Acta Psychiatrica Scandinavica
Volume 103 Issue 2 Page 84 - February 2001

Neuropsychological functioning in chronic fatigue syndrome: a review

Veronique Michiels, Raymond Cluydts


Objective: In this paper we review critically the current status of neurocognitive studies in patients with chronic fatigue syndrome (CFS).

Method: CFS literature was monitored as part of a large research project which involved several neuropsychological and psychopathological studies. The literature survey was the result of several consecutive searches on Medline and PsycInfo databases.

Results: The neurocognitive studies are reviewed in terms of scientificaly accepted aspects of attention and memory. In addition, we review possible explanations for cognitive dysfunction in CFS. This is preceded with a discussion of the methodological limitations that are considered to explain inconcistencies across neuropsychological studies in CFS.

Conclusion: The current research shows that slowed processing speed, impaired working memory and poor learning of information are the most prominent features of cognitive dysfunctioning in patients with CFS. Furthermore, to this date no specific pattern of cerebral abnormalities has been found that uniquely characterizes CFS patients. There is no overwhelming evidence that fatigue is related to cognitive performance in CFS, and researchers agree that their performance on neuropsychological tasks is unlikely to be accounted solely by the severity of the depression and anxiety.

Chronic fatigue syndrome (CFS) is characterized by a medically unexplained debilitating fatigue last-ing at least 6 months, reducing the patient's pre-morbid activity level for more than 50 ( 1, 2). This fatigue cannot be explained by any other medical condition. Typically, rest and sleep do not restore premorbid levels of energy. A prominent feature is its sudden onset in up to 85 of the patients. This is frequently characterized by an acute viral or infec-tious-like illness, but may also follow an injury or period of high stress ( 3, 4).

From a clinical point of view, these patients not only complain of persistent mental and physical fatigue at rest, but typically of prolonged exhaustion after relatively minor physical or mental exertion. CFS has a fluctuating nature with exacerbations and partial remissions ( 5, 6). Fatigue levels and the intensity of symptoms may fluctuate during daytime, and different CFS patients might show different patterns ( 7).
Their fatigue is frequently accompanied by somatic (e.g. sore throat, swollen lymph nodes, myalgias and arthralgias), psychopathological (e.g. depression and anxiety) and neuropsychological symptoms. Wessely ( 8) concluded that within four specific areas of enquiry such as neuropsychology, neuroimaging, neurophysiology and neuroendocrinology many similarities and differences emerge between CFS and neuropsychiatric disorders. A comparison has also been made between CFS and multiple chemical sensitivities subjects, who also report numerous somatic symptoms ( 9).

The cognitive complaints reported frequently by these patients are: poor concentration, decreased memory for recent events and poor word-finding ability ( 10). Studies have reported that 50-85 of CFS patients report cognitive problems and that these contribute considerably to their social and occupational dysfunction ( 11, 12). Neuropsycho-logical studies have attempted to describe cognitive functioning in CFS patients and to clarify the relation between cognitive complaints and cognitive performance. Given the nature of the cognitive complaints reported by these patients, neuropsychological studies have focused most frequently on attention and memory function.

The purpose of this review paper is to provide for a pattern of results with respect to attention and memory function in CFS patients based on the most consistent findings. Former reviews on neurocognitive function in CFS have discussed objective neuropsychological performance by grouping the results into broad categories such as: the use of comparison groups or test norms ( 13); cognitive tests showing significant difference or non-significance between CFS subjects and control subjects ( 14). Two have used neuropsychological categories. In their review Moss-Morris et al. ( 15) have used global intellectual functioning, receptive functioning, mental activity, memory and cognition, whereas Tiersky et al. ( 16) used intellectual functioning, attention/concentration/information processing, learning and memory and higher cognitive functions. It is our opinion that critically reviewing the literature in terms of widely accepted aspects of attention and memory has a greater potential for understanding the nature of cognitive functioning in CFS. As such this review is of specific interest for the clinical psychologist, who is offered a guideline as to which cognitive functions should be evaluated more thoroughly. A better knowledge of these cognitive (dys)functions will aid to disentangle similarities and/or differences with other patient groups; or else aid in differential diagnosis. Secondly, it may help to guide clinical neuropsychological evaluation, to establish a neuropsychological baseline estimating clinical fluctuations, symptom progression if any or rate of patient recovery. Thirdly, it may contribute to the management of these patients. As such, the following aspects of attention and memory ( 17) are used: alertness, span of attention and working memory, selective/focused attention, sustained attention, attentional flexibility, supervisory attentional control, processing speed, learning and retrieval of information, interference effects and incidental learning. In addition we discuss relevant neuropathological findings and the role of subjectively experienced fatigue and psychological factors that may affect cognitive performance of CFS patients. This review is preceded with a discussion of the methodological limitations that are considered to explain inconcistencies across neuropsychological studies in CFS.

Neuropsychological studies in CFS differ with respect to a number of procedural and methodological issues. There is considerable agreement among researchers with respect to these methodological constraints. Although most of them can apply for any neuropsychological study with other patient groups, some of them are more relevant in the research with CFS patients. Five main methodological limitations characterize neuropsychological studies in CFS: heterogeneity of subject samples, the use of comparison groups, neuropsychological functions in the study, procedures used and statistical power.

Heterogeneity of subject samples

Not all studies have used the CFS case definitions ( 1, 2, 18, 19) as a basis for subject inclusion. As a con-sequence neuropsychological studies may be evaluating cognition in heterogeneous illness groups ( 14, 16). Furthermore, it is thought that CFS samples may not only vary on diagnostic criteria for CFS but also on basic parameters as length of illness, illness onset, presence of prior and/or comorbid psychiatric illness and severity of their disability.

Another methodological concern is that few studies report on the presence of cognitive complaints or on the severity of cognitive malfunctioning in their CFS sample. Data from our first study ( 20) points to the existence of CFS subgroups with respect to the degree of their attentional and memory dysfunction.

Furthermore, researchers ( 15, 16, 21) have pointed to limitations of the case definition of CFS itself. That is, based on the criteria of the case definition, the CFS population does not represent a homogeneous illness group. Recent modifications of the definition ( 1, 19) have therefore emphasized the need of using subgrouping or stratification techniques in order to reduce heterogeneity. Only recently have some researchers considered taking into account the existence of subgroups within the CFS sample (e.g. 22, 23).

The use of comparison groups

A majority of neuropsychological studies on CFS have used matched healthy controls, and some studies have used patients with depression or multiple sclerosis as additional control groups. Matching variables frequently used are: age, education, intelligence and gender. However, in some studies controls were more closely matched to CFS subjects than others. On the other hand, in several studies (e.g. 24-28) performance was compared to published norms that may not have been appropriately matched for age, education, IQ and/or gender ( 14, 16).

Furthermore, in some studies (e.g. 29, 30) controls are poorly described and not screened thoroughly for neurological conditions and psychological status. Since these factors may influence cognitive functioning, the data of healthy controls may not always be reflective of optimal performance ( 16).

The neuropsychological functions in the study

According to Wearden and Appleby ( 14), some studies do not describe test procedures in sufficient detail to make the work replicable. Friedberg ( 31) questions the sensitivity of the neuropsychological tests used to the cognitive deficits of CFS subjects, as a possible explanation of differences in findings across studies. However, the issue of sensitivity is closely linked to the issue of specificity ( 32). In our opinion, and referring to the literature review of our three studies ( 20, 33, 34), neuropsychological studies used well-known clinical tests that have shown sufficient sensitivity to detect objective evidence of mild cognitive dysfunction in some CFS patients. How-ever, few studies have used tests with high specificity to examine specific aspects of neuropsychological function. Therefore we would like to question the specificity of the tests used, since those are useful in delineating the precise nature of cognitive dysfunction in CFS.

In addition, it is our opinion that in some studies specific test data are not presented, and data of separate variables of the same neuropsychological test are inadequately summed together, although they are considered to measure different aspects of cognition. For instance, while data on delayed recall are available, a measure of the rate of learning verbal information is not presented (e.g. 35, 36), and some studies do not present separate data on Digit Span forward and backward (e.g. 35, 37, 38). This makes it difficult to evaluate and clarify different aspects of attention or memory function, and hampers comparison across studies.

Procedures used

Other methodological difficulties neglected in neuropsychological studies to date are: the issue of controlling for the time of day of testing ( 15) and presence or absence of symptom exacerbation during the testing period ( 31). Few studies report on the time of day of testing. Usually they report that the test battery was administrated in one session, on the same day, but no further details are given. At this point there is no study that compared cognitive performance of CFS patients in the morning with their performance in the afternoon, or compared cognitive performance in a better period to a worse period. This may be an important issue to control since it is known that energy levels of CFS patients are usually lowest in the late afternoon and together with other clinical symptoms may fluctuate in time in terms of days and weeks ( 7). This raises the notion for a need for longitudinal studies using repeated examinations in CFS patients to clarify the factors involved in possible performance variability.

Statistical power

Most of the studies have used relatively small CFS sample sizes (a few less than n=20, the majority between n=20 and n=30). In addition, the pattern of scores of CFS patients on neuropsychological tests is found to be spread more widely than those of controls (reflected in larger standard deviations) and therefore it is thought that with the use of larger samples true differences are more readily detected ( 15), although it should be noted that the use of more homogeneous CFS samples would probably reduce the variability of scores on the measures.

Furthermore, most studies have carried out a large number of comparisons. This makes it more likely that differences emerge by chance on at least some of the tests (increasing the risk of making type I error) ( 14). However, it is important to realize that decreasing type I error (Bonferroni adjustment) will automatically increase type II error. Since the aetiology of CFS is far from clear, and given the heterogeneity of CFS samples, reducing type I error may be too stringent. One should also consider that conclusions are not reached on single investigations but are based on a body of evidence (consistency of results across studies). Again, using larger homogeneous groups and formulating very specific hypotheses will reduce probabilities of both types of error.


Throughout the day tonic arousal undergoes slow fluctuations which may be accounted for by factors such as food intake, sleep and endogenous neural and endocrine circadian rhythms. This circadian rhythm of tonic arousal leads to different performance changes in tasks of attention and memory ( 39). According to Whyte ( 40) phasic arousal refers to rapid fluctuations in alertness in response to task demands or warning stimuli. If a person is unable to sufficiently increase or sustain their phasic arousal, performance deficits may be observed on a variety of tasks such as word fluency ( 17), slowing of reaction time ( 39), encoding information into memory and any situation requiring the subject to deal rapidly and efficiently with incoming information ( 41).

Michiels et al. ( 34) showed that CFS patients were able to increase their phasic alertness, and were able to modify their phasic arousal level as needed to the same extent as the healthy controls.

Span of attention and working memory

Digit Span backward is considered an effortful activity that calls upon the working memory. It is distinct from Digit Span forward, which is considered a relatively effortless process, and gives an indication how much information the subject can grasp at once ( 32).

Some studies have not used this distinction, and have presented data for a Digit Span total score. Except for one study ( 37), CFS patients did not perform worse than healthy controls ( 27, 35, 38, 42). Of the nine studies differentiating between Digit Span forward and backward, seven found no significant difference in performance between groups on the forward condition ( 2, 9, 23, 43-46), and two noted a significantly worse performance of CFS patients relative to controls ( 20, 34). With respect to the backward condition two studies reported no significant differences relative to controls ( 9, 43) and three studies noted an impairment ( 20, 22, 34).

Two studies used different tasks to measure attention span and working memory. Using the Salthouse Reading Span Task, assessing working memory, Marshall et al. ( 36) found CFS patients to perform significantly worse relative to controls. Joyce et al. ( 29) reported that CFS patients showed a significantly reduced performance on a spatial span task (computerized version of Corsi's Block-tapping test) and on a spatial working memory task.

Selective/focused attention

A variety of attention tasks each requiring slightly different selective attentional abilities can be categorized broadly under focused attention. For instance, Cancellation tasks, Digit Symbol and the Trail-Making test part A are considered visuomotor search tasks and involve focused attention ( 32). According to Mirsky's taxonomy of attention ( 47, 48), we will subcategorize them as a focusing/execute element. The Stroop Color-Word test is known to require inhibition of interfering stimuli and requires warding off distractions ( 32, 49). We will refer to this test as a response interference element. Directing attention covertly to a visual location forms yet another aspect of focused attention known as visuospatial selective attention ( 50).


The majority of the studies evaluating performance on Cancellation and Digit Symbol tests have reported that CFS patients perform significantly worse than healthy controls ( 20, 26, 35, 42, 51). Since the number of correct digits was scored, these findings indicate that patients with CFS perform slower than controls on these visuomotor search tasks. One study ( 9) found that CFS patients performed equally well relative to controls. Altay et al. ( 24) (patients did not fulfill CFS definition according to Fukuda ( 1) or Holmes ( 2)) found that CFS patients performed significantly better than a comparison group (age-matched normative sample).

In contrast with the performance of CFS patients on the two former tests, their performance on the Trail-Making test parts A and B is reported to be within normal range by most studies ( 24, 30, 38, 42 (data for part B only), 52).

Response interference.

Most of the studies evaluating response interference using the Stroop Color-Word test have found that patients with CFS perform significantly slower than healthy controls in all three conditions, but no significant difference was found for the calculated inter-ference score or the group by condition interaction ( 33, 36, 53). It was concluded that CFS patients do not exhibit deficits with respect to this aspect of focused attention. The data of Fiedler et al. ( 9) using a different scoring method (number of correct responses minus incorrect responses) are in agreement with the former findings, no significant differences could be found. Smith et al. ( 46) reported that CFS patients were significantly slower on the third condition of the Stroop test, and interpreted this as indicative of a susceptibility to interference. However, this study did not control for slowness on the two other conditions, nor was there information on a calculated interference score.

Visuospatial selective attention.

In Michiels et al. ( 34), Posner's paradigm of covert orientation of attention ( 54) was used. Patients with CFS showed an overall slowing of response time on the three cued conditions, but no significant difference was found in response time benefit and cost as compared to healthy controls. This suggests that shifting of attention in the visuospatial field is intact in these CFS patients and that, according to Posner, visual selection of information for further processing is not impaired in CFS patients.

Sustained attention

The Continuous Performance test ( 55) or adapted versions, and other monitoring tasks (e.g. Gordon Diagnostic System Distractor Task, Mackworth clock vigilance task) measuring sustained attention have been administered to patients with CFS. Monitoring tasks involve the sequential presentation of stimuli over a period of time and require the subjects to respond when a given target stimulus is perceived ( 32).

The majority of the studies have reported no significant differences in performance of CFS patients relative to healthy controls on tests of sustained attention, whether using number of errors made or reaction time as the dependent variable ( 9, 36, 44, 51, 56). In two studies ( 46, 57) patients with CFS made significantly more errors relative to controls, and Vollmer-Conna et al. ( 57) also reported that CFS patients showed significantly slower responses.

However, it should be noted that poor performance (more errors, slow reactions) alone is not considered proof of poor sustained attention. The essential point is to evaluate whether subjects can maintain their original level of performance ( 58). Only one study ( 57) investigated effects of task duration on performance on the vigilance task. No significant differences could be found with regard to the rate of increase in error scores or reaction times (performance decrement) between CFS patients en healthy controls. However, on a pursuit-tracking task CFS patients were found to show significantly more fluctuation in the tracking performance as compared to healthy controls. This was interpreted by the authors as indicative of difficulties in maintaining attention.

In Michiels et al. ( 20) a prolonged visual search task (Symbol Digit Modalities test, parts 1 and 2) was used to evaluate effects of task duration on performance. No significant differences in the number of errors made on both the first and the second part were found. CFS patients showed no decline in performance relative to controls on the second trial of the Symbol Digit Modalities test. Scheffers et al. ( 59) evaluated time-on-task effects over the 2-hour duration of the Attention paradigm. There were no significant group differences in reaction time or error rates as a function of time on task. Therefore, CFS patients did not show more evidence of performance decrements than healthy controls.

Attentional flexibility

The Wisconsin Card Sorting test (WCST) or an adapted version is frequently used to evaluate attentional flexibility. This test involves sorting concepts and requires the subjects to shift sets ( 32). In general, Stankov ( 60) defined attentional flexi-bility as the ability to change sets over the course of a task. Studies investigating set-shifting using the WCST have found no significant differences between CFS patients and healthy controls with respect to perseverative responses (28-30, 52).

Supervisory attentional control

Closely related to the aspect of attentional flexibility is the concept of supervisory attentional control ( 61, 62). To date, there exist no clinically useful tests that directly assess supervisory control. However, response interference (Stroop test), the flexible reaction in response to feedback (WCST), performance on The Tower of London/Hanoi (planning and procedural knowledge skills) are considered to tap the integrity of the supervisory control ( 58). Only two neuropsychological studies ( 29, 63) reported on performance of CFS patients on planning tasks. Performance was assessed mainly by using measures of accuracy (e.g. number of problems solved correctly within a time limit). There were no significant differences between groups with respect to efficiency of planning ability.

Processing speed

Fine motor speed.

Two studies presented data on fine motor speed using the Finger Tapping test (FTT) ( 20, 42). Significantly reduced performance of CFS patients relative to controls was found in only one ( 20). Bastien ( 64) reported that 54 of patients with CFS were impaired on the FTT (dominant hand), using Halstead-Reitan cut-off scores. Further, Riccio et al. ( 30) did not find evidence of slowed psychomotor speed in these patients using a different test (Grooved Pegboard).

Simple and choice reaction time speed.

The majority of the studies evaluating simple and/or choice reaction time in patients with CFS have reported a significantly slowed performance relative to controls ( 33, 45, 46, 59, 65). Two studies did not find a difference between CFS patients and controls on these tasks ( 9, 63).

Speed of information processing independent of fine motor speed.

In addition to the former reaction time studies, but differentiating cognitive processing from motor speed, Michiels et al. ( 34) and Marshall et al. ( 36, 56) found evidence of cognitive slowing independent of psychomotor slowing in patients with CFS. Furthermore, impairment of information processing of complex auditory-verbal material (PASAT) was found in CFS patients compared to matched healthy controls in several studies ( 22, 33, 34, 36-38, 43, 66). However, Kane et al. ( 51) could not confirm impaired performance of CFS patients on the PASAT.

Two studies questioned whether this slowed processing of complex information is modality-specific or whether there is a general impairment in processing ( 34, 43). A visual analogue to the PASAT was used to address this issue in both studies. Johnson et al. ( 43) found some preliminary evidence to suggest that the information processing deficit may be more pronounced in the auditory domain. However, the results of Michiels et al. ( 34) showed that CFS patients were not differentially impaired on the auditory relative to visual information processing and performed poor on both.

In addition, neurophysiological studies using event-related brain potentials have been performed to evaluate cognitive processing time. Prasher et al. ( 65) found prolonged latencies of the cognitive potentials N2 and P3 in CFS patients compared to controls, whereas the sensory evoked potentials were found to be normal. Reaction times were also significantly prolonged. The studies of Polich et al. ( 67) and Scheffers et al. ( 59) confirmed the finding that processing of sensory information (i.e. N1, P2) was not impaired, but failed to find impaired N2 and P3 cognitive potentials in these patients.

Learning and retrieval of information

Learning refers to the acquisition of new information. It implies the process of storing information as long-term memory. Retrieval of learned information forms an essential part of the effectiveness of the memory system. Retrieval may occur through recall or by recognition. The latter is considered much easier than retrieval by recall. This is true for both healthy subjects and patients with cognitive dysfunction ( 32).

Verbal auditive information.

Word-list learning tasks have frequently been used to assess learning and memory in patients with CFS. Tests most often used are: the Buschke Selective Reminding test (SR) (or its procedure), and the California Verbal Learning test (CVLT). In addition, subtests of the Wechsler Memory Scale, such as Paired Associate Learning and Logical Memory (story recall) and adapted versions of these subtests, have been used. However, these latter do not have the advantage of most of the word-list learning tasks to evaluate different elements of learning and memory. That is, they lack specificity to clarify the nature of the memory dysfunction. Therefore, we will report on the findings of associate learning and story recall in a separate paragraph.

Although word-list learning tests allow determination of the nature of the learning problem when using the appropriate testing technique, most of the studies on patients with CFS simply report on recall measures such as total sum across learning trials and delayed recall (short-term and/or long-term). Some report only one measure. Eight studies found that patients with CFS learned at a lower rate than healthy controls ( 20, 22, 23, 33, 34, 38, 45, 52), five studies could not confirm this finding ( 9, 28, 42, 46, 66). In addition, seven studies reported that CFS patients recalled less information after a delay than healthy controls ( 20, 22, 23, 33, 36, 38, 44). Five studies did not find a deficit in delayed recall in these patients ( 28, 34, 35, 51, 66).

Except for two studies ( 46, 52), CFS patients are found to perform equally well relative to controls on recognition testing ( 20, 33, 34, 38, 66). Patients with CFS appear also to have intact word priming ( 68).

Reporting on these recall measures alone is not enough to clarify the nature of the memory problem. That is, it is not possible to know whether poor performance on these measures is due to a learning impairment or a retrieval problem. The few studies that tackled this issue found evidence to suggest that poor recall of verbal information is primarily due to poor initial storage of verbal information rather than to a retrieval problem ( 33, 34, 38). In addition, these studies report on preliminary evidence to suggest that slowed information processing may interfere with learning new information.

Few researchers have reported on the general memory index of the Wechsler Memory Scale (WMS). Except for one ( 63), impairment only on visual reproduction subtest), three studies found CFS patients not to be impaired on this index ( 44, 59, 68).

Paired-associate word learning (a measure of cued recall for words) of the WMS is considered to be sensitive to memory function in general ( 32). Studies using this subtest did not present separate scores for hard or easy associations. The six studies using this WMS subtest showed that patients with CFS were not impaired for associative word learning ( 26, 30, 35, 44, 63, 68). However, using different procedures, Grafman et al. ( 63) and Sandman et al. ( 48) found that patients with CFS recalled significantly fewer words than healthy controls.

The Logical Memory subtest of the WMS is a well-known test of story recall. This allows for a more naturalistic way of testing memory and provides a measure of supraspan and the contribution of meaning to the learning and recall of information. Except for one ( 30), five studies have found no significant differences between patients with CFS and controls on Logical Memory ( 35, 38, 42, 44, 63). Using an experimental paragraph recall task, Grafman et al. ( 63) reported that patients with CFS recalled significantly fewer story units than controls.

The use of different procedures (versions may differ in length, difficulty level, number of trials, scoring methods ( 32) and sample characteristics may account for the discrepancies found with associative learning and story recall.

Non-verbal information.

A variety of tests have been used to assess non-verbal memory in patients with CFS: Complex Figure test ( 22, 23, 38), Benton Visual Retention test ( 37), Visual Reproduction subtest of the WMS ( 9, 28, 44, 63), memory for unfamiliar faces ( 68), Memory for Location test ( 20), pattern-location associates learning test ( 29) and a non-verbal Selective Reminding test ( 51). Five studies have reported significant impairment of CFS patients on nonverbal memory measures ( 20, 22, 23, 44, 63), although seven studies could not confirm this finding ( 9, 28, 29, 38, 42, 51, 68).

Interference effects.

Very little is known on interference effects in patients with CFS. The Brown-Peterson technique ( 69) can be used for studying short-term memory deficits (i.e. rapid decay of memory trace). The purpose of this distractor task is to prevent active rehearsal of the material ( 32). Sandman et al. ( 52) used a computerized version of this technique. No significant group differences were detected in the number of errors for semantically similar material (proactive interference errors). However, they made significantly more errors in memory when rehearsal was prevented (more rapid decay).

Studies using word-list learning tasks have not provided data on proactive interference. The CVLT would allow to evaluate this element of memory. When proactive interference occurs, the recall of list B is much lower (two or three words) than immediate recall on trial 1 (list A). Very few studies have provided data on these measures of the CVLT. Close observation of the data of Johnson et al. ( 66) shows no apparent difference between both trials (mean score trial 1: 8.3 (SD=0.42), mean score list B: 7.8 (SD=0.42)). The results of Michiels et al. ( 33) are in accordance with this observation (mean score trial 1: 8.2 (SD=2.0), mean score list B: 8.4 (SD=2.4)).

Incidental learning

Incidental learning refers to the ability to acquire information without directed effort ( 32). Kane et al. ( 51) used the Digit Symbol test to have an indication on incidental learning in patients with CFS (subjects were not told that their memory would be tested). The CFS patients recalled significantly fewer symbol digit pairs than did healthy controls.

Three major explanations have been considered in the literature which may account for, or in part contribute to, the cognitive dysfunctions observed in some CFS patients: 1) a CNS involvement, 2) effect of subjectively experienced fatigue on performance and 3) effect of psychopathological factors (e.g. depression, anxiety) on performance. Furthermore, it should be noted that neither of these explanations are mutually exclusive, and that they may possibly interact.

What do neuro-imaging techniques tell us about a CNS involvement in these patients?

In general, MRI studies demonstrated abnormalities in the cerebral white matter of CFS patients, but similar abnormalities can be seen in normal controls or patients with depression. SPECT studies found generalized hypoperfusion as well as abnormalities in more focal regions. Overall, the MRI and SPECT findings are considered inconclusive and preliminary. To date no specific pattern of cerebral abnormalities has been found that uniquely characterizes CFS patients ( 70, 71).

Neuroendocrine studies found evidence to suggest abnormalities in hypothalamic-pituitary-adrenal axis functioning and neurotransmission. Hypocortisolism and increased serotonin neurotransmitter function have been reported in CFS. These abnormalities differ from those observed in clinical depression, where hypercortisolism and decreased serotonin neurotransmitter function was noted (e.g. 72). We hypothesize that cognitive dysfunction in CFS may be linked to these abnormalities since evidence exists that abnormal cortisol levels may be associated with decreased cognitive ability ( 73, 74, 75) and that serotonergic-cholinergic interactions may influence cognitive processes such as attention, learning and mnemonic function ( 76, 77).

Can fatigue as experienced by patients with CFS affect their cognitive performance?

There are currently no known studies that have compared cognitive performance among CFS subgroups differing in severity of fatigue level at the time of neuropsychological testing. Some studies have gathered information on subjective fatigue primarily as part of their CFS sample description, and have generally used correlational analysis in order to obtain preliminary data on this issue. Fatigue has been operationalized using several diff-erent measures: an overall mental and/or physical fatigue subtest score as part of a larger questionnaire, a fatigue rating scale reflecting the degree of subjective fatigue experienced on the day of testing, or weeks before testing, daily frequency of subjective fatigue, perceived workload, decline in objective performance, effect of subjectively experienced fatigue, resulting from participation in testing the first day, on cognitive performance the second day.

Several studies did not find evidence for an association between fatigue measures and neuropsychological performance ( 33, 53, 57, 63, 68) nor for a decline in objective performance ( 57) or for a decline in performance on measures from day 1 to day 2 (36; several patients refused to participate in this study because they felt they would become too fatigued). Three studies found small but significant relations between fatigue measures and some neuropsychological measures. One study found increasing fatigue to be related to decreased spatial recognition memory and verbal fluency ( 20). McDonald et al. ( 26) found fatigue to be adversely related to the Digit Span task. Subjectively experienced fatigue was negatively related to a measure reflecting modification of phasic arousal level and a measure reflecting engagement of attention ( 34).

Although these studies show no overwhelming evidence that fatigue is related to cognitive performance in CFS, it would be premature to conclude that CFS patients do not experience fatigue-related declines in cognitive function. For instance, cognitive performance of a homogeneous CFS sample (controlling, e.g. mood status, prevalence of sleep disturbances, illness onset) should be evaluated repeatedly on different occasions which reflect different levels of fatigue severity. Meanwhile one could monitor critical psychophysiological and neurobiological variables.

Can depression and/or anxiety account for the objectively observed cognitive difficulties in patients with CFS?

Patients with affective disorders may experience attentional and memory dysfunction ( 17). Since major depression, depressive symptoms and anxiety are frequently observed in patients with CFS (e.g. 78), the possibility that these may contribute to impaired cognitive performance must be considered. Again, most studies have used correlational analysis, or the use of a median split of the depression and anxiety scores, in order to have an indication on this issue. Although some studies did find small but significant associations between a minority of neuropsychological measures and symptoms of depression and anxiety ( 26, 33, 35, 53), the majority of neuropsychological studies did not find such a relationship ( 20, 28-30, 34, 37, 38, 45, 46, 63, 68). Possible explanations to account for these discrepancies are that different measures were used across studies to assess depression ( 16); some neuropsychological measures may be more sensitive (e.g. Digit Span backward) to show such a relationship. In clinical depression, tasks demanding high cognitive effort may be impaired (e.g. 79) and CFS patients may differ in severity of depression and anxiety across studies. Furthermore, it should be noted that correlation is not causation, a third variable may affect both depression and cognitive performance.

In addition, one study (22) examined the effect of the presence or absence of psychiatric illness on performance on memory, attention and information processing. The results showed that CFS patients without psychiatric disorder performed significantly worse on several neuropsychological measures when compared to CFS patients with psychiatric disorder and healthy controls.

Personality characteristics have been related to CFS. Buckley et al. ( 80) showed that CFS patients scored higher on neuroticism and lower on extraversion compared to healthy controls, but CFS patients had lower neuroticism and similar introversion levels to patients with major depressive disorder. They conclude that this finding should be interpreted as a reaction of the CFS patients to their chronic condition.

On the basis of these findings researchers agree that performance of CFS patients on neuropsychological tasks is unlikely to be accounted for solely by the severity of the depression and anxiety.

Overall, these results suggest that CFS patients show intact performance on several aspects of attention and memory function. The current available data show that slowed processing speed, impaired working memory and poor learning of information are the most prominent features of cognitive dysfunctioning in patients with CFS. Furthermore, preliminary evidence suggests that subjectively experienced fatigue is not related to cognitive performance in CFS and the findings are fairly consistent in showing that neuropsychological performance of CFS patients is unlikely to be accounted for solely by the severity of depression and anxiety. Although the findings on neuropsychological functions of patients with CFS are far from conclusive, it is our opinion that the current knowledge may prove useful to the management of these patients. It may be integrated as part of the general strategies which include the acknowledg-ment of the reality of the patient's symptoms and disabilities, and the provision of appropriate education about the nature of CFS to both patients and their families.

1. Fukuda K, Strauss SE, Hickie Iet al. The chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med 1994;121:953 959.

2. Holmes GP, Kaplan JE, Gantz NMet al. Chronic fatigue syndrome: a working case definition. Ann Intern Med 1988;108:387 389.

3. Buchwald D. Fibromyalgia and chronic fatigue syndrome: similarities and differences. Rheum Dis Clin North Am 1996;22:219 243.

4. Gonzalez MB, Cousins JC, Doraiswamy PM. Neurobiology of chronic fatigue syndrome. Prog Neuropsychopharmacol Biol Psychiatry 1996;20:749 759.

5. Farrar DJ, Locke SE, Kantrowitz FG. Chronic Fatigue Syndrome 1: etiology and pathogenesis. Behav Med 1995;21:5 16.

6. Smith AP. Chronic fatigue syndrome and performance. In: Smith AP et al. eds. Handbook of human performance . London: Academic Press, 1992;2::261 278.

7. Jason LA, Holden JG, Taylor SL, Melrose HJ. Monitoring energy levels in chronic fatigue syndrome. Psychol Record 1995;45:643 654.

8. Wessely S. The neuropsychiatry of chronic fatigue syndrome. Ciba Found Symp 1993;173:212 237.

9. Fiedler N, Kipen HM, Deluca J, Kelly-McNeil K, Natelson BA. Controlled comparison of multiple chemical sensitivities and chronic fatigue syndrome. Psychosom Med 1996;58:38 49.

10. Barrows DM. Functional capacity evaluations of persons with chronic fatigue immune dysfunction syndrome. Am J Occup Ther 1995;49:327 337.

11. Komaroff AL&Buchwald D. Symptoms and signs of chronic fatigue syndrome. Rev Infect Dis 1991;13 (Suppl. 1):8 11.

12. Abbey SE&Garfinkel PE. Chronic fatigue syndrome and depression: cause, effect, or covariate. Rev Infect Dis 1991;13 (Suppl. 1):73 83.

13. Dipino RK&Kane RL. Neurocognitive functioning in chronic fatigue syndrome. Neuropsychol Rev 1996;6:47 60.

14. Wearden AJ&Appleby L. Research on cognitive complaints and cognitive functioning in patients with chronic fatigue syndrome: what conclusions can we draw?J Psychosom Res 1996;41:197 211.

15. Moss-Morris R, Petrie KJ, Large RG, Kydd RR. Neuro-psychological deficits in chronic fatigue syndrome: artifact or reality?J Neurol Neurosurg Psychiatry 1996;60:474 477.

16. Tiersky LA, Johnson SK, Lange G, Natelson BH, Deluca J. Neuropsychology of Chronic Fatigue Syndrome: a critical review. J Clin Exp Neuropsychol 1997;19:560 586.

17. Cohen RA. The neuropsychology of attention . New York: Plenum Press, 1993.

18. Sharpe MC, Archard LC, Banatvala JEet al. A report chronic fatigue syndrome: guidelines for research. J R Soc Med 1991;84:118 121.

19. Schluederberg A, Straus SE, Peterson Pet al. Chronic fatigue syndrome research. Ann Intern Med 1992;117:325 331.

20. Michiels V, Cluydts R, Fischler B, Hoffman G, Le Bon O, De Meirleir K. Cognitive functioning in patients with chronic fatigue syndrome. J Clin Exp Neuropsychol 1996;18:666 677.

21. Mounstephen A&Sharpe M. Chronic fatigue syndrome and occupational health. Occup Med 1997;47:217 227.

22. Deluca J, Johnson SK, Ellis SP, Natelson BH. Cognitive functioning is impaired in patients with chronic fatigue syndrome devoid of psychiatric disease. J Neurol Neurosurg Psychiatry 1997;62:151 155.

23. Deluca J, Johnson SK, Ellis SP, Natelson BH. Sudden vs. gradual onset of chronic fatigue syndrome differentiates individuals on cognitive and psychiatric measures. J Psychiatr Res 1997;31:83 90.

24. Altay HT, Abbey SE, Toner BB, Salit IE, Brooker H, Garfinkel PE. The neuropsychological dimensions of post-infectious neuromyasthenia (CFS): a preliminary report. Int J Psychiatry Med 1990;20:141 149.

25. Daugherty SA, Henry BE, Peterson DL, Swarts RL, Bastien S, Thomas RS. Chronic Fatigue Syndrome in Northern Nevada. Rev Infect Dis 1991;13 (Suppl. 1):39 44.

26. McDonald E, Cope H, David A. Cognitive impairment in patients with chronic fatigue: a preliminary study. J Neurol Neurosurg Psychiatry 1993;56:812 815.

27. Millon C, Salvato F, Blaney Net al. A psychological assessment of chronic fatigue syndrome/chronic Epstein-Barr virus patients. Psychol Health 1989;3:131 141.

28. Schmaling KB, DiClementi JD, Cullum M, Jones JF. Cognitive functioning in chronic fatigue syndrome and depression: a preliminary comparison. Psychosom Med 1994;56:383 388.

29. Joyce E, Blumenthal S, Wessely S. Memory, attention and executive function in chronic fatigue syndrome. J Neurol Neurosurg Psychiatry 1996;60:495 503.

30. Riccio M, Thompson C, Wilson B, Morgan DJR, Lant AF. Neuropsychological and psychiatric abnormalities in myalgic encephalomyelitis: a preliminary report. Br J Clin Psychol 1992;31:111 120.

31. Friedberg F. Chronic Fatigue Syndrome: a new clinical application. Prof Psychol Res Pract 1996;27:487 494.

32. Lezak MD. Neuropsychological assessment . Oxford: Oxford University Press, 1995.

33. Michiels V, Cluydts R, Fischler B. Attention and verbal learning in patients with chronic fatigue syndrome. J Int Neuropsychol Soc 1998;4:456 466.

34. Michiels V, De Gucht V, Cluydts R, Fischler B. Attention and information processing efficiency in patients with chronic fatigue syndrome. J Clin Exp Neuropsychol 1999;21:709-729.

35. Krupp LB, Sliwinski M, Masur DM, Friedberg F, Coyle PK. Cognitive functioning and depression in patients with chronic fatigue syndrome and multiple sclerosis. Arch Neurol 1994;51:705 710.

36. Marshall PS, Forstot M, Callies A, Peterson PK, Schenck CH. Cognitive slowing and working memory difficulties in chronic fatigue syndrome. Psychosom Med 1997;59:58 66.

37. Deluca J, Johnson SK, Natselson BN. Information process-ing efficiency in chronic fatigue syndrome and multiple sclerosis. Arch Neurol 1993;50:301 304.

38. Deluca J, Johnson SK, Beldowicz D, Natelson BH. Neuro-psychological impairments in chronic fatigue syndrome, multiple sclerosis, and depression. J Neurol Neurosurg Psychiatry 1995;58:38 43.

39. Whyte J. Attention and arousal: basic science aspects. Arch Phys Med Rehabil 1992;73:940 949.

40. Whyte J. Neurological disorders of attention and arousal: assessment and treatment. Arch Phys Med Rehab 1992;73:1094 1103.

41. Nebes RD&Brady CB. Phasic and tonic alertness in Alzheimer's disease. Cortex 1993;29:77 90.

42. Gaudino EA, Coyle PK, Krupp LB. Post-Lyme syndrome and chronic fatigue syndrome. Neuropsychiatric similarities and differences. Arch Neurol 1997;54:1372 1376.

43. Johnson SK&Deluca J. Selective impairment of auditory processing in chronic fatigue syndrome: a comparison with multiple sclerosis and healthy controls. Percept Mot Skills 1996;83:51 62.

44. Marcel B, Komaroff AL, Fagioli LR, Kornish II RJ, Albert MS. Cognitive deficits in patients with chronic fatigue syndrome. Biol Psychiatry 1996;40:535 541.

45. Smith AP. Cognitive changes in myalgic encephalomyelitis. In: Jenkins R et al., eds. Post-viral fatigue syndrome . London: John Wiley & Sons Ltd, 1991:179 194.

46. Smith AP, Behan PO, Bell W, Millar K, Bakheit M. Behavioural problems associated with the chronic fatigue syndrome. Br J Psychol 1993;84:411 423.

47. Mirsky A. The neuropsychology of attention: elements of a complex behavior. In: Perecman E, ed. Integrating theory and practice in clinical neuropsychology . Hillsdale, New Jersey: Lawrence Erlbaum, 1989:75-91.

48. Mirsky AF, Anthony BJ, Duncan CC, Ahearn MB, Kellam SG. Analysis of the elements of attention: a neuropsychological approach. Neuropsychol Rev1991;2:109-145.

49. Stuss DT&Benson DF. The frontal lobes . New York: Raven Press, 1986.

50. Posner MI, Snyder CRR, Davidson BJ. Attention and the detection of signals. J Exp Psychol Gen 1980;109:160 174.

51. Kane RL, Gantz NM, Dipino RK. Neuropsychological and psychological functioning in chronic fatigue syndrome. Neuropsychiatry Neuropsychol Behav Neurol 1997;10:25 31.

52. Sandman CA, Barron JL, Nackoul K, Goldstein J, Fidler F. Memory deficits associated with chronic fatigue imunne dysfunction syndrome. Biol Psychiatry 1993;33:618 623.

53. Ray C, Phillips L, Weir WRC. Quality of attention in chronic fatigue syndrome: Subjective reports of everyday attention and cognitive difficulty, and performance on tasks of focused attention. Br J Clin Psychol 1993;32:357 364.

54. Posner MJ, Walker JA, Friedrich FJ, Rafal RD. Effects of parietal injury on covert orienting of attention. J Neurosci 1984;4:1863 1874.

55. Rosvold HE, Mirsky AF, Sarason I, Bransome E, Beck LH. A continuous performance test of brain damage. J Consult Clin Psychol 1956;20:343 350.

56. Marshall PS, Watson D, Steinberg Pet al. An assessment of cognitive function and mood in chronic fatigue syndrome. Biol Psychiatry 1996;39:199 206.

57. Vollmer-Conna U, Wakefield D, Lloyd Aet al. Cognitive deficits in patients suffering from chronic fatigue syndrome, acute infective illness or depression. Br J Psychiatry 1997;171:377 381.

58. Van Zomeren AH&Brouwer WH. Assessment of attention. In: Crawford JR et al., eds. A handbook of neuropsychological assessment . Hove, UK: Lawrence Earlbaum Assoc-iates, 1992:241-266.

59. Scheffers MK, Johnson R, Grafman J, Dale JK, Straus SE. Attention and short term memory in chronic fatigue syndrome patients: an event-related potential analysis. Neurology 1992;42:1667 1675.

60. Stankov L. Aging, attention, and intelligence. Psychol Aging 1988;3:59 74.

61. Shallice T. Specific impairments of planning. Phil Trans R Soc Lond 1982;298:199 209.

62. Ponsford J&Kinsella G. Attentional deficits following closed-head injury. J Clin Exp Neuropsychol 1992;14:822 838.

63. Grafman J, Schwartz V, Dale JK, Scheffers M, Houser C, Straus SE. Analysis of neuropsychological functioning in patients with chronic fatigue syndrome. J Neurol Neurosurg Psychiatry 1993;56:684 689.

64. Bastien S. Patterns of neuropsychological abnormalities and cognitive impairment in adults and children. In: Hyde B et al., eds. The Nightingale Research Foundation review of the clinical and scientific basis of myalgic encephalomyelitis/chronic fatigue syndrome . Canada: The Nightingale Research Foundation, 1992:453-463.

65. Prasher D, Smith A, Findley L. Sensory and cognitive event-related potentials in myalgic encephalomyelitis. J Neurol Neurosurg Psychiatry 1990;53:247 253.

66. Johnson SK, Deluca J, Fiedler N, Natelson BH. Cognitive functioning of patients with chronic fatigue syndrome (brief communication). Clin Infect Dis 1994;18 (Suppl. 1):84 85.

67. Polich J, Moore AP, Wiederhold MD. P300 assessment of chronic fatigue syndrome. J Clin Neurophysiol 1995;12:186 191.

68. Cope H, Pernet A, Kendall B, David A. Cognitive functioning and magnetic resonance imaging in chronic fatigue. Br J Psychiatry 1995;167:86 94.

69. Peterson LR&Peterson MJ. Short-term retention of individual verbal items. J Exp Psychol Learn Mem Cogn 1959;58:193 198.

70. Cope H&David AS. Neuroimaging in chronic fatigue. J Neurol Neurosurg Psychiatry 1996;60:471 473.

71. Greco A, Tannock C, Brostoff J, Costa DC. Brain MR in chronic fatigue syndrome. Am J Neuroradiol 1997;18:1265 1269.

72. Cleare AJ, Bearn J, Allain Tet al. Contrasting neuroendocrine responses in depression and chronic fatigue syndrome. J Affect Disord 1995;35:283 289.

73. Bohnen N, Houx P, Nicolson N, Jolles J. Cortisol reactivity and cognitive performance in a continuous mental task paradigm. Biol Psychol 1990;31:107 116.

74. Bohnen N, Jolles J, Twijnstra A, Mellink R, Sulon J. Coping styles, cortisol reactivity, and performance in a vigilance task of patients with persistent postconcussive symptoms after a mild head injury. Int J Neurosci 1992;64:97 105.

75. Lupien S, Roch Lecours A, Lussier I, Schwartz G, Nair NPV, Meany MJ. Basal cortisol levels and cognitive deficits in human aging. J Neurosci 1994;14:2893 2903.

76. Cassel J-C&Jeltsch H. Serotonergic modulation of cholin-ergic function in the central nervous system: cognitive implications. Neuroscience 1995;69:1 41.

77. Steckler T&Sahgal A. The role of serotonergic-cholinergic interactions in the mediation of cognitive behaviour. Behav Brain Res 1995;67:165 199.

78. Hickie I, Lloyd A, Wakefield D, Parker G. The psychiatric status of patients with chronic fatigue syndrome. Br J Psychiatry 1990;156:534 540.
79. Brand NA, Jolles J, Gispen-De Wied C. Recall and recognition memory deficit in depression. J Affect Disord 1992;25:77 86.

80. Buckley L, MacHale SM, Cavanagh JYO, Sharpe M, Deary IJ, Lawrie SM. Personality dimensions in chronic fatigue syndrome and depression. J Psychosom Res 1999;46:395 400.