ARTSCIENCE: THE ESSENTIAL CONNECTION
The Art and Science of Visualizing
Simulated Blood-Flow Dynamics
Dolores A. Hangan Steinman
and David A. Steinman
ABSTRACT
Looking at images across disciplines can help us
to think about the cross-fertilization that occurs
among the different kinds of visual media.
—Marita Sturken and Lisa Cartwright [1]
Beginning with Vesalius in the 16th century and continuing
through the 19th century, anatomists and physiologists relied
on the skill of artists to represent the unseen body. With the
advent of photography an important step was taken in terms
of scientific visual recording. Toward the end of the 19th century came the rudimentary forerunner of the moving pictures
used by Marey and Muybridge exclusively for scientific recordings [2]. Later, computer animation became an important tool
used to visualize chemical structures (such as the DNA helix)
and bonds, because it not only recorded but also modeled
changes in conditions and the local environment. Currently,
digital representations simulating real-life radiographs or angiograms are becoming commonplace. This is a result of the
medical profession’s increasing interest in developing less invasive visualization techniques, as well as a growing familiarity with such images among the general public.
Dolores A. Hangan Steinman (researcher), Imaging Research Laboratories,
Robarts Research Institute, PO Box 5015, London, Ontario, Canada N6A 5K8.
E-mail: <dhangan@imaging.robarts.ca>.
David A. Steinman (researcher), Department of Mechanical & Industrial Engineering,
University of Toronto, 5 King’s College Road, Toronto, Ontario, Canada M5S 3G8.
E-mail: <steinman@mie.utoronto.ca>.
T
he increasing use of
Over the centuries, efforts to recomputer enhancement and
veal the unseen body have taken the
simulation to reveal the unseen
physician from direct visualization
human body brings with it
challenges, opportunities and
(dissection/vivisection), to the reresponsibilities at the interface
cording of both macroscopic and
of art and science. Here they
microscopic images through cineare presented and discussed
matography, to the use of contrast/
in the context of efforts
radioactive agents and, finally, to
to understand the role of
blood-flow dynamics in
the notion of “computational imvascular disease.
aging,” whereby computers are used
to enhance or even simulate real-life
conditions [3]. It is interesting to
note that what initially advanced scientific visualization techniques was the clinician’s need to document any observed pathological change, along with the
laboratory researcher’s need to register and catalog observations. Today the goal is to translate data obtained in the laboratory or through clinical techniques into information easily
interpreted by the clinician. Thus the computer-generated images created by scientists working in this field function as a
metalanguage comprehensible to researchers and clinicians
alike (that is, to specialists who have, over the years, developed
a separate vernacular).
The phenomenon of interest to us is the relationship between blood-flow dynamics and vascular disease. Our task is to
establish ways to convert this complex phenomenon into a
mathematical model using computational fluid dynamics
(CFD) simulation techniques. With the simulation thus cre-
Fig. 1. (a) A typical textbook presentation of blood-flow patterns in an aneurysm [27], in which blood is shown to enter the aneurysm
at the distal neck (labeled “d”) and to exit through the proximal neck (labeled “p”). (Figure 6.11 from G. Hademenos and T. Massoud,
The Physics of Cerebrovascular Diseases [Heidelberg, Germany: Springer, 1998] p. 198. With kind permission of Springer Science and
Business Media.) (b) A still frame from a visualization of a canine vein pouch aneurysm simulation, which demonstrates blood-flow
dynamics broadly similar to that in (a). (© David Steinman) (c) A different perspective from the same visualization, now demonstrating
very different 3D blood-flow dynamics. Note: Dark streaks emphasize the fast-moving blood “particles”; lighter streaks highlight
increasingly slower flow. (© David Steinman)
©2007 ISAST
LEONARDO, Vol. 40, No. 1, pp. 71–76, 2007
71
ARTSCIENCE: THE ESSENTIAL CONNECTION
ated, we generate visual representations
of the CFD data intended to be both accurate and aesthetically pleasing. In the
process, we pay special attention to the
means by which this process can be manipulated or is manipulating the viewer
by altering his or her perception. So, imaging scientists face the task of taking
an intricate, multifaceted reality and presenting images of it that ultimately may
or may not alter the perception of that
reality.
PROGRESSION
OF REPRESENTATION
No serious aeronautical engineer
today would consider advancing
a new aircraft design without
extensive computational testing
and optimization. The potential
of CFD to play a similar role in
cardiovascular intervention is
very high.
—R.W. Metcalfe [4]
The study of hemodynamics (blood flow)
is important because of its existing relationship to vascular diseases such as aneurysms (the ballooning of blood vessels)
and atherosclerotic plaques (buildup of
fatty material in vessel walls), abnormalities that tend to develop at sites where
hemodynamics are most complex. A veritable menagerie of hemodynamic factors (e.g. wall shear stress, wall shear
stress gradients and blood particle residence times) has been implicated in
the development and progression of
these vascular diseases. Hemodynamic
forces likely play a role in the rupture of
a plaque or aneurysm, an event that
usually precipitates a heart attack or
stroke. It may be important for the clinician to consider hemodynamic information when deciding upon treatment, for
it has been seen in clinical practice that
judicious control of hemodynamic factors and positive “sculpting” of the hemodynamic environment can affect the
outcome of minimally invasive procedures employing balloons, stents or coils,
as well as surgeries such as aneurysm clipping, bypass grafting or plaque excision.
Although the putative links between vascular disease and hemodynamic forces
are almost universally accepted, they remain poorly understood. This may be attributed to a simple but unfortunate fact:
The hemodynamic factors of interest are
extremely difficult to measure in patients
Fig. 2. Detailed portrayal of blood-flow patterns on three orthogonal planes through a
patient-specific aneurysm simulation, shown at four different points during the cardiac cycle
(a–d). (Image from D.A. Steinman et al., “Image-Based Computational Simulation of Flow
Dynamics in a Giant Intracranial Aneurysm,” American Journal of Neuroradiology 24 [2003]
pp. 559–566. © American Society of Neuroradiology [ASNR]) Grayscales indicate the speed
of blood ( |V| , in cm/sec) through the planes; vectors indicate the magnitude and direction
of blood flowing within the plane; and white circles identify the centers of flow vortices [28].
72
Steinman and Steinman, Visualizing Blood-Flow Dynamics
using clinical medical imaging techniques. Instead, much of what we understand has been elucidated over the past
four decades by means of increasingly refined models.
In art, visual representations of the
human body have changed over the millennia. From the initial drawings and
carvings on stone (e.g. the Altamira
drawings and the Red Desert and Viking
petroglyphs, essentially described today
as “stick-people”), through the more
“evolved” representations encountered
on the walls of Mayan and Egyptian
tombs and Pompeii villas, for example,
to the “perfection” of Leonardo’s depictions, humans have faced a number of
challenges owing to changes in their ability to process the information acquired,
as well as in the skills required for representing this information. It is very
important to understand that, despite
the fact that the human body has not
changed radically, both its perception
and its portrayal are continuously changing [5]. Similarly, in the field of bloodflow simulation, mathematical models
and their visual representations have
evolved, paralleling the availability of increasingly sophisticated tools, skills and
knowledge.
Until about the 1950s, our understanding of blood-flow dynamics was limited to relatively simplistic mathematical
models, which assumed that blood vessels
were infinitely long, straight tubes and
that the blood moved steadily rather than
pulsating in response to heart contractions. Refinement of these mathematical
models to include the effects of blood
pulsatility led, in the 1960s, to the realization that the flow of blood in arterial
circuits follows the same general rules
that govern the flow of current through
electrical circuits. This led to the use of
real or simulated electrical circuits to predict, for example, how alterations to the
vascular trees (by disease) might affect
clinically measurable quantities such as
blood pressure waveforms.
Following the suggestion in the late
1960s that the complex forces exerted by
flowing blood near arterial branches and
bends might explain why atherosclerotic
plaques are usually found near these sites,
engineers began to translate their considerable skills in visualizing aerodynamic
and industrial flows to the challenge of
visualizing arterial blood flow. Often,
however, the branching or curved tube
models were necessarily idealized, sporting circular cross-sections and convenient planes of symmetry that facilitated
the construction of the physical models.
With the introduction of relatively in-
THE ART OF IMAGE-BASED
CFD SIMULATION
Cubism undertook a completely
two-dimensional transcription of
the three-dimensional phenomena.
—Clement Greenberg [8]
The process of creating clear and visually
pleasing representations is not a trivial
one, since our images need to portray
events that evolve in both time and space.
Artists have the option of expressing
these events in a manner that, for our
scientific purposes, lacks pedagogic clarity. If we take the case of Duchamp’s
series of Nude(s) Descending a Staircase
(1913), the difficulties of imaging a body
in motion in an accurate yet aesthetically
acceptable way are made clear. Unfortunately, despite Duchamp’s reference to
Marey’s scientific images, we cannot use
such depiction of time and motion. In
addressing the differences and similarities between science and art we walk a
very fine line between scientific rigor
(when defining and explaining concepts)
and flexibility of interpretation (when applying them to individual patients) [9].
Pedagogical simplicity is inevitably
achieved at some cost of verisimilitude.
—F.C. Holmes [10]
In our quest for clarity we strive to maintain the balance of making our data accessible without trivializing our efforts
[11]. Correspondingly, we must avoid
simplification to the detriment of accuracy. For example, in the classical repre-
of symmetry implicit in idealized models,
the choice of cut planes to render 3D flow
dynamics in 2D becomes somewhat arbitrary and capricious. Furthermore, time
as depicted here as a series of individual frames seemingly divorced from one
another lacks continuity. This apparent
wealth of information thus comes at the
expense of communication and understanding.
The problem of representing 3D motion in 2D print is of course not a new
one. As we have already mentioned, artists have struggled with this problem for
centuries. The need for clarity in this
field limits our choices; therefore we
strive to exploit representations of blood
flow that are familiar to the clinician. For
the case depicted in Color Plate G No. 1,
our solution was to consciously mimic the
experimental slipstream visualization
techniques [12] promoted by Charles
Kerber, an eminent neurosurgeon with a
keen interest in the study of blood-flow
Despite the fact that the human body has
not changed radically, both its perception
and its portrayal are continuously changing.
sentation of an aneurysm shown in Fig.
1a, blood is shown to enter the aneurysm
from the right (distal) end, swirl clockwise and exit through the left (proximal)
end. In still frames from a visualization
of one of our first patient-specific CFD
simulations we can see blood-flow patterns superficially similar to the idealized
case (Fig. 1b). However, when viewed
from a different perspective (Fig. 1c),
the blood is seen to enter the aneurysm
through both the proximal and distal
ends. Detailed knowledge of how blood
moves in and out of the aneurysm is crucial in selecting a treatment, and, as our
slightly more realistic representation
shows, this movement is more complicated than the simple pedagogical representation suggests.
Anatomic realism presents challenges
for displaying information, as accuracy
and rigor are not always guarantors of
clarity. As Fig. 2 demonstrates, the traditional engineering representation of flow
using flooded contour and vector field
plots is clear to scientists and perhaps aesthetically appealing to the layperson, but
not at all clear to medical students, physicians, technologists or the public at large.
In the absence of the convenient planes
dynamics. The 3D context now becomes
much clearer, and, although we have lost
the detailed depiction of the complex
vortical flow structures identified in Fig.
2, the complexity of flow in the aneurysm
is still evident in the mixing of the differently colored slipstreams. The temporal evolution of flow, depicted here in a
manner that echoes spontaneously the
classical Muybridgian approach, also becomes much clearer.
The scientist and the artist, each
at his own work, are in some real
sense pursuing the same aim and
by methods having more in common
than is usually admitted.
—Martin Johnson [13]
When challenged by the tension between
the novel concept and novel representation, we have chosen to promote the concept of a very complex patterned flow by
adopting traditional conventions and
adapting them to our needs. For example, when painters were hired to draw
anatomical atlases, the convention they
were supposed to obey was “red from the
heart, blue to the heart.” As illustrated in
Color Plate G No. 2a, we exploit this con-
Steinman and Steinman, Visualizing Blood-Flow Dynamics
73
ARTSCIENCE: THE ESSENTIAL CONNECTION
expensive desktop workstations in the
1980s, engineers and physicists began
turning toward CFD [6]. Paralleling the
evolution of physical models, early CFD
simulations still typically exploited simple geometries and planes of symmetry,
the former being easy to generate using
computer-aided design tools and the latter allowing only half of the model to be
simulated, with the other half merely represented as its mirror image.
Advances in the 1990s in high-resolution three-dimensional medical imaging
techniques such as magnetic resonance
imaging (MRI), computed tomography
(CT) and ultrasound (US) made it possible non-invasively to study vascular
structures with unprecedented accuracy
and detail. Still, concomitant accuracy
and detail in the measurement of bloodflow dynamics in vivo remained—and remains—elusive. Coupled with advances
in medical imaging processing, this decade has seen an explosion in imagebased CFD simulations whereby vessel
geometries derived from in vivo medical images are used to construct patientspecific CFD models for the purposes of
more clearly understanding the role of
blood-flow dynamics in the development
of vascular disease, as well as for planning
and optimizing vascular interventions.
The enormous amount of bloodvelocity data generated by even a single
image-based CFD simulation (often on
the order of gigabytes), the absence of
convenient planes of symmetry and the
need to present these simulations in a
clinical setting raise important conceptual and practical issues in their display
[7]. The resolution of these issues at the
interface of art and science is presented
and discussed below, using examples from
our research into elucidating the role of
hemodynamic forces in treatment and
prevention of cerebrovascular disease.
ARTSCIENCE: THE ESSENTIAL CONNECTION
Fig. 3. (a) Sequential frames in which a patient-specific CFD simulation has undergone
a “virtual angiogram,” whereby a dye or contrast agent has been virtually injected into
the CFD model, which at the same time undergoes virtual X-ray imaging to highlight the
blood vessel. (b) The corresponding frames from the clinical angiograms acquired for
this patient [29]. (© IEEE)
vention as a means of conveying to the
viewer the presence of retrograde flow
within the portion of the artery where
plaques tend to occur. Another convention we have exploited is that of color
Doppler ultrasound, a medical imaging
technique that traditionally uses cool
hues to depict reverse flow and warm
hues to depict forward flow. The use of
this color convention (Color Plate G No.
2b) provides a tentative link between the
rich detail of the blood-flow dynamics
simulated for this particular patient and
the relatively crude, but still clinically relevant, depiction offered by the clinical
ultrasound examination of the same patient.
In Fig. 3 we go one step further in
exploiting the links between computergenerated and routine clinical representations by “virtually” imaging a
patient-specific CFD model in the same
way that the patient was imaged clinically.
An image of a radiograph is very clearly
associated with the “idea” of a radiograph
in the minds of the public and experts
alike, and the use of “virtual imaging” is
meant to create the same intuitive connection. This particular convention has
the added bonus of allowing us to know
whether we are telling the truth with
CFD: Absent the ability to accurately
74
measure detailed blood-flow dynamics directly in vivo, indirect comparisons of the
real and virtually imaged patient are currently the only means available to us to
prove that our computer simulations reflect reality, at least as embodied in a clinical examination.
Taken together, these examples serve
to highlight the pursuit of aesthetic
agreement in the depiction of these
patient-specific CFD simulations, and the
consistency with traditional medical imaging expected by the clinicians and lay
people who ultimately “consume” these
images.
ADVANTAGES OF
COMPUTER-GENERATED
VISUAL REPRESENTATIONS
Science reorganized its conception of
the living body . . . to reflect the body’s
new status as a mobile, living system.
—Lisa Cartwright [14]
As we have shown, the use of computergenerated images derived from CFD simulations obviously has its advantages and
is definitely a step forward in the creation
of optimal “windows” into the unseen
human body. The principal advantage is
Steinman and Steinman, Visualizing Blood-Flow Dynamics
that they decipher the biological code
and visualize the invisible, as opposed to
merely mirroring reality. From strict observation of the physiological phenomenon (“noting everything normal and
regular” [15]), the translation is made to
the experiment (defined as the “variation or disturbance that an investigator
brings into the conditions of natural phenomena” [16]).
Another key advantage of these computer-generated images is the appearance of the organ as a whole (in the
context of its interactions with the rest of
the organism). A better understanding
comes from both the visualization of the
organ/tissue, as well as of the bodily context in which it is now placed. The blood,
for instance, and its flow patterns can neither be recorded nor regarded independently from the surrounding tissues
and organs.
In addition, computer simulations allow us to anticipate changes within tissues or organs without the restrictions
imposed by in vivo data collection, which
is usually limited by anatomical constraints [17]. Also important is the fact
that physiological processes can thus be
simulated without harming either human beings or experimental animals.
CHALLENGES FACED
BY THE IMAGER
Where cinematography enhanced or occasionally replaced sensory acuity, computer enhancement or simulation goes
further by replacing the recorded image
with its own creation. As alluded to
above, in the desire to be understood by
a wide audience and concomitantly make
the information available for both training and diagnosis/treatment purposes,
we are faced with challenges within two
essential issues: (a) scientific truth and
clarity; and (b) aesthetics and conventions.
Paul Hartal has written, “The painted
image has such an enormous power over
the observer that it can fool humans and
animals alike” [18]. In our society, where
images are used ubiquitously for their
ability to manipulate and alter perceptions, the generation of images for educational and diagnostic purposes bears
great responsibility. Scientists, and by extension their representations of medical
data, are in an exceptional position owing to the public’s implicit trust of their
profession [19]. In our field, data may
be used to make life-or-death decisions,
and so there is a real danger if the numerical data underlying a compelling
representation is false or incorrect. In
connection to the actual data collected
from the patient [25]. While they are occasionally intrinsic to the phenomenon
studied, we need to avoid the spectacular and carefully balance the shock value
that some images might carry. In the
1940s a popular scientific movie was the
recording by Goodhart and Balser of the
“woman-Oedipus,” an encephalitis patient who had, as a result of her ailment,
pulled out her own teeth and eyes. This
grim footage was acquired within a few
decades of Edison’s Manufacturing Company’s popular 1903 recording of an elephant’s death through electrocution. The
strong interest the public manifested
In the biomedical sciences, it is
becoming increasingly easy to create
powerful images that can trick the
observer into mistaking them for reality.
Aware of this power, one needs to be
very careful when dealing with these images and the concepts they represent. Science, whether its practitioners will it or
not, is in this regard akin to religion. Like
religious imagery, the scientific and medical images herewith promote a new set
of theories, principles, ideas and values.
In Marx’s terms, religion is “the opiate of
the masses” who are “coerced to mindlessly buy into a belief system” that is
not necessarily innate to them [21]. Over
the centuries, religious art was used to
convey religious myths and church doctrines, as well as to express abstract concepts while depicting seemingly accurate renditions of the world. Today mass
media are used to “disseminate the same
texts, images, and sounds to millions of
citizens—thus assuring the same ideological beliefs” [22]. One can therefore
ask, Is science a new type of religion, in
that it addresses the masses and tries to
“win them over”? Images thus become
of vital importance in this process of
accessing the imaginary/abstract realm.
We therefore need to be extremely careful not to turn our representations into
a mere propaganda of this new ideology [23].
Culture is a process, not a fixed
set of practices or interpretations.
—Sturken and Cartwright [24]
Another major challenge is to avoid creating eye-catching images that have little
in these cases (whether attributable to
curiosity, indirect participation in an
explicit ritual of death or simply the
shock value of repulsion) is at one end
of the scale. At the other extreme are
the weaker responses elicited from the
public to, say, images of tissue obtained
through electron microscopy, or a fixed
and stained tissue slide. The images we
generate and present to the public are
not repulsive or meant to incite morbid
curiosity. They are meant to allow both
the physician and the public to pierce
through the body’s carapace into the unseen and untouched using aesthetically
refined means. We do take into account
the visual satisfaction of the public while
using computer-generated images as substitutes for the human body (with the
goal of replacing invasive in vivo experiments).
It is difficult to generate images “acceptable” to an image-savvy audience.
The public is fascinated with both the
physiological process and the technology
used to explore it. Bombarded with visual
stimuli, the public also has the ability to
discern between professional and amateur images. This sophistication results
from exposure to science popularization
programs, which provide a false sense
of knowledge that deceptively seems to
empower the public to decide upon the
actual “reality” of the screened image. Additionally, cultural and aesthetic conventions, which influence the awareness and
appreciation of the viewer, continually
develop and change [26]. The process of
“computational” imaging thus becomes
a process of interpreting reality through
the creation of easily accessible images
following prevailing cultural trends and
consistent with the traditional medical
images that the public is used to seeing.
One additional risk is the extensive democratization of images without consideration for the actual data and message
that need to be transmitted. The images
generated by the computer but controlled by the scientist/researcher are following cultural trends while stimulating
both understanding and acceptance. Despite the perceived need for popular
recognition, scientists do confront and
educate the public and do not aim for the
lowest common denominator.
In summary, our computer-generated
simulations and subsequent representations must be: (a) an accurate and clear
reflection of the anatomical and physiological reality, as indicated by the raw
data; (b) in accordance with established
visual conventions in medical representation; and (c) in tune with both the style
of the day and the aesthetic demands and
visual background of the reader/viewer.
Acknowledgments
We thank Christine Sprengler (Department of Visual
Arts, University of Western Ontario) for advice and
encouragement, as well as the staff, students, collaborators and patients who have contributed to the examples presented here. This work was supported
by a grant from the Canadian Institutes of Health
Research (MOP-62934) and a Career Investigator
Award from the Heart & Stroke Foundation of Ontario. Finally, we thank Catherine Dhavernas and
Agnès Conacher (Department of French Studies,
Queen’s University) for getting us started.
References and Notes
1. M. Sturken and L. Cartwright, Practices of Looking:
An Introduction to Visual Culture (Oxford, U.K.: Oxford Univ. Press, 2002) p. 3.
2. The “moving pictures,” originally conceived by
Marey and Muybridge as a scientific tool, were borrowed by entertainers, whereas today advances in visualization are often dictated by the amusement
world of computer games and later used by scientists.
Also notable, as suggested by Cartwright, is the competition between the interest in the apparatus and
the interest in the image itself. L. Cartwright, Screening the Body: Tracing Medicine’s Visual Culture (Minneapolis, MN: University of Minnesota Press, 1997)
p. 81.
3. Computational imaging covers a broad spectrum,
from enhancement of medical images themselves to
computer simulations based on data derived from
those images. Our work focuses on the latter; however, the broader issues discussed in this paper apply
equally well to the former.
4. R.W. Metcalfe, “The Promise of Computational
Fluid Dynamics as a Tool for Delineating Therapeutic Options in the Treatment of Aneurysms,” American Journal of Neuroradiology 24 (2003) pp. 553–554.
5. It is equally important to recognize that visual representations have continued to change over the last
two centuries, owing in part to ever-evolving aesthetic
grounds (e.g. compare body representations of
Steinman and Steinman, Visualizing Blood-Flow Dynamics
75
ARTSCIENCE: THE ESSENTIAL CONNECTION
the biomedical sciences, it is becoming
increasingly easy to create powerful images that can trick the observer into
mistaking them for reality. Such misrepresentations may happen inadvertently;
however, as a facetious NASA technical
report pointed out, it is not difficult to
purposely divert public perception from
the actual data [20]. Our images, thus,
burden us with broader responsibility
because they carry two levels of interpretation: (a) our interpretation of the
raw data followed by the creation of a
simulation; and (b) the “reader”’s interpretation of the resulting visual representation.
ARTSCIENCE: THE ESSENTIAL CONNECTION
Moore versus Rodin, Picasso versus Modigliani or
Giacometti versus Botero).
6. Except for relatively trivial cases, equations governing fluid flow cannot be solved by hand. Computational fluid dynamics circumvents this by
dividing a complex domain into thousands of smaller
and simpler elemental shapes, each of which is then
governed by a simpler equation. However, because
these elements are connected to one another, thousands or even millions of equations must be solved
together, a feat achievable in practice only by the use
of computers.
7. “[It] is important to remember Linus Pauling’s
admonition that ‘modelling is a form of thinking’
and Rudolf Arnheim’s equally important insight,
that ‘artistic activity is a form of reasoning.’” R.S.
Root-Bernstein, “The Sciences and Arts Share a Common Creative Aesthetic,” in Alfred Tauber, ed., The
Elusive Synthesis: Aesthetics and Science (Dordrecht,
the Netherlands: Kluwer Academic Publishers, 1996)
p. 74.
8. C. Greenberg, Art and Culture: Critical Essays
(Boston: Beacon Press, 1961) p. 172.
9. “The interesting thing about art is that it doesn’t
have to have a utilitarian value. . . . [Artists] can indulge those things that are not rigorous.” M. Bond,
“The Art of Knowledge: An Interview with Keith
Tyson, Winner of Last Year’s Turner Prize,” New Scientist (20 September 2003) pp. 44–47.
10. F.C. Holmes, “Beautiful Experiments in the Life
Sciences,” in Tauber [7] pp. 83–102.
11. “Victor Weisskopf has described as a ‘destructive
element’ within the community of science the low
esteem in which clear and understandable presentation is held.” Root-Bernstein [7] p. 72.
12. In this in vitro approach, colored dyes having the
same density as the fluid medium are injected into
an optically transparent vascular model and filmed
as their trajectories evolve.
13. M. Johnson, Art and Scientific Thought; Historical
Studies Towards a Modern Revision of Their Antagonism
(New York: AMS Press, 1970) p. 23.
14. Cartwright [2] p. 82.
15. Cartwright [2] p. 26.
16. Cartwright [2] p. 26. It is important to appreciate the difference between the perception of scientific cinematic recording as “documentary” (i.e.
76
involving no manipulation) versus scientific moving
images created on the computer (i.e. realistic simulacra derived from real-life data having undergone
two conversions: from data collected from the individual patient to numerical data [binary code]; and
then from the numerical data into images). The concern we face is: Which one of the two is more accurate, reliable or closer to reality? The former is a
recording of the organ (albeit through the angle of
the “recorder”), while the latter involves the processing of objective data by a non-biased machine.
Hence, we could say that the issues at stake are: (a)
What is the message the “recorder” or “creator” of
images is trying to convey? (b) Who is the target audience? and (c) What is the final use of these images?
27. From G.J. Hademenos and T.F. Massoud, The
Physics of Cerebrovascular Diseases (New York: AIP Press,
1998).
28. From D.A. Steinman et al., “Image-Based Computational Simulation of Flow Dynamics in a Giant
Intracranial Aneurysm,” American Journal of Neuroradiology 24 (2003) pp. 559–566.
29. From M.D. Ford et al., “Virtual Angiography for
Visualization and Validation of Computational Fluid
Dynamics Models of Aneurysm Hemodynamics,”
IEEE Transactions on Medical Imaging 24 (2005) pp.
1586–1592.
30. From Steinman et al. [28].
17. That is, collection of organs/body parts is limited
to suit the practical needs of the microscopist: frog
legs, rabbit ears, chick embryo chorioallantoic membrane, mouse mesentery, etc.
31. From D.A. Steinman et al., “Reconstruction of
Carotid Bifurcation Hemodynamics and Wall Thickness Using Computational Fluid Dynamics and MRI,”
Magnetic Resonance in Medicine 47 (2002) pp. 149–159.
18. P. Hartal, The Brush and the Compass: The Interface
Dynamics of Art and Science (Lanham, MD: University
Press of America, 1998) p. 99.
32. From D.A. Vorp et al., “Computational Modeling
of Arterial Biomechanics,” Computers in Science and
Engineering 3 (2001) pp. 51–62.
19. Harris Poll #63, 27 November 2002.
20. A. Globus and E. Raible, “13 Ways to Say Nothing
with Scientific Visualization,” NASA Ames Research
Center Report RNR-92-006 (25 February 1992).
21. Sturken and Cartwright [1] p. 51.
22. L. Manovich, The Language of New Media (Cambridge, MA: MIT Press, 2002) p. 22.
23. Propaganda here refers to “false representations
to lure people into holding beliefs that may compromise their own interest.” Sturken and Cartwright
[1] p. 21.
24. Sturken and Cartwright [1] p. 4.
25. “There is a self-conscious limit to the artistic foray,
however, for the images addressed to a scientific audience are, in a sense, conservative: whereas colour
embellishes the dramatic effects used for popular
audiences, journal articles and professional presentations largely eschew such bold images and less
dramatic, monochromatic pictures are used.” Tauber
[7] p. 4.
26. Following discussions with our colleagues, we
noted a tension between two perceptions: For some,
beauty in science is equated with simplicity, harmony
and a certain quality of cleanness; others identify it
in science’s complex, unpredictable and imperfect
aspects.
Steinman and Steinman, Visualizing Blood-Flow Dynamics
Manuscript received 31 August 2005.
Dolores Steinman was trained as a pediatrician and went on to doctoral and postdoctoral
research in cancer cell biology. An accomplished photographer, she is particularly interested in the role of the arts in medicine and
devotes her efforts to bringing together the local scientific and artistic communities. Dr.
Steinman worked as a research associate at the
Robarts Research Institute and is currently a
Volunteer Art Docent at Museum London.
David Steinman spent the last decade as a scientist at the Robarts Research Institute, working to integrate the fields of computational
modeling and medical imaging. Necessity being the mother of invention, he spent much
of this time thinking of ways to present complex data to colleagues, collaborators and the
community at large. Dr. Steinman is currently
an associate professor of mechanical and
biomedical engineering at the University of
Toronto.
COLOR PLATE G
No. 1. Selected frames from a “virtual slipstream” visualization of blood-flow dynamics from the aneurysm
simulation depicted in Fig. 2. (Image from D.A. Steinman et al., “Image-Based Computational Simulation
of Flow Dynamics in a Giant Intracranial Aneurysm,” American Journal of Neuroradiology 24 [2003] pp. 559–566.
© American Society of Neuroradiology [ASNR].) The top and bottom rows show the aneurysm in lateral and
anterior-posterior views, chosen to match the view typically acquired during clinical angiography (see Reference
[30] in article by Dolores A. Hangan Steinman and David A. Steinman).
No. 2. Two different conventions used to
visualize blood-flow dynamics from anatomically realistic carotid bifurcation CFD simulations. (a) Virtual particles streaking through
the computed velocity field, color-coded
according to their speed (shown in units of
cm/sec according to the legend at upper right)
(see Reference [31] in article by Dolores A.
Hangan Steinman and David A. Steinman).
(© John Wiley and Sons) (b) Visualized axial
velocity component on selected planes through
another model, using a color scheme consciously meant to mimic that of color Doppler
ultrasound (see Reference [32] in Steinman
and Steinman). (© IEEE)
49