A New Wrap for Neuronal Activity? Marie E. Bechler and Charles ffrench-Constant Science 344, 480 (2014); DOI: 10.1126/science.1254446
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/344/6183/480.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/344/6183/480.full.html#related This article cites 13 articles, 6 of which can be accessed free: http://www.sciencemag.org/content/344/6183/480.full.html#ref-list-1 This article appears in the following subject collections: Neuroscience http://www.sciencemag.org/cgi/collection/neuroscience
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on May 4, 2014
The following resources related to this article are available online at www.sciencemag.org (this information is current as of May 4, 2014 ):
PERSPECTIVES An added bonus is that the wandering giant planets would have caused mayhem in the asteroid belt. In the Grand Tack model, most planetesimals and planetary embryos were ejected from the belt, and the survivors are a mixture of objects that formed interior and exterior to Jupiter’s original orbit. This may explain the paltry mass of the belt today and the diverse characteristics of asteroids’ spectra (7). However, this diversity may also be a weakness of the model because the wide range of implied formation locations may be hard to reconcile with the isotopic abundances seen in meteorites (8). Izidoro et al. propose a less traumatic alternative. They point to simulations of gas flow within the solar nebula (9) that suggest that material flowed toward the Sun, moving at different speeds at different distances from the star (see the figure, panel A). As a result, the nebula probably thinned out somewhere between 1 and 3 AU. If this partial gap survived long enough, it could have been preserved in the distribution of planetesimals and planetary embryos that formed subsequently. The simulations performed by Izidoro et al. show that reducing the number of planetary building blocks near Mars’ current orbit by 50 to 75% favors the formation of a puny red planet.
In both of these models, a Mars-like planet typically forms near 1 AU and is gravitationally scattered into the depleted region of the solar nebula at an early stage. Here the planet is starved of building material and stops growing. This suggests that Mars should be appreciably older than Earth, which is supported by radiometric dating (10), and that Earth and Mars should be made of similar stuff. It remains to be seen whether known differences in the two planets’ compositions (11) are small enough to be consistent with this prediction. The two models differ when it comes to the asteroid belt. The scenario described by Izidoro et al. predicts that the asteroids formed more or less where they are today, and the asteroid belt was depleted gradually over several hundred million years by long-range gravitational perturbations from the giant planets. In the Grand Tack model, the asteroid belt was purged at a very early stage, and the surviving members sample a much larger region of the solar nebula. These differences may help us distinguish between the models in future. Some caveats are in order. Like all current theories for planet formation, the models described here are incomplete. In particular, we lack a clear understanding of how and where planetesimals formed in the solar neb-
ula, which obviously has a bearing on subsequent events. The final stage of growth is also notoriously chaotic—a tiny change in initial conditions can completely change the final outcome. So, luck played a big role in shaping the solar system. Conventional planet-formation simulations (12) generate respectable Mars analogs in a few percent of cases without requiring any special measures. This leaves the slim possibility that Mars represents a bizarre statistical outlier, and its small size contains no deeper truths about our solar system. References 1. A. Izidoro, N. Haghighipour, O. C. Winter, M. Tsuchida, Astrophys. J. 782, 31 (2014). 2. C. J. Allegre, G. Manhes, C. Gopel, Earth Planet. Sci. Lett. 267, 386 (2008). 3. S. N. Raymond, D. P. O’Brien, A. Morbidelli, N. A. Kaib, Icarus 203, 644 (2009). 4. G. W. Wetherill, in Protostars and Planets, T. Gehrels, Ed. (Univ. of Arizona, Tucson, 1977), pp. 565–598. 5. B. M. S. Hansen, Astrophys. J. 703, 1131 (2009). 6. K. J. Walsh, A. Morbidelli, S. N. Raymond, D. P. O’Brien, A. M. Mandell, Nature 475, 206 (2011). 7. J. Gradie, E. Tedesco, Science 216, 1405 (1982). 8. C. M. O’D. Alexander et al., Science 337, 721 (2012). 9. L. Jin, W. D. Arnett, N. Sui, X. Wang, Astrophys. J. 674, L105 (2008). 10. N. Dauphas, A. Pourmand, Nature 473, 489 (2011). 11. I. A. Franchi, I. P. Wright, A. S. Sexton, C. T. Pillinger, Meteorit. Planet. Sci. 34, 657 (1999). 12. R. A. Fischer, F. J. Ciesla, Earth Planet. Sci. Lett. 392, 28 (2014). 10.1126/science.1252257
NEUROSCIENCE
A New Wrap for Neuronal Activity?
Neuron activity alters the state of white matter in the mammalian brain.
Marie E. Bechler and Charles ffrench-Constant
A
s we learn and experience the world around us, our brains are remodeling neuronal pathways. Is this remodeling only a property of neurons, or are other brain cell types also adapting and contributing to learning? A substantial proportion of cells in the mammalian brain are oligodendrocytes and their precursors (collectively called oligodendroglia). Oligodendrocytes generate multiple myelin sheaths, lipid-rich extensions of specialized plasma membrane that spirally coat condensed layers around neurons. This insulation facilitates rapid nerve conduction, increasing velocities of neuron signals 10-fold (1) by restricting current flow to the small gaps between sheaths— the nodes of Ranvier. More than half of the human brain is composed of myelinated
MRC Centre for Regenerative Medicine, University of Edinburgh, Edinburgh EH16 4UU, UK. E-mail: [email protected]; [email protected]
480
nerves (“white matter”). With such extensive myelinated neural networks, it raises the questions: Do the myelin-producing oligodendrocytes in the brain also adapt as neural circuits are modified in response to activity, and does adaptation of oligodendrocytes then affect the underlying neural circuits? On page 487 of this issue, Gibson et al. (2) provide new evidence to address how neuron activity may promote oligodendroglia changes. Brain magnetic resonance imaging studies of individuals engaging in long-term learning tasks, such as practicing piano, show white matter changes in brain regions associated with these experiences ( 3). Adults learning new skills, such as juggling, also have shown potential changes in myelin (4). However, this imaging technique cannot definitively pinpoint changes specific to myelin sheaths, as the measurements may also reflect changes in axon diameter, density, and organization (3). Therefore, direct
analysis of oligodendrocyte morphology is required. Such analysis in mice has shown a correlation between social isolation and the appearance of fewer and thinner myelin sheaths in brain regions associated with social behavior (5, 6). Previous studies have linked neuron activity to changes in oligodendrocyte precursor proliferation, differentiation, and myelination. Gibson et al. extend these findings using optogenetics, a technique that allows specific neurons expressing a lightsensitive ion channel to be excited by light. Surprisingly, short periods (2.5 min/day for 7 days) of selective stimulation of premotor cortex neurons in juvenile and adult mice, sufficient to induce turning behavior, led to precursor proliferation and increased oligodendrocyte numbers (see the figure). The authors also noted an increase in myelin thickness and improvements in motor function, neither of which was seen
2 MAY 2014 VOL 344 SCIENCE www.sciencemag.org Published by AAAS
PERSPECTIVES
CREDIT: C. BICKEL/SCIENCE
if stimulation was combined with an inhibitor of epigenetic changes required for cell differentiation. Gibson et al. infer that the effects on myelin and motor function are due to adaptive changes in oligodendroglia in response to activity, a conclusion with important implications for understanding brain plasticity. Verification of this interpretation should address some key questions. How does neuronal firing drive local increases in oligodendroglia proliferation and differentiation? Several secreted molecules from active neurons or responding glia have been reported to increase proliferation and/or differentiation, including platelet-derived growth factor, brain-derived neurotrophic factor, adenosine 5′-triphosphate, and glutamate. Oligodendroglia also have ion channels that may respond to neurotransmitters or local ion concentrations (7, 8). Further work should establish what neuronal firing frequencies and duration are sufficient to elicit physiological changes in these and other potential regulatory signals, and whether the responding oligodendroglia require direct contact with the axon in which activity is increased. Are the newly formed oligodendrocytes responsible for thicker myelin? Newer oligodendrocytes are normally generated throughout life, and these new cells can incorporate shorter sheaths into already myelinated neuron tracts, perhaps accounting for shorter myelin in older animals (9). It will be important to establish the fate of the additional new oligodendrocytes generated in response to neuronal activity. Do they generate the thicker myelin, or do existing oligodendrocytes increase the thickness of the already-formed sheaths? We know that existing sheaths may be dynamic, as they can increase thickness and be retracted once formed (10, 11). Equally, it will be important to consider whether the increased number of precursors (that may have homeostatic roles in their own right), rather than changes in myelination, may be contributing to any behavioral effects observed. Adaptive oligodendroglia. Activation of neurons (green) in the mouse premotor cortex increases the proliferation of oligodendrocyte precursors (orange), which differentiate into oligodendrocytes (red). The increased myelin thickness observed could have several possible explanations (not mutually exclusive): (A) myelination of unmyelinated axons by the new oligodendrocytes (red); (B) intercalation of new sheaths on myelinated axons; (C) thickening of existing sheaths (blue). The (unknown) signals responsible may be released by the active axons, leading to thicker myelin on neighboring axons even if their activity levels are not altered.
Unknown signals
Increase in myelin thickness
A
B
Does an increase in myelin thickness affect axonal conduction? Changes in myelin affect conduction velocity and could thereby influence the timing of signals throughout the nervous system—coordination required for synchronization or staggering to permit meaningful interpretation of sensory input and orchestration of fine movements (12). However, conduction velocity is also controlled by myelin sheath length, and the response of length to activity needs to be quantified. It cannot be assumed that any changes will speed conduction; although increasing myelin sheath length and thickness up to a limit will increase conduction velocities, there is a maximum beyond which these properties no longer increase velocity (1, 13). Whether thinning, shortening, or loss of myelin occurs, and whether this results from reduced activity, is also important to determine. The concept of myelin as an adaptable structure that changes to modulate the conduction of neurons, rather than a static structure with set properties, has many implications for understanding how neuronal circuits work in the adult, aged, healthy, and diseased brain. The use of optogenetics to probe adaptive myelination is an important technological advance and may be valuable for further experiments to explore this hypothesis. Further work should explore the dynamics of these changes, whether diminished activity causes reduction or retraction of existing sheaths, and whether myelin can be adapted in brain regions that, unlike the corpus callosum examined by Gibson et al., have very few unmyelinated axons. Determining mechanisms for modifying specific myelin sheath properties also will deepen our understanding of how glia contribute to learning and adapting from experience. References 1. W. A. H. Rushton, J. Physiol. 115, 101 (1951). 2. E. M. Gibson et al., Science 344, 1252304 (2014); 10.1126/science.1252304. 3. R. J. Zatorre, R. D. Fields, H. Johansen-Berg, Nat. Neurosci. 15, 528 (2012). 4. C. Sampaio-Baptista et al., J. Neurosci. 33, 19499 (2013). 5. J. Liu et al., Nat. Neurosci. 15, 1621 (2012). 6. M. Makinodan, K. M. Rosen, S. Ito, G. Corfas, Science 337, 1357 (2012). 7. R. Káradóttir, N. B. Hamilton, Y. Bakiri, D. Attwell, Nat. Neurosci. 11, 450 (2008). 8. P. P. Maldonado, M. Vélez-Fort, F. Levavasseur, M. C. Angulo, J. Neurosci. 33, 2432 (2013). 9. K. M. Young et al., Neuron 77, 873 (2013). 10. S. Goebbels et al., J. Neurosci. 30, 8953 (2010). 11. T. Czopka, C. ffrench-Constant, D. A. Lyons, Dev. Cell 25, 599 (2013). 12. A. H. Seidl, Neuroscience 10.1016/j.neuroscience. 2013.06.047 (2013). 13. L. M. N. Wu, A. Williams, A. Delaney, D. L. Sherman, P. J. Brophy, Curr. Biol. 22, 1957 (2012).
C
10.1126/science.1254446
www.sciencemag.org SCIENCE VOL 344 2 MAY 2014 Published by AAAS
481
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/344/6183/480.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/344/6183/480.full.html#related This article cites 13 articles, 6 of which can be accessed free: http://www.sciencemag.org/content/344/6183/480.full.html#ref-list-1 This article appears in the following subject collections: Neuroscience http://www.sciencemag.org/cgi/collection/neuroscience
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on May 4, 2014
The following resources related to this article are available online at www.sciencemag.org (this information is current as of May 4, 2014 ):
PERSPECTIVES An added bonus is that the wandering giant planets would have caused mayhem in the asteroid belt. In the Grand Tack model, most planetesimals and planetary embryos were ejected from the belt, and the survivors are a mixture of objects that formed interior and exterior to Jupiter’s original orbit. This may explain the paltry mass of the belt today and the diverse characteristics of asteroids’ spectra (7). However, this diversity may also be a weakness of the model because the wide range of implied formation locations may be hard to reconcile with the isotopic abundances seen in meteorites (8). Izidoro et al. propose a less traumatic alternative. They point to simulations of gas flow within the solar nebula (9) that suggest that material flowed toward the Sun, moving at different speeds at different distances from the star (see the figure, panel A). As a result, the nebula probably thinned out somewhere between 1 and 3 AU. If this partial gap survived long enough, it could have been preserved in the distribution of planetesimals and planetary embryos that formed subsequently. The simulations performed by Izidoro et al. show that reducing the number of planetary building blocks near Mars’ current orbit by 50 to 75% favors the formation of a puny red planet.
In both of these models, a Mars-like planet typically forms near 1 AU and is gravitationally scattered into the depleted region of the solar nebula at an early stage. Here the planet is starved of building material and stops growing. This suggests that Mars should be appreciably older than Earth, which is supported by radiometric dating (10), and that Earth and Mars should be made of similar stuff. It remains to be seen whether known differences in the two planets’ compositions (11) are small enough to be consistent with this prediction. The two models differ when it comes to the asteroid belt. The scenario described by Izidoro et al. predicts that the asteroids formed more or less where they are today, and the asteroid belt was depleted gradually over several hundred million years by long-range gravitational perturbations from the giant planets. In the Grand Tack model, the asteroid belt was purged at a very early stage, and the surviving members sample a much larger region of the solar nebula. These differences may help us distinguish between the models in future. Some caveats are in order. Like all current theories for planet formation, the models described here are incomplete. In particular, we lack a clear understanding of how and where planetesimals formed in the solar neb-
ula, which obviously has a bearing on subsequent events. The final stage of growth is also notoriously chaotic—a tiny change in initial conditions can completely change the final outcome. So, luck played a big role in shaping the solar system. Conventional planet-formation simulations (12) generate respectable Mars analogs in a few percent of cases without requiring any special measures. This leaves the slim possibility that Mars represents a bizarre statistical outlier, and its small size contains no deeper truths about our solar system. References 1. A. Izidoro, N. Haghighipour, O. C. Winter, M. Tsuchida, Astrophys. J. 782, 31 (2014). 2. C. J. Allegre, G. Manhes, C. Gopel, Earth Planet. Sci. Lett. 267, 386 (2008). 3. S. N. Raymond, D. P. O’Brien, A. Morbidelli, N. A. Kaib, Icarus 203, 644 (2009). 4. G. W. Wetherill, in Protostars and Planets, T. Gehrels, Ed. (Univ. of Arizona, Tucson, 1977), pp. 565–598. 5. B. M. S. Hansen, Astrophys. J. 703, 1131 (2009). 6. K. J. Walsh, A. Morbidelli, S. N. Raymond, D. P. O’Brien, A. M. Mandell, Nature 475, 206 (2011). 7. J. Gradie, E. Tedesco, Science 216, 1405 (1982). 8. C. M. O’D. Alexander et al., Science 337, 721 (2012). 9. L. Jin, W. D. Arnett, N. Sui, X. Wang, Astrophys. J. 674, L105 (2008). 10. N. Dauphas, A. Pourmand, Nature 473, 489 (2011). 11. I. A. Franchi, I. P. Wright, A. S. Sexton, C. T. Pillinger, Meteorit. Planet. Sci. 34, 657 (1999). 12. R. A. Fischer, F. J. Ciesla, Earth Planet. Sci. Lett. 392, 28 (2014). 10.1126/science.1252257
NEUROSCIENCE
A New Wrap for Neuronal Activity?
Neuron activity alters the state of white matter in the mammalian brain.
Marie E. Bechler and Charles ffrench-Constant
A
s we learn and experience the world around us, our brains are remodeling neuronal pathways. Is this remodeling only a property of neurons, or are other brain cell types also adapting and contributing to learning? A substantial proportion of cells in the mammalian brain are oligodendrocytes and their precursors (collectively called oligodendroglia). Oligodendrocytes generate multiple myelin sheaths, lipid-rich extensions of specialized plasma membrane that spirally coat condensed layers around neurons. This insulation facilitates rapid nerve conduction, increasing velocities of neuron signals 10-fold (1) by restricting current flow to the small gaps between sheaths— the nodes of Ranvier. More than half of the human brain is composed of myelinated
MRC Centre for Regenerative Medicine, University of Edinburgh, Edinburgh EH16 4UU, UK. E-mail: [email protected]; [email protected]
480
nerves (“white matter”). With such extensive myelinated neural networks, it raises the questions: Do the myelin-producing oligodendrocytes in the brain also adapt as neural circuits are modified in response to activity, and does adaptation of oligodendrocytes then affect the underlying neural circuits? On page 487 of this issue, Gibson et al. (2) provide new evidence to address how neuron activity may promote oligodendroglia changes. Brain magnetic resonance imaging studies of individuals engaging in long-term learning tasks, such as practicing piano, show white matter changes in brain regions associated with these experiences ( 3). Adults learning new skills, such as juggling, also have shown potential changes in myelin (4). However, this imaging technique cannot definitively pinpoint changes specific to myelin sheaths, as the measurements may also reflect changes in axon diameter, density, and organization (3). Therefore, direct
analysis of oligodendrocyte morphology is required. Such analysis in mice has shown a correlation between social isolation and the appearance of fewer and thinner myelin sheaths in brain regions associated with social behavior (5, 6). Previous studies have linked neuron activity to changes in oligodendrocyte precursor proliferation, differentiation, and myelination. Gibson et al. extend these findings using optogenetics, a technique that allows specific neurons expressing a lightsensitive ion channel to be excited by light. Surprisingly, short periods (2.5 min/day for 7 days) of selective stimulation of premotor cortex neurons in juvenile and adult mice, sufficient to induce turning behavior, led to precursor proliferation and increased oligodendrocyte numbers (see the figure). The authors also noted an increase in myelin thickness and improvements in motor function, neither of which was seen
2 MAY 2014 VOL 344 SCIENCE www.sciencemag.org Published by AAAS
PERSPECTIVES
CREDIT: C. BICKEL/SCIENCE
if stimulation was combined with an inhibitor of epigenetic changes required for cell differentiation. Gibson et al. infer that the effects on myelin and motor function are due to adaptive changes in oligodendroglia in response to activity, a conclusion with important implications for understanding brain plasticity. Verification of this interpretation should address some key questions. How does neuronal firing drive local increases in oligodendroglia proliferation and differentiation? Several secreted molecules from active neurons or responding glia have been reported to increase proliferation and/or differentiation, including platelet-derived growth factor, brain-derived neurotrophic factor, adenosine 5′-triphosphate, and glutamate. Oligodendroglia also have ion channels that may respond to neurotransmitters or local ion concentrations (7, 8). Further work should establish what neuronal firing frequencies and duration are sufficient to elicit physiological changes in these and other potential regulatory signals, and whether the responding oligodendroglia require direct contact with the axon in which activity is increased. Are the newly formed oligodendrocytes responsible for thicker myelin? Newer oligodendrocytes are normally generated throughout life, and these new cells can incorporate shorter sheaths into already myelinated neuron tracts, perhaps accounting for shorter myelin in older animals (9). It will be important to establish the fate of the additional new oligodendrocytes generated in response to neuronal activity. Do they generate the thicker myelin, or do existing oligodendrocytes increase the thickness of the already-formed sheaths? We know that existing sheaths may be dynamic, as they can increase thickness and be retracted once formed (10, 11). Equally, it will be important to consider whether the increased number of precursors (that may have homeostatic roles in their own right), rather than changes in myelination, may be contributing to any behavioral effects observed. Adaptive oligodendroglia. Activation of neurons (green) in the mouse premotor cortex increases the proliferation of oligodendrocyte precursors (orange), which differentiate into oligodendrocytes (red). The increased myelin thickness observed could have several possible explanations (not mutually exclusive): (A) myelination of unmyelinated axons by the new oligodendrocytes (red); (B) intercalation of new sheaths on myelinated axons; (C) thickening of existing sheaths (blue). The (unknown) signals responsible may be released by the active axons, leading to thicker myelin on neighboring axons even if their activity levels are not altered.
Unknown signals
Increase in myelin thickness
A
B
Does an increase in myelin thickness affect axonal conduction? Changes in myelin affect conduction velocity and could thereby influence the timing of signals throughout the nervous system—coordination required for synchronization or staggering to permit meaningful interpretation of sensory input and orchestration of fine movements (12). However, conduction velocity is also controlled by myelin sheath length, and the response of length to activity needs to be quantified. It cannot be assumed that any changes will speed conduction; although increasing myelin sheath length and thickness up to a limit will increase conduction velocities, there is a maximum beyond which these properties no longer increase velocity (1, 13). Whether thinning, shortening, or loss of myelin occurs, and whether this results from reduced activity, is also important to determine. The concept of myelin as an adaptable structure that changes to modulate the conduction of neurons, rather than a static structure with set properties, has many implications for understanding how neuronal circuits work in the adult, aged, healthy, and diseased brain. The use of optogenetics to probe adaptive myelination is an important technological advance and may be valuable for further experiments to explore this hypothesis. Further work should explore the dynamics of these changes, whether diminished activity causes reduction or retraction of existing sheaths, and whether myelin can be adapted in brain regions that, unlike the corpus callosum examined by Gibson et al., have very few unmyelinated axons. Determining mechanisms for modifying specific myelin sheath properties also will deepen our understanding of how glia contribute to learning and adapting from experience. References 1. W. A. H. Rushton, J. Physiol. 115, 101 (1951). 2. E. M. Gibson et al., Science 344, 1252304 (2014); 10.1126/science.1252304. 3. R. J. Zatorre, R. D. Fields, H. Johansen-Berg, Nat. Neurosci. 15, 528 (2012). 4. C. Sampaio-Baptista et al., J. Neurosci. 33, 19499 (2013). 5. J. Liu et al., Nat. Neurosci. 15, 1621 (2012). 6. M. Makinodan, K. M. Rosen, S. Ito, G. Corfas, Science 337, 1357 (2012). 7. R. Káradóttir, N. B. Hamilton, Y. Bakiri, D. Attwell, Nat. Neurosci. 11, 450 (2008). 8. P. P. Maldonado, M. Vélez-Fort, F. Levavasseur, M. C. Angulo, J. Neurosci. 33, 2432 (2013). 9. K. M. Young et al., Neuron 77, 873 (2013). 10. S. Goebbels et al., J. Neurosci. 30, 8953 (2010). 11. T. Czopka, C. ffrench-Constant, D. A. Lyons, Dev. Cell 25, 599 (2013). 12. A. H. Seidl, Neuroscience 10.1016/j.neuroscience. 2013.06.047 (2013). 13. L. M. N. Wu, A. Williams, A. Delaney, D. L. Sherman, P. J. Brophy, Curr. Biol. 22, 1957 (2012).
C
10.1126/science.1254446
www.sciencemag.org SCIENCE VOL 344 2 MAY 2014 Published by AAAS
481
