Applied Neuropsychology 2001, Vol. 8, No. 1, 23–30

Copyright 2001 by Lawrence Erlbaum Associates, Inc.

Quantitative Assessment of Cerebral Ventricular Volumes in Chronic Fatigue Syndrome

VOLUMETRIC ANALYSIS IN CHRONIC FATIGUE LANGE SYNDROME ET AL.

Gudrun Lange Departments of Psychiatry and Radiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA

Andrei I. Holodny Department of Radiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA

John DeLuca Departments of Neuroscience and Physical Medicine and Rehabilitation, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA, and Kessler Medical Rehabilitation Research and Education Corporation, West Orange, New Jersey, USA

Huey-Jen Lee Department of Radiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA

Xiao-Hong Michelle Yan Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania, USA

Jason Steffener Department of Psychiatry, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA

Benjamin H. Natelson Department of Neuroscience, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey, USA Previous qualitative volumetric assessment of lateral ventricular enlargement in chronic fatigue syndrome (CFS) has provided evidence for subtle structural changes in the brains of some individuals with CFS. The aim of this pilot study was to determine whether a more sensitive quantitative assessment of the lateral ventricular system would support the previous qualitative findings. In this study, we compared the total lateral ventricular volume, as well as the right and left hemisphere subcomponents in 28 participants with CFS and 15 controls. Ventricular volumes in the CFS group were larger than in control groups, a difference that approached statistical significance. Group differences in ventricular asymmetry were not observed. The results of this study provide further evidence of subtle pathophysiological changes in the brains of participants with CFS. Key words: chronic fatigue syndrome, lateral ventricle, volumetric assessment, morphometry, brain pathology This study was supported in part by Grant U01–AI–32247 establishing a Chronic Fatigue Syndrome Cooperative Research Center at the New Jersey Medical School. Requests for reprints should be sent to Gudrun Lange, UMDNJ-New Jersey Medical School, Department of Psychiatry, ADMC 1410, 30 Bergen Street, Newark, NJ 07107, USA. E-mail: [email protected]

23

LANGE ET AL.

Chronic fatigue syndrome (CFS) is a medically unexplained illness diagnosed by clinical case definition (Fukuda et al., 1994). Its characteristic features include severe fatigue, infectious and rheumatological symptoms, affective disturbances, and cognitive dysfunction. There is now converging evidence suggesting structural cerebral changes exist in persons with CFS. Specifically, small nonspecific MRI white matter lesions, primarily in the frontal lobes, have been reported in several studies. These white matter findings appear to occur most frequently in CFS patients without frank psychopathology (for a review, see Lange, Wang, DeLuca, & Natelson, 1998). In addition, a qualitative study reported ventricular enlargement in 36% of CFS participants with abnormal MRI scans (Natelson, Cohen, Brassloff, & Lee, 1993). The aim of this pilot study was to more rigorously examine whether lateral ventricular volume is increased in CFS patients compared to normal controls, utilizing a sensitive quantitative morphometric measurement technique. Lateral ventricular volume is one of the most common measurements in volumetric assessment of the brain. Enlargement of the ventricular system is nonspecific and generally regarded as an indirect measure of white matter loss, because much of the ventricular system is surrounded by white matter structures (i.e., Blatter et al., 1995; Coffey et al., 1993). As established in a variety of patient populations, quantitative volumetric analysis has been proven to be an excellent tool to measure even subtle volumetric changes (i.e., Gur et al., 1994; Holodny et al., 1998). Therefore, based on the initial qualitative findings by Natelson et al. (1993), it is hypothesized that participants with CFS will show quantitative evidence of volumetric enlargement of the lateral ventricle compared to healthy controls. Method For this pilot study, MRI data were obtained from 78 participants. The data of 25 participants had to be elimiTable 1.

nated from the subsequent volumetric analysis due to technically flawed MRI acquisitions. Successful quantitative volumetric analysis was conducted on the remaining 43 participants. Participants Participants were 28 CFS patients and 15 healthy controls who did not exercise regularly. CFS patients and controls were similar in age, years of formal education, gender distribution, and handedness (see Table 1). Participants with CFS were recruited either via self-referral based on media reports about the existence of the CFS Cooperative Research Center or by physician referral. All participants with CFS had a careful medical evaluation and were found to fulfill the most recent National Institute for Health and Centers for Disease Control case definition for CFS (Fukuda et al., 1994) with the following modifications: illness duration did not exceed 10 years at time of intake (M = 3 years, SD = 2.5 years, range = 1–8.5 years), there was no history of loss of consciousness for greater than 5 min, no history of psychiatric disorder in the 5 years prior to the onset of CFS, and no presence of current mania, schizophrenia, or eating disorders. Severity of CFS symptomatology was assessed by participants rating their level of discomfort on a Likert scale ranging from 0 (no discomfort at all) to 5 (extreme discomfort). Symptom ratings for 16 individual symptoms were then summed to express the overall degree of perceived severity of CFS symptomatology (M = 27, SD = 8.5, range = 14–46). Premorbid and current presence of Axis I psychiatric disorders as defined by the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; American Psychiatric Association, 1987) was established by administering the computerized version of the structured Diagnostic Interview Schedule (Markus, Robins, & Bucholz, 1990). Based on this structured interview, the presence of an Axis I disorder (other than

Demographic Characteristics of Participants With CFS and Healthy, Sedentary Controls Participants With CFSa

Age Years of Formal Education Gender (% Female) Handedness (% Right) Note: CFS = chronic fatigue syndrome. a n = 28. bn = 15. cStudent’s t test. dFisher’s Exact Test.

24

Controlsb

M

SD

M

SD

p

39.1 15.7 79% 88%

9.3 2.8

37.7 14.4 87% 93%

6.6 1.8

nsc nsc nsd nsd

VOLUMETRIC ANALYSIS IN CHRONIC FATIGUE SYNDROME

the ones listed as exclusions) at any time since the onset of CFS was observed in 11 of the 28 participants with CFS. Of these, 8 participants suffered from major depressive disorder only; 1 individual had major depressive disorder plus generalized anxiety disorder; 1 had major depressive disorder plus multiple anxiety disorders (generalized anxiety disorder, panic disorder, agoraphobia, social phobia); and 1 was diagnosed with multiple anxiety disorders (generalized anxiety disorder, panic disorder). Controls were recruited by advertising in the local community and were paid for their participation. All healthy participants also underwent medical, cognitive, and psychiatric assessments, including the Diagnostic Interview Schedule, and were excluded if any of the following were present: medical problems, premorbid or current psychiatric history, and medication use other than birth control pills. All participants gave informed consent prior to participating in the study.

Imaging Procedure and Quantitative Volumetric Data Analysis All MRI data analyzed in this study were acquired on a 1.0 Tesla magnet (Picker HPG, Highland Heights, OH). Three-dimensional gradient-echo sequences were used to generate axial T1 weighted images with a 25-cm field of view, a 256 × 256 matrix, TR of 36 msec, TE of 15 msec, and a flip angle of 30 degrees. Fifty serial slices were obtained at a thickness of 3 mm without interslice gap. The scanner was calibrated before each MRI acquisition. All identifying patient information was removed from scans prior to the volumetric analysis. Volumes of different tissue types were measured using an algorithm for segmenting 3-D MRI brain images requiring minimal user involvement (Yan & Karp, 1995). Brain volume was extracted by a presegmentation algorithm with optimal thresholding, morphological operations, and the Chamfer distance. An adaptive Bayesian algorithm was then applied for segmenting 3-D MRI images into brain and cerebral spinal fluid (CSF). The algorithm models MRI images as collections of regions with slowly varying intensity plus a white Gaussian noise. Compartments are modeled by a Markov random field with the 3-D second-order neighborhood system, where different potentials are used for in-plane and axial directions to account for anisotropical images. This incorporates spatial interactions among adjacent label voxels, which reduce degradation due to poor signal-to-noise ratio and feature contrast. A cubic B-spline function models slowly varying mean intensity

of each compartment through least squares fitting. The spline helps overcome “shading” effects and reduces bias against small isolated regions (sulcal CSF). The algorithm is implemented iteratively and adaptively. Each iteration consists of (a) estimating mean intensities of each compartment through least squares fitting of a spline expression to the entire image and (b) estimation of compartment by maximizing the a posteriori probability density using the iterative conditional mode algorithm. Adaptation is achieved by gradually increasing the number of control points of the spline function. This optimizes estimates for both compartments and mean intensity for fixed number of control points. Combining spline representation and adaptation makes the segmentation more accurate and robust.

Morphometric Measurements of Regional Cerebral Volumes All measurements of regional cerebral volumes were made on axial T1-weighted MRI images. Brain structures were identified and manually outlined with the use of a trackball. The areas corresponding to pixel values reflecting white matter, gray matter, and CSF within the outline were calculated automatically. Total brain, consisting of the sum of the pixel values of gray matter, white matter, and CSF, was measured automatically in every slice from the foramen magnum to the vertex. The right and left lateral ventricles, consisting of the body, the anterior horns, the occipital horns, and the temporal horns, were outlined and measured in each slice in which they were present. Pixel values for parts of the right and left lateral ventricles were summed separately for each slice and then over all slices in which they were present to determine the pixel value for the complete right and left lateral ventricles. Complete right and left lateral ventricles were then summed to represent the pixel value for the lateral ventricular system. Volume (mm3) was determined by the following formula: (field of view/number of matrix pixels) × slice thickness, where field of view = 250 × 250 mm2, number of matrix pixels = 256 × 256, and slice thickness = 3 mm, summed over all slices in which the brain structure under examination appeared.

Reliability of Ventricular Volumetric Measurements Two raters were trained to perform quantitative volumetric measurements under the direction of a 25

LANGE ET AL.

neuroradiologist. Both raters were blind to group membership and other participant characteristics. To establish interrater reliability, a sample of 10 consecutive scans was analyzed by both raters. The intraclass correlation coefficient was computed for the right as well as left hemisphere lateral ventricular volume and was .99 for each subcomponent. Remaining scans were then analyzed by a single rater.

tal brain size by calculating the ventricle-to-brain ratio ([log ventricular volume/log total brain volume] × 100) separately for the lateral ventricular system, as well as the right and left ventricles for each participant, and then compared groups with a unidirectional Student’s t test.

Results Statistical Analysis In this pilot study, the main hypothesis was unidirectional and predicted increased lateral ventricular CSF in the CFS group based on previously published data. Because the distribution of the raw volume data of the CFS group was significantly skewed in the positive direction (ratio of skewness to its standard error > 2.0 for each dependent variable), assumptions of normality were violated and, specifically by using a directional test, could increase the possibility of Type II error. Instead of using a less powerful nonparametric test statistic for this type of data, we chose to normalize the distribution by log transforming the data set to be able to conduct a more powerful parametric test. Thus, by using logn transformed data, the distribution was normalized (ratio of skewness to its standard error < 2.0 for each dependent variable) by rescaling the unit of measurement (mm3) without compromising power. Between-group differences in total lateral ventricular volume, as well as the volumes of the left and right lateral ventricles, were then analyzed using a one-tailed Student’s t test. The frequency of occurrence of ventricular asymmetry in CFS and control groups was assessed using a two-tailed Fisher’s Exact Test. In addition, we examined within-group differences in volume of the right versus left ventricles with a two-tailed Student’s paired t test. We chose bidirectional statistical analyses for these two asymmetry measures because, based on the literature, it could not be predicted that participants with CFS have either a greater likelihood of ventricular asymmetry or unequal ventricular size than controls. Using Pearson bivariate correlation with a significant level set at .05, log ventricular volumes were then correlated with presence of concurrent Axis I disorder, duration of illness, and overall severity of CFS symptoms. Because participant characteristics, such as age, handedness, and sex, did not differ statistically across the CFS and control groups, we did not control for these variables in the volumetric analysis of this pilot study. However, in a secondary analysis we did control for to26

Group differences in total ventricular volume are illustrated in Figure 1. Figure 2 provides an illustration of ventricular differences between participants with CFS and controls. Compared to control groups, the CFS group showed a larger mean volume of the lateral ventricular system (p < .057; see Table 2). The same trend was observed for the analysis of the right and left lateral ventricles with the CFS group showing ventricular enlargement relative to controls (see Table 2). Effect sizes were calculated for all three variables (Cohen, 1988) (a) to obtain another estimate of the degree of departure of lateral ventricular volume of participants with CFS from the null hypothesis and (b) to ascertain the degree of power to avoid Type II error. Effects for all three dependent measures were of medium size (d =. 50, r = .231); that is, expressed in terms of correlation, 54% of the variance for each variable was accounted

Figure 1. In the equivalent quantile–quantile plot shown, logn of total ventricular volume of CFS versus controls are ranked from smallest to largest and then plotted. If the two groups were drawn from the same population pool, data should lie along the line of identity (x = y). Note that the data lie to the right of the line of identity, corroborating the statistic that participants with CFS have greater total ventricular volume than controls.

VOLUMETRIC ANALYSIS IN CHRONIC FATIGUE SYNDROME

Figure 2. Examples of ventricular size in two contiguous slices in a participant with CFS (left) and a control participant (right) shown according to radiologic convention.

Table 2.

Logn Transformed Volumes of the Entire Lateral Ventricular System as Well as the Left and Right Lateral Ventricles Separately Participants With CFSa

Entire Lateral Ventricular System Left Lateral Ventricle Right Lateral Ventricle

Controlsb

M

SD

M

SD

t (df = 41)

p (One-Tailed)

10.45 9.77 9.71

.52 .49 .60

10.19 9.54 9.43

.46 .43 .53

–1.61 –1.51 –1.53

.057 .069 .066

Note: CFS = chronic fatigue syndrome. a n = 28. bn = 15.

for by group membership. The power to detect unidirectional group differences in this effect size range with unequal sample sizes was only 46%. The near statistically significant relation between CFS and control groups was maintained even after controlling for total brain volume by calculating ventricle-to-brain ratios (see Table 3).

We also considered whether groups would differ in the frequency of lateral ventricular asymmetry or whether the ventricular subcomponents within each group may have different mean volumes. The frequency of occurrence of ventricular asymmetry was similar between participants with CFS and controls (Fisher’s Exact Test, two-tailed, p = .322). As shown in 27

LANGE ET AL. Table 3. Logn Transformed Volumes of VBRs for the Entire Lateral Ventricular System as Well as the Left and Right Lateral Ventricles Separately Participants With CFSa VBR Entire Lateral Ventricular System Left Lateral Ventricle Right Lateral Ventricle

Controlsb

M

SD

M

SD

t (df = 41)

p (One-Tailed)

68.36 63.92 63.54

3.32 3.12 3.86

66.73 62.48 61.74

3.07 2.90 3.50

–1.57 –1.47 –1.50

.061 .074 .070

Note: VBR = ventricle-to-brain ratio; CFS = chronic fatigue syndrome. a n = 28. bn = 15.

Figure 3, both the CFS and control groups tended to have larger left than right ventricles, which is consistent with previous reports in normal adults (e.g., Zipursky, Lim, & Pfefferbaum, 1990). In addition, no significant mean differences were found between the size of right and left lateral ventricles within groups.

Discussion The results of this quantitative volumetric study support earlier qualitative findings of ventricular enlargement in individuals with CFS (Natelson et al., 1993). On average, there was a marginally significant trend toward enlargement of the lateral ventricles in the CFS group relative to age, gender, handedness, and education-matched controls. In addition, the statistical significance of the finding was not diminished by controlling for individual differences in total brain volume (ventricle-to-brain ratio). It has been argued that by removing this additional source of variance, the precision of the volumetric measurement is increased and the correlation with validity criteria, such as diagnostic status, is improved (Mathalon, Sullivan, Rawler, & Pfefferbaum, 1993). This is the first study to quantitatively document ventricular enlargement in CFS patients. Ventricular enlargement is a nonspecific finding. Its pathophysiological significance is unclear and has been linked to a variety of causes, including white matter loss in patients with vascular problems (Gorelick et al., 1992), Alzheimer’s disease (Murphy et al., 1993), and traumatic brain injury (e.g., Gale, Sterling, Bigler, & Blatter, 1995); gray matter loss in schizophrenics (e.g., Pfefferbaum & Marsh, 1995); and reversible metabolic abnormalities in participants suffering from anorexia nervosa and Cushing’s disease (Heinz, Martinez, & Haenggeli, 1977). Given the preliminary nature of this study, future volumetric studies in patients with CFS need to replicate and extend the scope of this initial 28

Figure 3. Frequency of ventricular asymmetry in controls and participants with CFS. The correlation between the mean volume of the lateral ventricular system, as well as left and right ventricles separately, with current Axis I disorder, duration of illness, and degree of severity of CFS symptoms was analyzed in participants with CFS only. None of the relations examined was significant.

work. Given the increasing number of studies showing cerebral pathology in CFS using structural (Buchwald et al., 1992; Lange et al., 1999; Natelson et al., 1993) and functional (SPECT [Costa, Tannock, & Brostoff, 1995; Schwartz et al., 1994] and PET [Tirelli et al., 1998]) neuroimaging tools, volumetric studies provide an additional technique for quantifying changes in the brain. Future volumetric studies should focus on understanding the possibly subtle involvement of specific gray and white matter structures in CFS. This preliminary investigation utilized a quantitative method for volumetric analysis (Yan & Karp, 1995). This type of probe is very sensitive in detecting even subtle changes in ventricular volume, and we were thus able to find the modest ventricular enlargement in participants with CFS versus controls. However, due to

VOLUMETRIC ANALYSIS IN CHRONIC FATIGUE SYNDROME

limitations in sample size, we only had 46% power to detect a true difference at the .05 level (one-tailed). Thus, this study was underpowered, failing to appropriately guard against false negative claims. To ensure appropriate power to detect a medium effect size at the .05 level of significance (one-tailed), roughly double the current sample size would be necessary (Cohen, 1988). Therefore, a follow-up study with a larger sample of participants with CFS and controls is necessary to confirm the important findings of this quantitative volumetric study. Because we were also interested in the functional significance of these findings, we examined the relation between increased ventricular volume in participants with CFS and factors such as presence of coexisting psychiatric diagnoses, duration of illness, and overall degree of perceived severity of CFS symptoms. None of these factors explained a significant portion of the variance in the volumetric measurements. It is possible, however, that an increase in ventricular CSF may be related to specific CFS symptoms or to objective cognitive function, which were not evaluated in this work. Again, these possibilities should be examined in a larger follow-up study. In conclusion, although the reasons for changes in ventricular size in CFS are unclear, findings of this carefully conducted study may have marked significance in the understanding of the pathophysiology of CFS. If these findings are replicated, this would suggest that at least a subset of CFS patients may have underlying brain pathology producing subtle cerebral loss. Significantly increased lateral ventricular volume in the group of participants with CFS would further add to the growing body of evidence suggesting the existence of an underlying neurological disease process (e.g., DeLuca, Johnson, Beldowicz, & Natelson, 1995; DeLuca, Johnson, Ellis, & Natelson, 1997; Natelson et al., 1993).

References American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author. Blatter, D. D., Bigler, E. D., Gale, S. D., Johnson, S. C., Anderson, C. V., Burnett, B. M., Parker, N., Kurth, S., & Horn, S. D. (1995). Quantitative volumetric analysis of brain MR: Normative database spanning 5 decades of life. American Journal of Neuroradiology, 16, 241–251. Buchwald, D., Cheney, P. R., Peterson, D. L, Henry, B., Wormsley, S. B., Geiger, A., Ablashi, D. V., Salahuddin, S. Z., Saxinger, C., Biddle, R., Kikinis, R., Jolesz, T. A., Folks, T.,

Balachandran, N., Peter, J. B., Gallo, R. C., & Komaroff, A. L. (1992). A chronic illness characterized by fatigue, neurologic and immunologic disorders, and active human herpes virus type 6 infection. Annals of Internal Medicine, 116, 103–113. Coffey, C. E., Wilinson, W. E., Weiner, R. D., Parashos, J. A., Djang, W. T., Webb, M. C., Figiel, G. S., & Spritzer, C. E. (1993). Quantitative cerebral anatomy in depression. A controlled magnetic resonance imaging study. Archives of General Psychiatry, 50, 7–16. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Costa, D. C., Tannock, C., & Brostoff, J. (1995). Brainstem perfusion is impaired in chronic fatigue syndrome. Quarterly Journal of Medicine, 88, 767–773. DeLuca J., Johnson, S., Beldowicz, D., & Natelson, B. H. (1995). Neuropsychological impairments in chronic fatigue syndrome, multiple sclerosis, and depression. Journal of Neurology, Neurosurgery, and Psychiatry, 58, 38–43. DeLuca, J., Johnson, S. K., Ellis, S. P., & Natelson, B. H. (1997). Cognitive functioning is impaired in patients with chronic fatigue syndrome devoid of psychiatric disease. Journal of Neurology, Neurosurgery, and Psychiatry, 62, 151–155. Fukuda, K., Straus, S. T., Hickie, I., Sharpe, M. C., Dobbins, J. G., & Komaroff, A. (1994). The chronic fatigue syndrome: A comprehensive approach to its definition and study. Annals of Internal Medicine, 121, 953–959. Gale, S. D., Sterling, C. J., Bigler, E. D., & Blatter, D. D. (1995). Trauma-induced degenerative changes in brain injury: A morphometric analysis of three patients with preinjury and postinjury MR scans. Journal of Neurotrauma, 12, 151–158. Gorelick, P. B., Chatterjee, A., Patel, D., Flowerdew, G., Dollear, W., Taber, J., & Harris, Y. (1992). Cranial computed tomographic observations in multi-infarct dementia: A controlled study. Stroke, 23, 804–811. Gur, R. E., Mozley, P. D., Shtasel, D. L., Cannon, T. D., Gallacher, F., Turesky, B., Grossman, R., & Gur, R. C. (1994). Clinical subtypes of schizophrenia: Differences in brain and CSF volume. American Journal of Psychiatry, 151, 343–350. Heinz, E. R., Martinez, J., & Haenggeli, A. (1977). Reversibility of cerebral atrophy in anorexia nervosa and Cushing’s syndrome. Journal of Computer Assisted Tomography, 1, 415–418. Holodny, A. I., Waxman, R., George, A. E., Rusinek, H., Kalnin, A. J., & de Leon, M. (1998). MR differential diagnosis of normal-pressure hydrocephalus and Alzheimer disease: Significance of perihippocampal fissures. American Journal of Neuroradiology, 19, 813–819. Lange, G., DeLuca, J., Maldjian, J. A., Lee, H., Tiersky, L. A., & Natelson, B. H. (1999). Brain MRI abnormalities exist in a subset of patients with chronic fatigue syndrome. Journal of Neurological Sciences, 171, 3–7. Lange, G., Wang, S., DeLuca, J., & Natelson, B. H. (1998). Neuroimaging in chronic fatigue syndrome. American Journal of Medicine, 105, 50S–53S. Markus, S., Robins, L. N., & Bucholz, K. (1990). Quick diagnostic interview schedule 3R version 1. St. Louis, MO: Washington University School of Medicine. Mathalon, D. H., Sullivan, E. V., Rawles, J. M., & Pfefferbaum, A. (1993). Correction for head size in brain-imaging measurements. Psychiatry Research: Neuroimaging, 50, 121–139.

29

LANGE ET AL. Murphy, D. G., DeCarli, C. D., Daly, E., Gillette, J. A., McIntosh, A. R., Haxby, J. V., Teichberg, D., Schapiro, M. B., Rapoport, S. I., & Horwitz, B. (1993). Volumetric magnetic resonance imaging in men with dementia of the Alzheimer type: Correlations with disease severity. Biological Psychiatry, 34, 612–621. Natelson, B. H., Cohen, J. M., Brassloff, I., & Lee, H.-J. (1993). A controlled study of brain magnetic resonance imaging in patients with the chronic fatigue syndrome. Journal of Neurological Sciences, 120, 213–217. Pfefferbaum, A., & Marsh, L. (1995). Structural brain imaging in schizophrenia. Clinical Neuroscience, 3, 105–111. Schwartz, R. B., Garada, B. M., Komaroff, A. L., Tice, H. M., Gleit, M., Jolesz, F. A., & Holman, B. L. (1994). Detection of intracranial abnormalities in patients with chronic fatigue syndrome: Comparison of MR imaging and SPECT. American Journal of Radiology, 162, 935–941.

30

Tirelli, U., Chierichetti, F., Tavio, M., Simonelli, C., Bianchin, G., Zanco, P., & Ferlin, G. (1998). Brain positron emission tomography (PET) in chronic fatigue syndrome: Preliminary data. American Journal of Medicine, 105, 54S–58S. Yan, M. X. H., & Karp, J. S. (1995). An adaptive Bayesian approach to three-dimensional MR brain segmentation. In Y. Bizais, C. Barillot, & R. DiPaol (Eds.), Information processing in medical imaging (pp. 201–213). Dordrecht, The Netherlands: Kluwer. Zipursky, R. B., Lim, K. O., & Pfefferbaum, A. (1990). Volumtric assessment of cerebral asymmetry from CT scans. Psychiatry Research, 35, 71–89.

Original submission July 13, 1999 Accepted September 29, 1999