VERBAL FLUENCY AND POSITRON EMISSION TOMOGRAPHIC MAPPING OF REGIONAL CEREBRAL GLUCOSE METABOLISM Michael J. Boivin 1•2 , Bruno Giordani\ Stanley Berent\ David A. Amato 3 , Shirley Lehtinen 1, Robert A. Koeppe 1 , Henry A. Buchtel 1•4 , Norman L. Foster1 , David E. Kuhl 1 CUniversity of Michigan Medical Center, Ann Arbor, MI; 2Spring Arbor College, Spring Arbor, MI; 3Harvard School of Public Health, Boston, MA; 4Veterans Administration Medical Center, Ann Arbor, Ml)
That frontal or temporal lobe lesions can impair word list generation (e.g., verbal fluency tasks) is well documented (Benson, 1979; Benton, 1968; Crock ett, Bilsker, Hurwitz et al., 1986; Miller, 1984; Milner, 1964; Stuss and Benson, 1986). Knowledge of these brain-behavior relationships, however, has been es tablished primarily with brain-injured patients. With the advent of new tech nologies that provide a dynamic functional view of the brain (e.g., positron e mission tomography or PET), there now exists the potential for evaluating the relationship between performance on various neuropsychological tasks and re gional brain activity in normal, intact persons, as well as in patient groups. PET/Neuropsychological research strategies include "activation" para digms, wherein subjects undergo PET scanning while simultaneously being ad ministered a cognitive task. Nonactivation or "resting state" procedures corre late subjects' cortical brain activity measured while they are at rest with per formance on cognitive tasks administered within a reasonable time frame of the actual PET scan. Studies in normal adults with the latter technique have failed to document consistent brain-behavior relationships (Chase et al., 1984b; Duara, Grady, Haxby et al., 1984; Grady, 1984; Haxby, Grady, Duara et al., 1986), though a positive association has been demonstrated between preserved cognitive ability and higher resting cortical activity in patients with dementia related disorders (Berent, Giordani, Lehtinen et al., 1988; Chase, Fedio, Foster et al., 1984a; Foster, Chase, Fedio et al., 1983; Foster, Chase, Mansi et al., 1984; Haxby, Duara, Grady et al., 1985; Karbe, Herholz, Szelies et al., 1989). PET "activation" strategies have been carried out with healthy adults simulta neously measuring both cerebral metabolic activity and performance on the Raven Progressive Matrices Test (Haier, Sieger, Nuechperlein et al., 1988) and an adaptation of a verbal fluency (VF) task (Duara, Chang, Barker et al., 1986; Parks, Loewenstein, Dodrill et al., 1988). In contrast to the resting state studies in neurological disorders, both of these studies found successful test perform ance to be negatively correlated with cortical brain activity, suggesting that the nature of observed brain/behavior relationships may not remain consistent for both diseased and normal brains. The expectation that VF (or some other cognitive performance task) would relate systematically to the brain's resting state, particularly in individuals for Cortex, (1992) 28, 231-239
232
M.J. Boivin and Others
whom the pertinent brain regions are not chronically impaired, may at first ap pear questionable. In support of this hypothesis, correlations between cognitive performance and other non-activation physiological measures in normals have been established using methods other than PET. For example, significant brain behavior interactions have been demonstrated with cognitive performance and evoked potentials (Haier, Robinson and Braden, 1983; Schafer, 1982) and 13 Hz EEG activity (Giannitrapani, 1987, 1988). In addition, in a PET study fo cusing on Huntington's disease patients, Berent and colleagues (1988) present ed the incidental observation that such relationships may exist in normal sub jects, though not in the same direction as might be expected from results with patients. That observation was not the primary focus of their study, and be cause of methodological constraints, could not be conclusively demonstrated. Yet, their preliminary finding was of sufficient interest to inspire this further investigation. Coming to a better understanding of the relationship that might exist in nor mals between a cognitive ability and resting metabolism will enable us to better understand the effects of disease on both behavior and the brain (Berent et al., 1988). The inability to demonstrate significant brain-behavior relationships us ing resting state PET procedures in normals has been attributed to several fac tors, including the restricted range of cognitive ability in normal samples, a re liance on more general measures of cognitive ability, and to the poor resolution in PET imaging of specific cortical regions expected to underlie those neuro psychological functions chosen for evaluation (Haxby et al., 1986). To address these concerns in this study, we prospectively chose to evaluate the relationship between resting brain metabolism in normals and a specific neuropsychological index of performance for which the mediating cortical regions can be clearly distinguished and defined. Thus, VF was chosen as the task, and it was hypoth esized that performance on this task would be reflected in changes in metabo lism of the frontal areas of the brain. MATERIALS AND METHODS
Subjects Thirty-three right-handed normal volunteers (14 males, 19 females) ranging in age from 21 to 71 years (males: mean= 46.9, S.D.= 19.6; females: mean= 42.6, S.D.= 13.5) were re cruited through local advertisements and by contacting spouses of patients included in other PET studies. Participants were thoroughly evaluated with respect to their medical history and physical health and were judged to have no significant medical difficulties or psychological illness (DSM-III-R criteria). The present sample represents a typical range of years of edu cation (mean= 15.5 years, S.D.= 2.0, range 12-18 years) and IQ as measured by the Wechs ler Adult Intelligence Scale-Revised (WAIS-R) (mean= 115.8, S.D.= 14.2, range 82-146). The research was approved by the University of Michigan Medical School Institutional Review Board, and informed consent was obtained from the subjects prior to their inclusion in the study. That same day following the PET scan, subjects were given a version of a VF task (Ben ton, 1968, Kimura, 1984), consisting of asking subjects to quickly name as many words as they could that began with the letter "d, as in dog". The examiner then noted the number of words said during a one minute period. Although VF was the principal measure of interest in the present study, subtests from the WAIS-R, not expected to be directly related to frontal
Verba/fluency and PET
233
cortical function, were also included. This was to provide a more rigorous statistical eval uation of the specificity of VF to the cortical regions examined.
Procedure Patients were scanned in a quiet, dimly lit room with minimal background noise, eyes blindfolded, ears unplugged, while resting, awake and supine. Scans were begun 30 minutes following the intravenous injection of a 18F-FDG bolus (10 mCi). 18F-FDG was synthesized by a method described by Hamacher, Coenen and Stocklin (1986), with a radiochemical pu rity of greater than 950Jo. PET scans were performed with a Cyclotron Corporation PCT 4600A tomograph having an in-plane resolution of II mm full width at half maximum (FWHM) and a Z-axis resolution of 9.5 mm FWHM. A laser was used to align the head along the cantho-meatalline with the head maintained in an immobile position throughout the stu dy. Five planes with 11.5-mm center-to-center separation were imaged simultaneously. For each patient, four sets of scans were taken and included two interleaved sets through the lower brain levels and two interleaved sets through higher brain levels (total of 20 slices). Each slice was separated by 5. 75 mm, and an attenuation correction was calculated by a standard ellipse method modified to account for attenuation from the head holder and skull. Blood samples were collected from the radial artery. Local cerebral metabolic rate for glucose (ICMRglc) was calculated using a three-compartment model and single scan approximation described by Phelps and associates (Phelps, Huang, Hoffman et al. 1979) with gray matter kinetic constants derived from normal subjects (Hawkins, Mazziotta, Phelps et al., 1983) and a lumped constant of 0.51. For this analysis, we selected the image frame where the dorsal portions of the caudate nuclei were metabolically most prominent. This level had the advantage of having anatom ical landmarks which allowed it to be reliably identified by three independent observers. It also encompassed portions of both the frontal and superior temporal cortical regions, areas that have been strongly implicated in the mediation of verbal behavior (Nauta, 1986). At this slice level of the brain, we also evaluated the metabolic activity of the occipital region as a means of control comparison for the brain-behavior relationships we examined, since this cortical area was not expected to be significantly involved in the mediation of verbal fluency. An automated regions of interest program was then used to define a symmetrical cortical ribbon approximately 1.5 em in width adjacent to the outer rim of the brain. The program then divided that ribbon into 16 equal sections, 8 for each hemisphere (LI to L8, Rl to R8), to obtain the average pixel value (absolute metabolic value given in mg/100 g/min) for each section. Cortical sections were then assigned to specific anatomical regions according to a comparison with the brain atlas developed by Matsui and Hirano (1978) (Table 1). Since the variance between subjects has been shown to be considerably less for normal ized regional metabolic values as opposed to absolute values (Duara, Gross-Glenn, Barker et al., 1987), the normalized values for the cortical sections were compared to the cognitive per formance measures. This is also consistent with other PET studies that have examined brain behavior relationships with regional cortical values in a manner similar to the present study (e.g., Haxby et al., 1986; Haier et al., 1988). The metabolic value for a given section was normalized by dividing it by the metabolic value for the entire cortex in that slice, and it is these relative values that have been compared to verbal fluency in the statistical analyses.
RESULTS
Table I contains the Pearson product-moment correlation coefficients of VF and the WAIS-R Vocabulary subtest with the normalized metabolic rate for glucose (CMRglc) for each of the 16 cortical sections for the slice used in the present analysis. Vocabulary was used as a comparison measure of general ver bal ability to demonstrate the specificity of the VF findings. Both the left and right superior (Ll, Rl) and middle frontal (L2, R2) corti
234
M.J. Boivin and Others TABLE I
Cognitive Performance and Regional CMRg/c Pearson Product-Moment Correlation Coefficients for Cortical Regions with Verbal Fluency and with WAIS-R Vocabulary Subtest Cortical regions Left hemisphere• L l. Superior frontal L2. Middle frontalb L3. Superior temporal L4. Superior temporal L5. Middle temporal L6. Inferior temporal L 7. Occipital L8. Occipital Right hemisphere R l. Superior frontal R2. Middle frontalb R3. Superior temporal R4. Superior temporal R5. Middle temporal R6. Inferior temporal R7. Occipital R8. Occipital Whole cortex
Verb. fluency
Vocabulary
-.35* -.35*
-.07
.09
.11
.14 .28 .31 .22 -.21 .06
-.53** - .47* -.18
-.02 -.18 -.31 -.21 -.10 .03 .01
.29 .50** .40* .09 .14
-.04
.08 .19 .29 .27 .10
.09
-.02
• The far left-hand column contains the cortical gyri that coincide with slice 3.40 em above the cantho meatalline for each of the cortical sections obtained with the automated Region of Interest program. b Also portions of the inferior frontal gyrus. * p < .05; ** p < .01.
Left Frontal Cortex 1.2
.
Right Frontal Cortex
.
y"' 1.0954- OJJ037x
.!!
1.2
y"' 1.1578- 0.0067x
f"'.
0.59
.!!
a:
a:
1.1
:::!
:::!
g
1.1
0
....
!I
.
~
i
~
a:
a: 0.9 10
20
0.9
30
0
Verbal Fluency
Left Temporal Cortex (Inferior)
.
1.1
y .. 0.9143 +0.0043X
.
a:
."~
1.1
30
. .. : ..
y .. 0.9589 + 0.0014x
.!!
a:
:::!
....
20
Right Temporal Cortex (Interior)
rz 0.56
.!! 0
10
Verbal Fluency
:::!
.
1.0
0 ....
1.0
i;;
!!
"
a:
I 0.9
f=
0.16
.. .. • •
a:
0.9 20
Verbal Fluency
30
0.8 10
20
30
Verbal Fluency
Fig. l - This figure contains a scattergram plot portraying the relationship between verbal fluency and relative CMRglc for the left and right frontal cortical regions (Sections 1 and 2 combined) as well as the left and right inferior/middle temporal cortical regions (Sections 5 and 6 combined) respectively for the scan slice 3:4 em above the can tho-meatal line.
Verba/fluency and PET
235
cal areas (defined by our regions of interest program outlined above) were found to have significant negative correlations with VF. The left middle (L5) and in ferior temporal (L6) areas had significant positive correlations with VF, while the corresponding regions in the right hemisphere did not. Elimination of in fluential ("outlier") points did not alter the direction of the correlations nor re duce the levels of significance. Correlations computed between the sixteen cor tical sections and the WAIS-R Vocabulary subtest did not result in any statist ically significant coefficients (Table 1). Also, theWAIS-R Full Scale IQ (FSIQ) did not correlate significantly with any of the cortical sections in the present an alysis. The scattergram plots for VF performance and CMRglc for the right frontal region (Sections R1 and R2 combined, see Table 1), left frontal region (Sections L1 and L2 combined) and left temporal region (L5 and L6 combined), respec tively, all depict strong correlations (Figure 1). The scattergram plot for VF and the right temporal region (Sections R5 and R6 combined) depicts a weak cor relation that is not statistically significant. Age did not significantly correlate with either cortical metabolic rate for CMRglc (r = - .02) or VF (r = - .06), suggesting that significant sectional cor relations were not age-related. Further, partialling out both age and FSIQ in a regression analysis only slightly reduced the coefficient value for VF and the left middle temporal cortical region (Table I) to r = + .48 (p < .01), while strength-
Fig. 2
236
M.J. Boivin and Others TABLE II
Principal Component Factor Analysis of Regional Cortical Activity and Cognitive Performance Measures: Orthogonal Transformation with a Varimax Solution Factors Variables Left cortical hemisphere Frontal region Superior temporal region Inferior temporal region Occipital region Right cortical hemisphere Frontal region Superior temporal region Inferior temporal region Occipital region Verbal fluency WAIS-R (Age corrected) Arithmetic Vocabulary Digit Symbol Picture Completion Block Design Picture Arrangement
2 -.78 .68
3
4
-.84 .85
-.83 .85
-.53 .75
.76 .77 .51 .73 .59 .76 .80
ening the negative correlation apparent for the right superior frontal cortical region (r = -.56, p < .01) and left superior frontal region (r = - .38, p < .05). Finally, a principal component factor analysis (orthogonal solution, Vari max rotation) was computed for a pool of PET cortical CMRglc measures and cognitive performance measures combined (Table II). These included the nor malized metabolic scores for eight cortical sections (four in each cortical hem isphere), verbal fluency, and the available WAIS-R subtests (Arithmetic, Vo cabulary, Digit Symbol, Picture Completion, Block Design, Picture Arrange ment). The final solution resulted in four factors having an eigenvalue of 1.50 or greater (Factor 1 = 4.25, Factor 2 = 3.02, Factor 3 = 2.21, Factor 4 = 1.54), and together these accounted for 78ct!o of the total variance. In the summary of this analysis presented below, all variables with a factor loading greater than .50 were said to load significantly for that factor. Although four factors were arrived at in the final solution, the first two fac tors were especially pertinent in corroborating the significant brain-behavior relationships for VF described above. Factor 1 could be labeled "Intellectual Performance" and included all of the Wechsler subtests, but not verbal fluen cy. It's important to note that none of the cortical regions loaded with this fac tor. However, Factor 2 ·included Verbal Fluency along with both left and right frontal and the left inferior temporal cortical regions. Verbal Fluency perform ance was unique among the present cognitive measures in its interaction with regional steady-state cortical metabolism hypothesized to be of import in ver bal performance. Factor 3 included left superior temporal and right inferior temporal cortex, while Factor 4 included right and left occipital cortical re gions.
Verba/fluency and PET
237
DISCUSSION
These results are supportive of recent studies that have used PET activation techniques to find negative correlations between performance on a cognitive task and regional brain metabolism in normal, healthy subjects (e.g., Duara et al., 1986; Haier et al., 1988; Parks et al., 1988). They confirm our initial find ings in normals with non-activation PET measures (Berent et al., 1988). Verbal fluency was correlated negatively with relative metabolic rate in the bilateral frontal cortical regions but correlated positively with relative metabolic rate in the left temporal lobe. No significant relationship was found with the right temporal cortical area. The preliminary nature of these findings and the num ber of computed correlations suggests caution in interpreting the results. On the other hand, as predicted, the correlational analyses indicated that these signi ficant relationships were highly specific and limited to performance on VF, and did not extend to other more global measures of language and intellect (e.g., WAIS-R subtests). Furthermore, the principal component factor analysis in dependently confirmed the correlational analyses. The negative relationship between VF and regional cortical metabolism found in this study might reflect a relative economy of effort and cognitive ef ficiency by the high VF performers, leading to their lower relative frontal lobe metabolic rates as seen most clearly during activation studies (Duara et al., 1986; Haier et al., 1988; Parks et al., 1988). The positive correlations noted for the left temporal areas, on the other hand, could reflect a verbal memory com ponent. That VF was negatively correlated with frontal cortical metabolism in the present study stands in contrast to the fact that bilateral frontal lesions are known to impair performance on such tasks. While this might, at first, seem surprising, predictions of how a dynamic measures of brain will relate to cog nitive performance measures are difficult (Haxby et al., 1986), and relation ships might well differ from normal to abnormal conditions. The nature of the relationship between steady-state regional brain metabolic measures and cog nitive performance may differ dramatically depending on the integrity of the neurological structures involved (Berent et al., 1988; Miller, de Leon, Ferris et al., 1987). One way of viewing the basis for the differences in brain-behavior relation ship expressed among patient and non-patient groups has been presented by Szechtman, Nahmias, Garnett et al. (1988), who compared 18F-FDG PET scans of schizophrenic patients to those of normals. The authors observed stable, dis tinctive cortical profiles or "cerebral metabolic tones" for both groups, sug gesting that the cognitive deficits seen in the schizophrenics could relate to a distortion in the balance of relative activity (metabolic tone) among cortical re gions. Such cortical "profiles" may characterize the relative tension or balance among various neural systems that reflect how well an individual is poised to engage in any of a variety of cognitive tasks. This parallels Luria's (1973, 1980) model of systems of cortical and subcortical working zones in concert with one another for the mediation of complex mental activity. These trait-like cortical
238
M.J. Boivin and Others
profiles may provide the brain-behavior link between PET and cognitive per formance measures taken at a different time and place. ABSTRACf
Impairment in verbal fluency (VF) has been a consistently reported clinical feature of fo cal cerebral deficits in frontal and temporal regions. More recent behavioral activation stu dies with healthy control subjects using positron emission tomography (PET), however, have noted a negative correlation between performance on verbal fluency tasks and regional cort ical activity. To see if this negative relationship extends to steady-state non-activation PET measures, thirty-three healthy adults were given a VF task within a day of their 18F-2-fluoro 2-deoxy-D-glucose PET scan. VF was found to correlate positively with left temporal corti cal region metabolic activity but to correlate negatively with right and left frontal activity. VF was not correlated significantly with right temporal cortical metabolic activity. Some previous studies with normals using behavioral activation paradigms and PET have reported negative correlations between metabolic activity and cognitive performance similar to that reported here. An explanation for the disparate relationships that were observed between frontal and temporal brain areas and VF might be found in the mediation of different task demands by these separate locations, i.e., task planning and/or initiation by frontal regions and verbal memory by the left temporal area.
Acknowledgements. Preparation of this article was supported in part by National Insti tutes of Health Grant NS 15655. We gratefully acknowledge the assistance of the University of Michigan Cyclotron/ P .E. T. facility in the Division of Nuclear Medicine, and of Ms. Patricia Bohland for her help in manuscript preparation. This study was completed by the senior author while serving as a visiting faculty member to the Department of Psychiatry, University of Michigan Medical Center. REFERENCES BENSON, D.F. Aphasia, Alexia, and Agraphia. New York: Churchill Livingstone, 1979. BENTON, A.L. Differential behavioral effects in frontal lobe disease. Neuropsychologia, 6: 53-60, 1968. BERENT, S., GIORDANI, B., LEHTINEN, S., MARKEL, D., PENNY, J.B., BUCHTEL, H.A., STAROSTA-RUB INSTEIN, S., HICHWA, R., and YouNG, A.B. Positron emission tomographic scan investigations of Huntington's disease: cerebral metabolic correlates of cognitive functiton. Annals ofNeurology, 23: 541-546, 1988. CHASE, T.N., FEDIO, P., FOSTER, N.L., BROOKS, R., DI CHIRO, G., and MANSI, L. Wechsler Adult In telligence Scale performance: cortical localization by fluorodeoxyglucose FIB-positron emission tomography. Archives ofNeurology, 41: 1244-1247, 1984a. CHASE, T.N., FOSTER, N.L., FEDIO, P., BROOKS, R., MANSI, L., and DI CHIRO, G. Regional cortical dys function in Alzheimer's disease as determined by positron emission tomography. Annals ofNeurol ogy, 15 (suppl.): Sl70-Sl74, 1984b. CROCKETT, B., BILSKER, D., HURWITZ, T., and KoZAK, J. Clinical utility of three measures of frontal lobe dysfunction in neuropsychiatric samples. International Journal ofNeuroscience, 30: 241-248, 1986. DUARA, R., GRADY, C., HAXBY, J., INGVAR, D., SoKOLOFF, L., MARGOLIN, R.A., MANNING, R.G., Cu TLER, N.R., and RAPOPORT, S.I. Human brain glucose utilization and cognitive function in relation to age. Annals of Neurology, 16: 702-713, 1984. DUARA, R., CHANG, J., BARKER, W., YOSHII, F., and APICELLA, A. Correlation ofregional cerebral me tabolic activation to performance in activating tasks. Neurology, 36 (suppl. 1): 349, 1986. DUARA, R., GROSS-GLENN, K., BARKER, W.W., CHANG, J.Y., APICELLA, A., LOEWENSTEIN, D., and BOOTHE, T. Behavioral activation and the variability of cerebral glucose metabolic measurements. Journal of Cerebral Blood Flow and Metabolism, 7:266-271, 1987. FOSTER, N.L., CHASE, T.N., FEDIO, P., PATRONAS, N.J., BROOKS, R.A., and DI CHIRO, G. Alzheimer's disease: focal cortical changes shown by positron emission tomography. Neurology, 33: 961-965, 1983. FOSTER, N.L., CHASE, T.N., MANSI, L., BROOKS, R., FEDIO, P., PATRONAS, N.J., and DI CHIRO, G. Cortical abnormalities in Alzheimer's disease. Annals ofNeurology, 16: 649-654, 1984. GIANNITRAPANI, D. The Electrophysiology of Intellectual Functions. New York: Karger, 1985.
Verba/fluency and PET
239
GIANNITRAPANI, D. The Role of 13-Hz Activity in Mentation. Basel: Karger, 1988. GILMAN, S., MARKEL, D.S., KOEPPE, R.A., JUNCK, L., KLUIN, K.J ., GEBARSKI, S.S., and HICHWA, R.D. Cerebellar and brain stem hypometabolism in olivopontocerebellar atrophy detected with positron emission tomography. Annals ofNeurology, 23: 223-230, 1988. GRADY, C.L. Neuropsychology and cerebral metabolism in normal aging. In N.R. Cutler (Moderator), Brain imaging: aging and dementia. Annals ofInternal Medicine, 101: 358-360, 1984. HAlER, R.J ., RoBINSON, D.L., and BRADEN, W. Electrical potentials of the cerebral cortex and psychom etric intelligence. Personality and Individual Differences, 4: 591-599, 1983. HAlER, R.J., SIEGER, E.V., NUECHPERLEIN, K.H., HAZLETT, E., Wu, J.C., PAEK, J., BROWNING, H.L., and BucHSBAUM, M.S. Cortical glucose metabolic rate correlates of abstract reasoning and attention studied with positron emission tomography.1ntelligence, 12: 199-217, 1988. HAMACHER, K., COENEN, H.H., and STOCKLIN, G. Efficient stereospecific synthesis of NCA-2(18F)-fluo ro-2-deoxy-D-glucose using aminopolyether supported direct nucleophilic substitution. Journal of Nuclear Medicine, 27: 235-238, 1986. HAWKINS, R.A., MAZZIOTTA, J.C., PHELPS, M.E., HUANG, S.-C., KUHL, D.E., CARSON, R.E., METTER, R.J., and RIEGE, W.H. Cerebral glucose metabolism as a function of age in man: influence of the rate constants in the fluorodeoxyglucose method. Journal of Cerebral Blood Flow and Metabolism, 3: 250-253, 1983. HAXBY, J.V., DUARA, R., GRADY, C.L., CUTLER, N.R., and RAPoPORT, S.l. Relations between neuro psychological and cerebral metabolic asymmetries in early Alzheimer's disease. Journal of Cerebral Blood Flow and Metabolism, 5: 193-200, 1985. HAXBY, J.V., GRADY, C.L., DUARA, R., ROBERTSON-TCHABO, E., KOZIARZ, B., CUTLER, N.R., and RA POPORT, S.I. Relations among age, visual memory, and resting cerebral metabolism in 40 healthy men. Brain and Cognition, 5: 412-427, 1986. KARBE, H., HERHOLZ, K., SZELIES, B., PAWLICK, G., WIENHARD, K., and HEISS, W.-D. Regional meta bolic correlates of Token test results in cortical and subcortical left hemispheric infarction. Neurol ogy,39: 1083-1088,1989. KIMURA, D. Neuropsychology test procedures. London, Ontario: DK Consultants, 1984. LURIA, A.R. The Working Brain: An Introduction to Neuropsychology. New York: Basic Books, 1973. LURIA, A.R. Higher Cortical Functions in Man (2nd ed.). New York: Basic Books, 1980. MATSUI, T., and HIRANO, A. An Atlas ojthe Human Brain for Computerized Tomography. New York: lgaku-Shoin, 1978. MILLER, E. Verbal fluency as a function of a measure of verbal intelligence in relation to different types of cerebral pathology: British Journal of Clinical Psychology, 23: 53-57, 1984. MILLER, J.D., DE LEON, J .J., FERRIS, S.H., KLUGER, A., GEORGE, A.E., REISBERG, B., SACHS, H.J., and WoLF, A.P. Abnormal temporal lobe response in Alzheimer's disease during cognitive processing as measured by ''C-2-deoxy-d-glucose and PET. Journal of Cerebral Blood Flow and Metabolism, 7: 248-251, 1987. MILNER, B. Some effects of frontal lobectomy in man. In J .M. Warren and K. Akert (Eds.), The Frontal Granular Cortex and Behavior. New York: McGraw-Hill, 1964, pp. 313-314. NAUTA, W. Fundamental Neuroanatomy. New York: Freeman Press, 1986. PARKS, R.W., LOEWENSTEIN, D.A., DoDRILL, K.L., BARKER, W.W., YOSHII, F., CHANG, J.Y., EMRAN, A., APICELLA, A., SHERAMATA, W.A., and DUARA, R. Cerebral metabolic effects of a verbal fluency test: a PET scan study. Journal of Clinical and Experimental Neuropsychology, 5: 565-515, 1988. PHELPS, M.E., HUANG, S-C, HOFFMAN, E.J., SELIN, C.S., SOKOLOFF, L., and KUHL, D.E. Tomographic measurement of local cerebral glucose metabolic rate in humans with (F-18)-2-fluoro-2-deoxygluco se: Validation of method. Annals ofNeurology, 6: 371-388, 1979. ScHAFER, E.W .P. Neural adaptability: a biological determinant of behavioral intelligence. International Journal ofNeuroscience, 17: 133-191, 1982. Sruss, D.T., and BENSON, D.F. The Frontal Lobes. New York: Raven Press, 1986. SZECHTMAN, H., NAHMIAS, C., GARNETT, E.S., FIRNAU, G., BROWN, G.M., KAPLAN, R.D., and CLEG HORN, J .M. Effect of neuroleptics on altered cerebral glucose metabolism in schizophrenia. Archives of General Psychiatry, 45: 523-532, 1988. Dr. Bruno Giordani, Neuropsychology Program, Rm. 480, Med Inn, Box 0840, University of Michigan Medical Center, Ann Arbor, Mi chigan 48109-0840, U.S.A.
That frontal or temporal lobe lesions can impair word list generation (e.g., verbal fluency tasks) is well documented (Benson, 1979; Benton, 1968; Crock ett, Bilsker, Hurwitz et al., 1986; Miller, 1984; Milner, 1964; Stuss and Benson, 1986). Knowledge of these brain-behavior relationships, however, has been es tablished primarily with brain-injured patients. With the advent of new tech nologies that provide a dynamic functional view of the brain (e.g., positron e mission tomography or PET), there now exists the potential for evaluating the relationship between performance on various neuropsychological tasks and re gional brain activity in normal, intact persons, as well as in patient groups. PET/Neuropsychological research strategies include "activation" para digms, wherein subjects undergo PET scanning while simultaneously being ad ministered a cognitive task. Nonactivation or "resting state" procedures corre late subjects' cortical brain activity measured while they are at rest with per formance on cognitive tasks administered within a reasonable time frame of the actual PET scan. Studies in normal adults with the latter technique have failed to document consistent brain-behavior relationships (Chase et al., 1984b; Duara, Grady, Haxby et al., 1984; Grady, 1984; Haxby, Grady, Duara et al., 1986), though a positive association has been demonstrated between preserved cognitive ability and higher resting cortical activity in patients with dementia related disorders (Berent, Giordani, Lehtinen et al., 1988; Chase, Fedio, Foster et al., 1984a; Foster, Chase, Fedio et al., 1983; Foster, Chase, Mansi et al., 1984; Haxby, Duara, Grady et al., 1985; Karbe, Herholz, Szelies et al., 1989). PET "activation" strategies have been carried out with healthy adults simulta neously measuring both cerebral metabolic activity and performance on the Raven Progressive Matrices Test (Haier, Sieger, Nuechperlein et al., 1988) and an adaptation of a verbal fluency (VF) task (Duara, Chang, Barker et al., 1986; Parks, Loewenstein, Dodrill et al., 1988). In contrast to the resting state studies in neurological disorders, both of these studies found successful test perform ance to be negatively correlated with cortical brain activity, suggesting that the nature of observed brain/behavior relationships may not remain consistent for both diseased and normal brains. The expectation that VF (or some other cognitive performance task) would relate systematically to the brain's resting state, particularly in individuals for Cortex, (1992) 28, 231-239
232
M.J. Boivin and Others
whom the pertinent brain regions are not chronically impaired, may at first ap pear questionable. In support of this hypothesis, correlations between cognitive performance and other non-activation physiological measures in normals have been established using methods other than PET. For example, significant brain behavior interactions have been demonstrated with cognitive performance and evoked potentials (Haier, Robinson and Braden, 1983; Schafer, 1982) and 13 Hz EEG activity (Giannitrapani, 1987, 1988). In addition, in a PET study fo cusing on Huntington's disease patients, Berent and colleagues (1988) present ed the incidental observation that such relationships may exist in normal sub jects, though not in the same direction as might be expected from results with patients. That observation was not the primary focus of their study, and be cause of methodological constraints, could not be conclusively demonstrated. Yet, their preliminary finding was of sufficient interest to inspire this further investigation. Coming to a better understanding of the relationship that might exist in nor mals between a cognitive ability and resting metabolism will enable us to better understand the effects of disease on both behavior and the brain (Berent et al., 1988). The inability to demonstrate significant brain-behavior relationships us ing resting state PET procedures in normals has been attributed to several fac tors, including the restricted range of cognitive ability in normal samples, a re liance on more general measures of cognitive ability, and to the poor resolution in PET imaging of specific cortical regions expected to underlie those neuro psychological functions chosen for evaluation (Haxby et al., 1986). To address these concerns in this study, we prospectively chose to evaluate the relationship between resting brain metabolism in normals and a specific neuropsychological index of performance for which the mediating cortical regions can be clearly distinguished and defined. Thus, VF was chosen as the task, and it was hypoth esized that performance on this task would be reflected in changes in metabo lism of the frontal areas of the brain. MATERIALS AND METHODS
Subjects Thirty-three right-handed normal volunteers (14 males, 19 females) ranging in age from 21 to 71 years (males: mean= 46.9, S.D.= 19.6; females: mean= 42.6, S.D.= 13.5) were re cruited through local advertisements and by contacting spouses of patients included in other PET studies. Participants were thoroughly evaluated with respect to their medical history and physical health and were judged to have no significant medical difficulties or psychological illness (DSM-III-R criteria). The present sample represents a typical range of years of edu cation (mean= 15.5 years, S.D.= 2.0, range 12-18 years) and IQ as measured by the Wechs ler Adult Intelligence Scale-Revised (WAIS-R) (mean= 115.8, S.D.= 14.2, range 82-146). The research was approved by the University of Michigan Medical School Institutional Review Board, and informed consent was obtained from the subjects prior to their inclusion in the study. That same day following the PET scan, subjects were given a version of a VF task (Ben ton, 1968, Kimura, 1984), consisting of asking subjects to quickly name as many words as they could that began with the letter "d, as in dog". The examiner then noted the number of words said during a one minute period. Although VF was the principal measure of interest in the present study, subtests from the WAIS-R, not expected to be directly related to frontal
Verba/fluency and PET
233
cortical function, were also included. This was to provide a more rigorous statistical eval uation of the specificity of VF to the cortical regions examined.
Procedure Patients were scanned in a quiet, dimly lit room with minimal background noise, eyes blindfolded, ears unplugged, while resting, awake and supine. Scans were begun 30 minutes following the intravenous injection of a 18F-FDG bolus (10 mCi). 18F-FDG was synthesized by a method described by Hamacher, Coenen and Stocklin (1986), with a radiochemical pu rity of greater than 950Jo. PET scans were performed with a Cyclotron Corporation PCT 4600A tomograph having an in-plane resolution of II mm full width at half maximum (FWHM) and a Z-axis resolution of 9.5 mm FWHM. A laser was used to align the head along the cantho-meatalline with the head maintained in an immobile position throughout the stu dy. Five planes with 11.5-mm center-to-center separation were imaged simultaneously. For each patient, four sets of scans were taken and included two interleaved sets through the lower brain levels and two interleaved sets through higher brain levels (total of 20 slices). Each slice was separated by 5. 75 mm, and an attenuation correction was calculated by a standard ellipse method modified to account for attenuation from the head holder and skull. Blood samples were collected from the radial artery. Local cerebral metabolic rate for glucose (ICMRglc) was calculated using a three-compartment model and single scan approximation described by Phelps and associates (Phelps, Huang, Hoffman et al. 1979) with gray matter kinetic constants derived from normal subjects (Hawkins, Mazziotta, Phelps et al., 1983) and a lumped constant of 0.51. For this analysis, we selected the image frame where the dorsal portions of the caudate nuclei were metabolically most prominent. This level had the advantage of having anatom ical landmarks which allowed it to be reliably identified by three independent observers. It also encompassed portions of both the frontal and superior temporal cortical regions, areas that have been strongly implicated in the mediation of verbal behavior (Nauta, 1986). At this slice level of the brain, we also evaluated the metabolic activity of the occipital region as a means of control comparison for the brain-behavior relationships we examined, since this cortical area was not expected to be significantly involved in the mediation of verbal fluency. An automated regions of interest program was then used to define a symmetrical cortical ribbon approximately 1.5 em in width adjacent to the outer rim of the brain. The program then divided that ribbon into 16 equal sections, 8 for each hemisphere (LI to L8, Rl to R8), to obtain the average pixel value (absolute metabolic value given in mg/100 g/min) for each section. Cortical sections were then assigned to specific anatomical regions according to a comparison with the brain atlas developed by Matsui and Hirano (1978) (Table 1). Since the variance between subjects has been shown to be considerably less for normal ized regional metabolic values as opposed to absolute values (Duara, Gross-Glenn, Barker et al., 1987), the normalized values for the cortical sections were compared to the cognitive per formance measures. This is also consistent with other PET studies that have examined brain behavior relationships with regional cortical values in a manner similar to the present study (e.g., Haxby et al., 1986; Haier et al., 1988). The metabolic value for a given section was normalized by dividing it by the metabolic value for the entire cortex in that slice, and it is these relative values that have been compared to verbal fluency in the statistical analyses.
RESULTS
Table I contains the Pearson product-moment correlation coefficients of VF and the WAIS-R Vocabulary subtest with the normalized metabolic rate for glucose (CMRglc) for each of the 16 cortical sections for the slice used in the present analysis. Vocabulary was used as a comparison measure of general ver bal ability to demonstrate the specificity of the VF findings. Both the left and right superior (Ll, Rl) and middle frontal (L2, R2) corti
234
M.J. Boivin and Others TABLE I
Cognitive Performance and Regional CMRg/c Pearson Product-Moment Correlation Coefficients for Cortical Regions with Verbal Fluency and with WAIS-R Vocabulary Subtest Cortical regions Left hemisphere• L l. Superior frontal L2. Middle frontalb L3. Superior temporal L4. Superior temporal L5. Middle temporal L6. Inferior temporal L 7. Occipital L8. Occipital Right hemisphere R l. Superior frontal R2. Middle frontalb R3. Superior temporal R4. Superior temporal R5. Middle temporal R6. Inferior temporal R7. Occipital R8. Occipital Whole cortex
Verb. fluency
Vocabulary
-.35* -.35*
-.07
.09
.11
.14 .28 .31 .22 -.21 .06
-.53** - .47* -.18
-.02 -.18 -.31 -.21 -.10 .03 .01
.29 .50** .40* .09 .14
-.04
.08 .19 .29 .27 .10
.09
-.02
• The far left-hand column contains the cortical gyri that coincide with slice 3.40 em above the cantho meatalline for each of the cortical sections obtained with the automated Region of Interest program. b Also portions of the inferior frontal gyrus. * p < .05; ** p < .01.
Left Frontal Cortex 1.2
.
Right Frontal Cortex
.
y"' 1.0954- OJJ037x
.!!
1.2
y"' 1.1578- 0.0067x
f"'.
0.59
.!!
a:
a:
1.1
:::!
:::!
g
1.1
0
....
!I
.
~
i
~
a:
a: 0.9 10
20
0.9
30
0
Verbal Fluency
Left Temporal Cortex (Inferior)
.
1.1
y .. 0.9143 +0.0043X
.
a:
."~
1.1
30
. .. : ..
y .. 0.9589 + 0.0014x
.!!
a:
:::!
....
20
Right Temporal Cortex (Interior)
rz 0.56
.!! 0
10
Verbal Fluency
:::!
.
1.0
0 ....
1.0
i;;
!!
"
a:
I 0.9
f=
0.16
.. .. • •
a:
0.9 20
Verbal Fluency
30
0.8 10
20
30
Verbal Fluency
Fig. l - This figure contains a scattergram plot portraying the relationship between verbal fluency and relative CMRglc for the left and right frontal cortical regions (Sections 1 and 2 combined) as well as the left and right inferior/middle temporal cortical regions (Sections 5 and 6 combined) respectively for the scan slice 3:4 em above the can tho-meatal line.
Verba/fluency and PET
235
cal areas (defined by our regions of interest program outlined above) were found to have significant negative correlations with VF. The left middle (L5) and in ferior temporal (L6) areas had significant positive correlations with VF, while the corresponding regions in the right hemisphere did not. Elimination of in fluential ("outlier") points did not alter the direction of the correlations nor re duce the levels of significance. Correlations computed between the sixteen cor tical sections and the WAIS-R Vocabulary subtest did not result in any statist ically significant coefficients (Table 1). Also, theWAIS-R Full Scale IQ (FSIQ) did not correlate significantly with any of the cortical sections in the present an alysis. The scattergram plots for VF performance and CMRglc for the right frontal region (Sections R1 and R2 combined, see Table 1), left frontal region (Sections L1 and L2 combined) and left temporal region (L5 and L6 combined), respec tively, all depict strong correlations (Figure 1). The scattergram plot for VF and the right temporal region (Sections R5 and R6 combined) depicts a weak cor relation that is not statistically significant. Age did not significantly correlate with either cortical metabolic rate for CMRglc (r = - .02) or VF (r = - .06), suggesting that significant sectional cor relations were not age-related. Further, partialling out both age and FSIQ in a regression analysis only slightly reduced the coefficient value for VF and the left middle temporal cortical region (Table I) to r = + .48 (p < .01), while strength-
Fig. 2
236
M.J. Boivin and Others TABLE II
Principal Component Factor Analysis of Regional Cortical Activity and Cognitive Performance Measures: Orthogonal Transformation with a Varimax Solution Factors Variables Left cortical hemisphere Frontal region Superior temporal region Inferior temporal region Occipital region Right cortical hemisphere Frontal region Superior temporal region Inferior temporal region Occipital region Verbal fluency WAIS-R (Age corrected) Arithmetic Vocabulary Digit Symbol Picture Completion Block Design Picture Arrangement
2 -.78 .68
3
4
-.84 .85
-.83 .85
-.53 .75
.76 .77 .51 .73 .59 .76 .80
ening the negative correlation apparent for the right superior frontal cortical region (r = -.56, p < .01) and left superior frontal region (r = - .38, p < .05). Finally, a principal component factor analysis (orthogonal solution, Vari max rotation) was computed for a pool of PET cortical CMRglc measures and cognitive performance measures combined (Table II). These included the nor malized metabolic scores for eight cortical sections (four in each cortical hem isphere), verbal fluency, and the available WAIS-R subtests (Arithmetic, Vo cabulary, Digit Symbol, Picture Completion, Block Design, Picture Arrange ment). The final solution resulted in four factors having an eigenvalue of 1.50 or greater (Factor 1 = 4.25, Factor 2 = 3.02, Factor 3 = 2.21, Factor 4 = 1.54), and together these accounted for 78ct!o of the total variance. In the summary of this analysis presented below, all variables with a factor loading greater than .50 were said to load significantly for that factor. Although four factors were arrived at in the final solution, the first two fac tors were especially pertinent in corroborating the significant brain-behavior relationships for VF described above. Factor 1 could be labeled "Intellectual Performance" and included all of the Wechsler subtests, but not verbal fluen cy. It's important to note that none of the cortical regions loaded with this fac tor. However, Factor 2 ·included Verbal Fluency along with both left and right frontal and the left inferior temporal cortical regions. Verbal Fluency perform ance was unique among the present cognitive measures in its interaction with regional steady-state cortical metabolism hypothesized to be of import in ver bal performance. Factor 3 included left superior temporal and right inferior temporal cortex, while Factor 4 included right and left occipital cortical re gions.
Verba/fluency and PET
237
DISCUSSION
These results are supportive of recent studies that have used PET activation techniques to find negative correlations between performance on a cognitive task and regional brain metabolism in normal, healthy subjects (e.g., Duara et al., 1986; Haier et al., 1988; Parks et al., 1988). They confirm our initial find ings in normals with non-activation PET measures (Berent et al., 1988). Verbal fluency was correlated negatively with relative metabolic rate in the bilateral frontal cortical regions but correlated positively with relative metabolic rate in the left temporal lobe. No significant relationship was found with the right temporal cortical area. The preliminary nature of these findings and the num ber of computed correlations suggests caution in interpreting the results. On the other hand, as predicted, the correlational analyses indicated that these signi ficant relationships were highly specific and limited to performance on VF, and did not extend to other more global measures of language and intellect (e.g., WAIS-R subtests). Furthermore, the principal component factor analysis in dependently confirmed the correlational analyses. The negative relationship between VF and regional cortical metabolism found in this study might reflect a relative economy of effort and cognitive ef ficiency by the high VF performers, leading to their lower relative frontal lobe metabolic rates as seen most clearly during activation studies (Duara et al., 1986; Haier et al., 1988; Parks et al., 1988). The positive correlations noted for the left temporal areas, on the other hand, could reflect a verbal memory com ponent. That VF was negatively correlated with frontal cortical metabolism in the present study stands in contrast to the fact that bilateral frontal lesions are known to impair performance on such tasks. While this might, at first, seem surprising, predictions of how a dynamic measures of brain will relate to cog nitive performance measures are difficult (Haxby et al., 1986), and relation ships might well differ from normal to abnormal conditions. The nature of the relationship between steady-state regional brain metabolic measures and cog nitive performance may differ dramatically depending on the integrity of the neurological structures involved (Berent et al., 1988; Miller, de Leon, Ferris et al., 1987). One way of viewing the basis for the differences in brain-behavior relation ship expressed among patient and non-patient groups has been presented by Szechtman, Nahmias, Garnett et al. (1988), who compared 18F-FDG PET scans of schizophrenic patients to those of normals. The authors observed stable, dis tinctive cortical profiles or "cerebral metabolic tones" for both groups, sug gesting that the cognitive deficits seen in the schizophrenics could relate to a distortion in the balance of relative activity (metabolic tone) among cortical re gions. Such cortical "profiles" may characterize the relative tension or balance among various neural systems that reflect how well an individual is poised to engage in any of a variety of cognitive tasks. This parallels Luria's (1973, 1980) model of systems of cortical and subcortical working zones in concert with one another for the mediation of complex mental activity. These trait-like cortical
238
M.J. Boivin and Others
profiles may provide the brain-behavior link between PET and cognitive per formance measures taken at a different time and place. ABSTRACf
Impairment in verbal fluency (VF) has been a consistently reported clinical feature of fo cal cerebral deficits in frontal and temporal regions. More recent behavioral activation stu dies with healthy control subjects using positron emission tomography (PET), however, have noted a negative correlation between performance on verbal fluency tasks and regional cort ical activity. To see if this negative relationship extends to steady-state non-activation PET measures, thirty-three healthy adults were given a VF task within a day of their 18F-2-fluoro 2-deoxy-D-glucose PET scan. VF was found to correlate positively with left temporal corti cal region metabolic activity but to correlate negatively with right and left frontal activity. VF was not correlated significantly with right temporal cortical metabolic activity. Some previous studies with normals using behavioral activation paradigms and PET have reported negative correlations between metabolic activity and cognitive performance similar to that reported here. An explanation for the disparate relationships that were observed between frontal and temporal brain areas and VF might be found in the mediation of different task demands by these separate locations, i.e., task planning and/or initiation by frontal regions and verbal memory by the left temporal area.
Acknowledgements. Preparation of this article was supported in part by National Insti tutes of Health Grant NS 15655. We gratefully acknowledge the assistance of the University of Michigan Cyclotron/ P .E. T. facility in the Division of Nuclear Medicine, and of Ms. Patricia Bohland for her help in manuscript preparation. This study was completed by the senior author while serving as a visiting faculty member to the Department of Psychiatry, University of Michigan Medical Center. REFERENCES BENSON, D.F. Aphasia, Alexia, and Agraphia. New York: Churchill Livingstone, 1979. BENTON, A.L. Differential behavioral effects in frontal lobe disease. Neuropsychologia, 6: 53-60, 1968. BERENT, S., GIORDANI, B., LEHTINEN, S., MARKEL, D., PENNY, J.B., BUCHTEL, H.A., STAROSTA-RUB INSTEIN, S., HICHWA, R., and YouNG, A.B. Positron emission tomographic scan investigations of Huntington's disease: cerebral metabolic correlates of cognitive functiton. Annals ofNeurology, 23: 541-546, 1988. CHASE, T.N., FEDIO, P., FOSTER, N.L., BROOKS, R., DI CHIRO, G., and MANSI, L. Wechsler Adult In telligence Scale performance: cortical localization by fluorodeoxyglucose FIB-positron emission tomography. Archives ofNeurology, 41: 1244-1247, 1984a. CHASE, T.N., FOSTER, N.L., FEDIO, P., BROOKS, R., MANSI, L., and DI CHIRO, G. Regional cortical dys function in Alzheimer's disease as determined by positron emission tomography. Annals ofNeurol ogy, 15 (suppl.): Sl70-Sl74, 1984b. CROCKETT, B., BILSKER, D., HURWITZ, T., and KoZAK, J. Clinical utility of three measures of frontal lobe dysfunction in neuropsychiatric samples. International Journal ofNeuroscience, 30: 241-248, 1986. DUARA, R., GRADY, C., HAXBY, J., INGVAR, D., SoKOLOFF, L., MARGOLIN, R.A., MANNING, R.G., Cu TLER, N.R., and RAPOPORT, S.I. Human brain glucose utilization and cognitive function in relation to age. Annals of Neurology, 16: 702-713, 1984. DUARA, R., CHANG, J., BARKER, W., YOSHII, F., and APICELLA, A. Correlation ofregional cerebral me tabolic activation to performance in activating tasks. Neurology, 36 (suppl. 1): 349, 1986. DUARA, R., GROSS-GLENN, K., BARKER, W.W., CHANG, J.Y., APICELLA, A., LOEWENSTEIN, D., and BOOTHE, T. Behavioral activation and the variability of cerebral glucose metabolic measurements. Journal of Cerebral Blood Flow and Metabolism, 7:266-271, 1987. FOSTER, N.L., CHASE, T.N., FEDIO, P., PATRONAS, N.J., BROOKS, R.A., and DI CHIRO, G. Alzheimer's disease: focal cortical changes shown by positron emission tomography. Neurology, 33: 961-965, 1983. FOSTER, N.L., CHASE, T.N., MANSI, L., BROOKS, R., FEDIO, P., PATRONAS, N.J., and DI CHIRO, G. Cortical abnormalities in Alzheimer's disease. Annals ofNeurology, 16: 649-654, 1984. GIANNITRAPANI, D. The Electrophysiology of Intellectual Functions. New York: Karger, 1985.
Verba/fluency and PET
239
GIANNITRAPANI, D. The Role of 13-Hz Activity in Mentation. Basel: Karger, 1988. GILMAN, S., MARKEL, D.S., KOEPPE, R.A., JUNCK, L., KLUIN, K.J ., GEBARSKI, S.S., and HICHWA, R.D. Cerebellar and brain stem hypometabolism in olivopontocerebellar atrophy detected with positron emission tomography. Annals ofNeurology, 23: 223-230, 1988. GRADY, C.L. Neuropsychology and cerebral metabolism in normal aging. In N.R. Cutler (Moderator), Brain imaging: aging and dementia. Annals ofInternal Medicine, 101: 358-360, 1984. HAlER, R.J ., RoBINSON, D.L., and BRADEN, W. Electrical potentials of the cerebral cortex and psychom etric intelligence. Personality and Individual Differences, 4: 591-599, 1983. HAlER, R.J., SIEGER, E.V., NUECHPERLEIN, K.H., HAZLETT, E., Wu, J.C., PAEK, J., BROWNING, H.L., and BucHSBAUM, M.S. Cortical glucose metabolic rate correlates of abstract reasoning and attention studied with positron emission tomography.1ntelligence, 12: 199-217, 1988. HAMACHER, K., COENEN, H.H., and STOCKLIN, G. Efficient stereospecific synthesis of NCA-2(18F)-fluo ro-2-deoxy-D-glucose using aminopolyether supported direct nucleophilic substitution. Journal of Nuclear Medicine, 27: 235-238, 1986. HAWKINS, R.A., MAZZIOTTA, J.C., PHELPS, M.E., HUANG, S.-C., KUHL, D.E., CARSON, R.E., METTER, R.J., and RIEGE, W.H. Cerebral glucose metabolism as a function of age in man: influence of the rate constants in the fluorodeoxyglucose method. Journal of Cerebral Blood Flow and Metabolism, 3: 250-253, 1983. HAXBY, J.V., DUARA, R., GRADY, C.L., CUTLER, N.R., and RAPoPORT, S.l. Relations between neuro psychological and cerebral metabolic asymmetries in early Alzheimer's disease. Journal of Cerebral Blood Flow and Metabolism, 5: 193-200, 1985. HAXBY, J.V., GRADY, C.L., DUARA, R., ROBERTSON-TCHABO, E., KOZIARZ, B., CUTLER, N.R., and RA POPORT, S.I. Relations among age, visual memory, and resting cerebral metabolism in 40 healthy men. Brain and Cognition, 5: 412-427, 1986. KARBE, H., HERHOLZ, K., SZELIES, B., PAWLICK, G., WIENHARD, K., and HEISS, W.-D. Regional meta bolic correlates of Token test results in cortical and subcortical left hemispheric infarction. Neurol ogy,39: 1083-1088,1989. KIMURA, D. Neuropsychology test procedures. London, Ontario: DK Consultants, 1984. LURIA, A.R. The Working Brain: An Introduction to Neuropsychology. New York: Basic Books, 1973. LURIA, A.R. Higher Cortical Functions in Man (2nd ed.). New York: Basic Books, 1980. MATSUI, T., and HIRANO, A. An Atlas ojthe Human Brain for Computerized Tomography. New York: lgaku-Shoin, 1978. MILLER, E. Verbal fluency as a function of a measure of verbal intelligence in relation to different types of cerebral pathology: British Journal of Clinical Psychology, 23: 53-57, 1984. MILLER, J.D., DE LEON, J .J., FERRIS, S.H., KLUGER, A., GEORGE, A.E., REISBERG, B., SACHS, H.J., and WoLF, A.P. Abnormal temporal lobe response in Alzheimer's disease during cognitive processing as measured by ''C-2-deoxy-d-glucose and PET. Journal of Cerebral Blood Flow and Metabolism, 7: 248-251, 1987. MILNER, B. Some effects of frontal lobectomy in man. In J .M. Warren and K. Akert (Eds.), The Frontal Granular Cortex and Behavior. New York: McGraw-Hill, 1964, pp. 313-314. NAUTA, W. Fundamental Neuroanatomy. New York: Freeman Press, 1986. PARKS, R.W., LOEWENSTEIN, D.A., DoDRILL, K.L., BARKER, W.W., YOSHII, F., CHANG, J.Y., EMRAN, A., APICELLA, A., SHERAMATA, W.A., and DUARA, R. Cerebral metabolic effects of a verbal fluency test: a PET scan study. Journal of Clinical and Experimental Neuropsychology, 5: 565-515, 1988. PHELPS, M.E., HUANG, S-C, HOFFMAN, E.J., SELIN, C.S., SOKOLOFF, L., and KUHL, D.E. Tomographic measurement of local cerebral glucose metabolic rate in humans with (F-18)-2-fluoro-2-deoxygluco se: Validation of method. Annals ofNeurology, 6: 371-388, 1979. ScHAFER, E.W .P. Neural adaptability: a biological determinant of behavioral intelligence. International Journal ofNeuroscience, 17: 133-191, 1982. Sruss, D.T., and BENSON, D.F. The Frontal Lobes. New York: Raven Press, 1986. SZECHTMAN, H., NAHMIAS, C., GARNETT, E.S., FIRNAU, G., BROWN, G.M., KAPLAN, R.D., and CLEG HORN, J .M. Effect of neuroleptics on altered cerebral glucose metabolism in schizophrenia. Archives of General Psychiatry, 45: 523-532, 1988. Dr. Bruno Giordani, Neuropsychology Program, Rm. 480, Med Inn, Box 0840, University of Michigan Medical Center, Ann Arbor, Mi chigan 48109-0840, U.S.A.
