Download Another Nail in the Plume Coffin?

PERSPECTIVES
GEOPHYSICS
Another Nail in the Plume Coffin?
At least one chain of hot-spot volcanoes is
not caused by a plume rising up from the
core-mantle boundary, calling for a
reexamination of the plume hypothesis.
Marcia K. McNutt
The author is at the Monterey Bay Aquarium Research
Institute, 7700 Sandholdt Road, Moss Landing, CA 95039,
USA. E-mail: [email protected]
1394
8 SEPTEMBER 2006
VOL 313
SCIENCE
Published by AAAS
Chain of volcanoes
A new kind of volcanism? In
this three-dimensional cartoon, a crack in a flexed plate
provides a conduit through
which buoyant pockets of partial melt in the upper mantle
can reach Earth’s surface.
volcanoes in the flexural
trough seaward of the
outer rise of the Japan
t
Crus
trench suggests that elastic bending of the plate
has opened up conduits
Plate
through which the partial
melt can rise to the surface
tle
(see the figure). A plume
an
m
en
source for these volcanoes is
olt
m
y
l
unlikely
in any case, because the
l
a
i
Part
nearby subduction of the Pacific
plate creates widespread downwelling
that would block the rise of a mantle plume.
The notion that the upper mantle beneath
the plates might contain small amounts of partial melt just about everywhere, and that all it
takes is a crack in the plate for that melt to rise
to the surface, has been previously proposed
as an alternative to the plume hypothesis (6).
However, prior examples of hot spots used to
illustrate this concept were not as compelling, because there was no obvious mechanism for cracking the plate and they were
located away from sites of obvious largescale downwelling.
Does the plate-cracking mechanism proposed by Hirano et al. only explain this one
small chain of volcanoes, or can it also explain
other chains of hot spots? Hirano et al. argue
that they have found a new and different kind
of hot-spot volcano. They base their argument
on the observed ratios of isotopes of inert
gases. According to plume theory, plumes tap
a deep reservoir rich in rare gases, whereas the
mid-ocean ridges obtain melt from a reservoir
in the upper mantle that is depleted in rare
gases. The volcanoes described by Hirano et
al. have more depleted isotopic ratios, such as
those found in mid-ocean ridge basalts;
according to plume aficionados, this property
would put them in a category different from
hot spots such as those in Iceland and Hawaii.
However, other interpretations of the rare gas
isotopes would not necessarily require distinct
mantle sources (7).
www.sciencemag.org
Downloaded from www.sciencemag.org on April 21, 2010
S
and surface-dominated
processes of plate tectonics. Second, the orientation and rate of age
progression of hot-spot
volcanoes reveal the
direction and speed of
absolute motion of the
tectonic plates, whereas only relative motion
could be inferred from
the properties of the
plates themselves. Scientists therefore immediately started to explore
the implications of the
plume model for properties
of Earth’s interior and for absolute plate motions, and the distinction between the predictions of a
hypothesis and observed facts quickly
became blurred.
But troubling inconsistencies began to
emerge. The absolute motions of plates inferred from hot spots on different plates did not
agree, prompting the proposal that the plumes
“blow in the mantle wind” (3). Rocks from hotspot volcanoes with different ratios of noblegas isotopes were thought to tap separate reservoirs in the mantle that were isolated from
intermixing. Unfortunately, as the number of
isotopic tracers grew, additional reservoirs
were required (4), taxing the ability of geodynamic models to keep them isolated over geologic time scales. Furthermore, because of their
small dimensions, plumes proved difficult to
detect in tomographic images of the mantle.
Advanced techniques do suggest the upwelling
of hot material from the lower to the upper
mantle beneath some hot spots (5), but not
beneath others. Are these problems simply the
maturing of a valid theory to deal with the complexity of the real planet, or are they the signs of
a paradigm in crisis?
Although insignificant in size on a planetary scale, the small seamounts described by
Hirano et al. are important because they are
young hot-spot volcanoes (1 to 8 million years
old) that erupt through a very old plate (135
million years old), away from any plate
boundaries. Petrologic data confirm that the
volcanoes tap a region of the mantle more than
100 km deep that contains very small percentages of melted material. The position of the
CREDIT: P. HUEY/SCIENCE
cientists love beautiful theories—the
kind that are elegant, predictive, and
have few free parameters. And they hate
it when theories like that prove to be wrong. It
is thus with much kicking, dragging, and
screaming that geoscientists are being brought
to the realization that all might not be well
with the concept of mantle plumes.
According to the plume hypothesis, linear
chains of volcanoes that form at “hot spots”
away from the boundary zones of tectonic
plates are caused
by hot, buoyant
Enhanced online at
jets that detach
www.sciencemag.org/cgi/
from the thermal
content/full/313/5792/1394
boundary separating Earth’s core
from its mantle. The latest attack on this
hypothesis comes from Hirano et al. on page
1426 of this issue (1). The authors describe a
small chain of hot-spot volcanoes off the
Japanese coast that almost assuredly cannot
have been formed by narrow, deep-Earth
upwellings.
The plume hypothesis was originally proposed by Jason Morgan (2), hot on the heels of
the plate tectonic revolution. According to the
latter theory, most geologic activity—such as
volcanic eruptions, earthquakes, and mountain
building—is concentrated at the margins of
the dozen or so major plates that move
together, move apart, or slip past each other on
the surface of our planet. Morgan and others
noted that island chains such as the Hawaiian
Islands and the Canary Islands consist of
young volcanoes in the middle of the tectonic
plates and therefore cannot be explained by
plate tectonics. Moreover, many of these island
chains, particularly those on the Pacific plate,
show a monotonic increase in age in one direction. Morgan hypothesized that these midplate
volcanoes, termed hot spots, are the surface
manifestation of mantle plumes that rapidly (to
a geologist’s way of thinking) rise up from the
core-mantle boundary.
Morgan’s hypothesis has several attractive
features. First, it suggests that the study of hot
spots will provide clues to the chemistry and
dynamics of Earth’s interior, which would otherwise be masked from scrutiny by the rigidity
PERSPECTIVES
cal, and geochemical inferences that have
been built on the plume paradigm.
References
1. N. Hirano et al., Science 313, 1426 (2006); published
online 27 July 2006 (10.1126/science.1128235).
2. W. J. Morgan, J. Geophys. Res. 73, 1959 (1968).
3. B. Steinberger, R. J. O’Connell, in The History and
Dynamics of Global Plate Motions, M. A. Richards, R. G.
Gordon, R. D. van der Hilst, Eds. (American Geophysical
Union, Washington, DC, 2000), pp. 377–398.
PSYCHOLOGY
4. A. Zindler, E. Jagoutz, S. Goldstein, Nature 298, 519 (1982).
5. R. Montelli et al., Science 303, 338 (2004); published
online 4 December 2003 (10.1126/science.1092485).
6. M. K. McNutt, D. Caress, J. Reynolds, K. Jordahl, Nature
389, 479 (1997).
7. D. L. Anderson, Proc. Natl. Acad. Sci. U.S.A. 95, 9087
(1998).
Published online 27 July 2006;
10.1126/science.1131298
Include this information when citing this paper.
Brain imaging reveals unexpected activity in
a patient clinically diagnosed as being in a
vegetative state, raising questions about the
properties of consciousness.
Is She Conscious?
Lionel Naccache
s she understanding my words and
feeling my caresses?” is a question
constantly asked by relatives and
caregivers of comatose and vegetative state
patients in intensive care units throughout
the world. Probing residual mental function
in such critical situations poses major medical and ethical issues. Our current answers
to this question are mainly based on detailed
behavioral and neurological observations,
but this approach may be blind to inner
active mental processes. Since the late
1980s, several neurophysiological correlates of cognitive processes have been proposed to better assess the existence of this
covert mental life (1). This approach has
aimed to discriminate among comatose or
vegetative state patients those who are still
cognitively active. On page 1402 of this
issue, Owen et al. (2) use functional magnetic resonance imaging (fMRI) to infer the
psychological processes at work in such
patients during mental imagery tasks that do
not elicit otherwise observable behavior.
The authors studied a 23-year-old woman
who sustained a severe traumatic brain injury
in a road traffic accident. After an initial comatose state (defined as an unarousable unresponsiveness state), she opened her eyes and
demonstrated sleep-wake cycles. However,
even during the waking periods, she was unresponsive (for example, to visual or auditory
stimuli) and did not manifest spontaneous
intentional behaviors. These signs are diagnostic of a vegetative state.
In a first experiment conducted 5 months
after the accident, Owen and colleagues presented spoken sentences to the patient while
CREDIT: MATHAUS GREUTER
“I
The author is in the Clinical Neurophysiology Department,
Hôpital de la Pitié-Salpêtrière, Paris, 75013 France and in
the Cognitive Neuroimaging Unit INSERM, Orsay, 91401
France. E-mail: lionel.naccache@ wanadoo.fr
recording neural responses with fMRI.
Speech-specific brain regions were clearly
activated while the patient listened to these
sentences, as compared to acoustically
matched noise sequences. Moreover, sentences containing ambiguous homophone
words (for example, creak/creek) activated
an additional left inferior frontal lobe region
known to subserve the selection of semantic
knowledge among competing alternatives (3).
In sharp contrast to her behaviour, which was
suggestive of poor cognitive abilities, this
patient could process external auditory information involving language.
In a second experiment, the authors engaged the patient in two mental imagery tasks
by asking her either to “imagine visiting the
rooms in your home” or to “imagine playing
tennis.” The result of this second fMRI investigation is quite spectacular. Patterns of brain
activation observed during the 30-s period of
each task were highly suggestive of an active
mental performance relevant to the task. In the
spatial mental imagery task, Owen et al.
observed neural activations within a network
including the parahippocampal gyrus, posterior parietal cortex, and lateral premotor cortex. When the patient was asked to imagine
playing tennis, strong activations were recorded in the supplementary motor areas that
control motor responses.
The observed brain activation patterns are
the classic neural correlates of these two
mental imagery tasks. Indeed, statistical parametric maps of brain activation observed in
the patient were indistinguishable from
those recorded from a group of conscious
control subjects.
Despite the patient’s very poor behavioral
status, the fMRI findings indicate the existence of a rich mental life, including auditory
language processing and the ability to perform mental imagery tasks. On one hand,
this single case makes a strong argument for
the development of fMRI and other neurophysiological tools (such as monitoring electroencephalogram brain responses to external stimuli) to evaluate cognition in such
patients. On the other hand,
we should not generalize from
this single patient, who suffered relatively few cerebral
lesions, to most other vegetative state patients, who typically have massive structural
brain lesions.
Is this woman conscious?
Most current behavioral and
neuroimaging evidence suggests that conscious processing is abolished during a vegetative state. Such patients do
not report mental states, nor do
they spontaneously engage in
intentional actions or interacImaging imagination. Shown is part of the engraving, “The Physician tions with their environment,
Curing Fantasy,” by Mathaus Greuter (1564–1638).
two key properties of con-
www.sciencemag.org
SCIENCE
VOL 313
Published by AAAS
8 SEPTEMBER 2006
Downloaded from www.sciencemag.org on April 21, 2010
It would be an overgeneralization at this
point to conclude that plume theory must be
discarded, but the study by Hirano et al.
proves, beyond a reasonable doubt, that at
least one chain of hot-spot volcanoes is not
produced by a plume. Therefore, it is time to
critically reexamine the widespread interpretation of midplate volcanism in terms of mantle plumes and all the geodynamic, geophysi-
1395