan, dihydroxyphenylalanine SE5 8AF, UK connection between the physical property (DOPA)) of enzymatic reactions that support P J Santosh measured (for example, radioactivity in some neurotransmitter synthesis.

Correspondence to: areas) and the underlying physiological phe- Dr Santosh nomenon (for example, metabolism) estab- Normal neurodevelopmental changes email: lishes and limits the utility of diVerent that inﬂuence paediatric neuroimaging [email protected] methods, and the technical constraints of each Knowledge of the normal patterns of brain Accepted 26 January 2000 technique. Anderson and Gore1 giveavery development in the clinically relevant ages Neuroimaging in psychiatric disorders 413

Table 1 Comparison of diVerent neuroimaging techniques

Neuroimaging technique Mechanism Advantages Disadvantages

Computed tomography Collimated beams of x rays are rotated Excellent images of skull, sulci, and Artefacts often arise in regions containing around the head and pass through the brain, ventricles. Volumetric and dynamic images very dense structures, especially in posterior losing energy in proportion to the density of can be obtained if spiral CT is used. fossa and areas close to bony interfaces. the various tissues (grey matter, white Patient is exposed to ionising radiation. matter, and CSF). Only transverse slices obtained. Magnetic resonance Brief radiofrequency pulses activate the Primary method of choice for brain Pacemakers, shell injury, plates, screws, or imaging inherent distribution of hydrogen atoms in imaging. Superior images with high spatial metallic implants are contraindications. the brain, following which the hydrogen resolution. Safe, no exposure to radiation. Subjects have to remain relatively still. atoms align themselves in the strong Repeated scans possible even in very young Scanner noise is quite loud. Constrained magnetic field generated by the children. Arbitrary planes, environment within receiver coil. superconducting magnet around the head. transverse/coronal/ sagittal images can be The diVerent realignment times after the generated. burst of radiofrequency perturbation, are used to delineate the diVerent tissues of the brain. Functional magnetic Changes in blood flow (characterised by Can map brain’s functional responses to Image analysis techniques need to improve resonance imaging altered levels of oxygen or oxygenated specific stimuli. Non-invasive and safe. Will in children. Artefacts near skull base limit its blood) establish a new equilibrium of the help us learn more about neurophysiology use. Constrained environment within oxygen dependent magnetic properties of in both disease and health. Reference receiver coil. haemoglobin, detected through high field anatomic images are simultaneously magnets. acquired with the functional data. Positron emission Based on the detection of annihilation Exact quantification of cerebral blood flow Radionuclide scanning technique. Cannot tomography photons arising from the decay of injected and metabolism. Whole head imaging is be used repeatedly, or in pregnancy. radiotracer. Arrays of scintillation detectors more reliable. Neuroreceptor concentration Cyclotron is necessary to provide the are used as an electronic collimation system, and aYnity can be measured. radiotracers with very short half life. which transforms the photons into visible Constrained by need to inject isotope for light; this is ultimately filtered and each new task. Anatomic data need to be reconstructed to form the image showing obtained separately. either the blood flow or glucose metabolism, receptor occupancy or neurochemical binding in the brain (depending on the tracer). Single photon emission Similar to PET. Measures changes in blood Available in most departments of nuclear Radionuclide scanning technique. Cannot tomography flow, receptor activity using appropriate medicine. Large numbers of radiotracers be used repeatedly, or in pregnancy. radiotracers. Detectors specialised for available. Cost eVective. Absolute quantification is not possible, and localising photons (ã rays) emitted by bilateral symmetrical reduction is diYcult to positron annihilation are used. The data are recognise. processed to create images of slices of brain in the transaxial, coronal, and sagittal planes. Magnetic resonance This exploits the slight diVerences in Provides a direct investigation of Long procedure if one is interested in spectroscopy resonant frequency of protons (usually 1H phosphorhylated intermediate metabolites quantification at molecular level. Only and 31P) bound to diVerent cell associated and neurotransmitters such as GABA and limited substrates are measurable. structures. It characterises the molecular glutamate. state of both bound and free tissue water and the chemical microenvironment of cells and provides a profile of the status of intermediary metabolism within a selected tissue volume. Magnetic Uses specialised superconducting detectors Can help to locate and measure the strength Bias favours only some neurones and not all encephalography and sensory coils to measure magnetic fields of electrical impulses from the brain. Safe in a particular field and hence may not be that surround the currents which give rise to procedure, no exposure to radioactive accurate. Structural scans necessary EEGs and ERPs. The magnetic fields reflect compounds. separately, to transpose findings onto a brain currents induced within the dendrites of map. neurones oriented parallel to the sulci.

from 4 to 18 years is necessary to interpret the profiles of developmental neuropsychiatric dis- subtle brain imaging findings reported in the orders could arise from complex interactions literature. Giedd and colleagues report the best between the gender specific diVerences in brain data to date, from the ongoing NIMH paediat- development and the child’s environment. For ric neuroimaging project.2–4 Cortical grey mat- example, the lower incidence of attention ter decreases with age, while white matter and deficit–hyperactivity disorder (ADHD) in girls CSF volumes (to a lesser extent), gradually might be related to having a relatively larger increase with age. Lateral ventricles, corpus caudate nucleus than boys. Also, the male only adolescent increase in lateral ventricular callosum, basal ganglia, amygdala, and the hip- 34 pocampus also increase in size with age. volume is intriguing in light of consistent findings of enlarged ventricular volume in ado- Although total brain volume appears to ap- lescent onset schizophrenia, which is more pre- proach adult size by school age, this belies an dominant in boys. active and sex specific dynamic balance achieved between growth and regression of Current status of neuroimaging in certain brain structures throughout childhood childhood psychopathology and adolescence. Similarly, brain metabolism Despite accumulating neuroimaging research (as shown by PET studies) rapidly increases in children, findings are often inconsistent and during the first years of life, peaks during not replicable across centres. Many problems childhood, and then declines to adult levels exist in the application of neuroimaging to during adolescence. children and these are detailed later in this Speculatively, the observed gender diVer- article. The important findings in childhood ences in incidence, age of onset, and symptom disorders are summarised below. 414 Santosh

Language impairment and reading disability— Autism has been associated with increased Children with specific language impairment regional brain volume, with the occipital, pari- have been reported to have a significantly etal, and temporal lobes (in decreasing fre- smaller left pars triangularis, with rightward quency) being enlarged.11 12 Reports that cer- asymmetry of language structures.5 Similarly, ebellar vermis lobules VI and VII are in those with specific reading disability, the specifically aVected in autism (either being degree of left cerebral asymmetry has been hypoplastic or hyperplastic),13 remains contro- found to correlate with both reading skills and versial because few others have been able to skills in phonemic analysis of spoken language.6 replicate it. Recently, the ability to attribute ADHD—Although methodological prob- independent mental states to self and others lems remain, there is increasing agreement on (theory of mind), has been associated with the the role of the prefrontal-striatal-thalamo- activation of the left medial frontal gyrus cortical circuit in ADHD, with a preponder- (Broadmann’s area 8) and the posterior cingu- 14 ance of the evidence suggesting that the right late cortex, areas which are not activated in 15 sided circuit is primary, at least at the level of Asperger’s syndrome during similar tasks. the basal ganglia. Recent reviews in this field Using the recently developed voxel based include those by Castellanos,7 Filipek,8 and whole brain analysis of structural MRI, Abell et 16 Overmeyer and Taylor.9 Neuroimaging in al report structural abnormalities in a distrib- ADHD has advanced greatly to unpick the uted system centred on the amygdala, in neural subsystems involved in its pathophysiol- autism. They report decreases of grey matter in ogy. Hyperactive adolescents have recently anterior parts of this system (right paracingu- been shown to have diVerent functional brain late sulcus, left inferior frontal gyrus) and abnormalities when tested on two tasks, a increases in posterior parts (amygdala/peri- motor inhibition task, and a motor timing amygdaloid cortex, middle temporal gyrus, inferior temporal gyrus) and in the regions of task.10 Adolescents with ADHD in comparison the cerebellum. These findings are appealing to matched controls, show lower power of because the structures involved have been pre- response in the right mesial prefrontal cortex viously implicated in social cognition by animal during both tasks, and in the right inferior pre- and histopathological studies. Using neuro- frontal cortex and left caudate during the stop chemical imaging, an increased serotonin syn- task. Figure 1 depicts the findings for the thesis capacity in autistic children (using response inhibition task (stop task) from this 11 10 á- C-methyl-L-tryptophan PET) has been study. reported, and is hypothesised to be related to Autism—Despite the presence of numerous the disruption of the normal developmental brain imaging studies attempting to isolate pattern of brain serotonin synthesis.17 brain regions or pathways specifically impli- Tourette’s syndrome—Increased dopamine cated, the literature remains inconclusive. transporter availability in the caudate in 123 18 19 STOP TASK I-BZM SPECT studies, and fMRI based activation of the orbitofrontal cortex, bilateral premotor regions, and the head of right caudate, and decreased activation of the Controls bilateral globus pallidus and thalamus during tic suppression tasks20 have been reported. Obsessive compulsive disorder (OCD) has been extensively studied and subjects show orbitofrontal and anterior cingulate hypermetabolism at rest. Successful attenuation of OCD symptoms results in the attenuation of the hypermetabolism of the orbitofrontal cortex, caudate ADHD nucleus, and anterior cingulate cortex, irre- spective of whether the treatment used is behaviour therapy or medication. Subjects with OCD show increased activation in bilateral orbitofrontal cortex, right caudate nucleus, and anterior cingulate cortex, during symptom induction procedures, with the degree of induced obsessionality being positively corre- Difference lated with the magnitude of activation within an anterior orbitofrontal locus. Furthermore, individuals with OCD fail to recruit normally the corticostriatal system during cognitive– behavioural activation paradigms, and instead +4 +20 +26 +42 activate the medial temporal lobe system Figure 1 fMRI diVerences between children with ADHD and controls, during a response (involved in conscious information process- inhibition task (stop task). The first two rows show generic brain activation maps ing). Rauch et al21 have provided an in depth (p < 0.003) for nine controls and seven adolescents with ADHD during performance of the stop task. The right side of the image corresponds to the left side of the brain. The last row review on the subject. shows the areas of significant diVerence (ANCOVA) in power of activation between Childhood onset schizophrenia (COS), defined ADHD and non-ADHD adolescents during task performance (p < 0.05). Controls show as schizophrenia with onset by 12 years of age, increased blood oxygenation level dependent (BOLD) response in right mesial frontal cortex, right inferior and middle frontal cortex, and left caudate (courtesy of Dr Katja provides a unique research opportunity to test Rubia PhD, London UK). the neurodevelopmental hypothesis of schizo- Neuroimaging in psychiatric disorders 415

Table 2 Brain regions involved (from neuroimaging data) in childhood psychiatric disorders

Parieto- Corpus Altered Disorder Frontal occipital Temporal Caudate Putamen Limbic callosum Cerebellum laterality ADHD Probable Possible Possible Probable Likely Likely Possible Possible Autism Possible Possible Probable Possible Probable Likely Tourette’s Probable Possible Probable Possible Likely OCD Probable Probable Likely Likely Schizophrenia Possible Probable Possible Likely Likely Possible Mania Possible Depression Possible Possible Dyslexia Possible Probable Probable Possible Possible

phrenia. Children meeting strict criteria for prefrontal dysfunction, which is similar to that COS show smaller cerebral volume and reported in ADHD). This is in keeping with thalamic area, and increased basal ganglia and the recently proposed model for the manage- lateral ventricular volumes,22 the size of the ment of psychopathology in neurodevelopmen- eVect being similar to that observed for adult tal disorders.32 Tourette’s syndrome, OCD, populations. In adolescents with COS, the depression, and ADHD frequently co-occur caudate enlargement appears to be secondary clinically, which may be related in part to com- to exposure to typical neuroleptics.23 Smaller mon elements in their pathophysiologies, volumes of cerebellar vermis (inferior posterior which at the level of the brain organisation may lobe) have also been reported, which is consist- involve particular cortical–subcortical circuits. ent with observations in adult schizophrenia.24 It is possible that putamen dysfunction leads to Patients with COS are reported to have a four- sensorimotor symptoms of Tourette’s syn- fold greater decrease in cortical grey matter drome, ventral caudate nuclear dysfunction volume during adolescence, with a disease leads to obsessions and compulsions, dorsola- specific reduction in the frontal and temporal teral caudate nuclear involvement leads to regions.25 hyperactivity and inattention, and predomi- AVective and depressive disorders—Reviews on nant involvement of the nucleus accumbens aVective disorders implicate the prefrontal cor- results in aVective or anxiety disorders.21 tex (especially orbital), basal ganglia, thalamus, and amygdala, with possible dopamine and serotonin underpinnings.26 27 Decreased frontal Limitations of neuroimaging studies in lobe volume and increased ventricular volume childhood It cannot be emphasised enough that it is very have been reported in children with depressive diYcult to obtain consistency in neuroimaging disorders, with a significant inverse relation findings in children and adolescents. One must between age and frontal lobe volume.28 Tempo- be cautious in generalising findings because the ral horn enlargement and deep white matter existing published literature reflects a publi- hyperintensities, with the absence of normal cation bias towards studies with positive frontal asymmetry have been reported in findings. With an increasing number of centres children and adolescents with bipolar illness. becoming involved in neuroimaging research, These findings in childhood aVective disorders findings being reported are not uniform. The have largely been replicated in adults with probable reasons for this are detailed below. similar disorders. Post-traumatic stress disorder (PTSD)—MRI based hippocampal atrophy, and PET based SUBJECT VARIABLES failure of hippocampal activation during the Gender performance of memory tasks have been Cerebral volume is about 9% larger in males, reported in women with childhood sexual and lateral ventricular volume increases at abuse related PTSD.29 Children and adoles- about twice the rate per year in males, cents with PTSD resulting from child maltreat- occurring mostly after 11 years of age (impor- ment, have recently been shown to have smaller tant when ventricular to brain ratios are calcu- intracranial and cerebral volumes than lated). While boys have larger globus pallidi, matched controls.30 Brain volume positively girls have larger caudate nuclei than boys. correlated with age of onset of PTSD trauma Similarly, while amygdala volumes increase and negatively correlated with duration of sharply in boys (about six to seven times that of abuse.30 This suggests that intense stress in girls), hippocampal volume increases more childhood can lead to long term structural and rapidly in girls (at about three times that of functional changes in the brain. On the other boys). hand, in anorexia nervosa, the observed mor- phological and functional cerebral alterations Handedness (enlarged CSF spaces especially of cortical As symmetry diVerences are often key features sulci), are interpreted to be consequences of in discriminating controls from patients with the anorexic state, which is at least partially disorders such as ADHD,33 dyslexia,34 or reversible with weight gain.31 Tourette’s syndrome,35 it is necessary to control After reviewing the neuroimaging findings in for handedness in paediatric neuroimaging the various disorders (summarised in table 2), studies. it seems possible that some symptoms (irre- spective of diagnosis) have common underlying Body size pathophysiology (for example, inattention in This is a very poor indicator of brain size in mania has been shown to be associated with humans.36 Children have a larger head to height 416 Santosh

ratio than adults, with large interindividual hypotheses can lead to confusion. Chance variations. associations are bound to occur from exhaus- tive analysis of small numbers of subjects, Intelligence receiving a disproportionate emphasis in the Intelligence has been reported to have a statis- literature. tically significant relation to brain size.37–39 Socioeconomic status and education have also IMAGING VARIABLES been reported to relate to brain size, although Image acquisition and anxiety the interdependence with factors such as Many children become anxious during scan- prenatal care, nutrition, and IQ is not clear. ning and become uncooperative, leading to inflated drop out rates and diYculty in unpick- Inherent variability in children ing the anxiety related findings (artefacts) dur- A striking feature of brain morphometric data ing functional neuroimaging. Familiarity and on normal children and adolescents, is the high comfort with the people acquiring the scan, degree of variability of brain structure size, undergoing scanning in the evening when even in well screened healthy cohorts, leading natural sleep is more likely, reading a bedtime to the need for larger sample sizes to detect sig- story or bringing in a favourite blanket or nificant diVerences. Most of the studies to date stuVed toy, and being allowed to stop the pro- do not meet the projected numbers necessary cedure at any time for any reason can all to rule out type II errors. increase the chances of acquiring adequate scans and make the experience more pleasant Developmental age for the child. Cortical and subcortical grey and white matter, and CSF volumes change rapidly during childhood and adolescence, resulting in problems Movement artefacts when children of a wide age range are studied. Movement during scanning produces significant artefacts and needs to be monitored and Cognitive style adjusted for. The advent of new collars to pre- Cognitive strategies being used to solve tasks vent movement will help improve the quality of during functional imaging may be diVerent at the scans. diVerent ages, leading to diVerent activation patterns in subjects. Scanning parameters Thicker slices result in less spatial resolution Ethical issues and greater partial volume eVects, a critical Ethical issues that need to be addressed include consideration for quantifying small but clini- the possibility of overprotection by policy mak- cally pertinent structures such as the caudate ers and institutional review boards—arising nucleus, putamen, or globus pallidus. from the recognition of children’s special vulnerability, without equal recognition of their need for research; assessment of the risk– Image analysis benefit ratio; the diYculty of justifying risk for Analysis of MR images has benefited enor- normal controls; development and use of age mously from advances in computer technology. graded consent; development of child friendly However, the absence of a “gold standard” imaging procedures; and disposition of un- hampers the validation of these techniques and wanted or unexpected knowledge about indi- comparison with results obtained from manual viduals, including the subject’s right not to tracing by expert human raters remains the know and parent’s right not to tell, among best standard. Developmentally correct child other things.40 brain maps are not yet freely available, resulting in the use of adult brain maps—Talaraich space, for analysis. This could result in compu- STUDY VARIABLES terised programs picking up wrongly identified Subject selection As subtle neuroimaging findings have been areas during analysis. Statistical threshold reported in many childhood disorders, it is adjustments for multiple comparisons, and important to have good normative data from a uncertainty regarding the heterogeneity of the control group. Ideally, normative data should condition under study also aVect the reliability be acquired from scans of community recruited of results. Nonetheless, imaging oVers distinct subjects who have been assessed prior to the advantages over non-imaging methods in scan. assessing function and structure.

Sample size/study design Interpretation of data The high variability of brain sizes and the non- The interpretation of data using “subtraction linear pattern of most developmental curves paradigms” has major limitations. It assumes call for large samples and longitudinal study that successive cognitive tasks lead to linear designs in order to adequately characterise cerebral activation, and discounts the current neuroanatomic patterns of development in understanding that the process is clearly more children. complex. Problems of averaging results across groups of subjects and unreliable identiﬁcation Lack of hypothesis driven neuroimaging research of boundaries, structures, sites of activation, It needs to be recognised that investigations and their changes over time further lower reli- using new techniques in the absence of guiding ability. Neuroimaging in psychiatric disorders 417

Measured parameter versus inference namic PET studies using receptor ligands (for The measured changes of physical parameters example, opioid or dopamine receptor ligands) in the various methods (for example, nuclear will help in understanding the site, and the MR signal decay time triggered by changes in neurochemical basis of neuronal activity, lead- the electron structure of iron, in MRI), is often ing to designing innovative pharmacological quite distant from the biological event that strategies. induced the change. This introduces doubt Judicious combination of complementary into the assumption that neuroimaging accu- methodologies (multimodal imaging) is neces- rately measures brain structure or function. It sary to understand the relation between struc- is possible that there is no clear quantitative ture and function, and will at the very least be relation between the biological change and the necessary to explore whether altered regional magnitude of the signal acquired for imaging. metabolism or receptor densities arise due to an underlying change in the volume of that Implication of abnormality structure. The demonstration of abnormality does not Studies on normal children will shed new necessarily indicate that it is of current light on the neurological underpinnings of nor- aetiological significance. Abnormalities in mal cognitive and emotional processes, helping brain structure can result from various early us to understand the deficits in children with experiences encountered by the subject, shap- specific problems (for example, the localisation ing its development including unstimulating of brain regions involved in phonologic environments, physical insults, and genetic processing, providing compelling evidence that alterations. Neuroimaging studies often neglect disturbances in phonologic processing is a core to appreciate the brain based adaptive capacity deficit in reading disability). and compensatory responses that accompany The recent findings in PTSD have aroused chronic childhood psychiatric disorders, when interest in the interrelation of psychosocial attributing findings to disease process, even stressors, brain function, and structure. It though this may only be compensatory. would be possible to harness functional neuroimaging techniques to test hypotheses Future trends and implications of based on the biopsychosocial models of neuroimaging in child psychiatry childhood psychopathology. Understanding The future will possibly see the increasing use the mechanism of trauma related alteration of of functional neuroimaging in treatment plan- brain function (and probably structure), is ning and monitoring response in psychiatry. important to plan appropriate preventive The functional methods will continue to evolve strategies, and to improve long term outcome and the primary challenge will be to develop in traumatised children. better computerised image analysis techniques Genetic information will increasingly be capable of handling the wealth of anatomic and used to leverage the probability of locating functional neuroimaging data in children. brain abnormalities, as the eVects of genetic Some of the important developments in the lesions can be mapped with reasonable cer- field are detailed below. tainty to specific brain regions. This powerful Research into new radiopharmaceuticals has methodology remains largely untapped to date. opened up the possibility of using SPECT and Knowing where in the brain these genes will PET, to study a wider range of clinically relevant express themselves will in the future permit neurotransmitters and receptors (for example, studies of the secondary eVects of these genes

D1 antagonists in prefrontal cortex, hippocam- on maturation, development, and adaptation of

pus, and amygdala; a D2 receptor agonist which brain structure and function. The approach of

localises mainly in the striatum; 5HT1A, 5HT2A, studying speciﬁc genetic/chromosomal disor-

5HT2C receptor ligands). Ligands which help ders (for example, fragile X syndrome, William quantify receptors such as N-methyl-D-aspartate syndrome, Rett syndrome) will also help eluci- (NMDA), á-amino-3-hydroxy-5-methyl-4- date the manner in which gene–brain– isoxazolepropionate (AMPA), ion channels, behaviour associations develop and vary across ã-aminobutyric acid B (GABA-B) etc, will help developmental disability. the understanding of the neurodevelopmental Transcranial magnetic stimulation (TMS), is disorders. a new, non-invasive technique for directly The ability to study in vivo pharmacokinetics stimulating cortical neurones, with the hope of of agonists and antagonists of dopaminergic, a therapeutic eVect.41 This technique is closely serotoninergic, noradrenergic, GABA, opioid, related to the MRI technique. Preliminary and muscarinic receptors has opened up new investigations using rapid rate TMS, to im- research avenues. PET activation studies using prove motor speed in Parkinson’s disease and radioligand tagged drugs (for example, 11C- mood in depression, have been encouraging. methylphenidate) will increasingly be used to This could soon prove to be an important neu- predict drug response. Probes investigating the ropsychiatric tool in the assessment and clinically relevant neuroreceptor subsystems management of neurodevelopmental disorders. will help in aetiological understanding of As in utero brain development is abnormal disorders such as ADHD, leading to more tar- in many neurodevelopmental disorders, meth- geted treatments. Appropriate pre- and post- ods of assessing brain development in utero drug treatment receptor occupancy data using will become a priority, to understand the rela- neuroreceptor scanning techniques, will help tion between observed in utero brain develop- predict side eVect propensity (for example, tar- ment, postnatal brain neurocognitive develop- dive dyskinesia) and clinical response. Dy- ment, and the subsequent development of 418 Santosh

neuropsychiatric disorders. Ultrasound based phy”, helping to understand the time course of mild ventriculomegaly in utero (in the absence minute diVerences between subjects, to diVer- of other abnormalities) is associated with mild entiate acquired from developmental abnor- developmental delay in about 20% of malities, and to understand cerebral reorgani- children.42 sation. Fluid attenuated inversion recovery (FLAIR), diVusion anisotropy imaging, and DEVELOPMENTS IN IMAGING TECHNOLOGY event related PET scanning (for example, EEG This will be ongoing and the areas of focus will spike related PET changes) are all strategies be: acquisition of ever faster images, improve- that will have a role in the near future in under- ment of activation paradigms, such as event standing complex neural mechanisms in neu- related task designs which oVer more flexibility ropsychiatry (especially the epilepsies). than block design paradigms, examining data In SPECT, there will be an improvement in beyond averaging. Co-registration of scan data image resolution; more novel radiotracers will across time will help define the diVerences in become available, and will become used widely subjects over time, leading to longitudinal as a cost eVective technique to which most changes being picked up. The future of brain departments will have access. imaging in child neuropsychiatry will probably be diVerent for each technique. In structural MR imaging, faster image Conclusion acquisition techniques may help improve the Neuroimaging in child psychiatry is a rapidly cooperation of distressed children, and better developing field and the diVerent techniques head restraint systems combined with software being used are increasing so quickly that no and hardware development that corrects each single individual will be able to be conversant acquisition for motion artefacts will help in all of the methodologies. As yet, no specific greatly and allow the study of developmentally and consistent abnormality has been detected delayed children and those with movement in childhood psychiatric disorders. Findings disorders. Higher field strength magnets used have frequently been inconsistent owing to in MR imaging could improve image quality, various factors that aVect neuroimaging in but may need to go through ethical committees children. Obsessive compulsive disorder has and may need the demonstration that higher shown the most consistent findings so far, with fields are a minimal risk for younger subjects. orbitofrontal cortex and the caudate nucleus Voxel based whole brain anaysis and connectiv- being implicated. Better understanding of the ity analysis will increasingly be used to corticostriatal neural networks will shed more understand neural circuitry. light on the neurodevelopmental disorders of Functional MR imaging will increasingly be childhood. used, with improvements in technology tack- The design of developmentally correct pae- ling the problem of movement artefacts, imag- diatric brain maps, for computerised analysis of ing data reduction, and postimaging data paediatric neuroimaging data should become a processing.43 The sound of the scanner will also priority. Identification of homogeneous sub- possibly decrease significantly, helping to groups in paediatric research using genetics, reduce the eVect of noise on functional MR molecular biology, or immunology could im- imaging. Rapid MR scanning (including prove the specificity of neuroimaging findings. gradient echo, fast spin echo, and planar Despite all the pitfalls of paediatric neuroimag- sequences), the recently developed event re- ing, refinements of the techniques and im- lated fMRI, contrast based fMRI techniques, provements in the field could help improve diVusion MR imaging, arterial spine labelling diagnosis, triage, and in predicting and moni- (ASL) techniques, and dynamic susceptibility toring medication response and side eVects. MR perfusion imaging of the brain oVer clinically relevant physiological data not obtainable 1 Anderson AW, Gore JC. The physical basis of neuroimaging by conventional MR imaging, and may be used techniques. Child Adolesc Psychiatr Clin N Am 1997;6:213– 64. in child psychiatric disorders in the near future. 2 Giedd JN, Castellanos FX, Rajapakse JC, et al. Sexual They are likely to be at least as sensitive and dimorphism of the developing human brain. Prog Neuropsy- chopharmacol Biol Psychiatry 1997;21:1185–201. specific as radionuclide based techniques, and 3 Giedd JN, Rumsey JM, Castellanos FX, et al. A quantitative oVer the added advantage of higher intrinsic MRI study of the corpus callosum in children and adolescents. Brain Res Dev Brain Res 1996;91:274–80. resolution, convenient co-registration with 4 Giedd JN, Snell JW, Lange N, et al. Quantitative magnetic conventional MR imaging, as well as time and resonance imaging of human brain development: ages 4–18. Cereb Cortex 1996;6:551–60. cost eVective imaging in patients who require 5 Gauger LM, Lombardino LJ, Leonard CM. Brain morphol- routine MR imaging. ogy in children with specific language impairment. J Speech 31 Hear Res 1997;40:1272–84. MR spectroscopy involving P for the evalu- 6 Dalby MA, Elbro C, Stodkilde JH. Temporal lobe asymme- ation of membrane lipids, and 13C for the try and dyslexia: an in vivo study using MRI. Brain Lang 1998;62:51–69. evaluation of glutamate neurotransmission and 7 Castellanos FX. Toward a pathophysiology of attention- excitotoxicity will be increasingly used. MRS deficit/hyperactivity disorder. Clin Pediatr 1997;36:381–93. 8 Filipek PA. Neuroimaging in the developmental disorders: technology has already improved to include the state of the science. J Child Psychol Psychiatry 1999;40: approaches voxel tailored to the needs of each 113–28. 9 Overmeyer S, Taylor E. Neuroimaging in ADHD—an over- individual and disorder. view. Pediatr Rehabil [in press]. In PET, the progress will probably involve 10 Rubia K, Overmeyer S, Taylor E, Brammer M, Williams S, Simmons A, Bullmore ET. Hypofrontality in ADHD improved scanning resolution with newer trac- during higher-order motor control: a study with functional ers, requiring lower radiation dose. An impor- MRI. Am J Psychiatry 1999;156:891–6. 11 Piven J, Arndt S, Bailey J, Andreasen N. Regional brain tant step in understanding cerebral function enlargement in autism: a magnetic resonance imaging will be the increasing use of “autoradiogra- study. J Am Acad Child Adolesc Psychiatry 1996;35:530–6. Neuroimaging in psychiatric disorders 419

12 Filipek PA. Structural variations in measures in the develop- sion in neurologic disease. Baltimore: John Hopkins Univer- mental disorders. In: Thatcher RW, Lyon G, Reid G, et al, sity Press, 1993:186. eds. Developmental neuroimaging: mapping the development of 28 Steingard RJ, Renshaw PF, Yurgelun-Todd D, et al. brain and behavior. San Diego: Academic Press, 1996:169– Structural abnormalities in brain magnetic resonance 86. images of depressed children. J Am Acad Child Adolesc Psy- 13 Courchesne E, Saitoh O, Townsend JP, et al. Cerebellar chiatry 1996;35:307–11. hypoplasia and hyperplasia in infantile autism. Lancet 29 Bremner JD. Neuroimaging of posttraumatic stress disor- 1994; :63–4. 343 der. Psychiatric Annals 1998;28:445–50. 14 Fletcher PC, Happe F, Frith U, et al. Other minds in the brain: a functional imaging study of “theory of mind” in 30 De Bellis MD, Keshavan MS, Clark DB, et al. Developmen- story comprehension. Cognition 1995;57:109–28. tal traumatology part II: brain development. Biol Psychiatry 15 Happe F, Ehlers S, Fletcher P, et al. “Theory of mind” in the 1999;45:1271–84. brain. Evidence from a PET scan study of Asperger 31 Herholz K. Neuroimaging in anorexia nervosa. Psychiatry syndrome. Neuroreport 1996;8:197–201. Res 1996;62:105–10. 16 Abell F, Krams M, Ashburner J, et al. The neuroanatomy of 32 Santosh PJ, Baird G. Psychopharmacotherapy in children autism: a voxel-based whole brain analysis of structural and adults with intellectual disability (seminar). Lancet scans. Neuroreport 1999;10:1647–51. 1999;354:231–40. 17 Chugani DC, Muzik O, Behen M, et al. Developmental 33 Castellanos FX, Giedd JN, March WL, et al. Quantitative changes in brain serotonin synthesis capacity in autistic and brain magnetic resonance imaging in attention-deﬁcit/ nonautistic children. Ann Neurol 1999;45:287–95. hyperactivity disorder. Arch Gen Psychiatry 1996;53:607– 18 Malison RT, McDougle CJ, Van Dyck CH, et al. [1231] 16. Beta-CIT SPECT imaging demonstrates increased striatal 34 Hynd GW, Hern KL, Novey ES, et al. Attention deﬁcit dopamine transporter binding in Tourette’s syndrome. Am hyperactivity disorder (ADHD) and asymmetry of the cau- 1995; :1359–61. J Psychiatry 152 date nucleus. J Child Neurol 1993;8:339–47. 19 Wolf SS, Jones DW, Knable MB, et al. Tourette’s syndrome: prediction of phenotypic variation in monozygotic twins by 35 Peterson BS, Gore JC, Riddle MA, et al. Abnormal magnetic caudate nucleus D2 receptor binding. Science 1996;273: resonance imaging T2 relaxation time asymmetries in 1225–7. Tourette’s syndrome. Psychiatry Research Neuroimaging 20 Peterson BS, Skudlarski P, Anderson AW, et al. A functional 1994;55:205–21. magnetic resonance imaging study of tic suppression in 36 Harvey PH, Krebs JR. Comparing brains. Science 1990;249: Tourette syndrome. Arch Gen Psychiatry 1998;55:326–33. 140–6. 21 Rauch SL, Bates JF, Grachev ID. Obsessive-compulsive dis- 37 Andreasen NC, Flaum M, Swayze V II, et al. Intelligence order. Child Adolesc Psychiatr Clin N Am 1997;6:365–82. and brain structure in normal individuals. Am J Psychiatry 22 Fraizer JA, Geidd JN, Hamburger S, et al. Brain anatomic 1993;150:130–4. magnetic resonance imaging in childhood-onset schizo- 38 Willerman L, Schultz R, Rutledge JN, et al. In vivo brain size phrenia. Arch Gen Psychiatry 1996;53:617–24. and intelligence. Intelligence 1991;15:223–8. 23 Fraizer JA, Geidd JN, Kaysen D, et al. Childhood-onset 39 Reiss AL, Abrams MT, Singer HS, et al. Brain development, schizophrenia: brain MRI rescan after 2 years of clozapine gender and IQ in children—a volumetric imaging study. maintenance treatment. Am J Psychiatry 1996;153:564–6. Brain 1996;119:1763–74. 24 Jacobsen LK, Giedd JN, Berquin PC, Krain AL, Ham- 40 Arnold LE, StoV DM, Cook E Jr, . Ethical issues in bio- burger SD, Kumra S, Rapoport JL. Quantitative morphol- et al ogy of the cerebellum and fourth ventricle in childhood- logical psychiatric research with children and adolescents. J 1995; :929–39. onset schizophrenia. Am J Psychiatry 1997;154:1663–9. Am Acad Child Adolesc Psychiatry 34 25 Rappoport JL, Giedd JN, Blumenthal J, et al. Progressive 41 George MS, Wassermann EM, Post RM. Transcranial mag- cortical change during adolescence in childhood-onset netic stimulation: a neuropsychiatric tool for the 21st cen- schizophrenia. A longitudinal magnetic resonance imaging tury. J Neuropsychiatry Clin Neurosci 1996;8:373–82. study. Arch Gen Psychiatry 1999;56:649–54. 42 Patel MD, Filly RA, Hersh DR, et al. Isolated mild fetal cer- 26 Drevets, W, Videen T, Price J, et al. A functional anatomical ebral ventriculomegaly: clinical course and outcome. Radi- study of unipolar depression. J Neurosci 1992;12:3628. ology 1994;192:759–64. 27 Mayberg HS. Neuroimaging studies of depression in neuro- 43 Zametkin A, Liotta W. The future of brain imaging in child logic disease. In: Starkstein SE, Robinson RG, eds. Depres- psychiatry. Child Adolesc Psychiatr Clin N Am 1997;6:447–60.