Influence of Exhaustive Treadmill Exercise on Cognitive Functioning Chronic Fatigue Syndrome --------------------------------------------------- Bron : The AMERICAN JOURNAL of MEDICINE. Recent Developments in Chronic Fatigue Syndrome Datum: September 28, 1998 Volume 105 (3A) John J. LaManca, PhD, Sue Ann Sisto, PhD, Tohn DeLuca, PhD, Susan K. Tohnson, PhD, Gudrun Lange, PhD, Jacqueline Pareja, BS, Sean Caok, BS, Benjamin H. Natelson, MD, Newark, New Jersey From the Chronic Fatigue Syndrome Cooperative Research Center, University of Medicine and Dentistry of New Jersey-New Jersey Medical School, Newark, New Jersey and the Research Department, Kessler Institute for Rehabilitation, West Orange, New Jersey. This work was supported by NIH Center grant #UO 1 AI-32247. Requests for reprints should be addressed to john J. LaManca, PhD, New Jersey Medical School, 88 Ross Street, East Orange, New Jersey 07018. The purpose of this study was to determine the effect ef exhaustive exercise on cognitive performance of patients with chronic fatigue syndrome (CFS) and sedentary healthy controls (CON). Subjects were 19 women with CFS and 20 CON. A test battery consisting of 4 cognitive tests (CTB) was given pre-, immediately post-, and 24 hours post-treadmill exercise to exhaustion. No differences were seen on the CTB pre-exercise. CFS patients improved at a slower rate than CON on the Symbol Digit Modalities Test (SDMT), Stroop Word Test (SWT), and Stroop Color Test (SCT). When compared with CON, a lower number of correct responses was seen for the CFS immediately postexercise on the SDMT (61 pm. 3 vs 66 pm. 2), SM (137 pm. 6 vs 146 pm. 6), and SCT (99 pm. 4 vs 107 pm. 3), and 24 hours postexercise on the SDMT (64 pm. 3 vs 69 pm. 2), SWT (134 pm. 7 vs 148 pm. 5), and SCT (101 pm. 4 vs 106 pm. 3). We conclude that after physically demanding exercise, CFS subjects demonstrated impaired cognitive processing compared with healthy individuals. Am J Med. 1998;105(3A): 59S-65S. 01998 by Excerpta Medica, Inc. Chronic fatigue syndrome (CFS) is an illness characterized by (I) new onset of fatigue that produces a substantial decrease in the patient's activity, and (II) a number of infectious, rheumatologic, and neuropsychiatric symptoms. Of these symptoms, 2 are especially prominent-the complaint of cognitive difficulties affecting attention and memory, and the complaint of exacerbation of CFS symptoms after exertion. [1] Komaroff and Buchwald [1] have remarked that in their experience "this postexertional malaise is unusual in healthy individuals or in those with diseases (such as the collagen vascular diseases) that have some clinical similarity to CFS." Survey results from their patient pool show that as many as 70% of CFS patients report symptom exacerbation, including worsening of cognitive function, after exertion. [1] This dramatic complaint has generated several studies examining the role of exertion in CFS. Peterson et al. [2] required patients to walk at the slow rate of 1 mile per hour (mph) for 30 minutes, after which they rated their symptom severity. CFS patients rated fatigue and muscle weakness significantly worse immediately and 40 minutes after cessation of exercise, whereas healthy controls did not show these changes. Lloyd et al. [3] asked subjects to perform a submaximal hand-grip task repeatedly for 30 minutes. Here, the result Was the opposite from expected-a frank decrease in fatigue, depression, and confusion for the CFS group after exercise. We have recently found in our laboratory that when asked to make a maximal effort on a graded treadmill test, CFS patients achieved a mean maximal oxygen consumption equal to 98% of the age-predicted value. [4] Consistent with the results of Peterson et al., [2] our CFS patients also reported a significant exacerbation of fatigue immediately after the stress test, which did not lessen over a 7-day post-treadmill observation period. We concluded that the impact of this postexertional fatigue needed further exploration. [4] Importantly, because no CFS patient reported a severe flare up of disease and because objective evidence of worsening fatigue was not striking, [4] we decided that a true maximal test could be performed and used as a probe of exercise-induced cognitive function and fatigue. We chose to assess cognitive function as the dependent measure in the present study because of the subjective complaints by CFS patients of cognitive problems after physical exertion. [1] Additionally, in contrast to the other symptoms of CFS, objective evidence of subtle but significant decreases in cognitive function exists. [5-10] Healthy subjects who did not exercise regularly were used to control for the inactivity inherent in having CFS. We hypothesized that exercise would result in a decrement in the cognitive function of the CFS patients relative to the healthy controls. PATIENTS AND METHODS -------------------- Patients with CFS were recruited for this study from the larger patient pool available through the CFS Cooperative Research Center. All CFS patients met the Centers for Disease Control and Prevention (CDC) working case definition of CFS [11-13] and had no known medical causes of their symytoms. Additionally, to decrease the heterogeneity of the CFS group, [14] only patients with illness duration <6 years, who reported at least substantial intensity on symptom severity scales in the month before recruitment and who had no major psychiatric diagnosis in the 5 years before illness as assessed by a computerized diagnostic psychiatric interview (Q Dis),[15] were included. Further exclusion factors included loss of consciousness for >5 minutes, use of steroids, antihyyertensives, antibiotics, daily inhalers, or benzodiazepines. Sedentary healthy control (CON) subjects were recruited from an approximately 30-mile radius of the CFS Center. These subjects were matched for gender, age, and education with the CFS group. "Sedentary" was defined as working in an occupation that did not require moderate-to-intense physical labor and not participating in physical exercise for >1 session per week. Subjects were also excluded if they had a history of medical illness or a major psychiatric diagnosis in the 5 years before the study as determined by Q DIS. Instrumentation --------------- Neuropsychological function was evaluated by utilizing a cognitive test battery (CTB) of standardized neuropsychological tests. Included in the CTB was the Stroop Color and Word Test, which is a measure of cognitive speed and disinhibition.[16] The 3 parts of the Stroop Test are: (I) the Stroop Word Test (SWT), which involves reading aloud the color names red, green, or blue that are printed in black on a sheet of paper; (II) the Stroop Color Test (SCT), which involves reading aloud the color of groups of XXX's printed in red, green, or blue; and (III) the Stroop Color/Word Test (SCWT), which involves saying aloud the color name red, green, or blue printed in contradictory ink colors of red, green, or blue (i.e., the word green printed in red ink would require the subject to respond by saying "red"). For each, the number of items correctly identified in 60 seconds was recorded. The second test in the CTB was the written form of the Symbol Digit Modalities Test (SDMT), a measure of psychomotor speed and vigilance, which is used as a screen for cerebral dysfunction. [17] This test requires the subject to look at a page filled with rows ofboxes. The top 2 rows contain the test key, which consists of 9 geometric designs in the first row each matched with a number 1-9 in the second row. The next series of rows contain geometric designs each matched with their own empty box. The subject must match numbers with the geometric design as they were paired in the test key and write their answer in the empty boxes below. The score is the number of responses and the number of correct responses. Accuracy was defined as a ratio of correct responses divided by completed responses. Also included in the CTR was an oral version of the Trail Making Test (TMT), which is often used as a measure of psychomotor speed and attention/concentra tion. [18] During this test the subjects were asked to count in ascending order, alternating between numbers and letters (i.e., 1-A-2-D-3-C ...) until the subject reaches 13. The latency to complete the task was recorded. The final part of the CTB was the Serial 13s Test (STT), where the subject was asked to serially subtract 13, starting at 100. Subjects were instructed that when an error was made the experimenter would indicate their last correct subtraction and the subject was to continue subtracting from that value until a correct response was given. The test was terminated after 7 correct subtractions and the latency to complete the task was recorded. A motorized treadmill was utilized during the exercise test. During this test, the subjects would breath through a 2-way breathing valve attached to a mask that covered their nose and mouth (Hans-Rudolph, Kansas City, Missouri). Expired air was analyzed for ventilatory and metabolic variables via a Q-PLEX I metabolic system, and cardiac activity was recorded utilizing the Q4000 electro-cardiographic (ECG) monitor (Quinton Instrument Company, Seattle, Washington). Ratings of perceived exertion were obtained by use of the Borg 15-point categorical scale. [19] Self-ratings of tiredness, energy, calmness, and tension were made via the Activation-Deactivation Adjective Checklist (AD ACL). [20] The Beck Depression Inventory (BDI) was used to ascertain level of depression. [21] Level of intelligence was estimated by the North American Adult Reading Test (NAART). [22] Self-ratings of fatigue, confusion, weakness, and "achiness" during the past 24 hours were obtained via a fatigue scale. [23] This fatigue scale was modified by giving weights (i.e., 0,1,2,3) to the optional responses for each item. These Likert scores were then summed, giving each subject a total fatigue score. Finally, CFS Symptom Severity was estimated via a questionnaire containing a Likert severity score for each of the symptoms listed in the 1988 case definition for CFS. [12] These ratings were summed to quantify the extent of the subjects' symptoms. Procedures ---------- The subjects reported to the human performance laboratory at the same time on 2 consecutive days. The subjects were asked to abstain from performing any heavy physical work or exercise and from ingesting caffeine for 24 hours before the test. Also, they were asked to refrain from eating or smoking for 3 hours before testing. On the first day, after signing informed consent, the subjects completed the NAART, BDI, CFS Symptoms Questionnaire, exercise fatigue scale, and the AD ACL. All cognitive testing was performed in a quiet temperature- controlled room. The subjects performed the CTB in the following order: the Stroop, SDMT, STT, and TMT. Standardized instructions were read to the subject by the test administrator before each section of the CTB. After the CTB, the subjects were given the AD ACL. The subjects were then instrumented to allow collection of heart rate and respiration data during the tread-mill exercise test. The treadmill protocol included 4 minutes of resting while seated in a chair for baseline metabolic data. The exercise stages were 3 minutes each, starting at 2.5 mph and no incline. After the initial stage, the treadmill speed was increased to 3.5 mph for a 3-minute period. The work intensity was then increased by raising the incline of the treadmill by 2% at each stage until the end of the test. The subjects were instructed to give a maximal effort by completing as many exercise stages as possible. The subjects were asked at the end of each stage to rate their perceived exertion and their ability to continue to the next work intensity. After reaching the peak stage where the subjects indicated they could go no further, the treadmill was stopped and lowered. The subjects again comyleted the AD ACL, and within 10 minutes of completing the treadmill test the subjects were again administered the CTB utilizing the same protocol described previously. After completing this post-treadmill CTB, the subjects again comyleted the AD ACL. On the second day, the subjects arrived at the laboratory at the same time as day 1. The subjects completed the CFS Symptoms questionnaire, exercise fatigue scale, and the AD ACL. They were also administered both the CTB and the AD ACL. Statistical Analysis -------------------- Each dependent variable from the CTB was analyzed by a 2 (group) X3 (time) repeated-measures analysis of variance (ANOVA). The dependent variables on the AD ACL were analyzed by a 2 (group) X 6 (time) ANOVA with repeated measures. Multiple comparisons were made with the Tukey's test utilizing a harmonic mean to adjust for unequal group size. Analysis of repeated measures with BDI scores, peak oxygen consumption (VO2peak), and exercise time-to-exhaustion as covariates were run on the independent variables of the CTB to ascertain if depression or aerobic power have an effect on cognitive performance. The Wilcoxon rank-sum test was used when the comparisons involved only 2 variables. Proportional changes on the fatigue scale were compared with Fisher's exact test. The sign test was used on each item of the CFS symptoms questionnaire to examine the probability of rating that symptom worse after exercise. RESULTS ------- Table 1 contains the means of the group physical characteristics, BDI, and NAART scores. No significant differences were seen for the groups on these variables (P >0.05) with the exception of the BDI, where the CFS group exhibited a significantly higher depression score when compared with CON (P <0.001). The CFS mean of 14.7 is indicative of mild depression. The healthy subjects ranged in age from 23-50 years and the CFS participants from 18-45 years with no significant differences in age between the 2 groups (P = 0.763). The CFS group ranged in their duration of illness between 8 months and 6 years. The mean duration of illness was 3.1 pm. 0.4 years. Group means for various peak metabolic variables are presented in Table 2. When comparing the groups, no difference was seen in VO2peak (P = 0.114) and peak heart rate (HRpeak; P = 0.152). However the CON had a higher mean time to exhaustion (P = 0.037) and respiratory exchange ratio (RER) (P = 0.019) when compared with the CFS group. Performance on the Stroop test is shown in Figure 1. A significant time (P = 0.002) and interaction between time and group (P = 0.002) was seen on the SWT (Figure 1, panel A). Multiple comparison tests revealed no significant group difference for the SWT pre-exercise (P >0.05); however, the CON showed significantly better performance represented by higher scores than CFS 24 hours postexercise (P <0.05). The CON scored significantly higher on the SWT immediately postexercise when compared with their own pre-exercise score (P <0.05), whereas the CFS score did not significantly improve across time (P >0.05). A significant time (P = 0.002) and interaction between time and group (P = 0.029) was seen on the SCT (Figure 1, panel B). The groups were not significantly different on the SCT pre-exercise, however the CON scored significantly better when compared with CFS immediately postexercise and 24 hours postexercise (P <0.05). The CON scored significantly higher on the SCT immediately postexercise when compared with their own pre-exercise score (P <0.05). CFS did not significantly improve their SCT score across time (P >0.05). There was a significant improvement across time on the SCWT scores (P <0.001) and the interaction P-value was Table 1. Subject Characteristics for the Groups (Means pm. SEM) --------------------------------------------------------------- CFS Group CON Group Variable (n - 19) (n = 20) ---------------------------------------------------------------------- Age (yrs) 34 pm. 2 34 pm. 2 Height (cm) 164.3 pm. 1.4 162.2 pm. 1.5 Weight (kg) 67.3 pm. 3.2 61.8 pm. 2.7 BDI 14.7 pm. 1.6* 1.6 pm. 0.7 NAART VIQ 111.1 pm. 1.6 109.6 pm. 1.8 PIQ 111.1 pm. 0.7 110.4 pm. 0.9 FSIQ 112.4 pm. 1.4 111.4 pm. 1.6 ------------------------------------------------------------- BDI = Beck Depression Inventory; CFS = chronic fatigue syndrome; CON = control; NAART = North American Adult Reading Test; VIQ = Estimated Verbal IQ (128.7-0.89 [NAART errors); PIQ = Estimated Performance IQ (119.4-0.42 [NAART errors); FSIQ = Estimated Full Scale IQ (127.8-e.78 (NAART errors). * P (0.001, contrast with CON group. Table 2. Variables at Peak Stage of Graded Treadmill Test (Means pm. SEM) ------------------------------------------------------------------------- CFS group CON grouy Variables (n = 19) (n 3 20) VO2peak (mL.kg-1.min-1) 27.72 pm. 1.6* 30.39 pm. 1.0 HR (bpm) 175 pm. 3 182 pm. 3 RER 1.07 pm. 0.047 1.13 pm. 0.01 Time to exhaustion (min) 15.7 pm. 1.3** 19.8 pm. 1.1 ------------------------------------------------------------------ CFS = chronic fatigue syndrome; CON = control; VO2peak = peak oxygen consumption; HR = heart rate; RER = respiratory exchange ratio. * Due to technical failure the metabolic data were not obtained for 2 CFS and 1 CON subject; therefore for VO2peak, and RER the CFS group = 17 and CON group = 19 for these variables. ** P = 0.037, contrast with CON group. 0.052, with the CON group improving to a greater degree when compared with the CFS group (Figure 1, panel C). The mean SDMT scores are presented in Figure 2. SDMT scores improved for both groups across time (P <0.001). An interactive effect of time and group was seen, with the CFS group improving at a slower rate than the CON (P = 0.004). No difference was seen at baseline (P >0.05); however, the CON scored significantly higher than CFS immediately postexercise and 24 hours postexercise (P <0.05). CFS did not significantly improve on the SDMT until 24 hours post-treadmill, whereas CON had significantly improved their score immediately post-treadmill when compared with baseline (P <0.05). Accuracy on the SDMT was measured by taking a ratio of correct responses to completed responses. The mean SDMT correct/complete ratio and SEM were: pre-exercise, 0.996 pm. 0.002 versus 0.991 + 0.003; immediately postexercise, 0.996 pm. 0.002 versus 0.985 pm. 0.005; and 24 hours postexercise, 0.996 pm. 0.002 versus 0.983 pm. 0.005, for CFS and CON, respectively. A significant effect of group was seen on the SDMT correct/complete ratio (P = 0.01), with the CFS group having the highest accuracy. No significant interaction or change over time was seen for the SDMT correctlcomplete ratio (P >0.05). The group means for the TMT and STT are presented in Table 3. No significant effect of group (P = 0.572), time (P = 0.087), or interaction (P = 0.498) was seen for the TMT. On the STT the groups did significantly improve across time (P <0.001); however no significant group difference (P = 0.183) or interaction (P = 0.610) was seen for this variable. The ANOVA with repeated measures when either BDI, VO2peak or exercise time-to-exhaustion was used as co-variates revealed no significant effects of these variables on CTB scores (P >0.05). A significant difference was seen on the fatigue scale (P <0.001), with CFS indicating more fatigue (15.6 pm. 1.6) than the CON group (2.3 pm. 0.5) pre-exercise. When comparing the frequency of subjects' reporting either a worsening or improvement of their fatigue 24 hours after the exercise test, more worsening than improvement was reported by CFS patients (13 vs 4) than for controls (4 vs 7, Fisher's exact test = 0.004). Table 4 contains the results for the 4 variables of the AD ACL. The CFS group reported that they had significantly less energy (P = <0.001) and were significantly more tired (P = <0.001) than the CON; however no significant effect of time (P >0.05) or interaction between time and group (P >0.05) was seen for tiredness and energy. No significant group difference or interaction between time and group was found for feelings of calmness or tension, but these ratings did change across time (P <0.05). No significant change in the total symptom severity score was seen for the CFS group on the CFS Symptom Questionnaire (P = 0.962). Sig, tests on individual symptoms revealed significant worsening 24 hours after the exercise test on only the questionnaire item which was concerned with extreme fatigue from mild exercise (P = 0.033). DISCUSSION ---------- To our knowledge, the findings of this study are the first to quantitatively support the common complaint of a decrease in cognitive functioning after physical exertion by CFS patients. These results indicate that compared with CON, CFS patients are impaired on rapid naming (SWT, SCT) and perceptual motor speed (SDMT) tasks after exhaustive exercise. It is important to note that the reduction in cognitive processing in CFS patients was seen only after exercise and not at baseline, since no differences were found on any of the CTB variables before the treadmill test. This is in contrast with data that have shown slowed cognitive processing in CFS patients while in a relatively basal metabolic state using the Paced Auditory Serial Addition Test (PASAT) in our laboratory [5-7] and on various cognitive tests in other laboratories. [8-10] One explanation for this discrepancy may relate to our previous finding that the cognitive impairment seen in CFS patients on the PASAT can be selectively restricted to tests that call for auditory processing of information and not tasks, as in the CTB utilized in this study, that involve primarily visual processing. [7] Another possible explanation is that the CTB used in this study was easier than those of prior studies and that CFS yatients show significant impairment only on the most challenging cognitive tasks. As the SDMT correct/complete ratio results indicate, accuracy was not affected across time. That is, the CFS group was consistently more accurate on the SDMT when compared with the CON group. Also, the lack of a significant difference in the groups' performances on the SCWT, TMT, and STT would indicate that the CFS group could focus attention pre- and postexercise. These data are consistent with other studies that found no impairment in the ability of CFS patients to focus attention. [24] Overall, the exercise challenge had no effect on the ability of subjects to focus attention or on the accuracy of their cognitive performance. In contrast, CFS patients showed a deficit in the speed of information processing on the more complex and challenging cognitive tasks after the exercise challenge. These results suggest that it is the speed of information processing, not overall accuracy of performance, that is affected by the exercise challenge that leads to decreased cognitive performance. One of the limitations of this study is the inability to delineate between the effects ofexercise and repeated test administration, since a nonexercising control group was not included. However, as stated above, our groups were not different on the CTB pre-exercise, and previous research has shown that repeated administration of cognitive tests alone does not result in differential performance between CFS and controls. For example, cognitive per- Table 3. The Group Performance on the Trail Making Test and the Serial 13's Test (Means pm. SEM) -------------------------------------------------------- Test Scores -------------------------------------- PRE-TM* POST-TM 24-hr POST-TM Test Group (sec) (sec) (sec) -------------------------------------------------------------------------- TrailMaking Test CFS (n = 19) 25 pm. 2 27 pm. 4 23 pm. 2 CON (n = 20) 25 pm. 2 23 pm. 4 22 pm. 2 Serial 13's Test ** CFS (n = 19) 56 pm. 9 47 pm. 9 42 pm. 10 CON (n = 20) 48 pm. 6 34 pm. 6 24 pm. 3 -------------------------------------------------------------------------- * PRE-TM, POST-TM, and 24-hr POST-TM, respectively, refer to the test given immediately prior to, immediately after, and 24 hours after the exercise test. ** P < 0.001, groups improved across time. Table 4. The Scores for the Variables Measured by the Activation-Deactivation Adjective Checklist (AD ACL, Means pm. SEM) ----------------------------------------------------- AD ACL* Scores ------------------------------------------- Variable Group AD ACL 1** AD ACL 2 AD ACL 3 -------------------------------------------------------------- Energy $ CFS 7.9 pm. 0.7 7.2 pm. 0.7 6.9 pm. 0.9 CON 13.4 pm. 0.7 13.3 pm. 0.8 12.2 pm. 0.9 Tirednes & CFS 14.8 pm. 1.0 14.5 pm. 0.9 15.0 pm. 1.0 CON 6.5 pm. 0.5 7.9 pm. 1.2 8.2 pm. 0.7 Tension @ CFS 8.4 pm. 0.7 8.4 pm. 0.9 8.3 pm. 0.9 CON 7.2 pm. 0.6 8.2 pm. 0.9 6.4 pm. 0.4 Calmness # CFS 12.0 pm. 1.0 10.4 pm. 0.9 13.0 pm. 0.9 CON 12.0 pm. 0.9 13.6 pm. 2.1 11.5 pm. 0.8 -------------------------------------------------------------- Table 4. The Scores for the Variables Measured by the Activation-Deactivation Adjective Checklist (AD ACL, Means pm. SEM) ----------------------------------------------------- AD ACL* Scores --------------------------------------------- Variable Group AD CL 4 AD ACL 5 AD ACL 6 --------------------------------------------------------------- Energy $ CFS 6.2 pm. 0.7 7.8 pm. 3.6 6.7 pm. 2.8 CON 11.5 pm. 0.8 12.7 pm. 0.8 12.3 pm. 1.0 Tirednes & CFS 15.3 pm. 0.9 15.5 pm. 0.9 16.1 pm. 0.8 CON 8.0 pm. 0.6 8.1 pm. 0.8 7.4 pm. 0.5 Tension @ CFS 7.5 pm. 0.8 7.1 pm. 0.8 6.7 pm. 0.5 CON 7.5 pm. 0.8 5.2 pm. 0.1 6.9 pm. 0.8 Calmness # CFS 13.2 pm. 0.9 12.9 pm. 0.9 12.9 pm. 0.9 CON 12.6 pm. 0.7 11.6 pm. 0.9 11.6 pm. 0.9 --------------------------------------------------------------- * AD ACL refers to the Activation-Deactivation Checklist ** AD ACL 1,2, 3, 4, 5, and 6 refer respectively to the AD ACL given prior to the pre-exercise test cognitive test battery (CTD), post pre-exercise CTB, prior to the immediate postexercise CTR, post immediate postexercise CTR, prior to the 24-hr postexercise CTB, and post 24-hr postexercise CTB. $ P < 0.001,CFS group reported significantly less energy in contrast to CON group. & P < 0.001, CFS group reported significantly more tiredness in contrast to CON group. @ P < 0.05, the groups' level of tension changed across time. # P < 0.05, the groups' level of calmness changed across time. CFS = chronic fatigue syndrome; CON = control. formance of CFS patients did not decrease on 2 consecutive days of testing, indicating that cognitive testing alone was not fatiguing enough to decrease performance. [9] Furthermore, we have shown that although CFS patients are impaired at baseline testing on the PASAT (complex concentration task) compared with CON, their performance did not deteriorate with repeated PASAT sessions embedded in a CTB designed to produce mental fatigue (unpublished observations). In the literature, conclusions concerning the effects of physical exertion on cognitive function in healthy subjects vary greatly due to the differing modes of exercise, exercise severity, and forms of cognitive testing utilized. Some data show a facilitation of cognitive function after physical exertion similar to the intensity used in the present study, [25] while others show a detrimental effect [26] and still others indicate no effect. [27] The study by Marshall et al., [9] which investigated the effects of physical exertion on cognitive processes in subgroups of their CFS patients and controls, revealed no significant differences between the groups. The present study differed from Marshall's in the level of physical exertion required, time of cognitive testing postexercise, and cognitive test utilized (with the exception of the Stroop). [9] Most studies concerned with the effect of physicai exertion on cognition consistently show that the more physically fit an individual is, the better she/he is able to perform cognitive tasks while exercising or immediately after physical stress. [28-31] Although our control group was able to exercise longer before reaching their peak level on the treadmill test, the O2peak scores indicated that the groups did not significantly differ in aerobic fitness levels. Therefore, this cannot explain the differences in cognitive performance between the CFS and CON groups observed in the present study. One theory proposed to explain the effect of physical exertion on cognition is that as the level of physical exertion increases, cognitive processes will improve to an optimal level of exertion and then deteriorate with increasing exertion levels. [31-32] The present data suggest that CFS patients, when compared with CON with comparable aerobic fitness levels, may have a lower optimal threshold where physical exertion becomes detrimental to cognitive performance. Subjective feelings of one's own level of energy, tiredness, calmness, and tension could have an effect on a subject's drive or motivation to perform cognitive tests. The CFS patients consistently indicated that they had less energy and were more tired than the sedentary healthy group throughout the study, yet this seemed to have little effect on baseline cognitive performance. The groups did not differ on feelings of tension and calmness at any time during the study, and the directional changes in all 4 of these subjective feelings were not different between the groups across time. This suggests that physical exertion had more influence on cognitive performance than on self-perceptions of arousal. In conclusion, when faced with a physically challenging exercise regimen, CFS patients demonstrate impaired cognitive processing compared with healthy individuals of equal aerobic fitness levels. These are the first objective data to examine this complaint in individuals with CFS. REFERENCES ---------- 1. Komaroff AL, Buchwald D. Symptoms and signs of chronic fatigue syndrome. Rev Infect Dis. 1991;13(suppl 1):S8-S11. 2. Peterson PK, Sirr S, Grammith FC, et al. Effects of mild exercise on cytokines and cerebral blood flow in chronic fatigue syndrome patients. Clin Diagn Lab Immunol. 1994; 1:222-226. 3. Lloyd A, Gandevia S, Brockman A, et al. Cytokine produc- tion and fatigue in patients with chronic fatigue syndrome and healthy control subjects in response to exercise. Clin Infect Dis. 1994;18(suppl):S1 42-S146. 4. Sisto SA, LaManca J, Cordero DL, et al. Metabolic and cardiovascular effects of a progressive exercise test in pa- tients with chronic fatigue syndrome. Am J Med. 1996;100: 634-640. 5. DeLuca J, Johnson S. Neuropsychological impairments in chronic fatigue syndrome, multiple sclerosis, and depres- sion. J Neurol Neurosurg Psychiatry. 1995;58:38-43. 6. DeLuca J, Johnson SK, Natelson BH. Information process- ing efficiency in chronic fatigue syndrome and multiple sclerosis. Arch Neurol. 1993;50:301-304. 7. Johnson SK, Deluca J, Diamond BJ, Natelson BH. Selective impairment of auditory processing in chronic fatigue syndrome: a comparison with multiple sclerosis and healthy controls. Percept Mot Skills. 1996;83:51-62. 8. Marshall PS, Watson D, Steinberg P, et al. Assessment of cognitive function and mood in chronic fatigue syndrome. Biol Psychiatry. 1996;39:199-206. 9. Marshall PS, Forstot M, Callies A, et al. Cognitive slowing and working memory difficulties in chronic fatigue syn- drome. Psychosom Med. 1997;59:58-66. 10. Grafman J, Schwartz V, Dale JK, et al. Analysis of neuro- psychological functioning in patients with chronic fatigue syndrome. J Neurol Neurosurg Psychiatry. 1993;56:684- 689. 11. Schluederberg A, Straus SE, Peterson P, et al. Chronic fatigue syndrome: definition and medical outcome assess- ment. Ann Intern Med. 1992;117:325-331. 12. Holmes GP, Kaplan JE, Gantz NM, et al. Chronic fatigue syndrome: a working case definition. Ann Intern Med. 1988; 108:387-389. 13. Fukuda K, Straus SE, Hickie I. et al. Chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med. 1994;121:953-959. 14. Natelson BH, Johnson SK, Deluca J. et al. Reducing het- erogeneity in chronic fatigue syndrome: a comparison with depression and multiple sclerosis. Clin Infect Dis. 1995;21: 1204-1210. 15. Marcus S, Robbins LW, Bucholz K. Quick Diagnostic Inter view Schedule III-R version 1. St. Louis, MO: St. Louis University School of Medicine, 1990. 16. Golden CJ. Stroop Color and Word Test: A Manual for Clinical and Experimental Uses. Wood Dale, IL: Stoelting Company, 1978:32. 17. Smith A. Symbol Digit Modalities Test Manual. Los Angeles: Western Psychological Services, 1982:22. 18. Ricker JH, Axelrod BN, Houlter BD. Clinical validation of the oral trail making test. Neuropsychiatry Neuropsychol Behav Neurol. 1996;9:50-53. 19. Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14:377-381. 20. Thayer RE. The Biopsychology of Mood and Arousal. New York: Oxford University Press, 1989:A178-A180. 21. Beck AT, Ward CH, Mendelson M. An inventory for mea- suring depression. Arch Gen Psychiatry. 1961;4:561-571. 22. Blair JR, Spreen O. Predicting premorbid 10: a revision of the National Adult Reading Test. Clin Neuropsychol. 1989; 3:129-136. 23. Chalder T. Berelowitz G, Pawlikowska T, et al. Develop- ment of a fatigue scale. J Psychosom Res. 1993;37:147- 153. 24. Scheffers MK, Johnson R, Grafman J, et al. Attention and short-term memory in chronic fatigue syndrome patients: an event-related potential analysis. Neurology. 1992;42: 1667-1675. 25. McGlynn GH, Laughlin NT, Rowe V. The effect of increasing levels of exercise on mental performance. Ergonomics. 1979;22:407-414. 26. Gutin B, DiGennaro J. Effect of a treadmill run to exhaustion on performance of simple addition. Res Exerc Sport. 1968; 39:958-964. 27. Flynn RB. Numerical performance as a function of prior exercise and aerobic capacity fbr elementary school boys. Res Q Exerc Sport. 1972;43:16-22. 28. Heckler B, Croce R. Effects of time of post-test after two durations of exercise on speed and accuracy of addition and subtraction by fit and less-fit women. Percept Mot Skills. 1992;75:1059-1065. 29. Pleban RJ, Thomas DA, Thompson HL. Physical fitness as a moderator of cognitive work capacity and fatigue onset under sustained combat-iike operations. Behav Res Meth- ods Instr Comput. 1985;17:86-89. 30. Gutin, B. Effect of increase in physical fitness on mental ability following physical and mental stress. Res Q Exerc Sport. 1966;37:211-220. 31. Tomporowski PD, Ellis NR. Effects of exercise on cognitive processes: a review. Psychol Bull. 1986;99:338-346. 32. Easterbrook JA. The effect of emotion on cue utilization and the organization of behavior. Psychol Rev. 1959;66:183- 201.