Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.protein.bio.msu.ru/~akula/JourVND1104007KulaichevLO.pdf Äàòà èçìåíåíèÿ: Wed Jun 8 15:26:32 2011 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 19:50:52 2012 Êîäèðîâêà:

, 2011, 61, 4, . 1­14


612.821.6


© 2011 . . .
, . .. e mail: akula @mail.ru 06.10.2010 . 20.12.2010 . , ( ). ; ; ; ; , ; , 100%. : , , , , , , , , , , , , , .

The Method of Correlation Analysis of EEG Synchronism and its Possibilities
A. P. Kulaichev
Department of Higher Nervous Activity, Moscow State University e mail: akula @mail.ru A new method for estimation of EEG synchronism based on the analysis of correlation between ampli tude modulation processes (envelopes) is considered. The efficiency of the given method is shown in dif ferent aspects: for detection of a topographical pattern, for analysis in frequency ranges, for estimation of stability of the functional state and its changes, for detection of heterogeneous subgroups of subjects and deviations, for analysis of individual and group distinctions and distinctions in synchronism be tween symmetrical pairs of derivations and hemispheric asymmetry, and for classifying division between mentally healthy subjects and deviations of schizophrenic spectrum with reliability approaching 100%. Key words: EEG synchronism, Pearson correlation, amplitude modulation, envelope, filtration, frequency range, topographical pattern, profile of synchronism, functional state, individual and group distinctions, in terhemispheric dominance, factor analysis, discriminant analysis, schizophrenia.

[15, 20], . (1936 .), . [2, 10]
1

0.1 4 [11]. () [27, 34].


2



, [9, 24] , (. [24, 25, 34, 35]), (. [4]) [5]. . 1. . ( , [8, . 227­ 230, 251­254]) [5], . , , . , , . , . . , () .
1

1

1. (, ), , . [8, . 220­221], ( [8, . 217­220]) . . 2. [8, . 221­223] , , "" (. 1)2. 3. 3. . , , , , ( 1), ( ) [9], (24 )4.
2

, . 1928­1930 . [38] , . , . 1950 , , " " [9].

() , , , , 90 . 3 . , 1 , ­ 0.3­0.4 , 1 . 4 4­8 , . , ( ), . 61 4 2011




39.5 40.0 40.5 41.0 41.5 42.0 42.5 43.0 43.5 44.0 44.5 45.0 45.5

3





. 1. : ­ ; ­ . Fig. 1. The envelope: ­ EEG with a high content of alpharhythm; ­ the result of filtering in alphadomain with overlay of the envelope.

. , ( ) . , , . . . , () 42% 0.6 0.99 , 0.42, 98.5% 0.6 0.17 ( 1404 1715 ) . (


0.4%), ( 0.14). , (1 ), , . , , . . , : 1020% 43 (. 2, ). , , , .
5

5

.. . [11]. . 4 2011

61


4

0.9 0.8 0.7 0.6 0.5 0.4

. 2. rxy ( ­ rxy > 0.2; ­ rxy > 0.6; ­ rxy > 0.8); ­ : ­ ; ­ , . Fig. 2. Three tomograms of EEG synchronism of one subject for standard grid of channels depending on cor relation values rxy ( ­ rxy > 0.2; ­ rxy > 0.6; ­ rxy > 0.8); : correlation profile of examinees: vertical axis ­ correlation values; horizontal axis ­ nearby pairs of electrodes arranged from the left to right and from the top to down depending on its arrangement on a scalp.

(. 2, , ), . ( ) () (. 2, ), . . (. 3): ­ , ­ . : 1) ; 2) ; 3) ; 4) , , .. (. 3): 1) ( ); 2) ( ); 3) ; 4)

. ( @ ), . . , . , [3, 14, 21, 22, 27, 29, 35]. [18], ­ [3, 22]. : [12], [19], [5], ­ , [25]. , , , . . : " ", 10­14 : 39
61 4 2011




( ) : m36y .txt
F73 s10W s12W s152 s153 s154 s155 s157 s158 s163 s164 s165 s167 s169 s170 s173 s174 67.43 53.91 95.50 44.19 83.70 71.21 75.43 76.78 75.66 76.34 60.55 77.79 60.10 69.82 89.43 92.54 F34 65.24 41.60 63.27 79.69 83.62 89.94 80.89 78.71 64.58 71.70 80.97 73.17 56.33 58.79 83.54 61.74 F48 58.89 52.82 51.56 F7T3 67.69 53.33 56.39 76 33 62

5



: .txt
s10W s12W s152 s153 s154 s155 s157 s158 s163 s164 s165 s167 s169 s170 s173 s174 F73 80.71 67.47 95.89 65.81 89.41 81.40 70.63 84.34 86.73 76.71 77.07 73.68 70.39 71.34 95.31 93.45 F34 80.14 56.56 61.69 91.60 90.90 92.79 77.87 84.94 75.67 70.35 87.38 77.31 55.83 74.98 90.42 55.34 F48 77.75 57.85 57.19 92.90 87.76 78.54 54.56 79.14 79.83 92.13 58.82 87.16 70.26 74.08 87.68 89.14 F7T3 81.42 66.20 65.28 82.37 96.38 84.76 75.35 96.87 85.43 61.69 80.89 96.03 79.30 74.94 96.73 94.94 F 8 50 67 98 75 70 52 83 84 63 81 80 68 65 93 5



. 3. . Fig. 3. The example of two matrixes of correlation profiles for different groups of examinees and three variants of their comparison.

( ) 45 , ( F20, F21, F25 10), ( ). 10­20% ( Fp1, Fp2, A1 + A2) 1 200 . , [1, 6]. . , ( ) (. [13]). . , : 1) , [16],


F T [32], Fp1 F7 F7 F3 F7 F3 [36]; 2) , [30], 1 2 5 6 [31], [37]. . 6 , 2­5. , ,
6

, 1 . 4 2011

61


6



1. , Table 1. Correlations between the profiles of EEG synchronization and identical pairs of electrodes calculated for neigh boring frequency domains 1 1 ± . . 0.76 0.81 0.81 0.68 0.73 0.68 ± ± ± ± ± ± 0.11 0.09 0.10 0.12 0.12 0.11 ± . . 0.72 0.82 0.84 0.64 0.65 0.64 ± ± ± ± ± ± 0.13 0.17 0.10 0.12 0.13 0.14 p p p p p p = = = = = = 0.053 0.045 0.14 0.068 0.0006 0.027



[7]. 2. . . , : , , , 1. rjj j (, . 3) : , , 1. rjj ( . 1). . , ( ) ( ). , . . ( = 0.045)7 . i ( rii , . 3). ( . 1)
7

1 . , rii ( ): ( ­ ) = 0.002. , , 8, . 3. . ( ­ ) , . 1 30 , . , rjj j . . . 2 rjj .
8

, ( ) ( ).

, 1, 2 ( ) ( 1 2), i 1(i) 2(i), j 1(j) 2(j). 61 4 2011






7

2. , Table 2. Correlation between the profiles of EEG synchronization and identical pairs of electrodes calculated for two neighboring EEG time interval in alpha domain . . 0.899 ± 0.052 0.876 ± 0.044 ± . . 0.895 ± 0.071 0.792 ± 0.084 = 0.385 = 0.00005

. i rii ( ) p < 0.00005 . rii , . ( ­ ) . 4. . ( , , ). rjj, , . rjk j,k . |rjk|. j Mj(rjk) (k ) . , Mj(rjk) = 0.505 ± ± 0.12 Mj(rjk) = =0.397 ± 0.084 p < 0.00005. , ,


( ), , ( "" , "" ). . Mj(rjk), ( ) . 4. : M(rjk) = 0.089 M(rjk) = 0.01 M(rjk) = 0.22. : M(rjk) 0.36 M(rjk) 0.44 , .
0.6 0.5 0.4 0.3 0.2 0.1 0 0 10 20 30 40
""



. 4. ; ­ ; ­ , . Fig. 4. Diagrams of distribution of average corre lation between the profiles of EEG synchronism in alpha domain for two groups of examinees: ver tical axis ­ correlation values; horizontal axis ­ examinees arranged according to increase of aver age correlation.
61 4 2011


8











. 5. : ­ < 0.05 ( ­ , ­ ); ­ > 0.6; ­ ( ) ; ­ . Fig. 5. Differences in EEG synchronism in alpha domain between two groups of examinees: a ­ with differences between the pairs of EEG chan nels at significance level < 0.05 (solid lines ­ more EEG synchronism in the norm group, dot ted lines ­ more EEG synchronism in the pathol ogy group); ­ statistically nondistinguished pairs of EEG channels at significance level > 0.6; ­ differences between the symmetric pairs of EEG channels (dominating pairs are shown) in the norm group; ­ differences between the symmet ric pairs of EEG channels in the pathology group.

(. 5.3). Mj(rjk) , , . 3.7, 2.4, 4.1, = 0.0001, 0.008, 0.00001 39 41 . . = 0.0014, 0.00005. . , . 5. .
5.1.

, 29 37 . . , 9. , . , ,
9

[7, . 74­78], , , . , , " " "" .

. 5, . , : 1) , , ; 2) , ; 3) : F3 F4: 0.70 > 0.63, C3 C4: 0.57 > 0.50, P3 P4: 0.47 > 0.39, O1 O2: 0.45 > 0.35; [7, . 186­192] p < 5 â 10­10. [1, 16].
61 4 2011






9

, (W ) (), . , , . . , (. 5, ), , , , . . . (. 5, ), (. 5, ) . .
5.2.









. 6. , (, ) (, ) (, ) (, ). Fig. 6. The pairs of EEG channels which correla tion between EEG envelopes has preferential pro jections on the first (, ) and the second (, ) fac tor for the norm (, ) and pathology (, ) groups.

. : 36 ( ) (), 66­67% . . , : 26.5­ 19.84­11.5­8.6% ; 31.2­16­ 11.6­8.4% . . , ( > 0.6)


(. 6, ), , , , (. 6, ). (. 6, , ), : 1) , ; 2) . ( 1 2 ), . . . . : 1) 0.106; 2) 0.807; 3) 0.227; 4) 0.328. , , , , , . . ( . 6, , ). ,
61 4 2011


10



3. ( ) Table 3. Number of wrong classified examinees for norm and deviations of schizophrenic spectrum 1 2 3 4 5 6 7 8 1 1 , . , , + 3 3 3 7 5 0 3 0 3 2 2 4 3 1 5 0 % 7.1 6.0 6.0 13.1 9.5 1.5 9.5 0 4 2 4 3 1 0 5 0 1 2 2 0 1 1 1 1 2 0 5 3 5 3 6 3 5 2 % 13.0 6.0 11.9 8.3 9.5 6.2 14.3 2.4

. 2 , .
5.3.

. , , [7, . 350­365]. , , . , . , . . , . , . ( 1­3).

[7, . 365­370], . . , , 1. 1. . , , ( ) . . 3 ( , 1­ 5) , 1, ( ), 6­7.1%. 2. . , . ­ , . 25 14.96
61 4 2011






11

3.47. (. 3) 1 5 , 2.58 14.96. ( 0.18­0.29 , ), , 2.58, 0.0049, 10 0.05 . . 3. . , , . (1 2) 2 (). ­1, 2. . 3, 6% 3 18, . 29 37 , (. 4), (. 3, 6): 1.5% 6.2% . . 14 , ­ . ( 2) 41 , 0.024 , 160 .


( 2­3.5 ), . : 10. . . 3 ( 7), ( 1) , , . , , (. 3, 8): 100% 2.4% . , , 11: 23% [39], 12.5% [23], 5.5­13.5% [28], 25­28.2% [33]. , ( 2, 3 ) , , . 6. . : : 1) ; 2) , ; 3.4) ( [35]) , : 5) ; 6) [5].
10

, . , 2.2­3.5 . 11 [6, 17, 26], 100% , [6, 17] : , . 61 4 2011


12



4. Table 4. Comparison of initial parameters of EEG synchronism 1 0.96 0.91 0.80 0.88 0.87 0.56 0.70 0.66 0.44 0.56 0.36 . . 0.21 0.18 0.18 0.17 0.18 0.12 10 18 13 21 13 12

, , . 12 . 84 , , 5 â 84, . . (. 4, 1) , . . . . ( 36 â â 84 = 3024) . . , (. 4, 2) 0­1. , . ,
12

. , . . . , . . . 4 "" ( " " ). , . . () ( ) ( ) : 1) ; 2) , ; 3) ; 4) ; 5) ; 6)
61 4 2011

41 ( 0.024 ) ( ) ( 160 ). 24 2.56 . .






13

; 7) ; 8) ; 9) ; 10) ; 11) 36 ; 12) , ; 13) , 100%; 14) . , , .
1. .., .., .., .. , . . 2005. 31(3). 1­8. 2. .., .. . . . . . 1979. 29(6). 10­71. 3. .., .., .., .. . . . . 2002. 88(10). 1330­ 1342. 4. .. . . 1999. 25(1). 125­133. 5. .., .., .., .. . . . . 2002. 4. 84­95. 6. .., .., .. . . . . . 2005. 55(4). 450­458. 7. .. . . .: ­ , 2006. 512 .


8. .. . . .: ­ , 2007. 640 . 9. .. . . . . . 2009. 59(6). 766­775. 10. .., .., .. . . . . . 1964. 14(2). 185­194. 11. .. . .: , 1972. 182 . 12. .. . . : .. . .: "", 2000. 187­196. 13. .., .., .. . . . . 2009. 19(1). 90­94. 14. .., .., .. . . . . . 2000. 50(1). 44­61. 15. .., .., .., .. . .: , 1987. 256 . 16. .., .., .., .. . . . . . 2005. 55(4). 496­ 504. 17. .., .., .., .., .. . . . . . 2007. 57(6). 699­706. 18. .., .. . . . . . 2003. 53(2). 151­155. 19. .., .., .., .. , . . 1990. 35(5). 850­854. 20. .., .. . .: , 1978. 238 . 21. .., .. ,
61 4 2011


14



22.

23.

24. 25.

26.

27. 28.

29.

30.

. . . 2008. 108(3). 58­66. .., .. . . 1996. 22(1). 11­21. Boostani R., Sadatnezhad K., Sabeti M. An efficient classifier to diagnose of schizophrenia based on the EEG signals. Exp. Syst. Appl. 2009. 36(3). 6492­ 6499. Bruns A. Fourier , Hilbert and wavelet based signal analysis: are they really different approaches? J. Neurosci. Methods. 2004. 137(2). 321­332. Bruns A., Eckhorn R., Jokeit H., Ebner A. Amplitude envelope correlation detects coupling among inco herent brain signals. Cogn. Neurosci. 2000. 11(7). 1509­1514. Fehr T., Wienbruch C., Moratti S., Rockstroh B., El bert T. Statistical discrimination of controls, schizo phrenics, depressives and alcoholics using local MEG frequency related variables. Biomed. Techn. 2001. 46(2). 242­244. Guevara M.A., Corsi Cabrera M. EEG coherence or EEG correlation? Int. J. Psychophysiol. 1996. 23(3). 145­153. Ince N.F., Stephane M., Tewfik A.H., Pellizzer G., McClannahan K. Schizophrenia classification using working memory MEG ERD/ERS patterns. Proc. of 3rd int. IEEE/EMBS conf. on neural eng. 2007. 457­460. Lastochkina N.A., Puchinskaya L.M. Correlation analysis of EEG rhythms and functional asymmetry of the hemispheres in children with the hyperdy namic syndrome. Neurosci. Behav. Physiol. 1992. 2(3). 251­258. Merrin E.L., Floyd T.C., Fein G. EEG coherence term in unmedicated schizophrenic patients. Biol. Psychiat. 1989. 25(1). 60­66.

31. Nagase Y., Okubo Y., Matsuura M., Kojima T., Torua M. EEG coherence in unmedicated schizo phrenic patients: topographical study of predomi nantly never medicated cases alert. Biol. Psychiat. 1992. 32(11). 1028­1034. 32. Norman R.M., Malla A.K., Williamson P.C., Morri son Stewart S.L., Helmes E., Cortese L. EEG coher ence and syndromes in schizophrenia. Br. J. Psychi atry. 1997. 170. 411­415. 33. Sakoglu U., Michael A.M., Calhoun V.D. Classifica tion of schizophrenia patients vs healthy controls with dynamic functional network connectivity . Neuroimage. 2009. 47(1). S39­S41. 34. Shaw J.C. Correlation and coherence analysis of the EEG: a selective turtorial review. Int. J. Psychophys iol. 1984. 1. 255­266. 35. Sterman M.B., Kaiser D.A. Comodulation: A new QEEG analysis metric for assessment of structural and functional disorders of the CNS. J. Neurothera py. 2001. 4(3). 7383. 36. Tauscher J., Fischer P., Neumeister A., Rappelsberger P., Kasper S. Low frontal electroencephalographic coher ence in neuroleptic free schizophrenic patients. Biol. Psychiatry. 1998. 44(6). 438­447. 37. Wada Y., Nanbu Y., Kikuchi M., Koshino Y., Hashim oto T. Photic stimulation in drug naive patients. Ab errant functional organization in schizophrenia: analysis of EEG coherence during rest. Neuropsy chobiology. 1998. 38(2). 63­69. 38. Wiener N. Generalized harmonic analysis. Acta Math. 1930. 55. 182­195. 39. Winterer G., Ziller M., Dorn H., Frick K., Mulert C., Wuebben Y., Herrmann W.M. Frontal dysfunction in schizophrenia ­ a new electrophysiological classifier for research and clinical applications. Eur. Arch. Psychiat. Clin. Neurosci. 2000. 250(4). 207­214.



61

4

2011