brief communications

Pattern of focal -bursts in chess players

Grandmasters call on regions of the brain not used so much by less skilled amateurs.

T
he brain’s medial temporal lobe struc- suggests that expert memory is based on a These marked differences in the distrib-
tures are thought to be important for large database of chunks in long-term ution of focal brain activity during chess
the initial formation of long-term memory. A chess grandmaster studies and playing point to differences in the mecha-
memory1,2, and active memory is indicated practises for at least 10 years to learn more nisms of brain processing and functional
by bursts of -band activity in these and than 100,000 patterns (memory chunks). brain organization between grandmasters
other areas of the association cortex3,4. Here Consequently, grandmasters can ‘recognize’ and amateurs. Lesions in structures that are
we use a new technique of magnetic the key elements in a problem situation activated in amateur players impair recent
imaging to compare focal bursts of -band much more rapidly than amateur players. memory while leaving remote memory
activity in amateur and professional chess Experts differ not only in the extent of their intact2. Grandmasters seem to rely more on
players during matches. We find that this knowledge, but also in its organization. remote than on recent memory.
activity is most evident in the medial High-level processing elements, such as High-level processing elements6 may
temporal lobe in amateur players, which is structuring knowledge and planning, assist also have contributed to the observed dif-
consistent with the interpretation that their in accessing the respective chunks6. ferences. The chunking theory of memory
mental acuity is focused on analysing We tested 20 male players (aged 4214 states that the number and nature of chunks
unusual new moves during the game. In years), each with more than 10 years of that chess experts can hold in long-term
contrast, highly skilled chess grandmasters tournament and training practice. Ten memory can be used to predict chess
have more -bursts in the frontal and professional grandmasters scored between performance6. Our results indicate that
parietal cortices, indicating that they are 2,400 and 2,600 on Elo’s chess-skill rating the activation of expert memory chunks
retrieving chunks from expert memory scale5; amateur players ranked 1,700 and produces focal -band activity in the
by recruiting circuits outside the medial above. Magnetoencephalographic record- neocortex, whereas amateur players primar-
temporal lobe. ings were made while subjects played ily encode and analyse new information,
The ‘chunking’ theory of chess playing5 against a computer and were scanned in the tasks that activate the medial temporal lobe
5 seconds after each move by the computer and the hippocampus. It is possible that
aa program for focal -bursts (20–40 Hz; these structures play only a transitional role
Fig. 1a). Examination of single slices during the establishment of expert memory
indicates pronounced activity in the region in the neocortex.
of the perirhinal and entorhinal cortex, Ognjen Amidzic, Hartmut J. Riehle,
hippocampus and related structures in Thorsten Fehr, Christian Wienbruch,
amateur players, but not in grandmasters. Thomas Elbert
There was a strong negative correlation University of Konstanz, Universitätsstrasse 10,
(r0.84) between the relative share of Fach-D30, 78457 Konstanz, Germany
dipoles in these structures and Elo chess e-mail: [email protected]
skill (Fig. 1b). The correlation was also sig- 1. Bontempi, B., Laurent-Demir, C., Destrade, C. & Jaffard, R.
Nature 400, 671–674 (1999).
nificant when only the 12 players who had
2. Squire, L. R. Psychol. Rev. 99, 195–231 (1992).
lost in the game were included in the analy- 3. Pulvermüller, F., Keil, A. & Elbert, T. Trends Cogn. Sci. 3,
sis (draws and wins were achieved only by 250–252 (1999).
grandmasters). There was no relationship 4. Singer, W. et al. Trends Cogn. Sci. 1, 252–261 (1997).
b 90 5. Elo, A. E. The Rating of Chess Players, Past and Present
between the length or complexity of the (Arco, New York, 1978).
80 game and the Elo score. 6. Gobet, F. Cognition 66, 115–152 (1998).
Proportion of dipoles (%)

70

60 anabolic, non-invasive stimulus may have

Anabolism
50 potential for treating skeletal conditions
40
Low mechanical signals such as osteoporosis.
A common perception of bone adapta-
30
strengthen long bones tion is that mechanical signals must be large
to influence morphology2. The peak signals

A
20
lthough the skeleton’s adaptability to
load-bearing has been recognized for that result from natural vigorous activity
10
1,500 1,800 2,100 2,400 2,700
over a century1, the specific mechani- cause microdamage to bone material and
cal components responsible for strengthen- require repair3. For example, peak strains of
Elo score
ing it have not been identified. Here we 2,000–3,000 microstrain are typically
Figure 1 Focal -band activity in the brains of chess players. show that after mechanically stimulating induced during locomotion4, stimulating
a, Determination of equivalent-current dipole density for inferior the hindlimbs of adult sheep on a daily osteoclasts and osteoblasts to remove and
‘slices’ through the hippocampus and medial temporal areas in an basis for a year with 20-minute bursts of then replace damaged tissue5.
amateur player’s brain. Increasing dipole density is indicated by a very-low-magnitude, high-frequency vibra- We have departed from this repair-
colour scale from dark to light blue, to yellow, to red. b, Relation- tion, the density of the spongy (trabecular) mediated hypothesis by proposing that
ship between chess-playing skill (Elo rating scale) and the relative bone in the proximal femur is significantly extremely small strains (for example, those
share of dipoles located in medial temporal lobe structures (black) increased (by 34.2%) compared to controls. that arise from muscle contraction during
and in the frontal and parietal cortices (red). Amateur players show As the strain levels generated by this treat- less vigorous but more frequent activities
more focal -bursts in the medial temporal lobe than grandmas- ment are three orders of magnitude below such as maintaining posture) are strong
ters, who show more activity in the frontal and parietal cortices. those that damage bone tissue, this determinants of bone morphology6.
NATURE | VOL 412 | 9 AUGUST 2001 | www.nature.com © 2001 Macmillan Magazines Ltd 603
brief communications
We examined the regulatory potential of Table 1 Proximal-femur parameters of control and stimulated sheep
extremely small (0.3g, where g is the Earth’s Control Experimental Difference P
gravitational field), high-frequency (30 Hz) Animal mass (kg) 71.17.1 70.39.4 1.1% n.s.
mechanical accelerations by subjecting the Total density (gm cm3) 46660 49653 +6.5% 0.1
hindlimbs of adult (6–8-year-old) female Trabecular density (gm cm3) 16937 22756 +34.2% 0.01
sheep (Warhill, intact ewes) to a ground- Bone volume/total volume (%) 15.24.1 20.14.8 +32% 0.04
based vertical oscillation7 for 20 min per day Trabecular spacing (m) 1,170124 75697 36% 0.02
for 5 days a week. When the animals were Trabecular number (trabeculae mm–2) 0.820.16 1.190.18 +45% 0.01
not being treated, they joined the controls to Bone-formation rate (m2 mm1) 8.412.7 17.916.3 +113% 0.2
roam freely over a pasture area. We used Mineralizing surface (%) 2.60.16 6.345.14 +144% 0.1
strain gauges attached to the animals’ tibia Animal mass and envelope-specific bone density (determined by quantitative computer tomography) of the proximal femur after 12 months of low-level
bone to calibrate the device: these showed mechanical stimulation. Also shown are indices of static and dynamic histomorphometry of the proximal femur. Although ‘whole-bone’ parameters of the
proximal femur show only a limited tendency to be influenced by mechanical stimuli (P0.1), the increase in treated animals compared with controls is over
that the peak-to-peak amplitude of the 30% for trabecular bone alone (P0.01). One control was lost over the course of study for reasons not associated with the protocol. Nine animals were
strain generated was about 5 microstrain, evaluated in the experimental group, with eight controls. All evaluations were made without knowledge of whether the animals were control or experimental.
which is 0.1% of the strain magnitude that is
known to cause yield-failure in bone8. We found that this low-level mechanical New York 11794-2580, USA
After 1 year of this mechanical-stimula- stimulation increased the rate of bone e-mail: [email protected]
tion regime, the density of trabecular bone formation 2.1-fold (P0.2) and the miner- †Department of Clinical Sciences, Colorado State
in the proximal femur, as quantified by alizing surface 2.4-fold (P0.1). This ana- University, Fort Collins, Colorado 80523, USA
computer tomography, was 34.2% greater bolic effect was highly specific to cancellous ‡Skeletech Inc., Bothell, Washington 98021, USA
in experimental sheep than in controls (porous) bone, as there was no significant 1. Wolff, J. The Law of Bone Remodeling (transl. Maquet, P. &
(P0.01; Table 1). This strong anabolic histomorphometric change in any of the Furlong, R.) (Springer, Berlin, 1986).
2. Frost, H. Anat. Rec. 26, 403–413 (1990).
response was substantiated by undecalcified cortical bone parameters. We detected 3. Burr, D., Martin, R., Schaffler, M. & Radin, E. J. Biomech. 18,
bone histology of the same region, which no difference in any bone index in the 189–200 (1985).
revealed a 32% increase in trabecular bone radius of either control or experimental 4. Rubin, C. & Lanyon, L. J. Exp. Biol. 101, 187–211 (1982).
5. Schaffler, M., Radin, E. & Burr, D. Bone 11, 321–326 (1990).
volume, a 45% increase in trabecular mesh animals (for example, mineral density was
6. Huang, R., McLeod, K. & Rubin, C. J. Gerontol. 54,
number (Fig. 1) and a 36% reduction in 0.6% less than controls; not statistically 352–357 (1999).
mesh spacing, indicating an increase in the significant), indicating that the anabolic 7. Fritton, J., Rubin, C., Qin, Y. & McLeod, K. Ann. Biomed. Eng.
mean width of each trabecular element and effect was specific to the region of 25, 831–839 (1997).
8. Carter, D., Harris, W., Vasu, R. & Caler, W. Am. Soc. Mech. Eng.
the addition of new trabeculae. the skeleton that was subjected to the 45, 81–95 (1981).
mechanical signal. 9. Hylander, W., Ravosa, M., Ross, C. & Johnson, K. Am. J. Phys.
Mechanical strain in the skeleton is a Anthropol. 107, 257–271 (1998).
product of functional load-bearing — as 10. Blob, R. & Biewener, A. J. Exp. Biol. 202, 1023–1046 (1999).
11. Fritton, S., McLeod, K. & Rubin, C. J. Biomech. 33,
seen, for example, in the mandible of the 317–326 (2000).
macaque9 and the tibia of the alligator10. In 12. Bain, S. & Rubin, C. J. Bone Miner. Res. 5, 1069–1075 (1990).
addition to the large-amplitude strains 13. Rubin, C., Gross, T., McLeod, K. & Bain, S. J. Bone Miner. Res.
10, 488–495 (1995).
typically associated with functional activity,
a strain signal, much less than 5 microstrain
in amplitude, arises through muscular
activity in the frequency band 10–50 Hz Vision
(ref. 11). Generation of this small-
amplitude, high-frequency muscle ‘vibra- Realignment of cones
tion’ persists through even such passive
activities as standing.
after cataract removal
Skeletal morphology may therefore be

T
hrough unique observations of an
sculpted by omnipresent, low-level muscle adult case of bilateral congenital
activity as well as by the peak impacts cataract removal1, we have found
inherent in load-bearing. In terms of clini- evidence that retinal photoreceptors will
cal relevance, the strong bone-generating swiftly realign towards the brightest regions
capacity of these small signals suggests that in the pupils of the eye. Cones may be
biomechanical intervention might help to phototropic, actively orientating themselves
strengthen bone in osteoporosis sufferers towards light like sunflowers in a field.
without the side-effects associated with For 40 years, one of us (P.D.) lived with
pharmacological treatment. In addition to an unusual pupil configuration caused by
being non-invasive and inducing a thera- congenital bilateral cataracts, first diagnosed
peutic response from the bone tissue itself, at three years of age. The cataracts were
low-intensity mechanical signals incorpo- managed by using a twice-weekly applica-
rate all aspects of a complex remodelling tion of atropine, which dilated P.D.’s pupils
cycle12 and ultimately stimulate formation to produce roughly annular (ring-shaped),
of lamellar bone13 to improve bone quantity clear regions around the dense nuclear
and quality. cataracts. Because his optics were so poor
Figure 1 Montages of photomicrographs of the proximal sheep Clinton Rubin*, A. Simon Turner†, (he was far-sighted, astigmatic and in-
femur used for static histomorphometric evaluation after 1 year of Steven Bain‡, Craig Mallinckrodt†, capable of accommodation), P.D. squinted
exposure (20 min per day) to a 0.3g, 30-Hz mechanical stimulus. Kenneth McLeod* continuously as an adaptation, particularly
a, b, There is 32% more trabecular bone in the proximal femur of *Musculo-Skeletal Research Laboratory, for close work. His eyelids thus horizontally
experimental animals (a) compared with age-matched controls (b) Department of Biomedical Engineering, cropped each annulus at the top and bot-
(P0.04). State University of New York, Stony Brook, tom, leaving two clear entry points for light
604 © 2001 Macmillan Magazines Ltd NATURE | VOL 412 | 9 AUGUST 2001 | www.nature.com
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.