Journal Logo

Clinical Sciences: Clinical Investigations

Spectral analysis of electroencephalography changes after choking in judo (juji-jime)


Author Information
Medicine& Science in Sports & Exercise: September 1998 - Volume 30 - Issue 9 - p 1356-1362
  • Free


Judo, which became an Olympic discipline in 1964, developed into a popular sport in Europe and North America and thus increasingly attracted the interest of sports medicine(13-15). In the 1950s, the first studies on physiologic effects of choking techniques were carried out in Japan (6,10,19). More recent investigations in this field applied the 133xenon-inhalation method to assess regional cerebral blood flow (rCBF) (17). In 1996 Raschka et al. (15) studied intracranial velocity of blood flow in the middle cerebral artery by means of transcranial Doppler ultrasonography.

The aim of the present study is to detect possible electroencephalography (EEG) -changes, induced by choking with a "shime-waza" technique: Can disturbances of brain function after choking be objectivized with electroencephalographic means and-if so-how long do these changes last?


Six male, healthy volunteers, without a cardiovascular or any other disease, experienced in judo or ju-jitsu, underwent a choking trial. Beforehand, written informed content was obtained by all probands. The average age was 29 yr (25-35 yr).

The EEG registration was performed with a 70-Hz, low-pass filter, an EEG time constant of 0.3 s, and with 19 scalp electrodes, distributed according to the international 10-20 system.

For the off-line analysis, analogous EEG data were digitalized at a digitalization frequency of 512. To avoid artifacts during EEG-recording, volunteers rested with eyes closed in a reclining chair inside a climatized, noiseless room. The trials were performed between 2 and 4 p.m.

First, EEG was recorded at resting conditions. After 40 s, choking was started, carried out by a black-belt judoka. Approaching the recumbent volunteer from the fronto-lateral position, the judoka crossed his arms, took hold of the judo-gi (judo-jacket), and started choking by turning his hands and forearms inwardly while pulling them against his body (see Fig. 1). Thus, the ventral parts of the neck were being compressed, in which the main vessels to the head are localized (carotid artery, jugular vein). Choking was interrupted immediately after the volunteer gave a sign, so fainting occurred in no cases. The average choking time was 8 s (SD = 1.8). After the end of choking, EEG registration was continued for at least 70 s.

Figure 1
Figure 1:
Technique of juji-jime.

Computerized EEG data at rest and after choking were compared for each person. Statistical analysis was applied to the average values of all six volunteers. Calculation for EEG data was performed with fast Fourier transformation (Fig. 2). With this method, frequency portions can be revealed that would not be visible in the common EEG recording (17). Values of amplitudes, frequencies, and phase relations were taken into account, this measure is called global field power (GFP).

Figure 2
Figure 2:
Typical brain maps before and after choking. Part 1 (this page): Increase of delta- and theta-power mostly above/over fronto-temporal scalp regions in phase BI and less prominently in phase BII. Part 2 (page 1359): Decrease of alpha-power in occipital regions after juji-jime. Part 3 (page 1360): Decrease of alpha-power in ranges of 11-12 Hz. Phase BI: 3-20 s after choking; phase BII: 40-70 s after the choking maneuver.

For a clear representation of GFP distribution, narrowfrequency ranges were defined. Because GFP in the resting EEG of adult persons is physiologically concentrated between 8 and 13 Hz, the alpha-range was divided into six 1-Hz bands with alpha-1 (7.5-8.5 Hz) up to alpha-6 (12.5-13.5 Hz). Thus, even slight focal changes may be detected, which on a larger scale would be obscured (20).

The low-frequency region was split up into delta-1 (0.1-1.5 Hz), delta-2 (1.5-3.0 Hz), theta-1 (3.0-5.0 Hz), and theta-2 (5.0-7.0 Hz). Beta-1 and beta-2 represent the high-frequency range of brain waves.

The topographic distribution of EEG data was displayed with brain maps (interpolated scalp/surface maps). A brain map represents a two-dimensional simplified picture of the scalp surface, from which EEG data are being recorded. We look at this idealized surface from above, the right side of the image corresponding to the right hemisphere and the black dot on top of it standing for the nose, indicating the front-side of the head. As we used the international 10-20-system, EEG data were registered by 19 scalp electrodes.

After computerized processing, these data were calculated and, by means of interpolation, intermediate data were attained to create a complete data surface as a basis for visual rendering: Each value of spectral power was assigned to a certain color on a color-coded scale, ranging from blue for the lowest (i.e., zero), to red for the highest values. Here, a black and white scale(respectively, a figure in black and white) is used, ranging from white for the lowest value to black for the highest. In addition to this, GFP was calculated for each brain map.

Numerous artifacts were excluded from the off-line analysis of EEG registrations. These artifacts were due to head movements during choking, contractions of face, neck, and head muscles, and eye movements and sweating. An interpretation of EEG, recorded during choking, is hardly possible, because brain waves are almost completely superimposed by artifactual elements (mostly muscle artifacts).

From the EEG at rest before choking, 10 1-s epochs were selected for calculation; data from resting EEG were called phase A. An average of 7 epochs within the time span between 3 and 20 s after choking were evaluated and denominated phase BI. Phase BII, ranging from 40 to 70 s after choking, comprised an average of six epochs for spectral analysis. The Wilcoxon test for pair differences was used for statistical analysis.


In this study, choking was performed for an average time of 8 s, neither leading to a loss of consciousness of volunteers nor provoking wave patterns specific for epilepsy in the EEG recording. The results of spectral analysis revealed a high rate of interindividual variability. Nonetheless, all persons showed a dominance of GFP in the physiological alpha-range with a normal topographic distribution.

For statistical evaluation, averaged GFP values of phase A were compared with those of phase BI and BII. Here, statistically significant changes could be found in the frequency ranges delta-1 to theta-1, as well as in the alpha-1-to alpha-6 bands. Immediately after choking(phase BI), power-in comparison to the EEG at rest-increased significantly in delta-1, delta-2, theta-1, and theta-2. At the same time, alpha-power decreased (Table 1). These temporary alterations of GFP values could be detected even in phase BII (40-70 s after choking), yet showed less statistical significance compared to the initial values in phase A. No such changes appeared within the theta-2 and alpha-1 bands, as well as in the beta-frequency range.

Average GFP values, including standard deviation of six persons from EEG at rest (phase A) and after choking in phase BI (3-20 s) and BII (40-70 s). For statistical analysis of GFP-differences between Phases A, BI, and BII, the Wilcoxon test was applied.

Changes of EEG data after choking appeared mostly above fronto-temporal scalp regions. A decrease of alphapower was accentuated in occipital regions. These changes are demonstrated by brain maps of one person, showing power distribution before and after choking (Fig. 2, 3a-3c). As mentioned above, an evaluation of EEG, registered during choking, naturally was not possible because of strongly pronounced artifacts.


Early Japanese studies on choking in judo. One of the first detailed studies on physiological effects during and after choking in judo was presented by Ikai et al. in 1958 (6). Besides blood pressure, pulse, oxygen saturation, and other parameters, EEG in one-lead recording was examined. In contrast to the present study, the loss of consciousness of choked volunteers, as well as nonepileptic seizures and convulsions were readily tolerated.

A variety of EEG-changes was described, which may be sketched as follows: After loss of consciousness, slow waves of low amplitude were reported to occur, interpreted as sign of epileptiform discharges. Immediately after spontaneous recovery, beta- and alpha-waves emerged. After approximately 20 s, brain waves returned to normal, in some cases after 2-3 min. During choking, an increase of amplitudes was described. From our experience, an interpretation of brain waves recorded during choking is hardly feasible.

Rather than examining the physiological effects of choking in general, Suzuki et al. (19) dealt with the question of how choking provokes unconsciousness. For this purpose, several persons were choked with a judo technique called "katajuji-jime." A control group underwent choking with a pneumatic tourniquet to provide a constant compression of the neck.

Although choking by katajuji-jime induced a loss of consciousness after 10 s, compression with a tourniquet caused fainting after 19 s. Here too, in some cases, epileptiform seizures occurred after the onset of unconsciousness. As to the one-lead EEG recording, Suzuki found in most cases slow waves of varying amplitude during unconsciousness and in the phase of recovery. Brain waves normalized within 1 min, in some cases only a few seconds elapsed after the subject regained consciousness until normalization of EEG took place.

Italian study on choking in judo. To detect possible long-term damages to the brain by repeated choking in judo, Rodriguez et al. (17) carried out a study with 10 judoka, examining EEG recording and regional cerebral blood flow (rCBF) with the 133xenon-inhalation method before and after unconsciousness induced by choking.

After approximately 10 s of choking, generalized 2- to 3-Hz delta-waves were reported to occur mainly over frontal regions. Unconsciousness set in after 15-16 s of choking and lasted for 15 s. Brain waves returned to normal as soon as the volunteers had recovered completely. As to the investigation of rCBF, Rodriguez et al. could not trace any significant changes after choking. In summary, the Italian authors did not find persisting alterations of cerebral function with either of the above mentioned methods.

Investigation with transcranial Doppler ultrasonography. Using the same juji-jime technique as in the present study, Raschka et al. (15) found by means of Doppler ultrasonography a highly significant reduction of the end diastolic flow in the midcerebral artery. In context with simultaneously decreasing values in oxymetry, these findings point at a cerebral hypoxia during choking.

Brain-mapping studies using hypoxia and vasovagal stimulation. In a study including 18 healthy volunteers, Saletu et al. (18) could prove an increase of GFP under standardized hypoxic conditions (inhalation of a defined gas mixture, containing 9.8% O2 and 90.2% N2). This effect was mostly due to a rise of activity within the delta- and theta-range. At the same time alpha-activity decreased parallel to the diminution of the state of vigilance.

Investigating topographic EEG-changes under hypobaric hypoxia (simulated high altitude), Ozaki et al. (12) found a decrease of GFP in the alpha-range at 3000 m of simulated height. At more than 5000 m, a significant increase of activity occurred in the theta-range, mainly above the frontal cortex.

Anoxia and alpha-rhythm. In an EEG-study on canine brains, Hogan and Fitzpatrick (5) described a sudden loss of alpha-activity due to an abrupt shift to hypoxic perfusion "with maintenance of other perfusion variables."

EEG and syncope. Already in 1957, Gastaut and Fisher-Williams (3) carried out studies on EEG changes accompanying syncopes with and without convulsions, elaborating a clear distinction between epileptic and syncopal events. In 71 patients with a history of syncope, cardiac arrest was precipitated by ocular compression. "No clinical or electrical abnormalities" occurred when cardiac standstill was no longer than 6 s. Only desynchronization of the EEG appeared because of the pain of ocular compression leading to a "blocking" of cortical rhythms. When asystole lasted 7-13 s, EEG showed slow waves, developing gradually or suddenly irrespective of disturbances of consciousness. These delta- and theta-waves appeared bilaterally, accentuated above the frontal regions. In most of these cases slow rhythms disappeared within a few seconds after heartbeat set in again and EEGs returned to normal.

Generalized myoclonic fits and tonic spasm appeared in patients with cardiac arrest lasting longer than 14 s. During these episodes EEG showed "flattening" and in no case spikes or spike-waves developed, which would be typical of epileptic fits. An EEG analysis in 14 patients with syncopes associated to" malignant ventricular arrhythmias" was carried out by Aminoff et al. in 1988 (1). In 13 patients, 22 episodes of ventricular fibrillation, tachycardia, or both were provoked. Definite loss of consciousness occurred in 15 cases. Here, a clear correlation between duration of arrhythmia and secondary loss of consciousness could be found.

EEG changes during both arrhythmia and syncopal events displayed a large variability, ranging from "no obvious change" in seven episodes, attenuation or loss of electrocerebral activity (in some cases, as the authors report, muscle artifacts might have veiled slowing), to visible slowing of background activity, which was progressively intermingled with low, irregular theta- and delta-activity.

After regular heart beat set in again, EEG normalized within 10-11 s in six episodes, while in one comatose patient after recovery electrocerebral activity remained attenuated for 25 s, followed by a span of 50 s with low generalized delta-activity before gradual development of normal brain waves began.

In the study by Aminoff et al., no brain waves typical of epilepsia were observed but rather a high variability of the intervals between onset of anoxia and first visible changes in the EEG. Moreover, these changes were by far not as uniform as had been described in preceding accounts. EEG changes during vasovagal syncopes were described by Grossi et al. in 1990 (4). In 28 of 271 patients with a transitory unconsciousness in their history, a vasovagal syncope was provoked by head-up tilt at an angle of 70°. EEG showed slowing of brain wave rhythm, followed by generalized, synchronous delta-waves of high amplitudes. It is quite obvious that vasovagal stimulation is also precipitated by choking with juji-jime.

EEG and concussive convulsions in Australian rugby players. McCrory et al. (9) examined 22 cases of concussive convulsions in Australian elite rugby footballers, applying clinical and neuropsychological methods as well as EEG recording and neuroimaging screening. After a collision, players lost consciousness within 2 s and developed convulsions instantly, lasting 10 to 150 s. All players recovered completely and no late seizures occurred, neuropsychological testing results returned to "baseline level" within 5 d. In no case did computed tomography or magnetic resonance imaging reveal any lesion of the brain. EEG yielded normal results, except for one player whose brain waves showed "transient left temporal slowing" immediately after the injury. The authors favor the hypothesis that convulsions in this particular context might result from "transient functional decerebration" caused by the abrupt and violent impact.

In the present study, an increase of GFP in the delta- and theta-range was observed. The results of our explorative brain-mapping investigation as to EEG-changes under choking are being confirmed by comparable studies with spectral analysis and earlier studies with conventional EEG recording and analysis.


The present explorative brain-mapping study showed longer lasting subclinical changes of electrical brain function after choking of healthy persons. Although conventional EEG did not display recognizable changes, spectral analysis of digitalized brain waves and calculation of GFP revealed a significant decrease of physiological alpha-activity and a significant increase of spectral power in the thetarange (which physiologically occurs when vigilance is reduced). Up to 20 s after choking, these changes reached statistical significance and could be traced within a time span of 40-70 s after choking, yet having decreased beneath a statistically significant level. Early studies of Japanese and Italian study groups described conspicuous EEG-alterations (using in part one-channel leads), in which a quantification or a temporal demarcation was not possible.

In our opinion, the results of our explorative study concerning EEG changes in the delta- and theta-range were due to a reduced cerebral perfusion, which would be comparable to studies carried out with spectral analysis of brain waves or with transcranial Doppler ultrasonography (12,15,18). It remains to be discussed whether in this explorative study decrease of alpha-power can merely be explained by alphablocking, e.g., because of the stress inflicted by choking or whether cerebral hypoperfusion is playing a major role, as several findings might suggest (5,12).


1. Aminoff, M. J., M. M. Scheinman, J. C. Griffin, and J. M. Herre. Electrocerebral accompaniments of syncope associated with malignant ventricular arrhythmias.Ann. Int. Med. 108:791-796, 1988.
2. Budo Study Group. Eastern Martial Arts: The Encyclopedia. Berlin: Werner Lind, 1996, p. 760.
    3. Gastaut, H., and M. Fisher-William. Electro-encephalographic study of syncope: its differentiation from epilepsy. Lancet 2:1018-1025, 1957.
    4. Grossi, D., C. Buonomo, F. Mirizzi, R. Santostasi, and F. Simone. Electroencephalographic and electrocardiographic features of vasovagal syncope induced by head-up tilt. Funct. Neurol. 5:257-260, 1990.
    5. Hogan, K., and J. Fitzpatrick. The cerebral origin of the alpha rhythm. Electroencephalogr. Clin. Neurophysiol. 69:79-81, 1988.
    6. Ikai, M., T. Ishiko, G. Ueda, et al. Physiological studies on "choking" in judo. Part I: studies in general. Bull. Assoc. Sci. Stud. Judo, Kodokan Report 1:1-12, 1958.
    7.Koiwai, E. K. Fatalities associated with judo. Physician Sportsmed. 9:61-66, 1981.
    8. Koiwai, E. K. Deaths allegedly caused by the use of "choke-holds" (shime-waza). J. Forens. Sci. 32:419-32, 1987.
    9. McCrory, P. R., P. F. Bladin, and S. F. Berkovic. Retrospective study of concussive convulsions in elite Australian rules and rugby league footballers: phenomenology, aetiology and outcome. Br. Med. J. 314:171-174, 1997.
    10. Ogawa, S., K. Akutsu, R. Sugimoto, H. Saiki, Y. Ikawa, and M. Tsuboi. Physiologic studies on "choking" in judo with reference to the hypophysio-adrenocortical system. Bull. Assoc. Sci. Stud. Judo, Kodokan Report 2:107-114, 1963.
    11. Owens, R. G., and E. J. Ghadiali. Judo as a possible cause of anoxic brain damage: a case report. J. Sportsmed. Phys. Fitness 31:627-628, 1991.
      12.Ozaki, H., S. Watanabe, and H. Suzuki. Topographic EEG changes due to hypobaric hypoxia at simulated high altitude. Electroencephalogr. Clin. Neurophysiol. 94:349-356, 1995.
      13. Raschka, C., M. Parzeller, and H. Glaser. Fatalities in club sport in the Federal Republic of Germany. Dtsch. Z. Sportmed. 47:17-22, 1996.
      14. Raschka, C., J. Roth, M. Parzeller, C. H. Hammar, and T. Bonzel. The cardiac incident in combat sports.Herz/Kreislauf 28:79-82, 1996.
      15. Raschka, C., A. Stock, K. Brunner, and K. Witzel. Investigation on changes of intracranial blood flow velocity induced by choking in judo (shime-waza) by means of transcranial Doppler-sonography. Dtsch. Z. Sportmed. 47:393-398, 1996.
      16. Reay, D. T., and R. L. Mathers. Physiological effects resulting from use of neck holds. FBI Law Enforcement Bull. 52:12-15, 1983.
        17. Rodriguez, G., S. Francione, M. Gardella, et al. Judo and choking: EEG and regional cerebral blood flow findings. J. Sports Med. Phys. Fitness 31:605-610, 1991.
        18. Saletu, B., G. Hitzenberger, J. Grunberger, et al. Double-blind placebo-controlled, pharmacokinetic and -dynamic studies with two new formulations of piracetam (infusion and sirup) under hypoxia in man. Int. J. Clin. Pharmacol. Ther. 33:249-262, 1995.
        19. Suzuki, K. Medical studies on "choking" in judo, with special reference to electro-encephalographic investigation. Bull. Assoc. Sci. Stud. Judo, Kodokan Report 1:23-47, 1958.
        20. Zschocke, S. Clinical Electroencephalography. Stuttgart: Springer-Verlag, 1995, pp. 580-675.


        ©1998The American College of Sports Medicine