ISSN 0362-1197, Human Physiology, 2007, Vol. 33, No. 2, pp. 146–156. © Pleiades Publishing, Inc., 2007. Original Russian Text © O.M. Razumnikova, 2007, published in Fiziologiya Cheloveka, 2007, Vol. 33, No. 2, pp. 23–34.
The Functional Significance of a2 Frequency Range for Convergent and Divergent Verbal Thinking O. M. Razumnikova Institute of Physiology, Siberian Division, Russian Academy of Medical Sciences, Novosibirsk, 630177 Russia Received August 1, 2005
Abstract—Mapping of the power and coherence of α2 bioelectric potentials was used to detect a common effect of rhythm desynchronization in two types of convergent verbal thinking (generation of any words beginning with the given letter or simple associations) with more pronounced left-hemispheric activation predominantly in the temporo-parieto-occipital cortical area. In addition, there were specific functional changes in α2 bioelectric potentials depending on the subjects’ sex and verbal creativity level, which was estimated by the capacity for divergent thinking (finding associations for triads of semantically remote words). The dynamics of the regional activation of the cortex depending on the verbal operation type was more characteristic of men compared to women and of creative persons compared to noncreative ones. In creative persons of both sexes, more original associations were accompanied by a decreased α2-rhythm coherence. In noncreative women, interhemispheric interaction was, conversely, increased. DOI: 10.1134/S036211970702003X
INTRODUCTION The terms divergent and convergent thinking were coined by Guilford [1] to contrast creative and standard thinking. Divergent (creative) thinking creates multiple solutions to a problem, each of which may be correct. Convergent (standard) thinking yields a single conclusion whose correctness is determined by existing knowledge. Considering the current notions on the specializations of the right and left hemispheres in the simultaneous and successive strategies of information processing, we may assume that this dichotomy corresponds to divergent and convergent thinking. Earlier, we demonstrated the specific features of the frequency– spatial organization of the cortical electrical activity during these two forms of thinking in models of arithmetic calculations and solution of a heuristic task [2, 3]. The purpose of this study was to determine the functional organization of α2 bioelectric potentials under other experimental conditions, namely, performance of a divergent verbal task as compared to convergent tasks. The special attention to the α2 range was accounted for by its specific roles both in the differentiation of verbal intellectual abilities and semantic processes [4–6] and in the analysis of individual differences in cortical activity related to the performance of creative tasks [7, 8]. In our study, the divergent verbal task was a remote association test where a subject had to find as original an association as possible to a triad of words [9, 10]. For example, a stereotypical association to the triad grandfather–eyeglasses–kind would be grandmother, while a more original one would be dwarf. The predetermined remoteness of semantic categories of words in a triad of stimulus words allowed us
to create conditions for divergent thinking by ensuring uncertainty and a wide choice of the possible answer from more than two alternatives. Convergent tasks were generation of words beginning with a given letter and chains of free associations where the first word was given and every subsequent word thought up by a subject was related to the preceding one, e.g., shore–sand– coarse–…, etc. It was inferred that the choice of answers in these tasks would involve the use of stable verbal stereotypes, i.e., the preference of convergent forms of verbal operations. Gender differences in cognitive activities are known to be the most consistently expressed in tests on verbal and visual–spatial functions [11]. For example, it has been demonstrated that women perform better on word generation and verbal creativity tests than men do [12, 13]. Fundamental gender differences have been found in lateralization of verbal functions and interhemispheric interactions in testing of verbal memory [14– 16]. In addition, the frequency–spatial organization of cortical electrical activity varies depending on the efficiency of mental activity [17]. Therefore, when analyzing functional changes in the α2 rhythm related to the performance of convergent and divergent verbal tasks, we took into account not only the sex of the subjects, but also the level of their creative abilities. EXPERIMENTAL The sample comprised 18 men and 21 women aged 17–20 years. The subjects were students; all of them were right-handed. The experimental protocol was in compliance with medical and ethical standards; it was approved by the Ethics Committee of the Institute of
146
THE FUNCTIONAL SIGNIFICANCE OF α2 FREQUENCY RANGE
147
Table 1. The main characteristics of verbal activity in men and women differing in the performance of the creative task Sex Men Women
GR0, inefficient
GR1, efficient
n
orig
tas
N1
N2
n
orig
tas
N1
N2
10 12
7.7** 8.7**
266.6 226.2
12.4* 15.6*
13.3 14.4
8 9
12.2** 13.5**
219.2 242.5
13.3 14.4
14.5 13.7
Note: n, number of subjects; orig, association originality (point scale); tas, total time spent on the search for remote associations (s); N1, number of words in Ex1; N2, number of words in Ex2. * Significant difference between the sexes (p < 0.05); ** significant difference between groups (p < 0.001).
Physiology of the Siberian Division of the Russian Academy of Medical Sciences. All subjects were familiarized with the experimental conditions and gave their informed written consent to participate. We used three verbal tasks: generating words beginning with a given letter (Ex1), composing chains of associations (Ex2), and finding remote associations for triads of words (Ex3). Each of the first two tasks was performed for 1 min; the time of Ex3 was not limited. We used a computer-aided version of the remote association test developed earlier [10] to estimate individual parameters of association originality and the time of generation of each answer. First, the subjects trained to perform the tests (without recording of EEGs), the answers being written on special forms. Then, they performed the tests simultaneously with EEG recording, the first two tasks being performed mentally while looking at a computer monitor displaying the title page of the association test. In the computer-aided variant of the remote association test, 20 triads of words were presented on the screen and a subject had to find an original association for each of them. The answers were recorded to a voice recorder; the time spent on generating the answer (tas) was recorded by pressing a key on the computer keyboard, after which the next triad of words was presented. During the tests, the subjects were in a soundproof room. The EEGs were recorded monopolarly from 16 sites located according to the international 10– 20 system at symmetric points of the right and left hemispheres: Fp1, Fp2, F3, F4, F7, F8, C3, C4, P3, P4, O1, O2, T3, T4, T5, and T6. Linked earlobe electrodes were used as a reference electrode. The EEGs in the state of quiet wakefulness and during these three tasks were recorded using a Galileo electroencephalograph at a time constant of 0.3 s and an upper frequency limit of 30 Hz. The signal discretization frequency was 256 Hz. The analysis epoch was 2 s. We analyzed 29 artifactfree epochs (taken with 50% overlap) with a total duration of 30 s. In the Ex3 test, we analyzed EEG fragments corresponding to the initial period of thinking up associative words (Ex3i) and the final stage of the test (Ex3f). The EEG recordings were treated using the Neirokartograf software (MBN, Russia) by means of an IBM PC. The fast Fourier transform was used to estiHUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
mate the EEG spectral densities in six frequency bands: θ1 (4–6 Hz), θ2 (6–8 Hz), α1 (8–10 Hz), α2 (10–13 Hz), β1 (13–20 Hz), and β2 (20–30 Hz). According to the goal of the study, we present here only data on α2 bioelectric potentials. The EEG coherences were determined for all 120 possible combinations of interhemispheric and intrahemispheric (BHcoh (“between hemispheres”) and WHcoh (“within hemispheres”), respectively) pairs of sites (the coherences of homologous and nonhomologous interhemispheric site pairs are denoted BHHcoh and BHNcoh, respectively). We also used averaged coherences of bioelectric potentials calculated for every one of the 16 sites. To normalize the data, we used the logarithms of the power ( log P ) and z transformation of coherences. Statistical analysis was performed using ANOVA/MANOVA followed by analysis of significant interactions via planned comparisons. The statistical significance was corrected by the Greenhouse–Geisser method. The relationship between the productivity of verbal activity and EEG parameters was determined using correlation analysis. RESULTS To determine functional changes in α2 bioelectric potentials related to the efficiency of divergent thinking depending on the sex of subjects, we performed ANOVA of the α2 power and coherence with the factors SEX (men–women) and GROUP (GR). For this purpose, all subjects were divided into groups with low and high verbal creativity (GR0 and GR1, respectively) according to the individual estimates of association originality (Table 1). As is evident from Table 1, GR0 and GR1 significantly differed from each other in the originality of the associations that the subjects thought up. In addition, women from GR0 generated more words in Ex1 than men did. Analysis of the a2 rhythm power. The ANOVA of the EEG power was performed according to the following scheme: SEX (2) × GR (2) × EXPERIMENT (EXP) (2) × SITE (DER) (8) × LATERALITY (LAT) (2). The two levels of the factor EXP were represented by pairs of all experimental situations (e.g., Ex1 and baseline, Ex2 and Ex1, Ex3f and Ex3i, etc.).
148
RAZUMNIKOVA
Table 2. The results of the ANOVA of the α2 rhythm power as dependent on the experimental situation Experimental situation Baseline
Ex1
Ex2
SEX × EXP* EXP × LAT** GR × EXP × DER*
Ex1
Ex2 Ex3f
Ex3i
LAT** GR × EXP × DER EXP*, LAT** EXP × LAT** EXP × DER**
EXP*, LAT* EXP × DER × LAT*
EXP**, LAT** SEX × EXP* EXP × LAT** EXP × DER × LAT** EXP**, LAT* EXP × DER** EXP**, LAT* SEX × EXP* SEX × GR × EXP SEX × GR × EXP × DER
Note: Here and in Table 3: * p < 0.05; ** p < 0.01; the interactions that are not marked with asterisks are close to significant (0.05 < p < 0.10) taking into account the Greenhouse–Geisser correction.
Table 2 shows the significant results of ANOVA obtained by this analysis of the α2 rhythm power. All effects of the factor EXP (Ex1–baseline, Ex3i–Ex1, etc.) were accounted for by a decrease in power in each subsequent experimental situation compared to the preceding one. The comparison between Ex3f and Ex3i was an exception. In this case, the power of the rhythm was, conversely, increased at the end of the creative task compared to its beginning. All effects of the factor LAT were related to the asymmetry of the α2 bioelectric potential power: the log P values were lower in the left hemisphere than in the right one. Analysis of the EXP × LAT interaction showed a higher degree of desynchronization in the left hemisphere than in the right one when we compared Ex1 and the baseline and an opposite effect in the case of the remote association task: the asymmetry of rhythm power was decreased in Ex3i and Ex3f compared to Ex1 because of greater desynchronization in the right hemisphere than in the left one. The regional specific features of changes in cortical activity (according to the results of the planned analysis of the EXP × DER and EXP × DER × LAT interactions) were that the aforementioned effects of α2 rhythm desynchronization in Ex3i and Ex3f were the most marked in the central–parietal–occipital areas of the cortex (p < 0.0001), less pronounced in sites F3/F4 and T5/T6 (p < 0.02), and nonsignificant in Fp1/Fp2, F7/F8, and T3/T4. Both Ex3i and Ex3f were characterized by a more local asymmetry of the α2 rhythm, with lower values in F3 and T5 than in, respectively, F4 and T6. In Ex1, significantly lower log P values in the left hemisphere were noted in all sites except Fp1/Fp2, F7/F8, and T3/T4. Situation Ex2 was also characterized by local asymmetry of log P with lower power in the left hemisphere than in the right one for sites O1–O2 and T5–T6. Comparison between Ex3f and Ex2 showed, in addition to the desynchronization of the rhythm in the central, pari-
etal, and occipital sites, that log P for í6 was lower in Ex3f than in Ex2. The planned analysis of the GR × EXP × DER interaction (F(7,245) = 2.07, p = 0.023) showed that the verbal creativity of the subjects affected the functional changes in the α2 rhythm related to the performance of simple verbal tasks in Ex1 and Ex2. The bioelectric potential power in Ex1 was decreased compared to the baseline in the central and frontal areas (F3/F4 and C3/C4) in GR0 and in the temporal (T3/T4 and T5/T6), parietal, and occipital areas in GR1. In comparison of Ex2 with Ex1 in GR0, the power of bioelectric potentials remained the same; in GR1, it was found to decrease in the frontal areas (Fp1/Fp2 and F7/F8) but increase in the parietal areas (p = 0.002). The SEX × EXP interaction was accounted for by two factors. First, the α2 rhythm in Ex1 was decreased compared to the baseline in women and increased in men. Second, bioelectric potentials did not change significantly in women performing the creative task, whereas, in men, the power of this rhythm in Ex3f was decreased compared to Ex1 (p = 0.0003) but increased compared to Ex3i (p = 0.0005). At the same time, gender differences in the changes in α2 rhythm during the divergent task depended on the degree of verbal creativity. The planned analysis of the marginal SEX × GR × EXP and SEX × GR × EXP × DER interactions showed an increase in the bioelectric potential power in Ex3f compared to Ex3i in both men and women from GR0 (p = 0.014). In GR1, log P was increased in men but decreased in women (p = 0.016). These gender differences displayed regional specificity: first, the power of α2 oscillations in the lateral–frontal sites was higher in women than in men in Ex3i and lower in Ex3f (p = 0.01); second, the α2 rhythm power in the central, parietal, and occipital sites was increased in Ex3f compared to Ex3i in men but unchanged in women (p = 0.02) and HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
THE FUNCTIONAL SIGNIFICANCE OF α2 FREQUENCY RANGE (a)
149
(b)
GR0
GR1
GR0
GR1
I
II
III
Fig. 1. Comparison of the maps of α2 rhythm power and coherence during different verbal operations as dependent on the subjects’ sex and creativity level: (a) men; (b) women; GR0, creative subjects; GR1, noncreative subjects. Comparisons of experimental situations: (I) Ex1 with baseline, (II) Ex3i with Ex1, and (III) Ex3f with Ex3i, where Ex1 is generation of words beginning with a given letter and Ex3i and Ex3f are, respectively, the initial and final stages of the remote association test. Solid and dashed lines connecting the corresponding pairs of sites indicate that the coherence in former situation was higher or lower, respectively, than in the latter (p < 0.05). Solid circles indicate the sites where the α2 rhythm power was higher in Ex3f than in Ex3i; double circles, the sites where the rhythm power was higher in men than in women (p < 0.05).
log P was increased in the temporal cortical areas in men but decreased in these areas for women (p = 0.005). Figure 1 shows the regional characteristics of the changes in α2 rhythm power at the final stage of the search for remote associations compared to the initial stage. Most of the significant gender differences in α2 rhythm reactivity in GR1 were found in posterior cortical areas and site F7. Analysis of the a2 rhythm coherence. The ANOVA of the coherence parameters averaged for each site was performed according to the same scheme as the analysis of power: SEX (2) × GR (2) × EXP (2) × DER (8) × LAT (2). Table 3 shows the results of the ANOVA of the α2 rhythm coherence as dependent on the experimental situation (BHNcoh and WHcoh are the coherences of bioelectric potentials averaged, respectively, HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
for all pairs of nonhomologous sites of the two hemispheres and in sites of the left or right hemisphere; BHHcoh is the coherence of bioelectric potentials in homologous sites; the factor LAT was not considered for the parameters of this type of coherence). The planned analysis of the effects shown in Table 3 for BHHcoh revealed a number of regional specific features of functional interhemispheric interactions, including ones related to the subjects’ gender and verbal creativity. The EXP × DER interaction was accounted for by a decrease in the BHHcoh in the anterior frontal sites in Ex1 compared to the baseline and an increase in this coherence in the central–parietal–occipital sites (p = 0.0007). An enhancement of BHHcoh in C3–C4 and T5–T6 (p = 0.003 and 0.004, respectively) was also found when we compared Ex3f with Ex1.
150
RAZUMNIKOVA
Table 3. The results of the ANOVA of the α2 rhythm coherence as dependent on the experimental situation Experimental situation Baseline
Ex1
Ex3i
BHHcoh: EXP × DER*, SEX × EXP × DER BHNcoh: EXP × DER*, SEX × EXP × DER WHcoh: SEX × GR*, EXP × DER**
Ex1
Ex3f
Ex2
BHHcoh: EXP × DER** BHNcoh: EXP*, EXP × DER*, SEX × EXP × DER × LAT*
BHHcoh: GR × EXP × DER, SEX × EXP × DER WHcoh: SEX × GR × EXP × LAT* BHNcoh: EXP × DER**
In accord with a marginal SEX × EXP × DER interaction, similar changes in the BHHcoh in the central– parietal–occipital sites in subjects of both sexes were accompanied by a decrease in the α2 rhythm coherence in frontal areas in women in Ex1 (sites Fp1–Fp2 and F7−F8) and its slight increase in men (p = 0.013). On going from Ex1 to Ex2, the BHHcoh increased in F7–F8, C3–C4, and T5–T6 and decreased in é1–é2 in men ( = 0.04), whereas no differences were found in women. The GR × EXP × DER interaction was accounted for by the fact that the transition from Ex2 to Ex3i in GR0 was accompanied by an increase in BHHcoh in the anterior frontal sites and its decrease in the occipital sites, whereas the changes observed in GR1 were opposite (BHHcoh was decreased in Fp1−Fp2 and increased in O1–O2) (p = 0.005). Conversely, in Ex3f compared to Ex3i in GR1, the BHHcoh for frontal areas (Fp1–Fp2, C3–C4, and P3–P4) was increased and that for O1−O2 was decreased; in GR0, the changes in BHHcoh were opposite. The factors SEX and GR were also significant in the case of WHcoh related to the performance of simple verbal tasks: in men from GR0 and GR1, the WHcoh was increased and decreased, respectively, in Ex2 compared to Ex1, whereas the WHcoh of women did not depend on either the experimental conditions or the group (GR0 or GR1). In addition, the left-hemispheric WHcoh in Ex1 was higher in women than men from GR0 but, conversely, higher in men than women from GR1 (p = 0.05). Maps of significant changes in the coherences of bioelectric potentials for individual pairs of sites (Figs. 1, 2) illustrate the regional differences in coherence changes as dependent on gender and verbal creativity. Figure 1 shows the results of comparisons of (I) Ex1 with the baseline, (II) Ex3i with Ex1, and (III) Ex3f with Ex3i for each of the four groups of subjects. Figure 2 shows maps of coherent relationships in situations Ex1, Ex3i, and Ex3f demonstrating significant
BHHcoh: GR × EXP × DER, SEX × GR × DER × LAT WHcoh: SEX × EXP × DER × LAT
differences (a) between men and women and (b) between creative and noncreative men (we did not find gender differences in coherence in GR1 for situation Ex3 or differences between women from GR1 and GR0). The observed foci of coherent relationships, i.e., the sites where coherent relationships were concentrated, show changes in the electrical activity of the corresponding cortical areas according to the observed effects of interactions between the factors SEX, GR, EXP, DER, and LAT for coherence parameters (Table 3). As can be seen in Fig. 1, Ex1 was accompanied by diffuse enhancement of interhemispheric coherence compared to the baseline in creative men, whereas only a few intrahemispheric relationships were increased in noncreative men. Creative women in this situation exhibited a local increase in α2 rhythm coherence in site C3; noncreative women, a decrease in interhemispheric interaction in the frontal cortex. Noncreative men exhibited an enhancement of coherent relationships in situation Ex3i compared to Ex1, mainly in right posterior cortical areas (the largest number of relationships was found in sites P4, T4, T6, and O2). In creative men, we observed an enhancement of interhemispheric interaction with a focus in T6 and a decrease in such interaction in P3 (in both cases, these were changes in the interaction with the frontal areas of the contralateral hemisphere). In women from GR0, situation Ex3i was characterized by weakening of some interhemispheric interactions compared to Ex1; in women from GR1, the right-hemispheric coherence was enhanced, with a focus in F4. We found multiple changes in coherent relationships in Ex3f compared to Ex3i only in men from GR1 and women from GR0: both left- and righthemispheric coherences were decreased in creative men, but only left-hemispheric relationships were decreased (with foci in Fp1, C3, and T5) in noncreative women. HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
THE FUNCTIONAL SIGNIFICANCE OF α2 FREQUENCY RANGE Ex1
Ex3i
151
Ex3f
(a)
GR1
GR0
(b)
Fig. 2. Comparison of the maps of α2 rhythm coherence during convergent thinking (Ex1) and the initial and final stages of divergent thinking (Ex3i and Ex3f, respectively) illustrating differences related to the subjects’ (a) sex and (b) creativity level. Solid and dashed lines connecting the corresponding pairs of sites indicate that the coherence parameters were higher or lower, respectively, (a) in women than in men and (b) in creative subjects than in noncreative ones (p < 0.05).
Difference maps of comparison between GR1 and GR0 with respect to coherence (Fig. 2b) show that successful performance of the verbal creative task by men in Ex3i was accompanied by an increase in interhemispheric interaction mainly in the central–parietal region of the cortex; in Ex3f, there are, in addition, relationships in the anterior region of the left hemisphere; hence, T3, C3, P3, and F4 may be considered to be foci of increased interhemispheric interaction. Note that focus T3 was formed in creative men even in situation Ex1 (Figs. 2a, 2b). In GR0, women exhibited higher BHNcoh values than men, with a focus in site P3 (Fig. 2a). Analysis of correlations between the verbal test performance and a2 rhythm parameters. We did not find significant correlations between the parameters of performance of the verbal tests and changes in the power of α2 bioelectric potentials in women. In men, we found a single negative correlation (k = –0.45, p = 0.06) between the number of words generated in Ex1 (N1) and the power reactivity (Ex1 minus baseline) in T5. Separate analysis of GR1 and GR0 showed numerous positive correlations between N1 and the power reactivity only in noncreative women (0.68 < k < 0.83, 0.003 < p < 0.05). This effect was characteristic of the anterior region of the left hemisphere and frontal–temporal areas of the right hemisphere (Fig. 3a). For N2, we HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
did not find significant relationships with the parameters of α2 rhythm power. An increase in the originality of remote associations corresponded to an increase in the rhythm reactivity in sites T3 and C3 in women (0.62 < k < 0.81, 0.01 < p < 0.04); this effect was distinct when we compared Ex3i with Ex1 in GR0 and Ex3f with Ex3i in GR1 (Figs. 3b, 3c). In men, we did not find significant correlations for the index of originality. The only significant negative correlation in them was that between the tas and the power reactivity in the comparison between situations Ex3f and Ex3i; this correlation was found in F8 for GR0 and in Fp1, Fp2, and T3 for GR1 (Fig. 3c). Figure 3 shows the results of the correlation analysis of coherence reactivities (Ex1 minus baseline, Ex3i– Ex1, and Ex3f–Ex3i) and, correspondingly, the number of generated words (I) or originality of remote associations (II, III). In men from GR0, N1 was positively correlated with WHcoh with a focus in O2 and BHNcoh with a focus in C3. In men from GR1, N1 was positively correlated with WHcoh in the left hemisphere but negatively correlated with this parameter in the right anterior cortex. In women from GR0, we found only a few positive correlations, whereas, in women from GR1, an increase in N1 was accompanied by an enhancement of BHNcoh in the anterior cortex and weakening of WHcoh with a focus in F7. The results of the correlation analysis of N2 and the difference in coherence between
152
RAZUMNIKOVA (a) GR0
(b) GR1
GR0
GR1
I
II
III
Fig. 3. Maps of correlations between the α2 rhythm reactivity and verbal activity parameters as dependent on the subjects’ sex and creativity level. (I) The number of words generated in Ex1; (II) comparison of Ex3i and Ex1 with respect to the originality of remote associations and coherence reactivity; (III) comparison of Ex3f and Ex3i. Solid and dashed lines show, respectively, the coherent relationships whose reactivity was positively or negatively correlated with the originality of associations (thin lines, p < 0.05; thick lines, p < 0.01). Solid circles indicate the sites where the power reactivity was positively correlated with the corresponding parameters of verbal activity; double circles, the sites with a negative correlation between the time spent on the search for original associations and the power reactivity (p < 0.05). The other designations are the same as in Fig. 1.
Ex2 and Ex1 are not shown in Fig. 3 because significant correlations were absent in GR0 and very few in GR1. In men from GR1, N2 was negatively correlated with the coherence reactivity in Fp1–F3, Fp1–P4, F7–F3, F7−T4, T3–Fp2, and T3–C3; in women, the correlation was positive and was observed in sites í5–ë4, T5−P4, C4–P3, C4–O1, O1–O2, and P3–P4. The originality of remote associations in men was negatively correlated with the change in interhemispheric interaction in situation Ex3i compared to Ex1 (Fig. 3, II). In men from GR0, an increased originality was related to weaker coherent relationships with foci mainly in Fp2, F4, and T5; in men from GR1, in sites F3, T3, and O2. In women, the sign of correlations depended on the success in performance of the creative test: in
GR0, an increase in originality was accompanied by a diffuse increase in coherence throughout the cortex; in GR1, by a decrease in coherence (with the main foci in F3 and C3 and Fp2, T6, and O2). The correlation analysis of word originality and changes in coherence during the divergent task (Ex3f minus Ex3i) (Fig. 3, III) showed coherence pattern dynamics correlated with the success of the search for remote associations. In men from GR0, an increase in originality was accompanied by an increase in coherence; in men from GR1, by a decrease in it. The focus of interhemispheric coherent relationships in creative men was located in the anterior temporal site of the right hemisphere; in noncreative men, in the parietal site of the left hemisphere. The only significant correlaHUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
THE FUNCTIONAL SIGNIFICANCE OF α2 FREQUENCY RANGE
tion in creative women was the P3–T5 correlation; in noncreative women, an increase in originality was accompanied by weakening of cortical interaction with a focus in O1. Correlation analysis of verbal activity parameters showed a positive relationship only between the originality and the number of words generated in Ex1 by subjects from GR0 (k = 0.45, p = 0.055 for GR0, whereas, for GR1, k = 0.12). DISCUSSION A decrease in the power of α2 oscillations with leftside lateralization and stronger activation in the posterior cortex was a common effect characteristic of convergent verbal tests (generation of words beginning with a given letter or a chain of associations). This agrees with the classic model of verbal processes based on functional maps of the cerebral cortex obtained by electroencephalographic and tomographic methods [18–20]. The domination of the left hemisphere during verbal activity when the strategies of assessment of semantic information are strictly defined is usually related to active inhibitory mechanisms in the right hemisphere [21]. In turn, the relatively strong activation of the right hemisphere that we observed in subjects performing the divergent verbal task agrees with data on the activity of the right hemisphere under experimental conditions requiring a wide network of associations [22]. Predominantly right-hemispheric activation was also demonstrated in a positron emission tomography study on creative verbal activity in another experimental model (composing a story) [23]. Therefore, we may conclude that, as expected, convergent verbal thinking is mainly underlain by the successive functions of the left hemisphere and divergent thinking, by the simultaneous strategies of the right hemisphere. In addition to these common effects, we found individual specific features of functional changes in the α2 rhythm during verbal activity that were related to the subjects' gender and verbal creativity level. Note that the ability for finding original verbal associations is reflected in the characteristics of the spatial organization of the electrical activity of cortical areas not only during the creative task, but also during simpler verbal operations. In the creative group, simple tasks (generation of words beginning with a given letter and chains of associations) were accompanied by α2 rhythm desynchronization predominantly in the posterior cortex; in the noncreative group, this effect of cortex activation was wider, covering also the central–frontal areas. It is conceivable that creative persons differ from noncreative ones in a more economical and efficient organization of cortical areas during both divergent and convergent forms of verbal thinking. This is evidenced by not only the observed specificity of the functional changes in the α2 rhythm during the verbal tests considered here, but also the corresponding differences in verbal activity parameters: GR1 was characterized by a HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
153
greater originality of associations, as well as the absence of the gender difference in the number of generated words that was observed in GR0. However, these parameters of word generation common for all of GR1 and the parameters of associative processes that are the same in both sexes are determined by different strategies of performing verbal operations. The existence of sexual dimorphism of hemispheric activation and functional interaction was earlier observed under different conditions of verbal activity [14, 15]. It is also known that women perform better on word generation and verbal creativity tests than men do [12, 13]. In our experiments, significant gender differences were found only with respect to the number of words generated. However, mapping of the functional changes in α2 bioelectric potentials showed such differences for all three types of verbal operations studied. Word generation was accompanied by significant desynchronization of the α2 rhythm only in women. In men, α2 oscillations were synchronized compared to the baseline. The relationship of this frequency band with semantic processes, including semantic memory [6, 24], allows us to assume that, in women, semantic processes entailing the activation of a cortical neuronal network widely distributed in both hemispheres are involved as early as at the first stage of verbal activity (generation of words). Afterwards, the interaction of these activated neuronal ensembles is modulated depending on the type of verbal operations to be performed: the thinking up of a chain of associations is accompanied by an increase in interhemispheric interaction, and the subsequent search for remote associations, only by local changes in coherent relationships, mainly in the anterior cortex. At all stages of verbal activity, creative women exhibited more localized changes in interhemispheric interactions than noncreative women did. We also observed this effect when comparing the maps of functional changes in α2 rhythm coherence and performing the correlation analysis of the parameters of association originality (Figs. 1, 3). This regional localization of the organization of cortical neurons may be regarded as confirmation of the hypothesis on the “neural efficiency” of interaction between cortical neuronal ensembles in persons with a high IQ. According to this hypothesis, a better cognitive performance is related to a weaker activation of cortical neurons in the α2 band [4, 5]. The patterns of α2 rhythm coherence correlations observed in our study indicate that a successful performance of the divergent task required of noncreative women, but not of creative ones, the involvement of practically all cortical regions, in contrast to the test with generation of words beginning with a given letter. The results of the comparisons of the maps of functional changes in bioelectric potential coherence and the correlation patterns in this group suggest that the low originality parameters in GR0 resulted from an “incorrect” organization of cortical areas (inadequate for high verbal creativity). In all experimental situa-
154
RAZUMNIKOVA
tions, including the creative test, the left frontal cortex was more involved in the activity in women from GR0 than in women from GR1. This suggests that the voluntary control of verbal activity was stronger in the noncreative women, with this control proving to be inefficient in the case of divergent thinking. Apparently, the creative women used a different strategy of searching for associations. The pattern of α2 rhythm coherence correlations indicates that, in contrast, hemispheric interaction decreased as originality increased in this group of subjects. The same effect was observed in noncreative women at a later stage of verbal divergent thinking. Earlier, we observed a similar decrease in interhemispheric interaction with respect to β2 bioelectric potentials in women successfully solving a heuristic problem, the α2 band displaying a trend toward a lower BHNcoh (in the α1 band, the decrease was significant) [2]. In general, women are characterized by stronger interhemispheric interaction and a weaker inhibitory effect of the left hemisphere on the right one compared to men [14, 25, 26]. Therefore, the relative “uncoupling” of hemispheres observed in the α2 band of creative women performing the verbal creative task may lead to longer right-hemispheric simultaneous processes that are necessary for an extended search for associations from remote semantic groups. The results of correlation analysis show that, in the group of creative men, more original associations also corresponded to weaker interhemispheric interaction at both the initial and final stages of testing for verbal creativity. Taking into account that the α2 rhythm desynchronization was increased in men in Ex3i compared to Ex1, we may assume that, in this case, the interaction between the cortex and subcortical structures was increased in creative men because the activation of cortical neurons accompanied by a decrease in their synchronization may be regarded as type D according to the classification of coherence patterns suggested by Petsche et al. [8]. Of special interest are the temporal and spatial changes in α2 rhythm patterns in men as dependent on their creativity. In the Ex1 convergent verbal thinking test, creative men exhibited higher coherence with a focus in T3 compared to both noncreative men and women. This suggests a local efficiency of the language areas of the left hemisphere in men from GR1, ensuring efficient generation of any words beginning with a given letter, so that gender differences in this parameter were absent in creative subjects and were significant in noncreative ones. In the divergent verbal test, differences between the groups were significant only for men, interhemispheric interactions being stronger in GR1 than in GR0. Therefore, the hemispheric organization in creative men performing verbal activity shows some similarity to the “female” thinking strategy based on interaction between the language areas of the left and right hemispheres, with the reserves of the right hemisphere being
used to fulfill verbal functions [14, 25, 27]. However, this combination of distant neuronal ensembles in both hemispheres inherent in creative men is likely to be only one of several conditions for successful verbal productivity. At the same time, the special role of the language zones of the left temporal areas in creative men is retained at all stages of verbal thinking. We can assume, on the basis of the associative theory of verbal processes [28, 29], that the wide integration of cortical areas in creative men underlies not only efficient productivity in word generation, but also the subsequent search for their associative links. The observed patterns of functional changes in the α2 rhythm demonstrate the importance of the frontal and temporoparietal cortices. Many researchers note that these areas are involved in performing verbal tasks [18, 29–31]. It is assumed that the frontal cortex plays the main role in word fluency [32], whereas the temporoparietal areas are involved in complex semantic processes, including word integration according to a given lexical stimulus or the organization of understanding and intentions of performing verbal tasks [33]. Gender differences in the α2 rhythm coherence changes during the divergent verbal test were observed only in GR0. The coherent relationship patterns in women were characterized by the presence of foci in the left anterior frontal area, which were more distinct in noncreative subjects; in creative subjects (of both sexes), the left central–parietal areas were of special importance. The changes in both the power and coherence of the α2 rhythm that we observed when comparing different stages of the divergent verbal test and the correlations between EEG parameters and association originality may reflect the search for an optimal strategy of verbal activity. It is still unclear whether the decrease in coherence during the divergent test (1) indicated the commitment of relevant brain regions ensuring the efficiency of verbal operations, (2) resulted from learning and the development of an optimal strategy for alternative choice between words from semantically remote groups in the network of associations, or (3) reflected an increased involvement of subcortical structures in the organization of a functional neuronal ensemble. It is probable that these effects are combined, thereby ensuring the diversity of individual strategies of verbal activity. The fact that coherent relationships were only slightly variable in creative women suggests that the first effect, i.e., maximal neural efficiency, was dominant in them. In creative men, the development of an optimal strategy for solving the divergent verbal task is likely to have played a more important role. Noncreative subjects largely relied on random generation of associations. This is evidenced by the positive correlation between the originality and the number of words generated in Ex1 observed only in GR0, as well as the diffuse correlations in this group between the parameters of verbal activity and the functional variability of the α2 rhythm (with respect to both power and coherence). Apparently, this random generHUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
THE FUNCTIONAL SIGNIFICANCE OF α2 FREQUENCY RANGE
ation of words was based on stereotypical semantic relationships. In the absence of sufficiently critical analysis of these relationships, the originality of the resultant associations was low. Creative subjects were more successful in combining the controlling functions of the frontal cortical lobes with associative verbal processes occurring in the temporoparietal cortical areas of both hemispheres. CONCLUSIONS (1) The performance of verbal tasks causes α2 rhythm desynchronization in the temporo-parietooccipital cortical area with a marked left-hemispheric asymmetry, which becomes weaker in the case of divergent verbal thinking. The regional and temporal patterns of functional changes in α2 bioelectric potentials determined by the type of verbal activity depend on the subjects' sex and verbal creativity level. (2) In women, the α2 rhythm desynchronization occurs during the first test (generation of words beginning with a given letter) and remains almost unchanged during the subsequent verbal operations. In men, generation of words beginning with a given letter is accompanied by α2 rhythm synchronization, which is replaced by desynchronization at the initial stage of the search for remote associations; by the end of this divergent task, the power of α2 bioelectric potentials increases in the posterior cortex. (3) Creative subjects differ from noncreative ones in a stronger dependence of regional cortical activation on the specific features of verbal operations. Functional variability of the α2 rhythm parameters depending on the type of verbal activity is more typical of creative men and the least typical of creative women. (4) An increase in the originality of remote associations is accompanied by weakening interhemispheric interaction in creative subjects of both sexes. In noncreative subjects, the relationship between the originality and the bioelectric potential coherence varies depending on the stage of testing: from negative relationships at the beginning of the test to positive ones at the end in men, and from positive to negative relationships in women. ACKNOWLEDGMENTS This study was supported by the Russian Foundation for Humanities Research, project no. 05-0606179a. REFERENCES 1. Guilford, Y.P., The Nature of Human Intelligence, New York: McGraw-Hill, 1967. 2. Razumnikova, O.M., Myshlenie i funktsional’naya asimmetriya mozga (Thinking and Functional Asymmetry of the Brain), Novosibirsk: Sib. Otd. Ross. Akad. Med. Nauk, 2004. HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007
155
3. Razumnikova, O.M., Functional Organization of Different Brain Areas during Convergent and Divergent Thinking: An EEG Investigation, Cogn. Brain Res., 2000, vol. 10, p. 11. 4. Doppelmayer, M., Klimesch, W., Stadler, W., et al., EEG Alpha Power and Intelligence, Intelligence, 2002, vol. 30, p. 289. 5. Neubauer, A., Fink, A., and Schrausser, D.G., Intelligence and Neural Efficiency: The Influence of Task Content and Sex on the Brain–IQ Relationship, Intelligence, 2002, vol. 30, p. 515. 6. Klimesch, W., Doppelmayr, M., Pachinger, Th., and Ripper, B., Brain Oscillations and Human Memory: EEG Correlates in the Upper Alpha and Theta Band, Neurosci. Let., 1997, vol. 238, p. 9. 7. Jauscovec, N., Differences in Cognitive Processes between Gifted, Intelligent, Creative, and Average Individuals while Solving Complex Problems: An EEG Study, Intelligence, 2000, vol. 28, no. 3, p. 213. 8. Petsche, H., Kaplan, S., von Stein, A., and Fitz, O., The Possible Meaning of the Upper and Lower Alpha Frequency Ranges for Cognitive and Creative Tasks, Int. J. Psychophysiol., 1997, vol. 26, p. 77. 9. Mednich, S.A., The Associative Basis of the Creative Process, Psychol. Rev., 1969, no. 2, p. 220. 10. Razumnikova, O.M., Sex and Training Specialization as Factors of Creativity in Students, Vopr. Psikhol., 2002, no. 1, p. 111. 11. Halpern, D.F., Sex Differences in Cognitive Abilities, Mahwah: Lawrence Erlbaum, 2000, 3rd edition. 12. DeMoss, K., Millch, R., and DeMers, S., Gender, Creativity, Depression, and Attributional Style in Adolescents with High Academic Ability, J. Abnorm. Child Psychol., 1993, vol. 21, no. 4, p. 455. 13. Dudek, S.Z., Strobei, M.G., and Runco, M.A., Cumulative and Proximal Influence on the social Environment and Children’s Creative Potential, J. Genet. Psychol., 1993, vol. 154, no. 4, p. 487. 14. Vol’f, N.V., Sex Differences of Interhemispheric Interferential Interactions during Memorizing Verbal Information, Zh. Vyssh. Nervn. Deyat., 1998, vol. 48, no. 3, p. 551. 15. Skrandies, W., Reik, P., and Kunze, Ch., Topography of Evoked Brain Activity during Mental Arithmetic and Language Tasks: Sex Differences, Neuropsychology, 1999, vol. 37, p. 421. 16. Vikingstad, E.M., George, K.P., Jonson, A.F., and Cao, Y., Cortical Language Lateralization in Right Handed Normal Subjects Using Functional Magnetic Resonance Imaging, J. Neurol. Sci., 2000, vol. 175, p. 17. 17. Gevins, A., Smith, M.E., and McEvoy, Yu.D., High-Resolution EEG Mapping of Cortical Activation Related to Working Memory: Effects of Task Difficulty, Type of Processing, and Practice, Cerebr. Cortex, 1997, vol. 7, no. 4, p. 374. 18. Ivanitskii, G.A., Nikolaev, A.R., and Ivanitkii, A.M., Interaction between the Frontal and Left Parietotemporal Cortices during Verbal Thinking, Zh. Vyssh. Nervn. Deyat., 2002, vol. 28, no. 1, p. 5. 19. Binder. J.R., Frost, J.A., Hammeke, T.A., et al., Human Brain Language Areas Identified by Functional Mag-
156
20. 21.
22.
23.
24.
25. 26.
RAZUMNIKOVA netic Resonance Imaging, J. Neurosci., 1997, vol. 17, p. 353. Neville, H.J. and Bavelier, D., Neural Organization and Plasticity of Language, Curr. Opin. Neurobiol., 1998, vol. 8, no. 2, p. 254. Federmeier, K.D. and Kutas, M., Right Words and Left Words: Electrophysiological Evidence for Hemispheric Differences in Meaning Processing, Cogn. Brain Res., 1999, vol. 8, no. 3, p. 373. Beeman, M.J. and Bowden, E.M., The Right Hemisphere Maintains Solution-Related Activation for Yet-toBe-Solved Problems, Mem. Cognit., 2000, vol. 28, no. 7, p. 1231. Bekhtereva, N.P., Starchenko, M.G., Klyucharev, V.A., et al., Study of the Brain Organization of Creativity: II. Positron-Emission Tomography Data, Fiziol. Chel., 2000, vol. 26, no. 5, p. 12 [Hum. Physiol. (Eng. Transl.), 2000, vol. 26, no. 5, p. 516]. Klimesch, W., Vogt, F., and Doppelmayer, M., Interindividual Differences in Alpha and Theta Power Reflect Memory Performance, Intelligence, 2000, vol. 27, no. 4, p. 347. Corsi-Cabrera, M., Ramos, J., Guevara, M.A., et al., Gender Differences in the EEG during Cognitive Activity, Int. J. Neurosci., 1993, vol. 72, no. 3/4, p. 257. Hetrick, W.P., Sandman, C.A., Bunney, W.E., Jr., et al., Gender Differences in Gating of the Auditory Evoked
27. 28.
29.
30. 31.
32.
33.
Potentials in Normal Subjects, Biol. Psychiatry, 1996, vol. 39, no. 1, p. 51. Hoffman, L.D. and Polich, J., P300, Handedness, and Corpus Callosal Size: Gender, Modality, and Task, Int. J. Psychophysiol., 1999, vol. 31, p. 163. Lutzenberger, W., Pulvermuller, F., and Birbaumer, N., Words and Pseudowords Elicit Distinct Patterns of 30Hz EEG Responses in Human, Neurosci. Lett., 1994, vol. 176, no. 1, p. 115. Bechtereva, N.P., Korotkov, A.D., Pakhomov, S.V., et al., PET Study of Brain Maintenance of Verbal Creative Activity, Int. J. Psychophysiol., 2004, vol. 53, no. 1, p. 11. Posner, M.I., Abdullaev, Y.G., McCaudliss, B.D., and Sereno, S.C., Anatomy, Circuitry and Plasticity of Word Reading, Neuroreport, 1999, vol. 10, no. 3, p. R12. Pulvermuller, F., Mohr, B., and Schleichert, H., Semantic or Lexico-Syntactic Factors: What Determines WordClass Specific Activity in the Human Brain?, Neurosci. Lett., 1999, vol. 275, no. 2, p. 81. Elfgren, C.I. and Risberg, J., Lateralized Frontal Blood Flow Increases during Fluency Tasks: Influence of Cognitive Strategy, Neuropsychology, 1998, vol. 36, no. 6, p. 505. Gallagher, H.I., Happe, F., Brunswick, N., et el., Reading the Mind in Cartoons and Stories: An fMRI Study of ‘Theory of Mind’ in Verbal and Nonverbal Tasks, Neuropsychology, 2000, vol. 38, p. 11.
HUMAN PHYSIOLOGY
Vol. 33
No. 2
2007