t

"

~'7.

,;il
ELSEVIER
Neuroscience Letters 198 (1995) 169-172

Non-dominant dorsal-prefrontal activation during chess problem solution evidenced by single photon emission computerized tomography (SPECT)
M. Onofrj ~,*, L. Curatola b, G. Valentini c, M. Antonelli b, A. Thomas a, T. F u l g e n t e a
aDepartment of Neurology, State University of Chieti, 66100 Chieti, Italy bDivisions of Neurology and Otolaringology, Hospital of S. Benedetto, S. Benedetto, USL 22, Italy CDepartment of Nuclear Medicine, Hospital of Macerata, Macerata, USL 15, Italy

Received 9 June 1995; revisedversionreceived 30 August 1995; accepted30 August 1995

Abstract

Expert chess players can recall meaningful chess positions with extraordinary precision in comparison with inexperienced players. We hypothesized, therefore, that their mental performance during chess deliberation could be an appropriate target for single photon emission computerized tomography (SPECT) studies. We studied cerebral activation with 1110 MBq 99mTc-Bicisate SPECT in five expert male chess players during mental solution of a complex chess problem. Region of interest (ROI) analysis, in comparison with average weighted cerebellar counts, showed activation by 10% or more, of the non dominant prefrontal area (fight in four dominant right handed players, left in one dominant left handed player) and by 2-6% in the non-dominant middle temporal areas. Maximum variability of ROI analysis versus cerebellar counts in test/retest evaluation is in our laboratory, as in others, 1.5%. Our results are in agreement with neuropsychological studies suggesting that the non-dominant hemisphere is specialized for chess skill, and show that non-dominant prefrontal and temporal lobe activation during chess deliberation can be detected by SPECT.
Keywords: SPECT; Cerebral cortex; Hemisphere dominance; Frontal lobe; Chess

Experienced chess players are able to recall meaningful chess positions and to find mentally the solution of chess problems [2,4,14] as they develop the ability to schematize chess position by 'chunking the basic elements' into relevant spatial relations [1]. We hypothesized, therefore, that brain activations of chess experts trying to solve mentally a complex chess problem [10] could be an appropriate target for single photon emission computerized tomography (SPECT) studies, as the effect of direct visual stimuli and motor activity will not interfere. Chess problem-solving was never studied previously with SPECT, and only a positron emission tomography study [15] was recently described, in average chess players, during a paradigm including direct checkerboard observation and active displacement of pieces. We studied, therefore, five expert chess players, including three authors of the present paper, all with an * Corresponding author, Clinica Neurologica, Ospedale ex Pediatrico, Via Martiri Lancianesi 6, 66100 Chieti, Italy. Tel.: +39 871 358814; fax: +39 871 355267/355395.

ELO score [5] above 1800, who volunteered for the study (with the international ELO rating system, on the average a good amateur player who reads chess manuals and plays in the local chess club has an ELO above 1200; players classified above 2005 ELO are regularly listed as chess masters). All subjects were able to play more games simultaneously with average amateurs, and could play the chess game turning their back to the checkerboard; the five chess players who were selected for the study were all males, two neurologists (42 (M.O.) and 44 (L.C.) years old, ELO 1850 and 1800, respectively), one pediatrician (37 years old, ELO 2000), one lawyer (53 years old, ELO 2200), and one otolaryngology surgeon (43 years old (M.A.), ELO 1950--2100). All subjects underwent Minnesota multiphasic personality inventory (MMPI), Wechsler adult intelligence scale (WAIS) IQ tests and LQ Edinburgh questionnaire [16]. MMPI results were inside normal limits in all; average WAIS IQ was 133 6. Four subjects are dominant right handed (95-100%) and one is 100% left handed: he is a surgeon who operates, writes and uses instruments with the left hand.

0304-3940/95/$09.50 1995 ElsevierScience IrelandLtd. All fights reserved SSD1 0304-3940(95)11985-3

170

M. Onofrj et al. I Neuroscience Letters 198 (1995) 169-172

1 2 3 4 5 6 7

8
a b c d e f g h

Fig. 1. The administered chess problem. This game was played between Lasker and Bauer in 1889 (Amsterdam). The white player attacks the castled black king and sacrifices both his bishops (15, B:h7 + !K:h7; 16, Q:h5 + K:g8; 17, B:g7 K:g7; 18, Q:g4 + K:h7; 19. R:f3 e5; 20, R:h3 + Q:h6; 21, R:h6 + K:h6; 22, Q:d7 +-). During the game, the black player will, anyhow, lose one of the bishops and be in definite disadvantage. The administered chess problem required the development of a complex strategy, with the sacrifice of two bishops, described in chess manuals [18] as the LaskerBauer game (1889). Fig. 1 shows the chess problem. All players observed the chess problem and mentally studied it for 15 min before and 15 min after the tracer injection. The correct solution was played by all at the end of the 15 min. SPECT brain activation was studied with 99mTcBicisate (Tc-Bi, ethyl cysteinate dimer-Neurolite) a new radiopharmacological agent, with a highly favorable dosimetry [13]. Tc-Bi SPECT had already been used in simple visual activation studies [12] and has shown percent activations similar to those observed with 99mTcH M P A O [19]. Brain uptake is extremely rapid and reaches its maximum (approximately 6% of the injection dose) within 5 min of intravenous injection.

All subjects underwent Tc-Bi SPECT scan in resting conditions (basal), and 4 8 - 7 2 h later, during stimulation. All players gave their informed consent to the study, according to forms and procedures approved by the Radiation Safety Committee and the National Institute for Nervous Disease and Stroke (NINDS) on Human Research Subjects of the National Institutes of Health (protocol no. 930077). Subjects fasted for at least 5 h prior to intravenous Tc-Bi injection, and did not consume alcohol, tobacco, coffee or tea on the day of SPECT. TcBi ( l l l 0 M B q ) were injected in both scans. SPECT acquisition was performed 90 min after injection with a rotating single head camera (GE) with high resolution collimator. Data were acquired from 64 projections (360 degree rotation), in 30 s steps, on a 64 x 64 matrix. Eight axial images were reconstructed with a thickness of two pixels by Butterworth back filtering (0.5 frequency/10 power factor); the attenuation correction was based on Changs algorithm (coefficient 0.09 cm-l). Axial images were resliced on coronal and sagittal planes by using the cantomeathal (CM) coordinate. The CM plane had been determined prior to 99mTc-Bicisate injection on a lateral planar projection using markers taped to the patient's head. The number of counts in each study was more than 6 x 106. Because of results obtained with the first procedure (Table 1), ROIs were analyzed in left and right hemisphere and, finally, by pooling ROIs of dominant versus non-dominant hemisphere. To obtain anatomical representation, the scans were scaled to standard stereotaxic coordinates, based on the Talairach Atlas as revised by the Damasio and Damasio Atlas [3]. Pre- and post-activation scans were compared with current ROI analysis [18]. ROIs were delineated over the peripheral rim of cortex in each slice and on caudate-lenticular nuclei. ROIs were pre frontal cortex (F04, F07 in Damasio and Damasio Atlas [3]), inferior frontal (F11-12), frontal operculum (F06), supplementary motor area (F03), premotor and rolandic region (F08), inferior parietal lobule (P01, P02), superior parietal lobule (P03, P04), middle temporal region (T03, T04), inferior temporal gyrus (T05, T06), mesial temporal lobe (T10, T l l , T12), occipital lobe (O01-O07) and a supplemen-

Table 1 Significant activation foci, measured as % increment in comparisonwith cerebellar activity Subject 1 Left Dorsal pre-frontal cortex Middle temporal cortex Superior parietal l o b u l e . Right 10 4.5 . Subject 2 Left . Right 11 4 Subject 3 Left 3 Right 10 5 . Subject 4 Left . . . Right 12 4 Subject 5 Left 12 6 Right -2 2

Percent activations of different ROIs obtained with the comparison versus cerebellar mean counts pre- and post-activation. Only ROI where percent changes exceeded the reported 1.5% coefficient of variation for test/retest reliability limits [7,11] are reported. Notice that subject 5 (M.A.) is left hand dominant, while subjects 1 to 4 are right hand dominant. Notice the right frontal and temporal activation in the dominant right handed subjects, and the mirror activation in subject 5 (dominant left handed).

M. Onofrj et al. I Neuroscience Letters 198 (1995) 169-172

171

Fig. 2. Cerebral regions associated with the cognitive activities involved in chess problem solutions, in 1, 3 and 5 subjects. The left hemisphere is displayed on the fight of SPECT images. Notice that the first two players are fight handed, and activation is observed in the fight hemisphere (dorsal prefrontal cortex and middle temporal cortex). In the third player, who is left hand dominant, the activation is in the same areas but on the left hemisphere. Notice nodules of activation (4 cm2) in non-dominant dorso-frontal cortex and in temporal cortex. The schematized hemispheres on the fight of the figure show projections onto brain coordinates corresponding to F07 and T03 of Damasio and Damasio Arias [3] (shaded areas). The scale at the bottom is 50 left, 428 fight, ml/10 g per min. tary, lenticular ROI was placed on caudate-lenticular nuclei. The first preactivation versus activation analysis was performed in comparison with cerebellar activity, with the following formula: (counts during activation of R O I / a v erage cerebellar c o u n t s ) - (counts in preactivation of same ROI/average cerebellar counts) = % activation. Although we are well aware that recent studies evidenced a relevant role of the cerebellum in cognitive activities [9], we used this first analysis procedure in order to have results comparable with previous studies on brain activation [7,8,11]. Secondly, we calculated the mean count density in each ROI, in order to obtain measures of relative regional cerebral blood flow (rCBF) for each region [13]. The formula of relative rCBF was: mean count density in each ROI/mean count density in the whole

slice. The data were analyzed using paired sample t-tests on each brain region, and with the Bonferroni method of multiple comparison correction. Two to 3 days before pre-activation scan a sham experiment was run in order to check for eye movements: electro-oculo-grams (EOG) were recorded from four derivations, with methods described in our previous paper [17], in all subjects, in resting conditions and while they mentally solved, with closed eyes, a chess problem. The pattern of eye movements was not consistent, and did not exceed the amount observed during quiet relaxation. EOG were monitored also during pre-activation and activation scans. Table 1 reports percent activations of different ROIs obtained with the first procedure (comparison versus cerebellar mean counts pre- and post-activation). In the comparison versus cerebellar average counts, ROI sampling methods have, in our laboratory, 1.5% coefficient of variation for test/retest reliability, the same as reported by other authors [7,11]. Other studies [8], based on split screen ROI comparison, showed that the average intrasubject difference between two studies was <3% for all brain regions. As shown in Table 1, chess problem solving induced in the four right handed expert chess players activation on the right prefrontal-dorsal cortex and of the right temporo-medial cortex. The percent increments were above 10% in the prefrontal dorsal cortex, and between 4 and 6% in right temporo-medial cortex. The activation in parietal and occipital areas ranged between - 2 and +3%, in the right and left hemisphere. In the lefthanded expert chess player (player 5), activation was instead observed in the left pre-frontal dorsal cortex (12%) and left temporo-occipital cortex (5%), thus confirming that brain activation was localized in the non-dominant hemisphere. Fig. 2 shows SPECT scans of the active areas in three subjects (1, 3, right handed; subject 5 (M.A.) left handed, as reported also in Table 1). The pixel at which there was a significant change (see Table 1) appeared as hyperintense nodules of approximately 4 cm 2 in dorsal prefrontal and temporal areas. The projection of these nodules onto sagittal renderings of the standard brain volume showed that nodules corresponded to dorsal prefrontal and temporal-medial areas, marked as F07 and T03 in Damasio and Damasio Atlas [3]. Relative rCBF in non-dominant dorsal frontal lobes was 1.263 _+0.035 during activation, 1.094 _+0.038 in pre-activation, 1.257 +_0.039 versus 1.140__. 0.028 in non-dominant medial temporal lobe, 1.142 _+0.016 versus 1.087 _+0.030 in parietal lobes, and 1.133 _+0.026 versus 1.136 _+0.019 in occipital lobes. The Bonferroni method comparison corrections test, with a corrected P value set at 0.0125, showed significant increments of pre-frontal dorsal non-dominant cortex and of non-dominant temporo-medial cortex. The mean rCBF was found to be significantly greater during chess deliberation in the nondominant prefrontal region (4 right, 1 left) and the non-

172

M. Onofrj et al. / Neuroscience Letters 198 (1995) 169-172

dominant temporo-medial region ( t - ( 1 . 5 ) = - 5 . 1 5 , P = 0.001; t-(1.5) = -3.98, P = 0.01). Fundamentally, our study shows the unprecedented evidence that the non-dominant frontal lobe is active in the brain of chess experts when they elaborate a solution of a complex chess problem. Activation was also detected in non-dominant temporal lobes. The amount of activation in prefrontal lobes was above 10%, comparable with activation detected in visual or motor areas during visual or motor performances [11].Surprisingly, little or no activation was observed in the parieto-occipital junction (Table 1), an area probably involved in the production of mental imagery [6]. Previous SPECT studies showed activation of dominant frontal lobes during the Tower of London task [14], and during the Stroop paradigm [12]. The activation of the non-dominant frontal lobe observed in our study corroborates the hypothesis that the non-dominant hemisphere is critical for chess skill: one previous study [1], performed with tachistoscopic split screen presentation, showed that chess problems were solved more efficiently if the presentation was on the left hemifield. The authors hypothesized that the chunking of basic elements into meaningful grouping and the parsing of visual stimuli were prominent functions of the non-dominant hemisphere, suggesting that the right non-dominant hemisphere may be relatively specialized for parsing simple visual stimuli according to default rules, such as the Gestalt laws of perceptual organization, whereas the left hemisphere may be relatively specialized for overriding such rules. Our findings show that the activity of the nondominant hemisphere is prominently localized in the dorsal prefrontal cortex. Classic descriptions of hemispheric function attribute prominent skills in non-verbal perception and paralinguistic competence to the non-dominant hemisphere. With our study we show, in the non-dominant hemisphere, a localized function, that is definitely nonlinguistic, but that consists, however, in complex categorization properties and strategic planning. [!] Chabris, C.F. and Hamilton, S.E., Hemispheric specialization for

skilled perceptual organization by chessmasters, Neuropsychologia, 30 (1992) 47-57. [2] Chase, W.G. and Simon, H.A., Perception in chess, Cognit. Psychol., 4 (1973) 55-81. [3] Damasio, H. and Damasio, A.R., Lesion Analysis in Neuropsychologia, Oxford UniversityPress, New York, 1989. [4] De Groot, A.D., Thought and Choice in Chess, 2nd Edn., Mouton, The Hague, 1978. [5] Elo, A.E., The Rating of Chess Players, Past and Present, Aroo, New York, 1978. [6] Farah, M.J., Is visual imagery really visual? - overlooked evidence from neuropsychology,Psychol. Rev., 95 (1988) 307-317. [7] Friberg, L., Andersen, A.R., Lessen, N.A., Holm, S. and Dam, M., Retention of 99mTc-bicisate in human brain after intracarotid injection, J. Cereb. Blood Flow Metab., 14 (1994) S19-$27. [8] George, M.S., Ring, H.A., Costa, D.C., Ell, P.J., Kouris, K., and Jarritt, P.H., Neuroaedvation and Neuroimaging with SPECT, Springer Verlag, London, 1991, pp. 158-171. [9] Grafman, J., Litvan, I., Massaqoui, S., Stewart, B.A., Sirigu, A. and Hallett, M., Cognitive planning deficit patients with cerebellar atriphy, Neurology, 42 (1992) 1493-1496. [10] Holding, D.H. and Pfau, H.D., Thinking ahead in chess, Am. J. Psychol., 98 (1985) 271-282. [11] Holm, S., Madsen, P.L., Sperling, B. and Lassen, N.A., Use of 99mTc-bicisate in activation studies by split dose technique, J. Cereb. Blood Flow Metab., 14 (1994) 115-120. [12] Larrue, V., Celsis, P., B6s, A. and Marc-Vergnes, J.P., The functional anatomy of attention in humans: cerebral blood flow changes induced by reading, naming, and the Stroop effect, J. Cereb. Blood Flow Metab., 14 (1994) 958-962. [13] Masdeu, LC., Brass, L.M., Holman, B.L. and Kushner, M.J., Brain single-photon emission computed tomogmphy, Neurology, 44 (1994) 1970-1977. [14] Morris, R.G., Ahmed, S., Syed, G.M. and Toone, B.K., Neural correlates of planning ability: frontal lobe activation during the Tower of London test, Neuropsychologia,31 (1993) 1367-1378. [15] Nichelli, P., Grafmann, J., Pietrini, P., Alway, D., Carton, J.C. and Miletich, R., Brain activity in chess playing. Nature, 369 (1994) 191. [16] Oldfield, R.C., The assessment and analysis of handedness: the Edinburgh inventory, Neuropsychologia,9 ( 1971) 97-113. [17] Onofrj, M., Fulgente, T., Nobilio, D., Malatesta, G., Bazzano, S, Colamartino, P. and Gambi, D., P3 recorded in patients with bilateral temporal lobe lesions, Neurology, 9 (1992) 1762-1767. [18] Paoli, E., Arte della Combinazione Scacchistica, Nursia, Turin, 1976, pp.115-116. [19] Woods, S.C., Hegeman, I.M., Zubal, I.G., Krystal, J.H., Koster, K., Smith, E.D. and Henninger, G.R., Visual stimulation increases technetium-99m-HMPAO distribution in human visual cortex, J. Nucl. Med., 32 (1991) 210-214.