interaction of hypocapnia, hypoxia, brain blood flow, and

Hyperventilation is a relevant problem in biology andmedicine. Its importance arises from the fact that respira-tion is vital in supplying the body with oxygen. However,excessive ventilation, exceeding the level required to cover

the body’s oxygen requirement [10], has negative biologicalvalue. This is because of a number of factors induced direct-ly or indirectly by impairment of central nervous systemactivity [6, 23, 29, 33]. Marked hyperventilation can lead tothe onset of serious impairments in the body’s acid-basebalance, which accompany hypocapnia [12]. The develop-ment of artificial life support systems including mechanicalventilation of the lungs increases the risk of hyperventila-tion [10, 28].

Despite the long history of this problem [6, 18, 23, 28,33], many questions remain to be answered. In particular,

Neuroscience and Behavioral Physiology, Vol. 38, No. 7, 2008

Interaction of Hypocapnia, Hypoxia, Brain Blood Flow,and Brain Electrical Activity in Voluntary Hyperventilation in Humans

É. A. Burykh

0097-0549/08/3807-0647 ©2008 Springer Science+Business Media, Inc.

647

Translated from Rossiiskii Fiziologicheskii Zhurnal imeni I. M. Sechenova, Vol. 93, No. 9, pp. 982–1000,September, 2007. Original article submitted March 16, 2007.

Changes in various physiological measures in voluntary hyperventilation lasting three minutes or more inhumans were studied and compared. Three-minute hyperventilation, in which the rate of external ventila-tion increased by an average factor of 4.5–5, produced similar phasic changes in central and brain hemo-dynamics. The rate of circulation, indicated by rheographic data, initially increased during hyperventila-tion, reaching a maximum at 1–2 min of the test; there was then a reduction, to a minimum 2–3 min afterthe end of the test; this was followed by a further slow increase. The rate of cerebral blood flow during all3 min of hyperventilation remained elevated in most subjects as compared with baseline and decreasedduring the 5 min following the end of the test. Transcutaneous carbon dioxide tension changed different-ly – there was a decrease to a minimum (about 25 mmHg) by the end of the test, lasting 1 min from theend of the test, this being followed by an increase to a level of 90% of baseline at 5 min after the test.Blood oxygen saturation remained at 98–100% during the test, decreasing to about 90% 5 min after thetest; this, along with the decrease in cerebral blood flow, was a factor producing brain hypoxia. In differ-ent subjects, changes in the spectral power of oscillations in different EEG ranges on hyperventilationwere “mirrored” to different extents by the dynamics of transcutaneous carbon dioxide tension. The dura-tion and repetition of hyperventilation were important factors for understanding the interaction betweenbrain hemodynamics, hypocapnia, hypoxia, and brain electrical activity. After several repetitions of 3-minhyperventilation over a period of 1 h, the increasing brain blood flow could decrease significantly on thebackground of relatively small changes in brain electrical activity. The data presented here were assessedfrom the point of view of the important role of brain tissue oxygen utilization mechanisms in adaptationto hypoxia and hypocapnia.

KEY WORDS: hyperventilation, hypocapnia, cerebral circulation, hypoxia, brain electrical activity, blood oxygen sat-uration, oxygen consumption.

Interinstitute Laboratory for Comparative Ecological-Physiological Studies, I. M. Sechenov Institute of EvolutionaryPhysiology and Biochemistry, Russian Academy of Sciences,44 M. Torez Prospekt, 194223 St. Petersburg, Russia; “Arktika”International Science Research Center, Far Eastern Branch,Russian Academy of Sciences, 13 K. Marx Street, 683000Magadan, Russia.

the roles of hypocapnia, changes in cerebral hemodynam-ics, and probable brain tissue hypoxia inducing abnormali-ties in brain electrical activity on hyperventilation remainunclear. The most difficult part of this question is therepeatedly demonstrated persistence of a normal level ofbrain oxygen consumption during hyperventilation despitethe reduction in cerebral blood flow [25, 28, 29]. Anotherrelatively poorly studied question is that of the influence ofthe duration of hypocapnia on cerebral blood flow in hyper-ventilation. The present study was performed to seekanswers to these and other questions.

METHODS

A total of 16 essentially healthy subjects of both gen-ders, aged 20–40 years, took part in this study. Thirteentook part in studies of 3-min voluntary hyperventilation.Two performed several repetitions of hyperventilation, withincreasing volumes, over the period of one hour. One sub-ject performed 30-min hyperventilation.

The following physiological parameters were assessed:external respiratory rate using a “Diamant” (Russia) com-

puterized spiroanalyzer, transcutaneous carbon dioxide ten-sion using a TSMZ transcutaneous monitor (Denmark),blood oxygen saturation using a Nonin pulsoximeter(USA), and heart rate; the central blood flow rate and cere-bral blood flow rate were monitored using an “Éntsefalan131-03” computerized encephalograph with a polygraphchannel, which was also used to record the EEG.

The central circulation rate was assessed on the basisof stroke and minute volumes as described by Tishchenko[16]. The rheoencephalogram was recorded in the occipito-mastoid and frontomastoid leads on the left and right(OML, OMR, FML, and FMR, respectively). This allowedassessment of parameters of brain blood flow in the basinsof the vertebral and internal carotid arteries on the leftand right.

The EEG was recorded using the international 10–20scheme from 16 leads (Fp1, Fp2, F3, F4, C3, C4, P3, P4,O1, O2, F7, F8, T3, T4, T5, and T6). The spectral power ofEEG oscillations was determined over the range 0–30 Hzand in the delta (0–4 Hz), theta (5–7 Hz), alpha (8–13 Hz),and beta (14–30 Hz) ranges.

Physiological parameters were assessed in baselineconditions, at each minute of hyperventilation, and at each

Burykh648

Fig. 1. Changes in rheoencephalogram shape during and after 3-min hyperventilation in subject I. A) Baseline; B, C, D) minutes1, 2, and 3 of hyperventilation, respectively; E, F, G, H) 1, 2, 3, and 5 min, respectively, from the end of hyperventilation. Triangleson curves show the dicrotic incisure; circles show the late systolic component.

minute after hyperventilation. This allowed assessment ofrelative changes in physiological parameters during andafter hyperventilation in relation to baseline.

Assessments of physiological parameters averaged bysubjects were analyzed using Statistica for Windows ver-sion 5.

RESULTS

Rheoencephalography (REG) is one of the most appro-priate methods for detailed evaluation of relative changes inbrain blood flow in response to different influences on thebody [21]. This method undoubtedly provides only indirectassessments of the characteristics of brain circulation. Theformulae developed for transforming measures to parame-ters generally used in relation to circulation in the organs interms of physiological units have not found wide use.However, this does not prevent REG from being used forassessment of relative changes in brain blood flow duringapplication of various treatments to the body. This resultsfrom the direct proportionality between variations in theelectrical impedance of brain tissue and variations in pulsefilling in the region of interest [21]. Thus, assessment of rel-

ative changes in the amplitudes of REG variations providesan adequate reflection of relative changes in pulse flow.

Figure 1 shows changes in the shape of the REG alongwith quantitative assessments during and after hyperventi-lation. This shows curves obtained by superimposition of50–70 REG traces corresponding to individual cardiaccycles in 1-min time intervals. The amplitude of the majorwave (which in this situation is the arterial component) inREG traces averaged by this method decreased from0.129 Ω in baseline to 0.91 Ω at the end of hyperventilation,subsequently increasing after the end of hyperventilationbut not reaching the baseline level (Fig. 1; Fig. 2, F).

It should be noted that use of the REG amplitude inisolation failed to characterize the rate of cerebral circula-tion, just as use of the stroke volume of the heart in isolationfails to characterize the central circulation. Pulse frequencyneeds to be known in both cases. Practical rheoencephalog-raphy uses measures calculated as the ratio of the REGamplitude or the amplitude of the arterial component (or thesum of the arterial and venous components) to the durationof the cardiac cycle [21]. In essence, these are measures ofthe rate of cerebral circulation, as they allow assessmentof the volume of blood entering the brain during a specifiedperiod of time. Brain blood flow rate (or, more precisely, the

Interaction of Hypocapnia, Hypoxia, Brain Blood Flow, and Brain Electrical Activity 649

Fig. 2. Dynamics of several physiological measures during and after 3-min hyperventilation in subject I. DCI REG = dicrotic index of therheoencephalogram; TCCO2T = transcutaneous carbon dioxide tension; AAC REG = amplitude of the arterial component of the rheoen-cephalogram; Saturation = blood oxygen saturation; BBFP = rheographic brain blood flow rate parameter in lead FMR; SV = cardiac strokevolume; MV = cardiac minute volume; MVR = minute volume of respiration. Circles on plots show values during hyperventilation.

relative changes in this measure) was assessed using theratio of the mean amplitude of the arterial component tothe mean duration of the cardiac cycle over a defined peri-od of time. This measure is identical to the parameter A60,i.e., the amplitude of the arterial component normalizedwith respect to a pulse rate of 60 bpm [21], though we pre-fer the term “rheographic brain blood flow rate parameter”(BBFP), as it reflects the physiological sense of the param-eter measured rheographically.

During hyperventilation by subject I, BBFP in leadFMR initially increased and then decreased, albeit remain-ing above baseline throughout the test (Fig. 2, I). Thisdecrease continued to 2 min after the end of hyperventila-tion, after which there was some increase. Similar changesoccurred on the background of opposite changes in theamplitude of the arterial component and heart rate. Duringhyperventilation, the increase in heart rate was greater thanthe decrease in the amplitude (in other words, the decreasein the mean cardiac cycle was greater than the decrease inamplitude). After hyperventilation, conversely, the decreasein heart rate was greater than the increase in amplitude.

Another important characteristic of brain blood flow isthe tone of the major vessels. In relation to hyperventilation,this is an important parameter, as many investigators see theincrease in the tone of the major vessels as the cause ofthe decrease in brain blood flow in hyperventilation [12, 26,28, 33]. In rheoencephalography, a number of quantitativeassessments are widely used in interpreting rheographiccurves, these providing indirect characterization of the tone

of the major vessels. First among these is the dicrotic index[21]. This is the ratio of the amplitude of the REG at thelevel of the dicrotic incisure to the maximum amplitude ofthe REG. Another parameter is the ratio of the REG at thelevel of the late systolic wave (Fig. 1) to the REG peak.The values of both parameters reflect the tone of the resis-tive vessels of the brain [21]. Our studies used the dicroticindex (DCI) as, unlike the late systolic wave, the dicroticincisure is always clearly detected.

The dicrotic index in subject I decreased on hyperven-tilation, increasing after hyperventilation to a level greaterthan baseline (Fig. 1; Fig. 2, C).

Comparison of changes in the dicrotic index and mea-sures of brain blood flow rate (Fig. 2, C, I) could give theimpression that the brain blood flow rate in subject I increasedduring hyperventilation because of a decrease in the tone ofthe major vessels (reduction in DCI), the post-hyperventila-tion decrease resulting from an increase in tone (increase inDCI). However, this conclusion would be premature, consid-ering the complex interaction in the whole-body situationbetween heart rate, cardiac stroke volume, tone of resistivevessels, arterial pressure in the system of one vessel or anoth-er, and the intensity of circulation in its basin [5].

Thus, MV (circulatory minute volume) and measuresof brain blood flow rate in lead FMR in the subject under-went similar changes during hyperventilation: an increase at1 min followed by sequential decreases at 2 and 3 min. Theminimal values of MV and the rheographic measure ofbrain blood flow rate were also seen at the same minute

Burykh650

TABLE 1. Statistical Assessments of Physiological Parameters During and After Hyperventilation for the Group of Subjects

Notes. MVR = minute volume of respiration, TCCO2T is transcutaneous carbon dioxide tension; HR is heart rate; MV is cardiac minute volume; BOS isblood oxygen saturation; BBFP is the rheographic brain blood flow rate parameter in leads FMR and OMR; DCI is the dicrotic index of the rheoen-cephalogram.

after the end of hyperventilation, i.e., at 2 min (Fig. 2, H, I).Thus, assessment of changes in brain blood flow requiresconsideration not only of changes in the tone of the majorvessels, but also changes in central circulation.

Consideration of mean values in different subjects dur-ing and after 3-min hyperventilation (Table 1) can yieldmore precise assessments of the role of changes in the cen-tral circulation and vascular tone in changes in brain bloodflow rate. Thus, MV at 1 min of hyperventilation increasedby an average of 8%, while the rheographic measure of thebrain blood flow rate in lead FMR increased by 42%. Thedicrotic index decreased by 29%, and was thus 71% ofbaseline. Thus, both the increase in MV and the decrease inthe tone of the major brain vessels play roles in the increasein the brain blood flow rate at 1 min of hyperventilation,though the latter is in all probability the more important.Five minutes after the end of hyperventilation, MVdecreased by an average of 15% of baseline, while the rheo-graphic measure decreased by a similar amount, i.e., by19%. The dicrotic index increased by 15%. Thus, here againwe can identify the need to consider both changes in the cen-tral circulation and the tone of brain vessels in the dynamicsof brain blood flow rate.

Assessment of the dynamics of the REG measure inlead OMR showed values very close to those seen in leadFMR. The mean correlation of the dynamics of rheoen-cephalogram measures in different leads was very high, atthe level of 0.90–0.95, which is evidence of the brain-widenature of changes in cerebral blood flow.

Comparison of the dynamics of transcutaneous carbondioxide tension and the rheographic measure of brain bloodflow rate showed their changes to be different in nature(Fig. 2, D, I; Table 1). In contrast to the biphasic nature of thedynamics of the transcutaneous carbon dioxide tension(decrease on hyperventilation, increase after hyperventila-tion), the dynamics of the rheoencephalographic measure ofbrain blood flow rate were triphasic, with an increase at thestart of hyperventilation, followed by a decrease after reach-ing a peak at 1–2 min of the test, which continued to 2–3 minafter the end of the test, and then a gradual increase. These dif-ferences were assessed quantitatively. Thus, the correlation

coefficient of the dynamics of transcutaneous carbon dioxidetension and the measure of the brain blood flow rate in leadFMR had a value close to zero: –0.03 ± 0.32 (Table 2).

The dynamics of blood oxygen saturation duringhyperventilation clearly warrant more detailed considera-tion than allowed by the remit of the present report. We noteonly that the decrease in blood oxygen saturation was toabout 90% at the end of hyperventilation. In the context ofthe present report, this point is interesting in relation to thepossible role of hypoxia in changes in brain electrical activ-ity on hyperventilation.

Table 2 shows quantitative assessments of the similar-ities of the dynamics of the various physiological measuresduring and after hyperventilation. Attention is drawn tothe relatively higher level of the correlation between thedynamics of heart rate and the rheographic measure of brainblood flow rate (0.86 ± 0.10) as compared with the correlationbetween the dynamics of heart rate and MV (0.67 ± 0.19). Thelevel of similarity between the dynamics of the dicroticindex of the REG and the measure of brain blood flow ratewas higher in terms of its absolute value (0.78 ± 0.15) thanthe similarity between the dynamics of MV and the measureof brain circulation rate (0.68 ± 0.23).

An extremely important factor in assessment of theinteractions between hypocapnia and brain blood flow inhyperventilation is the temporal aspect of these interactions.Thus, in 30-min hyperventilation in subject Kh (Fig. 3), therheographic measure of brain blood flow rate decreasedfrom baseline during hyperventilation itself. In most cases,this did not occur during 3-min hyperventilation. Attentionshould be drawn to the point that the brain blood flow ratecontinued to decrease from the 17th to the 29th minute ofhyperventilation on the background of a virtually unalteredlevel of transcutaneous carbon dioxide tension of 25 mmHg.It is also essential to note that about 20 min after the end ofhyperventilation, the brain blood flow rate became greaterthan baseline, while the transcutaneous carbon dioxide ten-sion remained below baseline.

Changes in brain electrical activity on hyperventilationdepended on the individual characteristics of the EEG [10]and increased as hyperventilation progressed. It was there-


TABLE 2. Statistical Assessments of the Ratios of Various Physiological Parameters in Hyperventilation

Notes. The table cells show group average correlation coefficients for the dynamics of the corresponding pairs ofmeasures during the period including the baseline, hyperventilation, and the post-hyperventilation period. For fur-ther details see Table 1.

fore necessary to present the dynamics of some of the spec-tral measures in several subjects (Fig. 4). In most of the sub-jects plotted (Fig. 4, A, C, E, G, I), the dynamics of the totalEEG spectral power in leads F3 and F4 were similar, withan increase during hyperventilation to a peak close to theend point of hyperventilation, followed by a reduction inthe post-hyperventilation period. However, some subjectscould also show deviations from this type of change in EEGspectral power. Thus, in subject P (Fig. 4, D), the spectralpower peak occurred in the middle rather than at the end ofhyperventilation. In subject V (Fig. 4, B, I), EEG spectralpower at the end of hyperventilation reached a plateau.

However, it follows from the results presented inTable 3 that all EEG ranges were dominated by a pattern ofchanges in spectral power consisting of a sequentialincrease in this measure during hyperventilation and a sub-sequent decrease at the end of the test. In this sense,changes in EEG spectral power “mirrored” changes in tran-scutaneous carbon dioxide tension. It also followed fromthe results presented in Table 3 that the changes in spectraldensity in the alpha range occurred more gradually than thechanges in this measure in other ranges. This was particu-larly characteristic for the post-hyperventilation period. It isentirely possible that the more gradual nature of thesechanges in spectral power in the alpha range resulted fromthe relatively greater dependence of variations in this rangeon another physiological parameter, changes in which werealso sequential, i.e., transcutaneous carbon dioxide tension.

We then evaluated the extent of similarity betweenchanges in EEG spectral power density in different rangesand changes in transcutaneous carbon dioxide tension dur-ing and after hyperventilation in each subject by analyzingcorrelation coefficients for the corresponding curves.Averaged correlation coefficients are shown in Table 4.

The results presented in Table 4 indicate that the great-est degree of similarity in the dynamics of transcutaneouscarbon dioxide tension and EEG spectral power was seenfor the alpha range. The negative sign of this correlation isevidence that the similarity is a “mirror” relationship. It isimportant to note that the degree of similarity between thedynamics of transcutaneous carbon dioxide tension andEEG spectral power could show marked differences notonly for individual ranges, but also for individual EEGleads. Thus, the correlation coefficient for the alpha range inlead T3 was –0.79, compared with –0.53 for lead O2.

Our attempt to answer the question of what might affectthe degree of similarity between changes in transcutaneouscarbon dioxide tension and the dynamics of EEG spectralpower density in one range or another involved comparisonof the EEG spectral characteristics at essentially identicallevels of transcutaneous carbon dioxide tension during andafter hyperventilation. Transcutaneous carbon dioxide ten-sion at 1 min of hyperventilation was comparable withthat 5 min after hyperventilation ended (Fig. 5, 1, A), whilethat at 2 min of hyperventilation was comparable with that2 min after the end of hyperventilation (Fig. 5, 1, B), that at3 min of hyperventilation being comparable with transcuta-neous carbon dioxide tension at 1 min after hyperventilation(Fig. 5, 1, C).

Despite the essentially identical transcutaneous carbondioxide tension levels, pulmonary ventilation, as expected,differed markedly during and after hyperventilation (Fig. 5,1–3, C). Differences were also seen for spectral power inthe EEG theta range, particularly in lead P3 (Fig. 5, 1–3, B).The mean value of this parameter showed no statisticallysignificant differences during and after hyperventilation(Fig. 5, 1–3, D). This supports the existence of a significantinterindividual spread in measurements of this parameter.

Burykh652

Fig. 3. Dynamics of the rheographic brain blood flow rate parameter (BBFP) in lead FMR andtranscutaneous carbon dioxide tension (TCCO2T) during 30-min hyperventilation in subject Kh.1) BBFP; 2) TCCO2T. Arrows show the beginning and end of hyperventilation.

However, on comparison of the number of subjects inwhom EEG spectral power density in the theta range duringhyperventilation was greater on hyperventilation than at thesame time after hyperventilation with the number of sub-jects in which the reverse relationship held, the signs crite-rion identified a clear statistical relationship. This was thatat different levels of transcutaneous carbon dioxide tension,EEG spectral power density in the theta range in lead P3was greater during hyperventilation than after hyperventila-tion. This was particularly clear on comparison of the thirdminute of hyperventilation with the first minute after hyper-ventilation (Fig. 5, 3, D).

Figure 5, 1–3, E shows the ranges of applicability andlevels of statistical significance of this relationship for thetheta range. These results lead to the conclusion that ashyperventilation continues, there is an increase in the sig-nificance of the extent to which the EEG spectral powerdensity in the theta range exceeded its value at the time atwhich the transcutaneous carbon dioxide tension showedthe same change after hyperventilation; the number of EEGleads in which this relationship was seen increased. Thismeans that the proportional influence on brain electricalactivity of some factor linked with hyperventilation itself,but other than hypocapnia, increased.


Fig. 4. Variations in the dynamics of EEG spectral power in the theta range in different subjects (A–J) during and afterhyperventilation. Thick lines show EEG lead F3; thin lines show EEG lead F4. Arrows on plots show the beginning andend of hyperventilation. The ordinates of all plots show EEG spectral power in the theta range, µV2.

With the aim of evaluating the relationship betweenhypocapnia, hypoxia, brain blood flow, and brain bioelec-trical activity at different levels of hyperventilation, we per-formed a study using dosed hyperventilation. The resultsare presented in Fig. 6. In subject V (Fig. 6, 1, A), therequired volume of hyperventilation was specified as multi-ples of the baseline level, i.e., 6–11 times greater than thebaseline minute volume of respiration. The feedback systemallowed the subject to maintain the rate of external respira-tion at the specified level to an error of 10%. In subject K,the specified volume of external respiration initiallyincreased the minute volume of respiration by a factor of2–4 and then by a factor of 7 compared with baseline(Fig. 6, 1, A). The regime in which this increasing hyper-ventilation (3 min “work,” 5 min “rest”) was performed ledto the following changes in transcutaneous carbon dioxidetension: values decreased during each episode of hyperven-tilation and increased after the end of each episode withoutreaching the initial level (Fig. 6, 2, A, B). Thus, each newcycle of hyperventilation started with somewhat smallervalues for transcutaneous carbon dioxide tension. Overall,in these ranges of hyperventilation, both subjects showed analmost linear relationship between the volume of hyperven-tilation and carbon dioxide tension. At the end of the cyclesof hyperventilation, transcutaneous carbon dioxide tensiongradually increased, though it did not reach the baselinelevel by 15 min.

The rheographic measure of brain blood flow duringcyclic hyperventilation decreased in parallel with the reduc-

tion in the level of transcutaneous carbon dioxide tension(Fig. 6, 3, A, B). On this background, both subjects showedmarked increases in BBFP, seen during performance ofhyperventilation, and decreases, seen during the “rest” peri-ods. Attention is drawn to the fact that brain blood flow rateafter cycles of hyperventilation showed some increase,nonetheless remaining at a markedly lower level comparedwith baseline, despite the increase in transcutaneous carbondioxide tension.

EEG spectral power in the theta range in lead F3(Fig. 6, 3, A, B) increased in both subjects during “work”periods in cyclic hyperventilation, rapidly reaching thebaseline level during the “rest” periods. Impairment of“mirroring” changes in spectral power in relation to changesin transcutaneous carbon dioxide tension was particularlynotable here, as shown in Fig. 5.

The increase in EEG spectral power in the theta rangewith the increase in the volume of hyperventilation becamegreater in both subjects, though this relationship was notstrictly followed in subject V (Fig. 6, 3, A).

Attention is drawn to the fact that 2 min after cyclichyperventilation in subject V and 3 min after cyclic hyper-ventilation in subject K (Fig. 6, 3, A), EEG spectral powerin the theta range decreased to baseline and remained essen-tially unchanged for 10 min after the test. As already noted,transcutaneous carbon dioxide tension and the rheographicmeasure of brain blood flow rate were significantly reducedcompared with baseline. It would appear that “normaliza-tion” of the EEG occurred long before recovery of each

Burykh654

TABLE 3. Dynamics of EEG Spectral Characteristics During and After Hyperventilation.

Notes. SP delta (etc.) = spectral power in the EEG delta (etc.) range. SP 0–30 = spectral power over the EEG range 0–30 Hz. Abs., µV2 = absolute values; Rel. = relative values, % of baseline.

subject’s characteristic carbon dioxide tension and brainblood flow rate.

DISCUSSION

EEG changes during hyperventilation were markedand were similar to those occurring in hypoxia [10, 11, 30].It is therefore not surprising that many authors are inclinedto explain these changes in brain electrical activity in termsof hypoxia. It is suggested that brain hypoxia occurs as aresult of the reduction in brain blood flow arising becauseof the increase in brain vessel resistance [6, 26, 32].

In fact, decreases in brain blood flow rate in hyperven-tilation have been demonstrated in many studies [12, 21, 25,28, 33]. It is difficult not to agree with the view that thismay be a factor in the reduction in the rate at which oxygenreaches the brain. However, the facts identified in the pre-sent study contradict the suggestion that circulatory hypox-ia plays the major role in producing EEG impairments.Firstly, EEG changes in the first minutes of hyperventilationcan occur on the background of even higher brain bloodflows than in conditions of normocapnia. Secondly, in pro-longed hyperventilation on the background of quite markeddecreases in brain blood flow rate, the EEG may not showsignificant changes in the form of increases in the spectralpower of those EEG ranges which react “sensitively” tohypoxia (Fig. 6).

In our view, it is often insufficient to consider the tis-sue and cellular mechanisms of adaptation to hypoxia whenassessing the mechanisms of the influences of any type ofhypoxia on the functions of the body and individual

organs [3]. In classical physiological experiments at thewhole-body level, of course, these mechanisms are difficul-ty to study with precision. However, there is an integralparameter of their outcomes – the oxygen consumption byan organ or its structural elements. If oxygen consumptionby an organ does not decrease in hypoxia, then hypoxia iscompensatory in nature. This is particularly clear in so-called loading hypoxia [4], when, at a decreased muscleoxygen tension, the tissue not only does not die, but is ableto perform a significant level of function.

In hyperventilation and its associated hypocapnia, manyauthors have demonstrated the persistence of [24, 25, 29] andeven some increase in the rate of oxygen consumption bythe brain [28] despite the reduction in the brain blood flowrate and the decrease in brain oxygen tension. In our view,this explains the absence of marked EEG changes whenthere are gradual decreases in the brain blood flow rate by afactor of almost two (Fig. 6).

Brain electrical activity reflects its functional state [15]and the functional state of its structural elements. In hypox-ia, there are decreases in the numbers of neurons showingspike activity on the background of a decrease in the domi-nant EEG frequency from levels in the alpha range (8–12 Hz)to that in the theta (5–7 Hz) and delta (0.5–4 Hz) ranges[1, 22]. Hypoxia also suppresses focal postsynaptic poten-tials in studies on cortical slices [14]. This is completelyexplicable given the generation of action potentials and thesubsequent repolarization of neurons associated with ATPcosts [19]; hypoxia threatens a deficiency of ATP, which ismainly synthesized by oxidation. Consequently, if brainoxygen consumption does not decrease in hyperventilation,then the energy provision of its organ-specific functions


TABLE 4. Extent of Similarity in the Dynamics of EEG Spectral Power in Different Ranges with theDynamics of Transcutaneous Carbon Dioxide Tension

Note. Table cells show group average correlation coefficients for the dynamics of EEG spectral powerin different frequency ranges and transcutaneous carbon dioxide tension.

may continue even in conditions of decreased brain bloodflow. This, in our view, explains the possible absence ofmarked changes in the EEG on gradual adaptation tohypocapnia (Fig. 6).

It is more difficult to explain why “hypoxic” shifts inthe EEG can arise during the first minutes of hyperventila-tion on the background of an increase in the brain bloodflow rate. However, before addressing this question, it isimportant to determine whether we are dealing with anincrease in the brain blood flow rate in the first minutes ofhyperventilation. There is now no doubt that hyperventila-tion and hypocapnia lead to decreases in the brain bloodflow rate [6, 12, 26, 28, 33]. Furthermore, it has beendemonstrated that within a certain range of blood carbondioxide tension, there is an almost linear relationshipbetween these parameters and the brain blood flow rate[33]. A decrease in brain blood flow rate was also seen inthe present study, as indicated by rheoencephalographicdata. However, it is important to note that this did not startimmediately and did not coincide with the peak of hypocap-nia. Furthermore, brain blood flow at 1 min of hyperventi-lation increased, despite the increase in the extent ofhypocapnia. The increase in blood flow, according torheoencephalographic data, was also seen in [17]. In [33],

which is key to understanding the influence of carbon diox-ide on brain blood flow, we found no indications as to thetiming of blood flow measurements in terms of the durationof hyperventilation at which a certain level of hypocapniawas reached. However, indirect data (duration of i.v. admin-istration of glucose over 4 h) indicated that there were timeperiods of 20–30 min between assessments of blood flow ateach stage of hypocapnia. It is likely that the authors foundthat the decrease in blood flow to a particular level at eachstage was not immediate, and that recovery after hyperven-tilation was also not immediate. This existence of this iner-tia in the response of brain blood flow to hypocapnia is indi-rectly demonstrated in Fig. 3. On the background of thisperiodicity and in the presence of other influences, brainblood flow is likely to increase despite the decrease in bloodcarbon dioxide tension. These influences, which overpowerthe influences of hypocapnia on brain blood flow, mayinclude an increase in the central circulatory rate, whichprobably arises as a response to physical loading – anincrease in the work performed by the respiratory muscles.With respect to this point we are in agreement with theauthor of [17].

The possibility that significant impairments in cellfunctions arise long before the critical decrease in oxygen

Burykh656

Fig. 5. Comparison of EEG spectral power at similar levels of transcutaneous carbon dioxide tension during and after 3-min hyperventilation. a) Duringhyperventilation; b) after hyperventilation. 1) Comparison of measures during the first minute of hyperventilation with measures during the fifth minuteafter hyperventilation; 2) second minute of hyperventilation versus second minute after hyperventilation; 3) third minute of hyperventilation versusfirst minute after hyperventilation. A–C) Transcutaneous carbon dioxide tension (TCCO2T), minute volume of respiration (MVR), and EEG spectralpower (EEG SP) in lead P3, averaged for subjects. D) Comparison of the number of subjects in whom EEG SP values in lead P3 were greater duringhyperventilation than in the comparison minute after hyperventilation. E) Leads showing statistically significant excesses of EEG SP during hyper-ventilation over levels in the corresponding post-hyperventilation period: small squares show p < 0.05; large squares show p < 0.01. Asterisks in plotsB and D show statistically significant differences: *p < 0.05; **p < 0.01.

tension, when the rates of oxygen consumption and ATP for-mation decrease, has been discussed in the literature sincepublication of the studies of Opitz and Schneider [31, 35].

In relation to this problem, which is termed the hypox-ic paradox [8], three aspects of this phenomenon will bediscussed here: 1) Should these functional changes betermed lesions or would it be more appropriate to call themreductions in or restrictions of functional activity? 2) Whatare the mechanisms of this phenomenon? 3) Why can thesefunctional changes occur on the background of unchangedoxygen consumption?

In [7], studies of the effects of hypoxia on isolated hep-atocytes showed that urea synthesis decreased by 40% at alevel of hypoxia at which the intracellular ATP leveldecreased by a total of 5%. We distinguish the views ofauthors who believe that in conditions of overload of theenergy-synthesizing system, restriction of the energy-con-

suming process of urea synthesis will facilitate the redistri-bution of energy to support other processes of greater vitalimportance. Restriction of neuron spike activity in condi-tions of hypoxia should in all probability be regarded as astrategy for redistributing energy flows to support the func-tions required to maintain the structural integrity of neu-rons. The mechanisms whereby the energy-consumingfunctions of cells are sensitive to decreases in oxygen ten-sion, to an extent not yet leading to decreases in ATP, havebeen widely discussed in the literature [7, 13, 35]. Examplesinclude changes in components of the adenylate pool [7]and changes in the redox state of neurons [13, 35].

As regards the answer to the third of these questions, itshould be noted that biochemical adaptation [20] to unfa-vorable external conditions requires increases in energyexpenditure associated with structural-functional reorgani-zation and the accompanying increase in biosynthesis.


Fig. 6. Dynamics of assessments of various physiological parameters during 3-min hyperventilation repeated sev-eral times during one day with increasing volume of hyperventilation. A, B) Subjects V and K. 1) MVR (minutevolume of respiration); 2) TCCO2T (transcutaneous carbon dioxide tension); 3) BBFP FMR (rheographic brainblood flow rate parameter in rheoencephalogram lead FMR); 4) EEG SP F3 (O) (EEG spectral power in the thetarange in lead F3).

Review [3] cited many references showing that on adapta-tion of animals and humans to mountain conditions, thereare prolonged increases in the oxygen consumption of thebody and individual organs despite hypoxia. Convincingdata on increases in oxygen consumption by the brain andother organs on adaptation of dogs to mountain conditionswere reported in [2]. It is entirely likely that during a certainperiod of time these energy costs of adaptation are ofgreater priority than performance of organ-specific func-tions to the maximal level.

We do not have hard evidence showing that transienthyperventilation triggers these energy-utilizing mechanismsof adaptation to hypocapnia. However, the fact remains thatthe decrease in the functional state of the brain, apparent asEEG changes, develops on the background of unalteredbrain oxygen consumption [25, 28, 29]. Thus, the structureof oxygen and energy utilization by the brain involves anincrease in the weighting of processes not associated withthe performance of organ-specific brain functions.

In all probability, triggering of these mechanisms issimilar to triggering occurring on adaptation to hypoxia.However, at the first minutes of hyperventilation, as demon-strated in the present studies, the situation is not one of cir-culatory hypoxia. We believe that the most likely factor inhypoxia is a shift in the hemoglobin dissociation curve to theleft as a result of hypocapnia. This is known to involve areduction in blood oxygen tension as compared with its levelin normocapnic conditions at the same level of blood oxygensaturation [34]. In other words, delivery of oxygen from theblood to the tissues becomes more difficult. This leads toincreases in the activity of tissue mechanisms directed, asdescribed by Barbashova [3], to the struggle for oxygen.

Our understanding of the molecular-genetic mecha-nisms of this process are increasingly well understood [27].The remit of this article does not allow the mechanisms oftissue adaptation to be discussed in more detail. However, itis useful to consider an important aspect which we believe isrelated to the present study. In all probability, the increase inthe activity of tissue adaptive mechanisms allows a reductionin the brain blood flow rate in hyperventilation and its accom-panying hypocapnia. If not, the possibility of a decrease inbrain oxygen consumption in conditions of decreased brainblood flow rate after transient 3-min (Table 1) or duringlonger-lasting moderate hyperventilation (Fig. 3) would haveto be considered. However, even in quite marked hypoxia,brain oxygen consumption did not decrease to below the crit-ical level [31].

Whatever the functional sense of the reduction in brainblood flow rate in hypocapnia (a decrease in the rate of“washing away” of carbon dioxide from the body [12] orsome other factor), it is important that this occurs to “cover”tissue compensatory mechanisms not involving markeddecreases in brain oxygen consumption.

A further fact obtained in the present study requiresconsideration. EEG changes at different levels of hypocap-

nia were more marked in the post-hyperventilation periodthan during hyperventilation itself. The published datanoted above provide evidence that the decrease in the func-tional state of the brain is coordinated with other mecha-nisms supplying energy for the vitally important functionsof the body and its structures at different levels. The load-ing on the energy-conserving mechanisms restricting func-tional activity may in all probability decrease with increas-es over time in the activity of the mechanisms providingoxygen and energy. From this point of view, the thirdminute of 3-min hyperventilation and the first minute afterthe end of hyperventilation were, respectively, the third andfourth minutes of hypocapnia overall. In the fourth minute,the need for presumptive energy-conserving mechanisms islower than in the third minute; consequently, the corre-sponding EEG changes should be less marked at the end ofhyperventilation. However, it is entirely likely that EEGchanges are not determined only by the humoral mecha-nisms mediating the influences of hypocapnia and hypoxiaon brain electrogenesis, but also by reflex changes associat-ed with the voluntary hyperventilation act itself.

Thus, the simplified “hypocapnia–circulatory hypox-ia–functional changes in the CNS” scheme does not in ourview explain many of the facts observed in studies of vol-untary hyperventilation. This is to a large extent becausethis scheme is one-sided and does not consider manyfeedback loops in the functional system of energy expen-diture and energy supply to individual neurons and thebrain as a whole.

This study was supported by the Russian HumanitarianScientific Foundation (Grant No. 05-06-06084).

REFERENCES

1. N. S. Akopyan, Electrophysiological Studies of Brain Activity inHypoxia [in Russian], Aiastan, Erevan (1987).

2. M. V. Balykin, Physiological Mechanisms of Oxygen Supply toVarious Internal Organs and the Skeletal Musculature in Dogs inHigh Mountain Conditions and Muscular Activity [in Russian],Author’s Abstract of Doctoral Thesis in Biological Sciences,Novosibirsk, (1994).

3. Z. I. Barbashova, Acclimation to Hypoxia and Its PhysiologicalMechanisms [in Russian], USSR Academy of Sciences Press,Moscow, Leningrad (1960).

4. N. I. Volkov and A. Z. Kolchinskaya, “‘Hidden’ (latent) loadinghypoxia,” Hypoxia Med. J., 1, No. 3, 30–35 (1993).

5. C. Heymans and D. Cordier, The Respiratory Center [Russian trans-lation], Medgiz, Leningrad (1940).

6. V. B. Grechin and Yu. D. Kropotov, Slow Nonelectrical Rhythms inthe Human Brain [in Russian], Nauka, Leningrad (1979).

7. A. M. Dudchenko and L. D. Luk’yanova, “The Trigger role of ener-gy metabolism in the cascade of functional-metabolic lesions inhypoxia,” in: Questions in Hypoxia: Molecular, Physiological, andMedical Aspects [in Russian], Istoki, Moscow (2004), pp. 51–83.

8. K. P. Ivanov, Basics of Body Energetics [in Russian], Nauka,St. Petersburg (1993), Vol. 1.

9. N. D. Luk’yanova, “Functional-metabolic characteristics of animalswith different individual levels of resistance to hypoxia,” in:

Burykh658

Questions in Hypoxia: Molecular, Physiological, and MedicalAspects [in Russian], Istoki, Moscow (2004), pp. 156–169.

10. V. B. Malkin and E. P. Gora, Hyperventilation [in Russian], Nauka,Moscow (1990).

11. V. B. Malkin and E. B. Gippenreiter, Acute and Chronic Hypoxia[in Russian], Nauka, Moscow (1977).

12. M. E. Marshak, The Physiological Significance of Carbon Dioxide[in Russian], Meditsina, Moscow (1969).

13. M. O. Samoilov, Responses of Brain Neurons to Hypoxia [in Russian],Nauka, Leningrad (1985).

14. M. O. Samoilov, D. G. Semenov, E. I. Tyulbkova, E. A. Rybnikova,L. A. Vataeva, T. S. Glushchenko, S. A. Stroev, and O. L. Miller,“Molecular mechanisms of the short- and long-term effects ofhypoxic preconditioning,” in: Questions in Hypoxia: Molecular,Physiological, and Medical Aspects [in Russian], Istoki, Moscow(2004), pp. 96–111.

15. E. N. Sokolov and N. N. Danilova, “Neuronal correlates of the func-tional state of the brain,” in: Functional States of the Brain [inRussian], Moscow State University, Moscow (1975), pp. 129–136.

16. M. I. Tishchenko, A. D. Smirnov, L. N. Danilov, and A. L. Aleksan-drov, “Characteristics and clinical application of integral rheography– a new method for measuring stroke volume,” Kardiologiya, 13,No. 10, 54–61 (1973).

17. V. L. Fantalova, “Characteristics of vasomotor changes in transientvoluntary hyperventilation by rheography and plethysmography,”Byull. Éksperim. Biol. Med., 75, No. 1, 9–14 (1973).

18. J. Holden and J. Priestley, Respiration [Russian translation],Biomedgiz, Moscow, Leningrad (1937).

19. A. Hodgkin, The Nerve Impulse [Russian translation], Mir, Moscow(1965).

20. P. Hochachka and J. Somero, Strategies of Biochemical Adaptation[Russian translation], Mir, Moscow (1977).

21. Kh. Kh. Rayullin, Clinical Rheoencephalography [in Russian],Meditsina, Moscow (1983).

22. O. Creutzfeld, A. Kassmatsu, and A. Vaz-Ferriere, “Aktivitätsä-nderung einzelner corticaler Neurone im akuten Sauerstoffmangelund ihre Beziehungen zum EEG bei Katzen,” Pflügers Arch. Ges.Physiol., 263, No. 6, 647–667 (1957).

23. S. G. Düsser de Barrene, C. S. Marshall, W. D. McCulloch, andL. F. Nims, “Observations of the pH of arterial blood, the pH and the

electrical activity of the cerebral cortex,” Amer. J. Physiol., 124,631–636 (1938).

24. J. F. Fazecas, L. C. McHenry, R. W. Alman, and J. F. Sullivan, “Cere-bral hemodynamics during brief hyperventilation,” Arch. Neurol., 4,No. 2, 132–138 (1961).

25. J. Grote, K. Zimmer, and R. Schubert, “Effects of severe arterialhypocapnia on regional blood flow regulation, tissue PO2 andmetabolism in the brain cortex of cats,” Pflügers Arch., 391, No. 3,195–199 (1981).

26. A. M. Harper, “Physiological control of the cerebral circulation. Ce-rebral blood flow and metabolism,” Physiological Society StudyGuide, 4, 4–26 (1990).

27. Q. Ke and M. Costa, “Hypoxia-inducible factor-1 (HIF-1),” Mol.Pharmacol., 70, No. 5, 1460–1480 (2006).

28. S. S. Kety and C. F. Schmidt, “The effects of active and passivehyperventilation on cerebral blood flow, cerebral oxygen consump-tion, cardiac output, and blood pressure of normal young men,”J. Clin. Invest., 25, No. 1, 107–119 (1946).

29. X. Ma, R. Bay-Hansen, J. Hauerberg, F. M. Knudsen, N. V. Olsen,and M. Juhler, “Effect of graded hyperventilation on cerebralmetabolism in a cisterna magna blood injection model of subarach-noid hemorrhage in rats,” J. Neurosurg. Anesthesiol., 18, No. 1,18–23 (2006).

30. J. S. Meyer and A. G. Waltz, “Diagnostic value of the EEG mani-festation provoked by inhalation of nitrogen or of gases having a lowoxygen content,” in: Cerebral Anoxia and Electroencephalogram,Springfield (1961), pp. 324–325.

31. E. Opitz and M. Schneider, “Über die Sauerstoffversorgung desGehirn,” Eergebn. Physiol., 46, 126–208 (1950).

32. O. B. Paulson and F. W. Scharbrough, “Physiologic and pathophys-iologic relationship between the electroencephalogram and theregional cerebral blood flow,” Acta Neurol. Scand., 50, No. 1,194–220 (1974).

33. M. Reivich, “Arterial pCO2 and cerebral hemodynamics,” Amer. J.Physiol., 206, No. 1, 25–35 (1964).

34. J. W. Severinghaus, “Oxyhemoglobin dissociation curve correctionfor temperature and pH variation in human blood,” J. Appl. Physiol.,9, No. 2, 197–200 (1956).

35. B. K. Siesjö, Brain Energy Metabolism, Chichester, etc. (1978).


interaction of hypocapnia, hypoxia, brain blood flow, and

Documents