| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the College of Health Sciences (K.B.S.), University of Texas, El Paso, Texas, and the Departments of Psychiatry and Behavioral Sciences (K.B.S., R.M.), Radiology (D.H.L.), and Medicine (J.I.F., D.S.B.), University of Washington, Seattle.
Address reprint requests to: Karen B. Schmaling, PhD, College of Health Sciences, University of Texas at El Paso, 1101 N. Campbell St., El Paso, TX 79902. Email: schmaling{at}utep.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
METHODS: Single-photon emission computerized tomography scans were completed on 15 subjects with CFS and 15 healthy persons twice: at rest and when performing the Paced Auditory Serial Addition Test (PASAT).
RESULTS: No group differences were found for performance on the PASAT despite CFS subjects’ perceptions of exerting more mental effort to perform the task than healthy subjects. Inspection of the aggregate scans by group and task suggested a pattern of diffuse regional cerebral blood flow among subjects with CFS in comparison with the more focal pattern of regional cerebral blood flow seen among healthy subjects. Between-group region-of-interest analysis revealed that although CFS subjects showed less perfusion in the anterior cingulate region, the change in CFS subjects’ activation of the left anterior cingulate region during the PASAT was greater than that observed for healthy subjects. The differences were not attributable to lesser effort by the subjects with CFS, confounding effects of mood perturbation, or to poorer performance on the experimental task.
CONCLUSIONS: Further research regarding CFS subjects’ diffuse cerebral perfusion and its relationship to inefficient neuropsychological performance is warranted.
Key Words: chronic fatigue syndrome • neurocognitive function • single-photon emission computerized tomography • working memory • information processing.
Abbreviations: ANCOVA = analysis of covariance; CFS = chronic fatigue syndrome; MANCOVA = multivariate analysis of covariance; MASQ = Multiple Ability Self-Report Questionnaire; MRI = magnetic resonance imaging; PANAS = Positive and Negative Affectivity Scale; PASAT = Paced Auditory Serial Addition Test; rCBF = regional cerebral blood flow; SF-36 = 36-item short form of the Medical Outcomes Study; SPECT = single-photon emission computerized tomography.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Single-photon emission computerized tomography (SPECT) provides dynamic, functional measures of brain function based on regional cerebral blood flow (rCBF) compared with the static, structural data provided by conventional computerized axial tomography or magnetic resonance imaging (MRI). SPECT scanning can be done under resting or experimentally activated conditions. Activated SPECT involves providing a specific challenge to the patient; the challenge is chosen for its relationship to either the functional task or brain region that is hypothetically implicated in the illness. Neuroimaging research in CFS generally has found white matter lesions in the frontal area on MRI and cerebral hypoperfusion in the brain stem on SPECT (6) and positron emission tomography (7). Michiels et al. (8) hypothesized that the presence of comorbid depression may aggravate the cognitive condition of patients with CFS. However, CFS patients without comorbid psychopathology have been found to have more white matter lesions on MRI (9), more brain stem hypoperfusion (10), and worse neuropsychological test performance (11) than patients with major depression or CFS patients with major depression.
In the research using SPECT in CFS, evidence for abnormalities has been inconsistent. In a cotwin control study of 99mTc-hexamethylpropyleneamine oxime (HMPAO) SPECT of monozygotic twins discordant for CFS, no statistically significant differences were found between the CFS and healthy twins by either blinded expert readers or automated, operator-independent analytic software quantification of rCBF (12). A recent review of neuroimaging research in CFS by Lange et al. (6) noted that brain stem hypoperfusion on SPECT or hypometabolism on positron emission tomography seems to be a potentially reliable finding in CFS patients (6, 7), allowing the differentiation of patients with CFS and no history of psychiatric disease from those with primary depressive or other psychiatric disorders. However, brain stem evaluation on SPECT may be at the margin of spatial resolution of this technique. The inconsistent SPECT results could be due to the heterogeneity of the CFS samples, choice of control group, or because studies have been done at rest. Activated SPECT scans may be needed to more consistently unmask the deficits associated with CFS.
Furthermore, pairing neuropsychological testing with functional neuroimaging would be useful to examine the association of behavioral and physiological functioning among patients with CFS. Fischler et al. (13) found positive associations between frontal blood flow, objective and subjective measures of cognitive function, and depressive symptoms. Asymmetrical tracer uptake (R > L among patients with CFS) differentiated patients with CFS (10% of whom had comorbid depression) from patients with major depression. Subtyping CFS patients may be crucial to identify the relationship between rCBF and behavior as measured by neuropsychological testing or self-reported complaints.
Taken together the studies cited above have several major methodological drawbacks. Critiques of previous neuropsychological and neuroimaging research (5, 6) in CFS have noted the lack of 1) correlation of neuropsychological performance with rCBF; 2) stratification of the CFS samples by presence of current psychiatric illness, chronicity of CFS, or adequate consideration of other confounding factors (eg, head injury); 3) comparison samples matched by age and gender to the CFS sample; and 4) use of activated scanning conditions. Activated scanning conditions may provide more useful data than scans under resting conditions and reflect patient subjective reports of impairments with increases in mental activity. Activation tasks must be controllable and replicable; performance must be monitored and must generate a detectable and reproducible change in rCBF (14).
The purposes of the present study were to compare global and regional CBF between patients with CFS and healthy control subjects during a demanding cognitive task and to examine the association between task-induced rCBF changes and task performance. We chose the Paced Auditory Serial Addition Test (PASAT) for the experimental condition based on reviews of previous research that suggest it is sensitive to the cognitive deficits experiences by patients with CFS (5). The PASAT is an information-processing task that places high demands on working memory and attention (15). We focused on rCBF changes of the anterior cingulate because of the association of this region with effortful information processing (16), which is required to perform the PASAT, and because of the specific association of this region with PASAT performance based on a previous SPECT study (17).
We hypothesized that patients with CFS would show relatively less resting-state cerebral perfusion, both globally and in the anterior cingulate region, and would show less rCBF change related to the activation task than the healthy control subjects.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Potential healthy participants were selected from individuals who responded to advertisements or recruitment posters and were matched to the participants with CFS by gender, age, ethnicity, and years of education.
All potential participants underwent a medical and psychiatric evaluation that consisted of a short telephone screening interview to assess medical history, neuropsychological trauma (eg, loss of consciousness, chemical exposure, and head injury), and alcohol use behaviors; a self-report questionnaire that probed for the presence and absence of all the major and minor CFS criteria; a physical examination; and a computer-assisted structured psychiatric diagnostic interview, the Diagnostic Interview Schedule, version III-A (18). Women of child-bearing age were tested to ensure that they were not pregnant before their participation.
Potential subjects with CFS were excluded for psychiatric disorders that are exclusionary for CFS (1): psychosis, bulimia or anorexia nervosa, major depression with melancholia, alcohol or drug abuse or dependence within 2 years of the diagnosis or anytime afterward. Additionally, although not exclusionary for CFS, persons with major depressive disorder or any anxiety disorder (posttraumatic stress disorder, panic disorder, obsessive-compulsive disorder, or generalized anxiety disorder) in the past month also were excluded because active mood or anxiety disorders have been linked to changes in rCBF (19, 20). The use of antidepressant medications was not exclusionary for subjects with CFS because of their widespread use for CFS symptoms, often at doses considered subtherapeutic for depression. To exclude CFS subjects taking antidepressants would have been pragmatically difficult and potentially would have selected patients unrepresentative of the larger CFS population.
Potential healthy participants who reported chronic fatigue, met criteria for CFS or fibromyalgia, had a major medical illness (such as high blood pressure, diabetes, or severe asthma), or had any current or historic psychiatric diagnosis (including alcohol or drug abuse) were excluded.
Other exclusionary factors for all participants included medications known to potentially affect cerebral functioning (eg, narcotics, steroids, benzodiazepines, or antihypertensives), previous history of neuropsychological trauma (see above), previous neuropsychological testing, perimenopausal symptoms, left-handedness, history of fever of 104°F or higher for several hours or longer, severe claustrophobia (because of the nature of the scanner), learning disability or special education program, and exposure to radiation above 2 rads in the last year.
One hundred sixteen people responded to the advertisements for healthy control subjects. Of these, 51 declined to participate after hearing about the study because of a lack of interest, 30 were excluded because they did not meet study criteria (the most common reason was because they were too young to match to the CFS patients), 10 were excluded for medical reasons (eg, history of loss of consciousness), and 10 were excluded for psychiatric reasons. One hundred twelve patients with presumptive CFS were recruited from the fatigue clinic. Of these, 58 declined to participate after hearing about the study because of a lack of interest, and 39 were excluded because they did not meet study criteria (mostly for psychiatric reasons).
Questionnaires
Physical functioning.
The physical functioning subscale of the Medical Outcomes Study Short Form-36 (SF-36) (21) was used to evaluate physical functional status. The scores on this 10-item subscale range from 0 to 100. Persons without chronic mental or physical conditions usually score well above 80 on the scale (21).
Cognitive complaints.
The Multiple Ability Self-Report Questionnaire (MASQ) (22) assesses subjects’ perceptions of difficulty in performing 38 specific cognitive tasks. The MASQ has five scales: 1) language, 2) visuoperceptual ability, 3) verbal memory, 4) visual memory, and 5) attention/concentration. A higher score indicates a greater degree of self-reported cognitive impairment. Validity coefficients between MASQ self-ratings and composite neuropsychological scores corresponding to each scale have been reported as ranging from 0.20 to 0.36 (22). The attention-concentration score was retained for analysis because this score purportedly reflects the skills most closely related to those needed to perform the activation task (see below).
Present state affect.
Subjects completed the Positive and Negative Affectivity Scale (PANAS) (23) on six occasions (see below) to assess current positive and negative affect. The PANAS consists of 10 adjectives that describe positive affect and 10 adjectives that describe negative affect. Each adjective is rated on a five-point scale, and the ratings are summed across adjectives for a total score.
Procedures
The study was reviewed and approved by the University of Washington Human Subjects Committee; all participants gave written informed consent. After qualifying for the study and before the first SPECT scan, participants completed questionnaires regarding descriptive and demographic information, physical functioning, and cognitive complaints. Then each participant underwent SPECT twice. Intravenous catheters were placed at least 10 minutes before injections. For the first scan, the control scan, the subject was injected with the radioactive tracer while they were in a seated, resting state. For the second scan, the experimental scan, subjects were injected while they were performing the PASAT.1
More specifically, on arrival at the nuclear medicine department, subjects completed the PANAS. Then a nuclear medicine technologist inserted an intravenous catheter into their forearm. The participants completed the PANAS again. After a 10-minute desensitization period, the participants were asked to close their eyes and rest for another 10 minutes. They were given the option of using an eye mask to assist them in keeping their eyes closed. Participants were not instructed to think of anything in particular and were encouraged not to speak. Subjects’ ears were unplugged, and they were exposed to the ambient noise in the area so that auditory stimulation was present in both conditions.
After 30 seconds with their eyes closed, participants were injected intravenously with 30 mCi of 99mTc-ECD (Neurolite, DuPont, North Billerica, MA), a radioactive tracer. A saline push was given after the radiotracer injection. After 10 minutes passed, the subjects were asked to open their eyes and complete a final PANAS to assess their mood during the previous 10 minutes. After the subject completed the PANAS, the intravenous catheter was removed. The SPECT scan was started 45 minutes after the injection of the radiotracer to allow background clearance to decay to produce a better image. Participants were scanned with a Prism 3000 triple-headed gamma camera equipped with Ultra-High Resolution Fan Beam collimators (Marconi Medical Systems, Cleveland, OH). Images were produced by means of low-pass filtered back projection with Chang uniform attenuation corrections at one ellipse per slice (0.11 cm-1). Approximately 40 transverse slices were obtained at single-pixel thickness (2 mm) with a system spatial resolution of 7 mm (full width at half-maximum).
The experimental scan was completed 48 hours after the rest scan. All procedures were the same except that, instead of sitting quietly, the participants completed the PASAT. The PASAT involves listening to an audiotape that presents a random sequence of single digits. The subject mentally adds the last two numbers he or she has heard and vocalizes the response before the next number is spoken on the audiotape. This test is demanding because the subject simultaneously must not only add the numbers but also continually monitor in working memory the results of the previous calculation (15). The test consists of four trials of 50 digits; the interval between the presentation of digits decreases with each trial. For the experimental scan, the subject was injected 11 digits (24 seconds) into the first trial so that the cerebral radiotracer uptake, which occurs from 20 seconds to 1.5 minutes, would be accomplished during the remaining 2 minutes, 6 seconds, that the subject completed the trial. The PASAT is scored for the total number of correct responses by trial and across trials. The duration of this task is approximately 10 minutes for all four trials. As with the rest scan, the SPECT scan started 45 minutes after injection.
After the PASAT, participants were asked to rate their level of mental exertion during the task. The Borg scale of perceived physical exertion (24) was adopted for use as a measure of perceived mental exertion. Subjects were asked to indicate how hard they worked mentally during the task by using the 15-point Borg scale, which ranges from 6 ("no exertion at all") to 20 ("maximal exertion"). Subjects were paid $50 for completing the study.
Data Reduction and Analytic Plan
SPECT image analysis.
Image analysis was performed with Brain Registration and Analysis of SPECT Studies (Nuclear Diagnostics, Stockholm, Sweden). To outline activated voxels, a threshold difference of 5% (using maximal pixel uptake for normalization; visual cortex) was used to determine voxel-by-voxel difference images for the comparison of control vs. experimental SPECT. Voxel size was 8 mm3, and full width at half-maximum was 7 mm for the instrument. Summed state images (control vs. experimental) were created by Multi-Modality Volume Math Software (Nuclear Diagnostics) for the 15 subjects in each group (CFS or healthy) and compared for an overall difference image, yielding a location and volume of activated voxels for CFS and healthy subjects. In addition, summed images for each group of 15 subjects were automatically fitted to a template with 46 volumes of interest based on an MRI atlas for measurement of counts in each volume of interest for both control and experimental conditions. As described above, the anterior cingulate region was chosen a priori as the focus for the regional difference analysis.
Data analysis.
The questionnaires were scored for their subscales of interest. CFS and healthy control subjects were compared on SF-36 physical functioning and MASQ attention-concentration scores for using univariate t tests. Multivariate analysis of covariance (MANCOVA) was used to compare positive and negative affect scores from the PANAS by group, type of scan, and time, using years of education as a covariate. MANCOVA was used to compare the number of correct responses to the PASAT by group and trial, using years of education as a covariate. Borg scores of perceived mental exertion were compared by group with years of education as a covariate, using univariate ANCOVA. Z scores reflecting cerebral blood flow to the right and left anterior cingulate regions relative to whole brain were compared by group and type of scan. Differences in Z scores were considered significant on the basis of qualities of the Z-score distribution: Z scores are distributed so that differences >1.96 are significant at p < .05 and differences >2.58 are significant at p < .01.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
As would be expected, subjects with CFS reported significantly poorer physical functioning on the SF-36 (mean = 46.67, SD = 27.75) than did healthy subjects (mean = 96.00, SD = 5.73) (t for unequal variances (15.2) = -6.74, p < .001). Subjects with CFS reported poorer global attention-concentration abilities on the MASQ (mean = 26.4, SD = 4.13) than did healthy subjects (mean = 32.25, SD = 3.50) (t(28) = -4.18, p < .001).
Positive and Negative Affect by Scan and Over Time
A four-way (2 x 3 x 2 x 2) MANCOVA was performed to examine the effects of the scan (control vs. experimental), time (baseline vs. after catheter insertion vs. after scan), group (CFS vs. healthy subjects), and type of affect (positive vs. negative), with years of education as a covariate. No significant main effects or interactions were found. Across time, scans, and groups, subjects averaged positive affect scores of 27.28 and negative affect scores of 12.55. For comparison purposes, each score is about two points less than the respective scores of a normative sample of college undergraduates (23).
Experimental Task: Performance and Perceived Difficulty
As noted above, the injection of the tracer was done while the subjects were performing the first of four trials of the PASAT. A two-way (4 x 2) MANCOVA was performed to compare the number of correct responses on each trial of the PASAT by group (CFS vs. healthy), controlling for years of education. No significant main effects for trial or group (F(3,25) = 2.27, NS; F(1,27) = 1.27, NS, respectively) or their interaction (F(3,25) = -0.39, NS) were found. The estimated marginal means for the number of correct responses by trial and subject group, controlling for education, are shown in Table 1.
|
Regional Cerebral Blood Flow by Task and Group
Figures 1 and 2 depict aggregate group scans for CFS and healthy subjects, respectively, by task. In each figure, the upper rows show scans from the experimental task with color-coded voxels indicating activation greater than the 5% threshold, and the lower rows show scans from the control task. Widespread, diffuse activation was seen in CFS subjects in both the frontal and temporal lobes and thalamic nuclei (Fig. 1), whereas healthy subjects activated small regions in the right frontal and right temporal cortex (Fig. 2). The total volume of threshold activation was greater among CFS subjects than among healthy control subjects (mean = 317.67 and 181.95 ml, respectively; t(14) = -2.29, p < .05).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
No group differences were found in performance on a demanding task of information processing, the PASAT, despite CFS subjects’ perceptions of exerting more mental effort to perform the task than healthy subjects (and finding moderate associations between perceived exertion and PASAT performance among CFS subjects). Compared with subjects in an earlier study, CFS subjects in this study had more total correct responses to the PASAT (130 vs. 124.2) and healthy subjects had fewer correct responses (141 vs. 146.4) (26). Thus, although the PASAT was perceived as requiring more mental exertion by subjects with CFS, they did not have significantly fewer correct responses on the PASAT than did healthy subjects when controlling for years of education. Because we did not find significant differences between the groups on PASAT performance, differential performance was unlikely to contribute to between-group error in the results of the SPECT scans.
Our hypotheses regarding the results of the SPECT scans were partially supported. We had hypothesized that CFS subjects would evidence relatively less rCBF perfusion at rest and during the experimental task, both globally and in the cingulate region, compared with healthy subjects, which was supported. We also predicted that CFS subjects would show less brain activation in response to the PASAT. However, total volume of threshold brain activation during PASAT performance was significantly greater for the CFS subjects than for the healthy control subjects. Visual inspection of the aggregate scans by group and task strongly suggested a pattern of diffuse rCBF among subjects with CFS, compared with the more focal pattern of rCBF seen among healthy subjects. In addition, the change in CFS subjects’ activation of the left anterior cingulate region during the PASAT compared with rest was significantly greater than that observed for healthy subjects.
The increase in global and regional perfusion observed during the experimental condition may be associated with several factors. Recruitment of additional brain regions during experimental tasks has been reported for other effortful conditions, including working memory (27) and dual task performance (28) in healthy subjects, and finger tapping after stroke (29). Greater perceived effort may be associated with relatively greater activation, although we could not test this speculation directly with the data in the present study. However, our data do suggest that such differences probably are not attributable to lesser effort by the subjects with CFS, confounding effects of mood perturbation, or poorer performance on the experimental task.
The specific implications of greater activation in the left anterior cingulate for CFS subjects are complex. The anterior cingulate is implicated in a variety of cognitive functions, including response selection, attention, nociception, premotor control, avoidance learning, affect modulation, and memory (16). Deary et al. (17) also found anterior cingulate effects during administration of the PASAT. In contrast to our results, these authors reported significant decreases in anterior cingulate activity for the total subject sample. However, the subjects in that study were insulin-treated diabetic outpatients with or without multiple hypoglycemic episodes. No differences were found between the two groups in PASAT performance or in regional brain perfusion, making it difficult to apply the findings to those from our study.
There are several considerations that may limit the generalizability of our results. The majority of subjects with CFS who were initially approached about participating in the study declined to participate or were excluded, so our resulting sample may not validly reflect the characteristics and abilities of the population of persons with CFS. Although we tried to match CFS and healthy subjects carefully on characteristics that could affect working memory performance or rCBF, the healthy subjects had more years of education than did the subjects with CFS. This difference may have resulted in unknown effects on rCBF. In addition, although we tried to carefully control for the potentially confounding effects of other conditions so that we would be able to attribute our results to the effects of having CFS, it is possible that there were unidentified systematic differences between subjects in the CFS and healthy groups that may account for the observed differences. For example, we chose to include CFS subjects taking antidepressant medications since the use of these medications is common among patients with CFS. However, although we did not find any differences between subjects with CFS taking and not taking antidepressants,2 the effects of antidepressants may have exerted undetected effects on our variables of interest. As another example, although our subjects with CFS were selected for meeting CFS case definition criteria but not meeting criteria for psychiatric disorders, subsyndromal psychiatric symptoms and fatigue levels at the time of the scans should be assessed in future studies. Variability in fatigue levels may affect cognitive function and CBF. We noted previously that our sample of subjects with CFS was predominantly female, as is common among clinical samples. The results of our study may have limited generalizability to population-based samples, to men, or to those who differ from our sample, such as significantly older or younger persons or those with cooccurring psychiatric disorders. We did not counterbalance the order of the control and experimental condition for reasons enumerated earlier,1 and it is possible that order effects accounted for some of the observed results.
Further research regarding CFS subjects’ apparent diffuse brain activation during a demanding cognitive task and its effects on other aspects of neurocognitive function is warranted. The mechanisms that underlie the differences await the results of future research efforts, and hypotheses related to cognitive efficiency and cerebral perfusion will be fruitful areas to pursue.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
2 Age, years of education, functional status, perceived cognitive ability, affect, and total volume of threshold activation were compared for the subjects with CFS taking vs. not taking antidepressant medications (N = 12 and 3, respectively). Although the small number of subjects not taking medications limits power to detect differences, none of the comparisons approached a value of t that would be significant with an infinitely large sample (all t values < 1.645, NS), suggesting that antidepressant use did not influence outcomes among CFS subjects.
Received for publication June 26, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  X. Caseras, D. Mataix-Cols, V. Giampietro, K. A. Rimes, M. Brammer, F. Zelaya, T. Chalder, and E. L. Godfrey Probing the Working Memory System in Chronic Fatigue Syndrome: A Functional Magnetic Resonance Imaging Study Using the n-Back Task Psychosom Med, November 1, 2006; 68(6): 947 - 955. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. P. de Lange, J. S. Kalkman, G. Bleijenberg, P. Hagoort, S. P. vd Werf, J. W. M. van der Meer, and I. Toni Neural correlates of the chronic fatigue syndrome--an fMRI study Brain, September 1, 2004; 127(9): 1948 - 1957. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Matthews, M. P. Paulus, and J. E. Dimsdale Contribution of Functional Neuroimaging to Understanding Neuropsychiatric Side Effects of Interferon in Hepatitis C Psychosomatics, August 1, 2004; 45(4): 281 - 286. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
