Heart and Circulatory Physiology

Postural neurocognitive and neuronal activated cerebral blood flow deficits in young chronic fatigue syndrome patients with postural tachycardia syndrome

Julian M. Stewart, Marvin S. Medow, Zachary R. Messer, Ila L. Baugham, Courtney Terilli, Anthony J. Ocon


Neurocognition is impaired in chronic fatigue syndrome (CFS). We propose that the impairment relates to postural cerebral hemodynamics. Twenty-five CFS subjects and twenty control subjects underwent incremental upright tilt at 0, 15, 30, 45, 60, and 75° with continuous measurement of arterial blood pressure and cerebral blood flow velocity (CBFV). We used an n-back task with n ranging from 0 to 4 (increased n = increased task difficulty) to test working memory and information processing. We measured n-back outcomes by the number of correct answers and by reaction time. We measured CBFV, critical closing pressure (CCP), and CBFV altered by neuronal activity (activated CBFV) during each n value and every tilt angle using transcranial Doppler ultrasound. N-back outcome in control subjects decreased with n valve but was independent of tilt angle. N-back outcome in CFS subjects decreased with n value but deteriorated as orthostasis progressed. Absolute mean CBFV was slightly less than in control subjects in CFS subject at each angle. Activated CBFV in control subjects was independent of tilt angle and increased with n value. In contrast, activated CBFV averaged 0 in CFS subjects, decreased with angle, and was less than in control subjects. CCP was increased in CFS subjects, suggesting increased vasomotor tone and decreased metabolic control of CBFV. CCP did not change with orthostasis in CFS subjects but decreased monotonically in control subjects, consistent with vasodilation as compensation for the orthostatic reduction of cerebral perfusion pressure. Increasing orthostatic stress impairs neurocognition in CFS subjects. CBFV activation, normally tightly linked to cognitive neuronal activity, is unrelated to cognitive performance in CFS subjects; the increased CCP and vasomotor tone may indicate an uncoupling of the neurovascular unit during orthostasis.

  • orthostatic stress
  • cognition
  • functional hyperemia

chronic fatigue syndrome (CFS) is associated with orthostatic intolerance (OI), especially in the young (21, 49). OI is defined by signs and symptoms such as dizziness, fatigue, tachycardia, hypotension, visual disturbances, hypocapnia, headache, cognitive deficits, and nausea while in the upright position that is relieved by recumbence (46, 55). In CFS, OI often takes the form of postural tachycardia syndrome (POTS) (27, 54) or neurally mediated hypotension (4, 47). In our studies in adolescents and young adults, POTS has been uniformly present in CFS subjects. POTS is well defined in adults by an increase in heart rate (HR) of >30 beats/min upon upright posture for 10 min or a maximum HR of >120 beats/min, with signs and symptoms of OI (50). Blood pressure (BP) is usually maintained during orthostasis and may even increase (26); larger HR increments and late BP decreases may occur in the young (31). POTS is more common in the adolescent and young adult subgroup of CFS patients but may also affect older patients (27, 53).

CFS is associated with neurocognitive impairment (30) of concentration, working memory, and information processing speed (13). A pathophysiological cause for the neurocognitive impairment in CFS with POTS (CFS/POTS) has not been found.

It has been hypothesized that impaired cerebral perfusion contributes to the neurocognitive dysfunction in CFS/POTS subjects. Evidence for impaired cerebral perfusion is unclear, as studies have both supported (3, 11, 28) and refuted (19, 36) this; however, our prior studies have supported the impaired perfusion hypothesis. Using transcranial Doppler ultrasound (TCD), we (38) demonstrated reduced cerebral blood flow (CBF) and cognitive loss in CFS/POTS patients during 70° upright tilt but not in the supine position.

We hypothesized that the postural cognitive loss in young CFS/POTS patients is due to orthostatic reductions in CBF related to abnormalities in cerebral vasomotor tone (16) and reduced neuronal activation of CBF reactivity (48). We propose that during cognitive tasking, the blood flow response to neurocognitive tasking is suboptimal and is related to increased vasomotor tone, suggesting deficits in the neurovascular unit (comprising functional interactions among the neurons, blood vessels, and glia) (32).


To investigate these hypotheses, we tested neurocognition during a graded head-up tilt table test using an n-back memory task. The n-back task is a test of working memory, concentration, and information processing of progressive difficulty (5) and has been used to probe the working memory in CFS subjects (9). We combined n-back testing (n values ranging from 0 to 4) with bilateral TCD during incremental upright tilt at 0, 15, 30, 45, 60, and 75°. We measured changes in cerebrovascular function by measuring CBF velocity (CBFV) using a two-parameter analysis that incorporates critical closing pressure (CCP; corresponding to vasomotor tone) (1, 7, 16, 42) and the resistance-area product (corresponding to proximal resistance) and compared these with the Gosling pulsatility index (PI) (6, 24). We assessed the response of CBFV during cognitive activation (34, 43, 48) to model blood flow linkage to neuronal activation (20, 45).


We enrolled 25 subjects diagnosed with CFS using the Fukuda criteria (22), which included some subjects from a prior investigation (40). All CFS subjects were found to have POTS during a 70° head-up tilt table test performed on another day. OI was defined by the presence of dizziness, fatigue, exercise intolerance, headache, memory problems, palpitations, nausea, blurred vision, pallor, and abnormal sweating while in the upright position that was relieved by recumbence and had no other medical explanation. The diagnosis of POTS was made in these patients during the screening tilt. POTS was diagnosed by symptoms of OI during tilt associated with an increase in sinus HR of >30 beats/min for ages of 21–29 yr and with an increase in HR of >45 beats/min for ages of 15–21 yr or to a HR of >120 beats/min during the first 10 min of tilt (35, 52). Twenty healthy volunteer control subjects with no history of illness or OI were also enrolled. The age range for all subjects was 15–29 yr. Control subjects were matched by sex and age to CFS patients so that approximately the same female-to-male ratio was obtained in control subjects as in CFS/POTS patients.

All subjects had normal physical examinations and normal electrocardiographic and echocardiographic evaluations. Only those free from systemic illness were eligible. This excluded patients with illnesses associated with cardiovascular dysfunction such as diabetes, renal disease, congestive heart failure, systemic hypertension, inflammatory diseases, neoplasm, immune-mediated disease, trauma, morbid obesity, and peripheral vascular disease. Tobacco smoking was exclusionary. Control subjects were recruited through a healthy volunteer database. Use of medication was exclusionary for control subjects. If CFS patients used medication, it was discontinued at least 2 wk before the study.

Normal vision or vision corrected by glasses or contact lenses and intact depth perception were necessary to complete the protocol.

This research was approved by the Institutional Review Board and was performed in accordance with the Declaration of Helsinki. Subjects over 18 yr old signed informed consent. Subjects of <18 yr old as well as their legal guardians read and signed the consent form.


Subjects refrained from eating for at least 4 h and did not consume caffeinated beverages for at least 12 h before being tested. Subjects initially rested in a supine position on the tilt table (Colin Medical) with a footboard. We used an oscillometric calibrated Finometer (FMS) to assess beat-to-beat arterial BP (ABP) and electrocardiography to measure HR and rhythm. Arterial O2 saturation was measured by pulse oximetry. End-tidal CO2 was measured by a nasal cannula connected to a capnograph (CapnoCheck 9004, Waukesha, WI). Bilateral TCD (Multigon, Yonkers, NY) was used to assess CBFV and CBFV changes in left and right middle cerebral arteries (MCAs) to testing. MCA data did not lateralize, and the reported CBFV values represent an average of both sides. BP, ECG, TCD, respiration, end-tidal CO2, and O2 saturation were acquired continuously to a computer throughout the experiment via analog-to-digital conversion.

Subjects were also instrumented with video screen eyewear (I-O Display Systems, Menlo Park, CA) and earplugs. A single operator made all measurements, and a second person performed the analyses.

N-Back Task

A parametric n-back test (5, 41) with 0-, 1-, 2-, 3-, and 4-back levels presented a load-dependent, progressively difficult mental task. The n-back test was viewed on the video screen eyewear. The stimulus duration was 1 s, the interstimulus duration was 1 s, and there was a 10-s pause between each n-back level (5). The video screen displayed which n-back level was next during this pause. Custom software generated a sequence of 29 capital letters, excluding vowels, for each n-back level. These were generated as distinct random sequences for each n-back level and angle of tilt. All subjects saw identical sequences. Subjects responded to the n-back task by pressing a button attached to a handgrip placed in their dominant hand.


After instrumentation, subjects remained in the supine position with their eyes open for 5 min to accommodate. Subjects then practiced responding with the button/handgrip to the beat of a metronome, rested for 5 min, and underwent three n-back practice sessions in which the n value varied from 0 to 4 in sequence.

Subjects rested for 15 min. Baseline measurements were taken during the last 3 min. Subjects completed n-back tasks in the supine position and were then tilted upright to 15° for a total of 10 min. The first 1 min of data was excluded to allow for hemodynamic stabilization. Minutes 2–4 of the tilt were used to obtain baseline values for that angle. The progressive n-back task started at minute 4. Once 10 min had elapsed, subjects were progressively tilted to 30, 45, 60, and 75°, and stabilization, baseline data collection, and the n-back task were repeated at each angle.

Stopping criteria (end test) during incremental upright tilt were symptoms of presyncope (defined as a decrease in systolic BP to 80 mmHg); a decrease in systolic BP to 90 mmHg associated with symptoms of lightheadedness, nausea, sweating, or diaphoresis; or progressive symptoms of OI accompanied by a request to discontinue the test. Presyncopal subjects were immediately returned to the supine position, and the test was ended. If subjects completed the procedures during all angles of tilt, they were returned to the supine position.

Data Analysis

All data were continuously sampled at 500 Hz. NCSS 2007 (LCC) statistical software was used in the analysis. Demographic data were analyzed by two-sided independent Student's t-tests. Mean CBFV for each pulse was computed as a time average over a cardiac cycle.

The initial analysis of n-back outcome, CCP, and neuronal activation of CBFV during the n-back test and orthostasis used three-way ANOVA with Student-Newman-Keuls post hoc tests to determine main effects and two-way interactions. Whenever between-group differences were established, data were reassessed using mixed-model repeated-measures two-way ANOVA for each n-back test with tilt angle as the repeated factor. Data are depicted and reported as means ± SE. Significance was set at P < 0.05.

N-back outcome measures.

Correct responses and time to response were used to measure n-back outcomes. Correct responses were defined as the subject appropriately responding to an n-back repeat. Absolute reaction time was calculated (in ms) as the difference between the time that the letter first appeared on the video screen and the time that the subject responded. For each level, data were reported as normalized reaction time, which was calculated as the mean absolute reaction time per total number of correct responses and took into account the number of times that a subject responded correctly and incorrectly.

Neuronal activated CBFV: changes in CBFV during cognitive activation.

Baseline resting CBF and metabolism relate to an intrinsic state of neuronal activity (57). Increased neuronal activity during mental tasks increases glutamate production, initiating a metabolism-driven interaction of the neurovascular unit, which comprises glia, especially astrocytes, and the local vasculature and results in increased blood flow (32), which we refer to as neuronal activated CBFV. Some cortical areas decrease CBFV because of decreased neuronal activity. Such functional hypoperfusion has been associated with cognitive deficits (37) during n-back tests in CFS subjects (9). Measuring changing CBFV during neuronal activation or “functional TCD” can provide regional CBFV data that accurately correlate with 15H2O positron emission tomography scans (48). An increase in TCD blood flow during n-back activation is the expected response to a net increase in blood flow in the area perfused by the MCA, which supplies the majority of cortex affected by the n-back study (10). CBFV measured by TCD in the MCA typically increases during cognitive activation (34, 43, 48).

Thus, the change of CBFV (in units of cm·s−1·min−1) serves as an index of task-related neuronal activation during n-back testing. We quantitated this as the rate of change (slope) of the CBFV during each n-back task at each angle of tilt for every subject, as shown in Fig. 1. Because CBFV varied from subject to subject, we normalized this slope to the average CBFV during measurement and expressed the result as the percent change in CBFV per minute.

Fig. 1.

Changes in cerebral blood flow velocity (CBFV; top) during the sequence of n-back tasks (n = 0 to 4; bottom). The n-back responses were to alphabetic characters (consonants only) encoded as numbers 2 (= letter “B”) through 26 (= letter “Z”). Randomized n-back sequences were repeated at each angle of tilt. Changes in CBFV were obtained by linear least-squares fits to the data (expressed in units of cm·s−1·min−1). Slopes were subsequently normalized to CBFV for purposes of comparison.

Thus, the index was determined by the following equation: 100 × (1/CBFV)(ΔCBFV/Δt), where Δt is the change in time.

This quantity is positive for a net increase in CBFV (increased regional neuronal activity) during mental activation or negative for a net decrease in neuronal activity. Figure 1 shows representative changes in the slope during an n-back activation sequence (0–4 back) in a control subject in the supine position.


CCP can be defined as follows (7): The arterial blood pressure (ABP) level at which flow stops is defined as the critical closing pressure (CCP) or, in cardiac literature, the zero-flow pressure. Above the CCP, an approximately linear slope, sometimes referred to as the inverse flow resistance, defines the relation between pressure and flow. When these variables are plotted as an x-y function it follows that flow is linearly (but not proportionally) related to pressure and that it can be regulated by changes in both CCP (the x intercept) and slope. This is also known as the “resistance-area product” (1). Thus, CBFV = (arterial pressure − CCP)/resistance-area product (1).

We estimated the CCP of the cerebral vasculature at the level of the MCA for each heart beat using the equivalent linear regression between the mean CBFV in the MCA on the x-axis and arterial pressure on the y-axis. Thus, arterial pressure = resistance-area product × CBFV + CCP. The resistance-area product is thought to reflect proximal cerebrovascular resistance at frequencies exceeding the inverse of the R-R interval. CCP is the y-intercept (the pressure where CBFV = 0). Figure 2 shows three fitted beats for illustrative purposes. In practice, individual beats were fit by linear regression. However, if R2 < 0.5, that beat's values of CCP and resistance-area product were interpreted as “too noisy” and were replaced by interpolations based on cubic spline fitting with adjacent beats. CCP was estimated as the y-intercept (the intercept on the arterial pressure axis where mean CBFV = 0). Typically, CBFV leads arterial pressure in time. Thus, before regression, the waveforms of arterial pressure and CBFV were aligned for each beat, such that peaks and troughs occurred at the same time for both signals. The cerebral vasculature functions as a high-pass filter at frequencies exceeding ∼0.07 Hz. Conductance is therefore not equivalent to the inverse of 0 Hz (direct current) resistance but relates to hydraulic impedance terms, which are beyond the scope of the present investigation. Custom written software calculated CCP for each cardiac cycle. While subjects were in upright position, BP was corrected for the hemostatic pressure difference between pressure at the brain and at the heart to allow for an appropriate estimate of CCP. CCP may be determined by various fitting procedures (44). For that reason, we repeated the fitting procedure using the first harmonic method of Aaslid et al. (1), again obtaining values that yielded the best fit in the least squares sense. The superiority of the two-parameter relationship of arterial pressure = resistance-area product × CBFV + CCP in place of the one-parameter Ohmic formulation for cerebrovascular resistance [mean arterial pressure (MAP) = cerebral vascular resistance × mean CBFV] is based on physiological experiments showing the collapse of small arterioles in the cerebrovascular circulation and the linearity of the relationship once CCP is taken into account (7, 16). Direct experiments on primate brains have indicated that the resistance-area product may represent the resistive properties of the proximal vasculature (16).

Fig. 2.

Computation of the two-parameter fit using phasic arterial pressure (AP; top) and the corresponding phasic CBFV (middle). Bottom: fits used to obtain the critical closing pressure (CCP; y-intercept) and the resistance-area product (slope).

Gosling PI.

PI was calculated using the following formula: PI = (maximum CBFV − minimum CBFV)/(mean CBFV) (6, 24), where maximum CBFV is the maximum systolic velocity, minimum CBFv is the minimum diastolic velocity, and mean CBFV is the CBFV averaged over the entire cardiac cycle. Averaging over multiple cardiac cycles was performed. An increase in PI indicates increased cerebral vascular resistance.



Mean age (22 ± 1 vs. 23 ± 1 yr), height (172 ± 2 vs. 167 ± 2 cm), weight (65 ± 4 vs. 66 ± 4 kg), and body mass index (21 ± 1 vs. 23 ± 1 kg/m2) were not different between CFS/POTS and control subjects, respectively. Twenty CFS/POTS subjects were right handed and five CFS/POTS subjects were left handed. Eighteen control subjects were right handed and two control subjects were left handed. Dropout during incremental tilt table testing due to presyncope was similar for CFS/POTS and control subjects and has been separately reported (33).

CBFV, End-Tidal CO2, MAP, Pulse Pressure, and HR in CFS/POTS Compared With Control Subjects

Supine mean CBFV at any given tilt angle was not significantly different by unpaired t-test for CFS/POTS subjects (73.9 ± 1.0 cm/s) compared with control subjects (75.1 ± 1.3 cm/s) and was unaffected by n value for both groups. As shown in Fig. 3, CBFV and end-tidal CO2 decreased significantly and in parallel during incremental tilt and were highly correlated (R2 = 0.82). Changes in CBFV and end-tidal CO2 were not different between groups and were not associated with n-back responses. However, using repeated-measure ANOVA, there was a small yet significant between-group difference in the absolute values of mean CBFV and end-tidal CO2 throughout incremental tilt (P < 0.001), with larger values found in control subjects. No between-group differences were detected in systolic or diastolic CBFV.

Fig. 3.

CBFV as a function of tilt angle during incremental upright tilt. Solid lines show data from control subjects; shaded lines show data from chronic fatigue syndrome (CFS)/postural tachycardia syndrome (POTS) subjects. CFS/POTS patients had significantly decreased CBFV by repeated-measures ANOVA. There were, however, no differences in changes of CBFV at any angle.

Supine HR was higher for CFS/POTS subjects (80.2 ± 1.4 beats/min) compared with control subjects (70.8 ± 0.9 beats/min). HR was unaffected by n value for both CFS/POTS and control subjects. As shown in Fig. 4, HR increased with tilt for all subjects but was larger for CFS/POTS subjects compared with control subjects at every angle of tilt (P < 0.001).

Fig. 4.

Heart rate [in beats/min (bpm); top] and pulse pressure (bottom) as a function of tilt angle during incremental upright tilt. Solid lines show data from control subjects; shaded lines show data from CFS/POTS subjects. CFS/POTS patients had significantly increased heart rates by repeated-measures ANOVA. Post hoc heart rates at each tilt angle were also different.

MAP was similar in CFS/POTS (82.8 ± 0.4 mmHg) and control (83.0 ± 0.4 mmHg) subjects and was unaffected by tilt angle or n value in the n-back task. MAP did not increase with the n-back task. Pulse pressure was unaffected by n value for both CFS/POTS and control subjects. As shown in Fig. 4, pulse pressure decreased with tilt for all subjects but was smaller for CFS/POTS subjects compared with control subjects (P < 0.05).

N-Back Outcome During Orthostasis

N-back outcomes during orthostasis are shown in detail in Fig. 5, which shows both the number of correct answers and the reaction time for CFS/POTS patients and healthy controls as a function of n value and tilt angle.

Fig. 5.

N-back outcome measures during orthostasis. Top: numbers of correct responses. Bottom: reaction times (in ms). Data are plotted as a function of the incremental tilt angle, which is the repeated-measure factor for each n value. Solid lines show data from control subjects; shaded lines show data from CFS/POTS subjects. Control outcomes worsened (decreased number of correct responses and increased reaction times) with n value but were unaffected by tilt angle. CFS/POTS outcomes were inferior to control outcomes and deteriorated progressively with tilt angle.

Untasked (0-back) responses were similar for control and CFS/POTS subjects and were independent of tilt angle.

Control vs. CFS/POTS subjects.

The responses of control subjects were dependent on n value (P < 0.001) and were independent of tilt angle for both the number of correct responses and reaction times. Thus, cognitive and mechanical tasking were unrelated to orthostasis but dependent on task difficulty.

In contrast, responses of CFS/POTS subjects, while dependent on n value (P < 0.001), were also dependent on tilt angle, such that the responses progressively worsened as orthostasis progressed. Both between-group differences and within-group differences for CFS/POTS subjects increased monotonically with the angle of tilt, as shown in Fig. 5. Thus, in contrast to healthy control subjects, the outcomes of cognitive and mechanical tasking progressively deteriorated with orthostasis in CFS/POTS subjects.

Relation of CBFV During Cognitive Activation to N-Back Tasking and Orthostasis

Neuronal activated CBFV is shown for n-back tasks at each angle of tilt in Fig. 6.

Fig. 6.

Cognitive activated mean CBFV during orthostasis. Activated blood flow was measured by the following equation: slope = change in CBFV per minute during n-back. Data are plotted as a function of the incremental tilt angle, which is the repeated-measure factor for each n value. Solid lines show data from control subjects; shaded lines show data from CFS/POTS subjects. Control slopes increased with n value but were unaffected by tilt angle. CFS/POTS slopes statistically increased with n value, deteriorated progressively with tilt angle, and were often <0, signifying a decrease in CBFV in the middle cerebral artery.

Control activated CBF was independent of tilt angle and increased progressively with n value. CFS/POTS activated CBF also decreased with tilt angle but was significantly reduced compared with control. In contrast to control subjects, activated CBF in CFS/POTS subjects decreased with the angle of tilt.

Averaged over all n values and tilt angles, activated CBF in CFS/POTS subjects was not significantly different from n = 0. Thus, on average, there was no overall net increase in blood flow during a cognitive task in CFS/POTS subjects.

CCP, Resistance-Area Product, and PI

CCP, resistance-area product, and PI for CFS/POTS and control subjects are shown in Fig. 7, where mean CBFV = (MAP − CCP)/resistance-area product and PI = (maximum CBFV − minimum CBFV)/(mean CBFV).

Fig. 7.

Resistance-area product (top), Gosling pulsatility index (middle), and CCP (bottom) during orthostasis. Data are plotted as a function of the incremental tilt angle, which is the repeated-measure factor for each n value. Results were independent of n value for all subjects. Solid lines show data from control subjects; shaded lines show data from CFS/POTS subjects. The control resistance-area product was unaffected by tilt angle but increased in CFS/POTS subjects. The control pulsatility index was unaffected by tilt angle but decreased in CFS/POTS subjects. CCP decreased with tilt angle in control subjects. CCP was increased above control for CFS/POTS subjects and was unaffected by tilt angle.

CCP, resistance-area product, and PI were independent of n value for both subject groups.

The resistance-area product was significantly lower for CFS/POTS subjects compared with control subjects when in supine position and during 15° upright tilt (P < 0.01). The resistance-area product increased at higher tilt angles in CFS/POTS subjects so that it was not different from control subjects at 30, 45, 60, and 75°.

Changes in PI mirrored the changes in resistance-area product. PI was significantly lower for control subjects compared with CFS/POTS subjects when in the supine position and during low-angle upright tilt (P < 0.025). PI increased in control subjects at higher-angle upright tilt.

CCP in CFS/POTS subjects was always increased compared with control subjects (P < 0.001) and was not dependent on tilt angle (P = 0.18). CCP monotonically decreased in control subjects with angle of tilt, as described in the literature (8, 39).


There are three new findings in the present study.

First, n-back outcome is impaired in CFS/POTS subjects during orthostasis. There was a progressive decrease in the number of correct identifications and an increase in the reaction time of the n-back task during incremental orthostasis. Thus, orthostasis results in neurocognitive impairment in CFS/POTS subjects but not in control subjects.

Second, the expected increase of CBFV during cognitive neuronal activation, termed “neuronal activated CBFV,” is absent in CFS/POTS subjects. This may be construed as an uncoupling of CBFV from neuronal activation because cognitive activity is occurring without an increase in CBFV (and often with a decrease in CBFV).

Finally, CCP, which is directly related to vasomotor tone and thus inversely related to the efficacy of neurovascular coupling, is increased in CFS/POTS subects and fails to decrease as expected with orthostasis.

Impaired N-Back Outcome in CFS/POTS Subjects During Orthostasis

Prior investigations have examined cognitive performance in CFS/POTS subjects. While some studies were performed with subjects in the supine position and sometimes while seated, posture was never altered. However, the use of different positions (seated or supine) in prior studies could account for different results. Thus, Caseras et al. (9) used an n-back test of n = 0–3 with blood O2 level-dependent function MRI while subjects performed the test in the supine position. While they found no n-back performance differences in CFS subjects compared with control subjects, similar to our supine data shown in Fig. 5, brain activation was reduced by 2-back and 3-back, similar to our findings shown in Fig. 6. On the other hand, cognitive deficits have not been demonstrated by other investigators using a variety of cognitive testing tools (13, 17). Most of these cognitive testing tools are unsuitable for time-delimited orthostatic testing. Differences of CFS/POTS from control subjects appear to segregate with posture such that supine testing shows no difference in cognition compared with control subjects in the supine position but does show a difference in cognition compared with control subjects while seated (14, 15, 59). This is consistent with orthostatic-dependent results.

Neuronal Activated CBFV Is Paradoxically Reduced by N-Back Testing and Further Reduced During Orthostasis

In the present study, the reduction in cognitive performance in CFS/POTS subjects compared with control subjects was unrelated to change in CBFV (40). Although absolute CBFV differed by a small amount between groups, this could be accounted for by small differences in end-tidal CO2 as there is a direct correlation between CO2 and CBF (29). Also, we found no differences in the changes of CBFV between groups during tilt, nor did we find any correlation between CBFV averaged over n values and the n-back responses themselves (40). This is inconsistent with our earlier work in POTS subjects, which showed decreased CBFV (∼2-fold decrease) compared with control subjects during a nonincremental tilt to 70° (38). Moreover, estimates of cardiac output decreased in parallel with CBFV and were significantly reduced in POTS subjects compared with control subjects. The present findings suggest that incremental tilt experiments cause an acclimation of reflex responses in CFS/POTS subjects during gradual orthostatic changes. Acclimation to gradual, progressive tilting might result from nonautonomic compensation such as the venoarteriolar reflex (58).

Also, transcranial Doppler data should be interpreted cautiously because it measures global hemispheric CBFV and is likely insensitive to small localized changes. However, its use is complementary to tilt methodology because of its rapid temporal resolution. More elaborate imaging modalities cannot provide temporally continuous measures of CBF and are not feasible during tilt. Nevertheless, previous studies (3, 11, 28) using computed tomography (CT)-, single-positron emission CT-, or MRI-based methods have shown abnormalities in regional cerebral perfusion in CFS patients, including regions of increased blood flow and areas of decreased blood flow during cognitive tasking within the cortical domain perfused by the MCA.

Our data concerning the cognitive activation of CBFV is consistent with CBF heterogeneity reported in CFS/POTS. Overall, our data show that blood flow failed to increase during an n-back task for all n values, with clear reductions in task-dependent neuronal activated CBFV as the orthostatic stress increased. This is in contrast with control subjects, in whom the progressive task difficulty enhanced task-dependent blood flow, which is also orthostasis independent.

Neuronal activity and CBF are tightly coupled in both the resting and activated brain. Thus, neuronal activity causes increased CBF (2), termed as “functional hyperemia.” The coupling is sufficiently tight such that local neuronal activity can be assessed by measuring regional blood flow (23).

Thus, increased CBF during mental tasks is expected and reflects enhanced neuronal activity. The present data imply that net increases in blood flow in the area perfused by the MCA is absent during n-back testing in CFS/POTS subjects; this absence of task-related hyperemia may relate to a disruption in neurovascular coupling in CFS/POTS, as would occur if there were altered O2 extraction in CFS/POTS compared with control subjects or reduced neuronal activity. Disruption in neurovascular coupling is known to occur in diabetes, depression, hypertension, stroke, and Alzheimer's disease (23). The hypothesis of disrupted neurovascular coupling receives support from the reduction in n-back performance with angle of tilt in CFS/POTS subjects but not in control subjects.

CCP (and Therefore Peripheral Vasomotor Tone) Is Greatly Increased in CFS/POTS Subjects

A loss of neurovascular coupling is consistent with finding that CCP is increased in CFS/POTS subjects, which reflects an increase in vasomotor tone and a loss of ability for the neurovascular unit to dilate appropriately during progressive larger angles of tilt and orthostatic stress. While the decreased resistance-area product at low tilt angles may in part compensate for this finding, the resistance-area products increased to equal control values while CCP remained elevated. Vasomotor tone was elevated at the level of the neurovascular unit, which may therefore be malfunctioning. The decrease in CCP in control subjects during upright tilt has been observed previously in healthy volunteers (8, 39) and comprises a small contribution from decreasing intracerebral pressure during orthostasis and a larger contribution that compensates for the reduced perfusion pressure at the level of the MCA as part of the autoregulatory response. The Gosling PI, designed to measure vascular resistance, was also increased in CFS/POTS subjects compared with control subjects at lower angles of tilt.

In summary, an n-back task was used to test cognitive function in CFS/POTS subjects compared with healthy control subjects during orthostatic stress. Performance decreased with difficulty (n value) in CFS/POTS and control subjects and was reduced in CFS/POTS subjects but in not control subjects during orthostasis. While CBF is normally tightly linked to neuronal activation, as proved true for control subjects, CBF was essentially unrelated to cognitive performance in CFS/POTS subjects. Combined with persistently increased CCP/vasomotor tone, this indicates uncoupling of the neurovascular unit during orthostasis in CFS/POTS subjects and results in cognitive loss, often called “mental cloudiness,” which can be remediated by supine posture.


Neurocognition cannot be assessed by any single test, and its definition remains a subject for consensus (25). The n-back test was chosen as an acceptable means to assess working memory, attention, reaction time, and information processing that could be step wise increased in difficulty and could be rapidly delivered in the supine and upright position.

Graded tilt table testing is not identical to standing. However, CFS patients often do not remain upright for protracted periods of time, preferring to sit (21). An incremental tilt may be representative of varied postural circumstances throughout a normal day; each angle of tilt may be pertained to graded orthostatic stress during daily life.

TCD only measures blood flow through a particular cerebral blood vessel but has good temporal accuracy. The MCA was used because it is the main vessel that perfuses the area of the brain activated by the n-back test (5). TCD does not have regional accuracy, and therefore the values of CBFV obtained may reflect an average over areas perfused. In some areas perfusion may increase with orthostasis, whereas in other areas perfusion may decrease, particularly during cognitive tasks, where a small but definite increment in blood flow is expected in association with cognitive neuronal activity. Also, we did not separately display the left and right MCA velocity data. We found that there were no differences in data from the right and left MCAs and therefore averaged their signals. It is possible that subsequent analysis would have shown some disparity between these signals.

The pressure-blood flow relationship defining CCP is believed to be curvilinear at low pressures (42), but a linear approximation is generally used. Such low pressures are neither safe nor feasible in patient-oriented research, and a linear approximation is the rule. When the two-parameter model was used in anesthetized primates, the linear fit was determined to be statistically satisfactory (16).

Young patients with CFS may not all have POTS. However, the consensus among investigations of pediatric and adolescent subjects with CFS suggests that almost all young CFS patients have OI. The present results may not pertain to CFS patients without POTS, as often occurs in older CFS patients.

Methods to assess neurocognitive behavior during tilt may be dependent on subjective influences. Undergoing head-up tilt or n-back testing can be more potent psychological stressors for CFS/POTS patients than for control subjects and may therefore affect neurocognitive performance. More complex experiments with sham interventions and perhaps time controls for tilt could be performed to overcome this limitation.

MCA blood flow may depend on cardiac output and thus on stroke volume, although the results remain controversial (12, 56). A Finometer was used to measure ABP, and this includes a module called “ModelFlow,” which computes a Windkessel-based estimate of cardiac output. These methods have not always been accurate (18, 51) and were excluded from the study.

The Gosling PI was developed and validated for use in the peripheral vasculature (24, 33), and its use in the cerebral circulation was adopted thereafter. Results must therefore be interpreted cautiously in the present analysis. However, the key findings were not strictly dependent on PI but rather on the validated use of CCP analysis (16).


This work was supported by National Heart, Lung, and Blood Institute Grants 1-F30-HL-097380 (to A. J. Ocon), 1-RO1-HL-074873 (to J. M. Stewart), and 1-RO1-HL-087803 (to J. M. Stewart) and by a grant from the Chronic Fatigue and Immune Deficiency Syndrome Association of America (to M. S. Medow).


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: J.M.S. and A.J.O. conception and design of research; J.M.S., Z.R.M., I.L.B., C.T., and A.J.O. performed experiments; J.M.S., M.S.M., Z.R.M., and A.J.O. analyzed data; J.M.S., M.S.M., C.T., and A.J.O. interpreted results of experiments; J.M.S. and A.J.O. prepared figures; J.M.S. drafted manuscript; J.M.S., M.S.M., Z.R.M., I.L.B., C.T., and A.J.O. approved final version of manuscript; M.S.M. and A.J.O. edited and revised manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
View Abstract