Книги по МРТ КТ на английском языке / Functional Neuroimaging in Child Psychiatry Ernst 1 ed 2000

254B. S. Peterson and P. Thomas

(7±65 years) and the inability to include children in the control group to match on this variable (presumably because of the ethical concerns about exposing normal children to radioactivity) likely inXuenced the group comparisons. In addition, the TS and control groups were not matched by sex, and apparently handedness was not assessed, further confounding the interpretation of the Wndings. Sixteen of the patients were on medication at the time of the scan, and an additional 10 had received medications in the past. Another possible confound concerns the use of two diVerent SPECT scanners. All normal controls were scanned on a single machine, whereas subjects with TS were imaged on two diVerent scanners. Group membership was not blinded for scans acquired on the machine that scanned both subject groups, eVectively unblinding the entire study. Even though the investigators did not Wnd evidence for systematic diVerences in the results from the two scanners, they could not exclude the possibility of greater partial volume eVects in data from the patient group owing to use of the second scanner.

The same investigators undertook another HMPAO SPECT study on parents or siblings of Wve children with TS (Moriarty et al., 1997). Twenty family members were scanned: eight subjects had TS (eight males: age range 8±57 years); four had OCD symptoms; one had OCD symptoms and tics, and seven had no symptoms. The ROIs were visually inspected for qualitative abnormalities. No control group was included in the study, and ªabnormalityº was not deWned. In the seven scans from patients with TS that had acceptable quality, hypoperfusion was seen in the caudate in Wve subjects (although again, this should read ªbasal gangliaº owing to limitations in scanner resolution). In addition, hypoperfusion was seen in the parietal or temporal lobe in Wve, in the thalamus in two and in the frontal lobe and brainstem in one. Abnormalities were not seen in any of the family members who were symptom free. Despite the limitations of small subject numbers and the reliance on qualitative data, these results again suggest the presence of reduced blood Xow to the basal ganglia in TS.

Yet another HMPAO SPECT study of six adults with TS (three males; three females; mean age 36 years) and nine normal elderly control subjects (four males, Wve females; mean age 70 years) reported an increased asymmetry of basal ganglia blood Xow in the TS group (reported as a right/left ratio, p#0.001) (Klieger et al., 1997). The abnormally large asymmetry of basal ganglia blood Xow seemed to be driven by a relatively large (12.6%) decrease in Xow to the right-sided structures, a decrease that in itself was not statistically signiWcant (normalized to cerebellar blood Xow; p5 0.11). The limitations of this study include a small number of subjects, poor matching of patients and con-

trols (especially on age), failure to report comorbidities in the patient group, use of only a single transaxial slice for image analysis, and nonstandard ROI deWnition (left-sided ROIs were deWned as mirror images of those on the right). Finally, it seems likely that the increased asymmetry of basal ganglia blood Xow seen in the patients with TS in this study was a post-hoc Wnding.

An FDG PET study of 10 right-handed adults withTS (nine men, one woman; mean age 41.5"12.7 years) and 10 normal right-handed control subjects (Wve men, Wve women; mean age 42.5"11.5 years) failed to detect any global or regional metabolic diVerences between groups (Eidelberg et al., 1997). An analysis of the covariation between regional metabolic rates using a variant of principle components analysis identiWed two factors, or regional metabolic groupings, that together accounted for 29% of the subject-by-region variance. Regions loading most strongly on the Wrst factor included lateral frontal, medial frontal, cuneate, and paracentral cortices, as well as the midbrain. The authors interpreted these loadings as representing cortical motor circuits, in particular the lateral premotor and supplementary motor area systems. This interpretation is open to question, since the lateral frontal regions at a Talairach z-coordinate of 132mm would indicate not motor systems but prefrontal and medial frontal regions that subserve attentional functions. Factor scores for the TS subjects were signiWcantly elevated compared with those of the normal controls (p#0.01), which the authors attributed to motor activity. Regions loading most heavily on the second factor included the midbrain; the lenticular nucleus and thalamus in the subcortex; the medial, lateral, and superior temporal regions; and the lateral frontal and inferior parietal portions of the cortex. Regional loadings were predominantly right-sided. The authors did not interpret what circuits the loadings on this factor may represent, and the TS subjects did not diVer signiWcantly from controls on their loadings for this factor. The scores for the subjects with TS did, however, correlate positively and signiWcantly with one measure of symptom severity from the (now nonstandard) Tourette Syndrome Global Scale (r5 0.85; p#0.005). Although these Wndings are interesting, they must be regarded as preliminary and in need of rigorous replication, since they are undoubtedly post hoc and exploratory. Interpretation of the Wndings is also potentially confounded by the failure to assess cormorbid illness in the TS group, the unequal sex compositions of the two groups, and the greater degree of motion artifact in the TS group. Finally, the abnormal correlation patterns seen in the TS group could have been a consequence of the need to suppress tics during tracer uptake, an implicit behavioral demand not required of the normal controls.

A functional MRI study of voluntary tic suppression

Our recent functional MRI (fMRI) study investigated the eVects of voluntary tic suppression on fMRI signal changes (Peterson et al., 1998). These signal changes are produced by a change in regional deoxyhemoglobin content, which is itself induced by a local change in neuronal activity and metabolic demand (see Chapter 3). In the study, 22 patients with TS (11 males and 11 females; mean age 35.7"10.9 years; 10 comorbid OCD; 14/22 right-handed) were asked to alternate 40 s epochs of allowing themselves to tic freely with 40 s epochs of suppressing their tics voluntarily, for a total of eight tic/suppression cycles. Signal intensity decreased signiWcantly during tic suppression epochs in the ventral globus pallidus, ventral putamen, midbody of each hemithalamus, right posterior cingulate, left hippocampus and parahippocampus, the cuneus bilaterally, left sensorimotor cortex, and left inferior parietal region. Increased signal intensity during tic suppression occurred in the ventral portion of the right caudate nucleus, right midfrontal cortex, right middle temporal gyrus, superior temporal gyrus bilaterally, right anterior cingulate cortex, and bilateral inferior occipital regions (Fig. 14.3, p. 242).

The severity of tic symptoms outside of the scanner correlated with the change in signal intensity associated with tic suppression inside of the scanner much more frequently in subcortical (8/8 regions) than in cortical regions (2/15 regions). Moreover, the direction of correlation between symptom severity and signal change in the subcortex suggested that as symptom severity increases, the absolute value of the magnitude of signal change in the subcortex decreases. This suggests that the subcortical signal changes ± increases in the right caudate and decreases in the rest of the subcortex ± participate in the suppression of these unwanted behaviors. When these braking mechanisms begin to fail and subcortical signal change is reduced, tics are progressively more likely to escape inhibition. Motion artifact, assessed in several diVerent ways, did not seem to account for the observed correlations. These Wndings suggest that the pathogenesis of TS involves dysfunction of subcortical portions of CSTC circuitry. An additional interpretation is that the most important determinant of clinical phenotype and symptom severity may not be the genetic vulnerability to tic disorders. Rather, the stronger determinant of symptom severity may be the degree to which the control system either suppresses or unmasks the inherent diathesis to tic. The functional integrity of this control system itself may be under genetic control, or it may be a manifestation of other nongenetic pathophysiologic determinants.

Radioligand studies

Other preliminary functional studies have focused on quantifying various functional indices of the dopaminergic neurotransmitter system in the striatum in subjects with TS (Table 14.4). An [18F]-Xuorodopa study of three patients with TS, six with OCD, and 30 normal control adults found no diVerence between patient and control groups for the inXux constants of Xuorodopa (Brooks et al., 1992). A [11C]- raclopride PET investigation from the same laboratory studied three patients with TS and eight normal adults and found no group diVerences in [11C]-raclopride binding (Singer et al., 1992). These PET studies were then expanded to include 10 patients with TS and 34 controls in the Xuorodopa study, and Wve patients with TS and nine controls in the [11C]-raclopride study (Turjanski et al., 1994). Again, no group diVerences were found. The negative Xuorodopa Wndings indicate normal Xuorodopa uptake and decarboxylation, while the negative [11C]-raclopride Wndings indicate normal numbers of available dopamine D2 and D3 postsynaptic receptor-binding sites in these patients with TS. The patient and control groups in these studies, however, were inadequately matched (Table 14.4); 7 of the 10 TS subjects in the Xuorodopa study were on neuroleptic medication at the time of scanning, and the patient numbers in each study were still relatively small. In addition, Xuorodopa has limited signal-to-noise properties (#2:1) mainly because the peripheral radiolabeled Xuorodopa metabolite (3-O-methyl-6-F-DOPA) is present in the striatum (Garnett et al., 1983). All of these considerations may have limited the statistical power to detect potential group diVerences (Table 14.4).

Although these initial studies suggested normal synaptic dopamine transmission in TS, a subsequent pilot study of striatal [123I]-(-CIT (2(-carboxymethoxy-3(-[4- iodophenyl]tropane) uptake suggested the presence of elevated presynaptic dopamine transporter levels in TS (Malison et al., 1995). Transporter levels were 37% higher in Wve patients with TS (mean age 27"8 years) than in Wve normal controls who were individually matched by age (28"8 years) and sex: 10.4"2.3 versus 7.7"1.4 (mean"SD) ratio of striatal to occipital uptake, respectively. The limitations of the study included small subject numbers, the use of paired t-tests (ageand sex-matching do not account for all confounding biological variability), and the inclusion of TS subjects who had prior neuroleptic exposure. Despite these limitations, the Wndings are intriguing, and they are consistent with one autopsy study of three adults with TS and six normal controls that reported increased numbers of presynaptic dopamine-uptake sites in the caudate nucleus and putamen of the former (Singer

et al., 1991). In this same study, levels of dopamine and its metabolites in the subjects with TS were not signiWcantly diVerent from control levels. Whereas the autopsy study found higher transporter levels in the putamen, the SPECT study suggested that transporter levels were particularly elevated in the caudate nucleus.

A subsequent [123I]-iodobenzamide (IBZM, a D2-like receptor ligand) SPECT study of Wve monozygotic twin pairs (four male and one female twin pairs; mean age 29"4 years; all currently medication free) reported an increase of IBZM binding in the caudate of all Wve of the more severely aVected co-twins (Wolf et al., 1996). No diVerences in IBZM binding were seen in the putamen. The group diVerences did not appear to be a result of diVerences in blood Xow between groups, since blood Xow values from HMPAO SPECT scans in the same twin pairs did not reveal any group diVerences. Intrapair diVerences in IBZM binding correlated highly with the corresponding diVerences in overall symptom severity between co-twins (r5 0.99; p#0.001). The investigators suggested that a variation in D2 receptor availability in the caudate nucleus might have accounted for the diVering severity of symptom expression between siblings who have the same genetic vulnerability. Although these initial Wndings are exciting, it is important to remember that they are based on a sample of Wve twin pairs, rendering the statistical tests highly unstable. In addition, the possible discordance of comorbid illnesses is not considered as a possible confound. Neither is the possibility considered that the greater D2 receptor availability in the more severely aVected twin could represent either a compensatory response to the presence of more severe tics or an epiphenomenon owing to greater neuroleptic exposure in the more severely aVected twin.

The D2-like receptors were again studied with [11C]-3-N- methylspiperone ([11C]-NMSP) in two diVerent PET protocols (Wong et al., 1997). All patients were medication free for at least 6 months. The Wrst protocol measured only caudate-to-cerebellar ratio of radioligand binding at 45 min postinjection in nine patients with TS (mean age 27 years; Wve with prior neuroleptic use) and 44 neurologically normal control subjects (22 males, 22 females; age 19±73 years). The second, applied in 20 TS subjects (16 males, 4 females; mean age 36.2"8.9 years) and 24 normal adults (mean age not provided; range 18±83 years) was designed to provide an absolute measure of available receptor density (B9max), which they hypothesized would suggest the presence of elevated D2-like dopamine receptors in TS. After their Wrst scan, this second group of subjects received a dose of unlabeled haloperidol, which in part displaced [11C]-NMSP from its D2-like receptors. They then underwent a second scan to determine the degree of

radioligand displacement. In the Wrst protocol, the caudate-to-cerebellar ratio of radioligand binding did not diVer signiWcantly between the TS and control groups. In the second protocol, Bmax binding potential values also did not diVer between groups. These results are consistent with prior studies that show the postsynaptic D2-like receptors inTS to be similar to those of normal control subjects. If the Wndings of the previously described IBZM monozygotic twin study are correct, then these NMSP Wndings would have two primary implications. First, D2- like receptor levels must determine a relatively small portion of between-subject variance in clinical symptom severity in singletons with TS, even though the levels may account for much of the variability in symptom severity between monozygotic co-twins. Second, nongenetic factors must account for a small portion of the normal individual variability in D2 receptor levels.

Finally, a relatively recent Xuorodopa PET study of 11 adolescents with TS (mean age 15.2"1.9 years) and 10 sexmatched normal controls (mean age 14.8"1.7 years) found higher Xuorodopa accumulation in the left caudate nucleus (25%) and the right midbrain (53%) of the former (Ernst et al., 1999). Subjects watched a videotape during the Wrst 80min of tracer uptake and were not speciWcally told to suppress tics. It should be noted that activities that engage attention, even relatively simple ones such as watching videotapes, often alter tic frequency (typically reducing it). The investigators, however, recorded motor activity of the patients during tracer uptake and scanning, but they found no correlation with Xuorodopa levels.While intriguing and consistent with the growing number of studies implicating the right caudate nucleus in TS pathophysiology (Table 14.3), these results must nevertheless be interpreted with caution. Had the p values been correlated for multiple comparisons, the analyses for Xuorodopa accumulation would not have retained their statistical signiWcance. In addition, if the Wndings are correct, it is still unclear whether the regionally speciWc increase in DOPA decarboxylase enzyme activity is a consequence of a larger number of dopaminergic synapses or enzyme upregulation in those regions. Altered levels of dopamine (from medication blockade, increased presynaptic uptake, or altered release and catabolism) can produce enzyme upregulation. The prior preliminary Wndings of increased presynaptic transporter in singletons with TS and increased postsynaptic D2 receptor availability in more severely aVected monozygotic co-twins could both be consistent with reduced levels of extracellular dopamine leading to upregulation of DOPA decarboxylase. Reduced extracellular dopamine, in turn, can reXect downregulation of dopamine release from either increased GABAergic

258B. S. Peterson and P. Thomas

or reduced glutamatergic input to the striatum (Fig. 14.4). Thus, the explanations consistent with the known circuitry or the basal ganglia are many, complex, and varied.

Implications of imaging studies

Which brain systems are most implicated in pathophysiology?

A careful review of the Wndings of each of the metabolism and blood Xow-related studies (Table 14.3) shows the consistency with which the basal ganglia are implicated in the pathophysiology of TS. Only the HMPAO SPECT study of George and colleagues (1992) failed to show abnormalitites in this region. Nearly all other PET and SPECT studies demonstrated reduced metabolism or blood Xow to the basal ganglia in TS subjects relative to controls, most frequently in the ventral striatum and most often in the left hemisphere. Two studies failed to show regional group diVerences in basal ganglia blood Xow or metabolism but did detect other evidence for abnormal asymmetry in basal ganglia blood Xow (Klieger et al., 1997) or correlations between measures of activity in basal ganglia circuitry and symptom severity (Eidelberg et al., 1997). Most studies lacked the spatial resolution needed to diVerentiate individual nuclei within the basal ganglia, and two of the studies that reported hypometabolism in the inferior insular cortex (Chase et al., 1986; Braun et al., 1993) may actually have inadvertently measured instead the activity in the lateral aspect of the putamen, which is in close proximity to the insular cortex.

The radioligand studies also generally implicate the basal ganglia in TS pathophysiology, though perhaps less consistently than do the metabolism and blood Xow studies. This is likely because of the speciWcity of the information that these radioligand studies provide regarding particular neurochemical characteristics of the basal ganglia and because of the complexity of the physiologic regulation of dopamine metabolism and dopamine receptor system. Radioligand studies that have examined synaptic

dopamine activity and D2-like receptor levels in TS have been variously hampered by the inclusion of medicated

subjects, small numbers of participants, and age and sex mismatches between groups. Nevertheless, the positive Wndings thus far indicate abnormalities of dopaminergic

innervation to the caudate nucleus. Increased D2 receptor levels in the caudate nucleus appear to account for some of

the nongenetic determinants of variance in symptom severity. DOPA decarboxylase levels and presynaptic dopamine transport levels may be elevated in this region.

These preliminary results may indicate a hyperinnervation of the striatum by dopaminergic neurons. This striatal hyperinnervation could result in increased thalamocortical activity by stimulating the direct and further inhibiting the indirect pathway through CSTC circuits, as follows (Leckman et al., 1997). Evidence from both rodents (Girault et al., 1986) and primates (Filion and Tremblay, 1991; Filion et al., 1991) indicates that stimulation of D1 and D2 receptors produces opposing behavioral eVects, with D1 stimulation increasing and D2 stimulation decreasing behavioral activity. In addition, stimulation of D1 receptors facilitates the release of GABA whereas stimulation of D2 receptors inhibits it. Figure 14.4b may help to understand the model supporting these opposite behavioral eVects based on the anatomic connectivity within CSTC circuits. In the direct pathway, dopaminergic hyperinnervation activates D1 receptors (located postsynaptically, particularly on GABA postsynaptic neurons). GABA neurons, which project from the internal segment of the globus pallidus/substantia nigra pars reticulata to the thalamus will in turn block inhibitory interneurons, resulting in enhanced the glutamatergic (GLU) tone of thalamic glutamatergic neurons projecting to the cortex. This enhanced thalamocortical GLU tone will result in an excessive motor output. The role of the indirect pathway and D2 receptors in motor dyscontrol with respect to dopaminergic hyperinnervation is unclear. Presumably, the inhibitory eVect of GABAergic projection neurons from the external segment of the globus pallidus to the reticular nucleus, internal segment of the globus pallidus/substantia nigra, and subthalamic nucleus would be reinforced. This increased GABAergic tone of the projections to these other three basal ganglia nuclei would have a threefold eVect (Fig. 14.4b), all of which would tend to produce a net increase in thalamic cortical excitation and an increased propensity for movement. First, it would suppress activity in the GABAergic projections from the thalamic reticular nucleus to other thalamic nuclei, thereby disinhibiting thalamocortical excitation. Second, it would suppress activity in the GABAergic inhibitory projections from the internal segment of the globus pallidus/substantia nigra to thalamic nuclei, further disinhibiting thalamocortical excitation. Third, it would suppress excitatory activity in glutamatergic projections from subthalamic nuclei to the internal segment of the globus pallidus/substantia nigra, which again would ultimately produce thalamocortical disinhibition. It can be appreciated that, in this scheme, striatal hyperinnervation by dopaminergic neurons would overdetermine, through multiple eVects on the direct and indirect pathways, a functional and neurochemical syndrome of disinhibition in TS. This could account for much

260B. S. Peterson and P. Thomas

of the phenomenology of TS-related conditions ± not just tics, but also recurrent obsessional thoughts, compulsions, and impulsive behaviors ± depending upon which portions of the striatal, pallidal, or thalamic nuclei were disinhibited.

The blood Xow and metabolism studies, along with the radioligand Wndings, therefore, quite strongly implicate in TS pathophysiology the basal ganglia portions of CSTC circuitry, in particular the caudate nucleus portions of the ventral striatum. As we have seen, however, the basal ganglia are conduits for information-processing streams that serve multiple and diverse functions. The ventral striatum, and particularly the ventral caudate nucleus, tends to subserve temporolimbic and orbitofrontal portions of CSTC circuitry (Alexander et al., 1990). Of these systems, regions belonging to the temporolimbic system ± the anterior cingulate, parahippocampal, and possibly insular cortices in particular ± appear to diVer most consistently between groups in studies of metabolism and blood Xow (Table 14.3). The other CSTC system in which regions diVer frequently in TS from normals is the one that involves numerous association cortices, particularly frontal, parietal, and superior temporal regions.

What were the functional imaging studies really imaging?

This overlap in Wndings between diVerent studies using various imaging modalities provide converging evidence that CSTC circuits are probably involved in TS pathophysiology. The usual and most obvious interpretation of this overlap is that dysfunction in the basal ganglia and related cortical portions of the CSTC circuit gives rise to the symptoms of TS. Our study of the brain regions involved in tic suppression, however, oVers an alternative interpretation of the overlap between imaging studies. Tic suppression produces a decrease in signal intensity (which is indirectly related to changes in neuronal activity) in ventral basal ganglia regions, particularly the ventral putamen and globus pallidus. PET and SPECT studies furthermore suggest that ªrestingº metabolism and blood Xow are reduced in subjects with TS relative to normal controls in the ªventral striatumº, and this tends to be used by the respective investigators as synonymous with ªventral caudate nucleusº. These imaging modalities, however, rarely if ever have provided suYcient spatial resolution to discriminate ventral caudate from the lenticular nucleus, or to discriminate ªinferior insular cortexº from ªinferior lateral putamenº. Furthermore, subjects in most PET and SPECT studies may have either intentionally or unintentionally suppressed their tics during tracer uptake. This is a

time when the novelty of the situation or the engagement of attention in the procedure may itself have been suYcient to suppress tics, which is a well-recognized clinical phenomenon. Tic suppression would then have produced a hypometabolism in the subjects with TS relative to the control subjects, who did not have the burden of suppressing these unwanted behaviors during the scan.

The cortical regions that activate during tic suppression have also been reported to diVer in subjects with TS from that seen in normal controls in ªrestingº blood Xow and metabolism studies. These PET and SPECT Wndings in the cortex are more variable than are those in the basal ganglia (Fig. 14.2). In our fMRI study, we have interpreted many of the cortical regions involved in tic suppression as deriving from the intensive attentional requirements that tic suppression requires. This may also explain the variable cortical Wndings in the PET and SPECT studies, particularly those involving frontal, superior temporal, and anterior cingulate cortices. The decrease in metabolism and blood Xow to hippocampal and parahippocampal cortices, posited to result from reallocation of attentional resources during active task conditions, may be a consequence of the preoccupation with contemporaneous somatosensory information during tic suppression (Shulman et al., 1997b). The lower hippocampal and parahippocampal activity would, in turn, reduce the retrieval of memory traces from long-term storage in these regions.

Broader implications of impulse control demands in functional imaging studies

The implications of the fMRI tic suppression study reach beyond the investigations of only TS. Just as prior TS imaging studies may have unwittingly studied the neural correlates of tic suppression, imaging investigations of other disorders and conditions may have studied similar or identical impulse control circuits. OCD imaging studies, for example, have either implicitly or explicitly required their subjects to suppress the enactment of compulsive urges during the scan. The most consistent Wndings in OCD imaging studies thus far have been hypermetabolism and increased blood Xow to the inferior frontal cortex, right caudate nucleus (Baxter et al., 1987, 1988; Rauch et al., 1994; Breiter et al., 1996), and anterior cingulate cortex (Swedo et al., 1989; Rauch et al., 1994; Breiter et al., 1996): regional changes; these were also induced by tic suppression. If these Wndings in OCD were, in fact, a result of the suppression of obsessions and compulsions during the scan, then it is of little surprise that these diVerences normalized after successful antiobsessional therapy (Benkelfat et al., 1990; Baxter et al., 1992; Swedo et al., 1992;

Rubin et al., 1995; Schwartz et al., 1996), when the need to suppress less-severe compulsions would have attenuated.

It seems likely that other functional neuroimaging protocols may also have implicitly studied the suppression of unwanted impulses. The subjective experience of pain, for instance, has of late become an area of great interest and intensive investigation. Painful stimuli typically produce widespread areas of cerebral activation in the perigenual portion of the anterior cingulate cortex, related mesial frontal cortex, midprefrontal and inferior parietal areas, and somatosensory cortex (Talbot et al., 1991; Derbyshire et al., 1997, 1998). Decreases may be seen in the amygdala, hippocampus, or thalamus (Hsieh et al., 1995; Derbyshire et al., 1997). We believe that these diverse areas of activation represent not only the subjective experience of pain but also the suppression during the scan of the powerful impulse to withdraw from the painful stimulus. The possibility that these diverse imaging protocols implicitly studied the suppression of unwanted impulses is bolstered by a PET study that purported to investigate the sensation of itch and the urge to scratch in response to the subcutaneous injection of histamine.We postulate that the study did not only image the urge to scratch, it also imaged the concomitant cerebral activation that is the neural correlate of the suppression of the scratch (Hsieh et al., 1994). This postulate is supported by the similiarities of many of the activations ± which include anterior cingulate cortex, supplementary motor and premotor areas, and inferior parietal cortex ± with the areas that activate during the suppression of tics, compulsions, and the withdrawal to painful stimuli.

The possibility that diverse functional imaging paradigms may have unwittingly studied the suppression of unwanted impulses underscores a profound and fundamental diYculty with functional imaging studies in general. The CNS is a dynamic entity whose raison d'être is to sense and react. These two domains of experience ± sensing both internal and external stimuli and reacting appropriately (or not) to them ± are not so easily separated as we would want and suppose them to be for our imaging protocols. The subtraction paradigms that are now so popular in PET and fMRI may not be able to subtract out completely sensation from response to the sensation. And the between-group comparisons of resting brain blood Xow and metabolism in disease processes all assume that the disease process in question is an entity isolated from the rest of experience and CNS functioning. Studies of schizophrenia are a prime example. The subjective responses to the presence of positive symptoms, such as hallucinations, delusions, and thought disorganization, must involve a wide range of compensatory mechanisms

that may confound the interpretation of functional imaging data. The patients but not their controls are likely during the scanning procedure to attempt to make sense of the fragmentation of their sensory experience, to test the reality of hallucinations and delusions, and to cope with the intense anxieties during the procedure that normal subjects usually do not have. Imaging studies of subjects with TS may have similar confounds. The patients usually have to suppress their tics during the scan, which normal subjects do not have to do. Even if not suppressing their tics, the TS subjects at least have the experience of them during the scan, whereas normal subjects do not. And of course they will have the premonitory urges associated with their tics. Finally, subjects with TS are likely to have more thoughts during the scan about whether they have complied or not with instructions to hold their head and neck still during the procedure, even if they are allowed to tic.

Chronic adaptive responses are likely to have diverse eVects on structural and neurochemical features of the brain that aVect the Wndings from our functional imaging studies. Learning and experience also have long-term plastic eVects on brain structure and function (Merzenich et al., 1983; Black et al., 1990; Elbert et al., 1995). It is possible, then, that the radioligand Wndings in TS may not represent a process central to the pathophysiology of TS in the sense of causing symptoms. They could instead represent a compensatory response to the presence of tics.

We have already suggested several ways to help to determine what is cause and what is compensatory eVect in functional imaging data in TS. Imaging protocols must include young children, and perhaps even genetically vulnerable children prior to disease onset. If adults are studied exclusively, the patient groups should be representative of the general population of subjects who have lifetime histories of TS, not just those who continue to be symptomatic. Imaging protocols should increasingly be longitudinal in design; they should include studies of agerelated changes in normal children and adults, and they should collect in the same individuals data from other imaging modalities. Studies of trait vulnerabilities in family members will also be valuable in sorting out what is cause (i.e., a trait) and what is compensatory (i.e., a state) eVect. Thorough characterization of comorbid neuropsychiatric diagnoses and the accounting for their eVects in the imaging data is now mandatory. Finally, it is necessary in study design to assess both the implicit and explicit demands of the scanning protocols. This will require isolating to the degree possible those features of the imaging data that are a result of tic behaviors and those (such as inhibitory control mechanisms in tic suppression) that

262B. S. Peterson and P. Thomas

compensate for them. Distinguishing what in our data is causal to tics and what is an epiphenomenon or compensatory for them will perhaps be our greatest future methodologic and scientiWc challenge.

iReferencesi

Alexander, G. and Delong, M. (1985). Microstimulation of the primate neostriatum. II. Somatotopic organization of striatal microexcitable zones and their relation to neuronal response properties. J. Neurophysiol., 53, 1417±30.

Alexander, G., Crutcher, M. and DeLong, M. (1990). Basal gangliathalamocortical circuits: parallel substrates for motor, oculomotor, `prefrontal', and `limbic' functions. Prog. Brain Res., 85, 119±46.

Apter, A., Pauls, D. L., Bleich, A. et al. (1993). An epidemiologic study of Gilles de la Tourette's syndrome in Israel. Archiv. Gen. Psychiatry, 50, 734±8.

Baldwin, M., Frost, L. L. andWood, C. D. (1954). Investigation of the primate amygdala. Movements of the face and jaws. Neurology,

4, 596±8.

Balthazar, K. (1956). Uber das anatomishe substrat der generalisierten tic-krankeit (maladie des tics, Gilles de la Tourette): Entwicklungshemmung des corpus striatum. Arch. Psychiatr. Nervenkr., 195, 531±9.

Bates, J. F. and Goldman-Rakic, P. S. (1993). Prefrontal connnections of medial motor areas in the rhesus monkey. J. Comp. Neurol., 336, 211±28.

Baxter, L., Schwartz, J., Mazziotta, J. et al. (1988). Cerebral glucose metabolic rates in nondepressed patients with obsessive-com- pulsive disorder. Am. J. Psychiatry, 145, 1560±3.

Baxter, L., Jr, Schwartz, J. M. Bergman, K. S. et al. (1992). Caudate glucose metabolic rate changes with both drug and behavior therapy for obsessive-compulsive disorder. Arch. Gen. Psychiatry, 49, 681±9.

Baxter, L. R. Jr, Phelps, J. M., Mazziotta, J. C., Guze, B. H. and Schwartz, J. M. (1987). Local cerebral glucose metabolic rates in obsessive-compulsive disorder: a comparison with rates in unipolar depression and normal controls. Arch. Gen. Psychiatry, 44, 211±18.

Benkelfat, C., Nordahl, T. E., Semple, W. E., King, A. C., Murphy, D. L. and Cohen, R. M. (1990). Local cerebral glucose metabolic rates in obsessive-compulsive disorder. Patients treated with

clomipramine. Arch. Gen. Psychiatry, 47, 840±8.

Black, J. F., Isaacs, K. R., Anderson, B. J., Alcantara, A. A. and Greenough, W. T. (1990). Learning causes synaptogenesis, whereas motor activity causes angiogenesis, in cerebellar cortex of adult rats. Proc. Natl. Acad. Sci. USA, 87, 5568±72.

Braun, A. R., Stoetter, B., Randolph, C. et al. (1993). The functional neuroanatomy of Tourette's syndrome: an FDG-PET study. I. Regional changes in cerebral glucose metabolism diVerentiating patients and controls. Neuropsychopharmacology, 9, 277±91.

Breiter, H. C., Rauch, S. L., Kwong, K. K. et al. (1996). Functional

magnetic resonance imaging of symptom provocation in obses- sive-compulsive disorder. Arch. Gen. Psychiatry, 53, 595±606.

Brooks, D. J., Turjanski, N., Sawle, G. V., Playford, E. D. and Lees, A. J. (1992). PET studies on the integrity of the pre and postsynaptic dopaminergic system in Tourette syndrome. Adv. Neurol., 58, 227±31.

Bruun, R. D. (1988). The natural history of Tourette's syndrome. In

Tourette's Syndrome and Tic Disorders: Clinical Understanding and Treatment, eds. D. J. Cohen, R. D. Bruun and J. F. Leckman, pp. 21±39. New York: Wiley.

Burd, L., Kerbeshian, J., Wikenheiser, M. and Fisher, W. (1986). A prevalence study of Gilles de la Tourette syndrome in North Dakota school-age children. J. Am. Acad. Child Psychiatry, 25, 552±3.

Castellanos, F. X., Giedd, J. N., Eckburg, P. et al. (1994). Quantitative morphology of the caudate nucleus in attention deWcit hyperactivity disorder. Am. J. Psychiatry, 151, 1791±6.

Chappell, P., Riddle, M., Anderson, G. et al. (1994). Enhanced stress responsivity of Tourette syndrome patients undergoing lumbar puncture. Biol. Psychiatry, 36, 35±43.

Chase, T. N., GeoVrey, V., Gillespie, M. and Burrows, G. H. (1986). Structural and functional studies of Gilles de la Tourette syndrome. Rev. Neurol., 142, 851±5.

Comings, D. E., Himes, J. A. and Comings, B. G. (1990). An epidemiologic study of Tourette's syndrome in a single school district. J. Clin. Psychiatry, 51, 463±9.

Cooper, I. S. (1969). Involuntary Movement Disorders. New York:

Harper and Row.

Craig, A. D., Reiman, E. M., Evans, A. and Bushnell, M. C. (1996). Functional imaging of an illusion of pain. Nature, 384, 258±60. de Divitiis, E., D'Errico, A. and Cerillo, A. (1977). Stereotactic surgery in Gilles de la Tourette syndrome. Acta Neurochirurg., 24,

73.

Derbyshire, S. W. G., Jones, A. K. P., Gyulai, F., Clark, S., Townsend, D. and Firestone, L. L. (1997). Pain processing during three levels of noxious stimulation produces diVerential patterns of central activity. Pain, 73, 431±45.

Derbyshire, S. W. G., Vogt, G. A. and Jones, A. K. P. (1998). Pain and Stroop interference tasks separate processing modules in anterior cingulate cortex. Exp. Brain Res., 118, 52±60.

Diamond, A. and Goldman-Rakic, P. S. (1989). Comparison of human infants and rhesus monkeys on Piaget's AB task: evidence for dependence on dorsolateral prefrontal cortex. Exp. Brain Res., 74, 24±40.

Divac, I., Rosvold, H. and Szwarcbart, M. (1967). Behavioral eVects of selective ablation of the caudate nucleus. J. Comp. Physiol. Psychol., 63, 184±90.

Drewe, E. (1975). Go-no-go learning after frontal lobe lesions in humans. Cortex, 11, 8±16.

Eidelberg, D., Moeller, J. R., Antonini, A. et al. (1997). The metabolic anatomy of Tourette's syndrome. Neurology, 48, 927±34.

Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B. and Taub, E. (1995). Increased cortical representation of the Wngers of the left hand in string players. Science, 270, 305±7.

Erenberg, G., Cruse, R. P. and Rothner, A. D. (1987). The natural

history of Tourette syndrome: a follow-up study. Ann. Neurol.,

22, 383±5.

Ernst, M., Zametkin, A. J., Jons, P. H., Matochik, J. A., Pascualvaca, D. and Cohen, R. M. (1999). High presynaptic dopaminergic activity in children with Tourette's disorder. J. Am. Acad. Child Adolesc. Psychiatry, 38, 86±94.

Filion, M. and Tremblay, L. (1991). Abnormal spontaneous activity of the globus pallidus neurons in monkeys with MPTP-induced parkinsonism. Brain Res., 547, 147±51.

Filion, M., Tremblay, L. and Bedard, P. J. (1991). EVects of dopamine agonists on the spontaneous activity of the globus pallidus neurons in monkeys with MPTP-induced parkinsonism. Brain Res., 547, 152±61.

Fried, I., Katz, A., McCarthy, G., Sass, K., Spencer, S. and Spencer, D. (1991). Functional organization of human supplementary motor cortex studies by electrical stimulation. J. Neurosci., 11, 3656±66.

Gadow, K. D., Sverd, J., Sprafkin, J., Nolan, E. E. and Ezor, S. N. (1995). EYcacy of methylphenidate for attention-deWcit hyperactivity disorder in children with tic disorder. Arch. Gen. Psychiatry, 52, 444±55.

Garnett, E., Firnau, G., Nahmias, C. and Chirakal, R. (1983). Striatal dopamine metabolism in living monkeys examined in positron emission tomography. Brain Res., 280, 169±71.

George, M. S., Trimble, M. R., Costa, D. C., Robertson, M. M., Ring, H. A. and Ell, P. J., (1992). Elevated frontal cerebral blood Xow in Gilles de la Tourette syndrome: a 99Tcm-HMPAO SPECT study.

Psychiatr. Res., 45, 143±51.

Girault, J. A., Spampinato, U., Glowinski, J. and Besson, M. J. (1986). In vivo release of [3H]-gamma-aminobutyric acid in the rat neostratum. II. Opposing eVects of D1and D2 dopamine receptor stimulation in the dorsal caudate putamen. Neuroscience, 19, 1109±17.

Goetz, C. G., Tanner, C. M., Stebbins, G. T., Leipzig, G. and Carr, W. C. (1992). Adult tics in Gilles de la Tourette's syndrome: description and risk factors. Neurology, 42, 784±8.

Golden, G. S. (1974). Gilles de la Tourette's syndrome following methylphenidate administration. Dev. Med. Child Neurol., 16, 76±8.

Goldman-Rakic, P. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. In

Handbook of Physiology, The Nervous System, eds. V. Mountcastle, F. Plum and S. Geiger, pp. 373±414. Bethesda, MD: American Physiological Society.

Gratton, G. (1997). Attention and probability eVects in the human occipital cortex: an optical imaging study. Neuroreport, 8, 1749±53.

Hall, M., Costa, D. C., Shields, J., Heavens, J., Robertson, M. and Ell, P. J. (1990). Brain perfusion patterns with Tc-99m-HMPAO/SPET in patients with Gilles de la Tourette syndrome. Eur. J. Nucl. Med.,

16, WP18.

Hassler, R. and Dieckmann, G. (1970). Stereotaxic treatment of tics and inarticulate cries or coprolalia considered as motor obsessional phenomena in Gilles de la Tourette's disease. Rev. Neurol.

123, 89±100.

Heinze, H. J., Mangun, G. R., Burchert, W. et al. (1994). Combined

spatial and temporal imaging of brain activity during visual selective attention in humans. Nature, 372, 543±6.

Hsieh, J. C., Hagermak, O., Stahle-Backdahl, M. et al. (1994). Urge to scratch represented in the human cerebral cortex during itch.

J. Neurophysiol., 72, 3004±8.

Hsieh, J.-C., Belfrage, M., Stone-Elander, S., Hansson, P. and Ingvar, M. (1995). Central representation of chronic ongoing neuropathic pain studied by positron emission tomography. Pain, 63, 225±36.

Iverson, S. and Mishkin, M. (1970). Perseverative interference in monkeys following selective lesions of the inferior prefrontal cortex. Exp. Brain Res., 11, 376±86.

Jadresic, D. (1992). The role of the amygdaloid complex in Gilles de la Tourette's syndrome. Br. J. Psychiatry, 161, 532±4.

Jankovic, J., Glaze, D. G. and Frost, J., Jr (1984). EVect of tetrabenazine on tics and sleep of Gilles de la Tourette's syndrome.

Neurology, 34, 688±92.

Karp, B. I., Porter, S., Toro, C. and Hallett, M. (1996). Simple motor tics may be preceded by a premotor potential. J. Neurol. Neurosurg. Psychiatry, 61, 103±6.

Kelley, A. E., Lang, C. G. and Gauthier, A. M. (1988). Induction of oral stereotypy following amphetamine microinjection into a discrete subregion of the striatum. Psychopharmacology, 95, 556±9.

Kirwan, C. (ed.) (1971). Aristotle's Metaphysics. Oxford: Clarendon

Press.

Klieger, P. S., Fett, K. A., Dimitsopulos, T. and Kurlan, R. (1997). Asymmetry of basal ganglia perfusion in Tourette's syndrome shown by Technetium-99m-HMPAO SPECT. J. Nucl. Med., 38, 188±91.

Korzen, A. V., Pushkov, V. V., Kharitonov, R. A. and Shustin, V. A. (1991). Stereotaxic thalamotomy in the combined treatment of Gilles de la Tourette's disease. Zh. Nevrotpatol. Psykhiatri, 91, 100±1.

Kurlan, R., Kersun, J., Ballantine, H., Jr and Caine, E. D. (1990). Neurosurgical treatment of severe obsessive-compulsive disorder associated with Tourette's syndrome. Movement Disord., 5, 152±5.

Larumbe, R.,Vaamonde, J., Artieda, J., Zubieta, J. L. and Obeso, J. A. (1993). ReXex blepharospasm associated with bilateral basal ganglia lesion. Movement Disord., 8, 198±200.

Lawrence, M. and Redmond, D. (1985). MPTP lesions and dopaminergic drugs alter eye blink rate in African green monkeys. Pharmacol. Biochem. Behav., 38, 869±74.

Leckman, J. F., Knorr, A. M., Rasmusson, A. M. and Cohen, D. J. (1991). Basal ganglia research and Tourette's syndromes. [Letter]

Trends Neurosci., 14, 94.

Leckman, J. F., Walker, D. E. and Cohen, D. J. (1993). Premonitory urges in Tourette's syndrome. Am. J. Psychiatry, 150, 98±102.

Leckman, J. F., Peterson, B. S., Anderson, G. M., Arnsten, A. F. T., Pauls, D. L. and Cohen, D. J. (1997). Pathogenesis of Tourette's syndrome. J. Child Psychol. Psychiatry, 38, 119±42.

Leckman, J. F., Zhang, H., Vitale, A. et al. (1998). Course of tic severity in Tourette's syndrome: the Wrst two decades. Pediatrics, 102, 14±19.