Microswimmer
Part of a series on |
Microbial and microbot movement |
---|
Microswimmers |
Molecular motors |
A microswimmer is a microscopic object with the ability to move in a fluid environment.
Given the recent nature of the field, there is yet no consensus in the literature for the nomenclature of the microscopic objects this article refers to as "microswimmers". Among the many alternative names such objects are given in the literature, microswimmers, micro/nanorobots and micro/nanomotors are likely the most frequently encountered. Other common terms may be more descriptive, including information about the object shape, e.g., microtube or microhelix, its components, e.g., biohybrid, spermbot,[2] bacteriabot,[3] or micro-bio-robot,[4] or behavior, e.g., microrocket, microbullet, microtool or microroller. Researchers have also named their specific microswimmers e.g., medibots,[5] hairbots,[6] iMushbots,[7] IRONSperm,[8] teabots,[9] biobots,[10] T-budbots,[11] or MOFBOTS.[12][1]
Background
In 1828, the British biologist
Ever since Newton established his equations of motion, the mystery of motion on the microscale has emerged frequently in scientific history, as famously demonstrated by a couple of articles that should be discussed briefly. First, an essential concept, popularized by
Here, ρ represents the density of the fluid; u is a characteristic velocity of the system (for instance, the velocity of a swimming particle); l is a characteristic length scale (e.g., the swimmer size); and μ is the viscosity of the fluid. Taking the suspending fluid to be water, and using experimentally observed values for u, one can determine that inertia is important for macroscopic swimmers like fish (Re = 100), while viscosity dominates the motion of microscale swimmers like bacteria (Re = 10−4).[15]
The overwhelming importance of viscosity for swimming at the micrometer scale has profound implications for swimming strategy. This has been discussed memorably by
Microorganisms have optimized their
where is the velocity of the fluid and is the gradient of the pressure. As Purcell noted, the resulting equation — the Stokes equation — contains no explicit time dependence.[16] This has some important consequences for how a suspended body (e.g., a bacterium) can swim through periodic mechanical motions or deformations (e.g., of a flagellum). First, the rate of motion is practically irrelevant for the motion of the microswimmer and of the surrounding fluid: changing the rate of motion will change the scale of the velocities of the fluid and of the microswimmer, but it will not change the pattern of fluid flow. Secondly, reversing the direction of mechanical motion will simply reverse all velocities in the system. These properties of the Stokes equation severely restrict the range of feasible swimming strategies.[16][15]
As a concrete illustration, consider a mathematical scallop that consists of two rigid pieces connected by a hinge. Can the "scallop" swim by periodically opening and closing the hinge? No: regardless of how the cycle of opening and closing depends on time, the scallop will always return to its starting point at the end of the cycle. Here originated the striking quote: "Fast or slow, it exactly retraces its trajectory and it's back where it started".[16] In light of this scallop theorem, Purcell developed approaches concerning how artificial motion at the micro scale can be generated.[15] This paper continues to inspire ongoing scientific discussion; for example, recent work by the Fischer group from the Max Planck Institute for Intelligent Systems experimentally confirmed that the scallop principle is only valid for Newtonian fluids.[17][15]
Types
Different types of microswimmers are powered and actuated in different ways. Swimming strategies for individual microswimmers
Often, researchers choose to take inspiration from nature, either for the entire microswimmer design, or for achieving a desired propulsion type. For example, one of the first bioinspired microswimmers consisted of human
A different nature-inspired approach is the use of
Strategies for the fabrication of microswimmers include two-photon polymerisation 3D printing, photolithography, template-assisted electrodeposition, or bonding of a living component to an inanimate one by exploiting different strategies. More recent approaches exploit 4D printing, which is the 3D printing of stimuli-responsive materials.[49][50][51][52] Further functionalization is often required, either to enable a certain type of actuation, e.g., metal coating for magnetic control or thermoplasmonic responses, or as part of the application, if certain characteristics are required for e.g., sensing, cargo transport, controlled interactions with the environment, or biodegradation.[53][54][55][56][1]
Natural microswimmers
Motile systems have developed in the
Some of the smallest known natural motile systems are
Apart from motor proteins, enzymes, traditionally recognized for their catalytic functions in biochemical processes, can function as nanoscale machines that convert chemical energy into mechanical action at the molecular dimension. Diffusion of various enzymes (e.g. urease, and catalase), measured by fluorescent correlated spectroscopy (FCS), increases in a substrate-dependent manner.[63][64] Moreover, when enzymes are membrane-bound, their catalytic actions can drive lipid vesicle movement. For instance, lipid vesicles integrated with enzymes such as transmembrane adenosine 5’-triphosphatase, membrane-bound acid phosphatase, or urease exhibit enhanced mobility correlating with the enzymatic turnover rate.[65]
Bacteria can be roughly divided into two fundamentally different groups,
Specifically, for microorganisms that live in aqueous environments, locomotion refers to swimming, and hence the world is full of different classes of swimming microorganisms, such as bacteria,
The following table, based on Schwarz et al., 2017,[68] lists some examples of natural or biological microswimmers.
Name | Image | Size (μm2)a | Speed (μm/s)b | Propulsion mechanism | Natural swimming habitat | Sources | |
---|---|---|---|---|---|---|---|
bacterial swimmers (prokaryotes) |
Escherichia coli | 0.5 × 2 | 30 | Peritrichous bundles |
Intestinal flora | [69] | |
Serratia marcescens | 1 × 2 | 50 | Peritrichous bundles | Respiratory and urinary tracts (parasitic) | [70] | ||
Salmonella typhimurium |
0.5 × 2 | 30 | Peritrichous bundles | Intestines (parasitic) | [71] | ||
Bacillus subtilis | 1 × 3 | 20 | Peritrichous bundles | Intestinal flora | [72] | ||
Aliivibrio fischeri | 1 × 2 | 50 | Lophotrichous flagella |
Mucus (symbiotic) | [73] | ||
Vibrio alginolyticus | 2 × 3 | 40 | Monotrichous flagellum |
Blood (parasitic) | [74][75] | ||
Listeria monocytogenes | 0.5 × 1.5 | <1 | Peritrichous or amphitrichous bundles |
Inter- and intracellular (parasitic) | [76][77] | ||
Magnetococcus marinus | 2 × 2 | 200 | Two lophotrichous bundles | Marine water | [78][79] | ||
Magnetospirillum gryphiswaldense |
0.5 × 2 | 60 | Two amphitrichous flagella | Freshwater sediments | [80] | ||
Mycoplasma mobile | 0.5 × 0.5 | 5 | Gliding via protrusions | Fish gills (parasitic) | [81] | ||
Chlamydomonas | 10 × 10 | 150 | Two lophotrichous flagella | Freshwater, soil | [82] | ||
Tetrahymena | 25 × 50 | >500 | Holotrichous cilia | Freshwater | [83] | ||
Trypanosome | 3 × 20 | 30 | Monotrichous flagellum | Blood (parasitic) | [84][85] | ||
sperm cells |
Human |
3 × 5 | 50 | Monotrichous flagellum |
Reproductive tract | [86][87] | |
Bovine |
5 × 10 | 100 | Monotrichous flagellum | Reproductive tract | [87][88][89] | ||
Murine |
3 × 8 | 120 | Monotrichous flagellum | Reproductive tract | [86][88] |
Synthetic microswimmers
"An artificial microswimmer is a cutting-edge technology with engineering and medical applications. A natural microswimmer, such as bacteria and sperm cells, also play important roles in wide varieties of engineering, medical and biological phenomena. Due to the small size of the microswimmer, the inertial effect of the surrounding flow field may be negligible. In such a case, reciprocal body deformation cannot induce migration of a swimmer, which is known as the scallop theorem. To overcome the implications of the scallop theorem, the microswimmer needs to undergo a nonreciprocal body deformation to achieve migration. The swimming strategy is thus completely different from macro-scale swimmers...".[90]
One of the current engineering challenges is to create miniaturized functional vehicles that can carry out complex tasks at a small scale that would be otherwise impractical, inefficient, or outright impossible by conventional means. These vehicles are termed nano/micromotors or nano/microrobots, and should be distinguished from even smaller molecular machines for energy, computing, or other applications on the one side and static microelectromechanical systems (MEMS) on the other side of this size scale. Rather than being electronic devices on a chip,
There are a variety of sensing, actuating, or pickup-and-delivery applications that scientists are currently aiming for, with local drug targeting for cancer treatment being one of the more prominent examples.[102][5] For applications like this, a micromotor needs to be able to move, i.e., to swim, freely in three dimensions efficiently controlled and directed with a reliable mechanism.[68]
It is a direct consequence of the small size scale of microswimmers that they have a low
Microorganisms have adapted their locomotion to the harsh environment of low Reynolds number regime by invoking different swimming strategy.[106] For example, the E. coli moves by rotating its helical flagellum,[107][108] Chlamydomonas flagella have a breaststroke kind of motion.[109] African trypanosome has a helical flagellum attached to the cell body with a planar wave passing through it.[110][111] Swimming of these kind of natural swimmers have been investigated for the last half-century.[112] As a result of these studies, artificial swimmers have also been proposed, like Taylor sheet,[113] Purcell's two-hinge swimmer,[16][114] three-linked spheres swimmer,[115][116][117] elastic two-sphere swimmer [118] and three-sphere with a passive elastic arm,[119] which have further enhanced understanding about low Reynolds number swimmers. One of the challenges in proposing an artificial swimmer lies in the fact that the proposed movement stroke should not be reciprocal otherwise it cannot propel itself due to the Scallop theorem. In Scallop theorem, Purcell had argued that a swimmer with one-hinge or one degree of freedom is bound to perform reciprocal motion and thus will not be able to swim in the Stokes regime.[106][16][112]
Purcell proposed two possible ways to elude from Scallop theorem, one is 'corkscrew' motion [107][104] and the other is 'flexible oar' motion.[120][121] Using the concept of flexible oar, Dreyfus et al reported a micro swimmer that exploit elastic property of a slender filament made up of paramagnetic beads.[31] To break the time inversion symmetry, a passive head was attached to the flexible arm. The passive head reduces the velocity of the flexible swimmer, bigger the head, higher is the drag force experienced by the swimmer. The head is essential for swimming because without it the tail performs a reciprocal motion and the velocity of the swimmer reduces to zero.[122][112]
Another way microswimmers can propel is through catalytic reactions. Taking inspiration from Whitesides, who used the decomposition of hydrogen peroxide (H2O2) to propel cm/mm-scale objects on a water surface,[123] Sen et al. (2004) fabricated catalytic motors in the micrometer range.[92] These microswimmers were rod-shaped particles 370 nm in diameter and consisted of 1 µm long Pt and Au segments. They propelled via the decomposition of hydrogen peroxide in solution which would be catalyzed into water and oxygen. The Pt/Au rods were able to consistently reach speeds of up to 8 µm/s in a solution of 3.3% hydrogen peroxide. The decomposition of hydrogen peroxide in the Pt side produces oxygen, two protons and two electrons. The two protons and electrons will travel towards the Au, where they will be used to react with another hydrogen peroxide molecule, to produce two water molecules. The movements of the two protons and the two electrons through the rod drag the fluid towards the Au side, thus this fluid flow will propel the rod in the opposite direction. This self-electrophoresis mechanism is what powers the motion of these rods.[93] Further analysis of the Pt/Au rods showed that they were capable of performing chemotaxis towards higher hydrogen peroxide concentrations,[94] transport cargo,[95] and exhibited steerable motion in an external magnetic field when inner Ni segments were added.[95]
Responding to stimuli
Reconfigurable synthetic or artificial microswimmers need internal feedback[125] Self-propelling microparticles are often proposed as synthetic models for biological microswimmers, yet they lack the internally regulated adaptation of their biological counterparts. Conversely, adaptation can be encoded in larger-scale soft-robotic devices but remains elusive to transfer to the colloidal scale.[125]
The ubiquity and success of motile bacteria are strongly coupled to their ability to autonomously adapt to different environments as they can reconfigure their shape, metabolism, and motility via internal feedback mechanisms.[126][127] Realizing artificial microswimmers with similar adaptation capabilities and autonomous behavior might substantially impact technologies ranging from optimal transport to sensing and microrobotics.[128] Focusing on adaptation, existing approaches at the colloidal scale mostly rely on external feedback, either to regulate motility via the spatiotemporal modulation of the propulsion velocity and direction [129][124][130][131] or to induce shape changes via the same magnetic or electric fields,[132][133][134] which are also driving the particles. On the contrary, endowing artificial microswimmers with an internal feedback mechanism, which regulates motility in response to stimuli that are decoupled from the source of propulsion, remains an elusive task.[125]
A promising route to achieve this goal is to exploit the coupling between particle shape and motility. Efficient switching between different propulsion states can, for instance, be reached by the spontaneous aggregation of symmetry-breaking active clusters of varying geometry,[135][136][137][138] albeit this process does not have the desired deterministic control. Conversely, designing colloidal clusters with fixed shapes and compositions offers fine control on motility [139][140][141] but lacks adaptation. Although progress on reconfigurable robots at the sub-millimeter scale has been made,[142][143][144][145][146] downscaling these concepts to the colloidal level demands alternative fabrication and design. Shape-shifting colloidal clusters reconfiguring along a predefined pathway in response to local stimuli [147] would combine both characteristics, with high potential toward the vision of realising adaptive artificial microswimmers.[125]
Biohybrid microswimmers
The so-called biohybrid microswimmer can be defined as a microswimmer that consist of both biological and artificial parts, for instance, one or several living microorganisms attached to one or various synthetic parts. The biohybrid approach directly employs living microorganisms to be a main component or modified base of a functional microswimmer.[150][151] Initially microorganisms were used as the motor units for artificial devices, but in recent years this role has been extended and modified toward other functionalities that take advantage of the biological capabilities of these organisms considering their means of interacting with other cells and living matter, specifically for applications inside the human body like drug delivery or fertilisation.[152][153][68]
A distinct advantage of microorganisms is that they naturally integrate motility and various biological functions in a conveniently miniaturised package, coupled with autonomous sensing and decision-making capabilities. They are able to adapt and thrive in complex in vivo environments and are capable of self-repair and self-assembly upon interaction with their surroundings. In that sense, self-sufficient microorganisms naturally function very similar to what we envision for artificially created microrobots: They harvest chemical energy from their surroundings to power molecular motor proteins that serve as actuators, they employ
In fact, the biohybrid approach can be conceived in a dualistic way, with respect to the three basic ingredients of an in vivo microrobot, which are motility, control, and functionality. Figure 1 illustrates how these three ingredients can be either realized biologically, i.e., by the microorganism, or artificially, i.e., by the synthetic component. For example, a hybrid biomicromotor based on a sperm cell can be driven by the flagellum of the sperm or by an attached artificial helical flagellum.[154][155] It can orient itself autonomously via biological interactions with its surroundings and other cells, or be controlled and supervised externally via artificial sensors and actuators. Finally, it can carry out a biological function, like its inherent ability to fertilize an egg cell, or an artificially imposed function, like the delivery of synthetic drugs or DNA vectors. A biohybrid device may deploy any feasible combination of such biological and artificial components in order to carry out a specific application.[68]
Hydrodynamics can determine the optimal route for microswimmer navigation[156] Compared to the well explored problem of how to steer a macroscopic agent, like an airplane or a moon lander, to optimally reach a target, optimal navigation strategies for microswimmers experiencing hydrodynamic interactions with walls and obstacles are far-less understood.[156] Furthermore, hydrodynamic interactions in suspensions of microswimmers produce complex behavior.[157][158] The quest on how to navigate or steer to optimally reach a target is important, e.g., for airplanes to save fuel while facing complex wind patterns on their way to a remote destination, or for the coordination of the motion of the parts of a space-agent to safely land on the moon. These classical problems are well-explored and are usually solved using
There is growing interest in optimal navigation problems and search strategies [162][163][164][165][166][167] of microswimmers [58][103][168][169] and "dry" active Brownian particles,[170][99][171][172][156] The general problem regarding the optimal trajectory of a microswimmer which can freely steer but cannot control its speed toward a predefined target (point-to-point navigation) can be referred to as "the optimal microswimmer navigation problem". The characteristic differences between the optimal microswimmer navigation problem and conventional optimal control problems for macroagents like airplanes, cruise-ships, or moon-landers root in the presence of a low-Reynolds-number solvent in the former problem only. They comprise (i) overdamped dynamics, (ii) thermal fluctuations, and (iii) long-ranged fluid-mediated hydrodynamic interactions with interfaces, walls, and obstacles, all of which are characteristic for microswimmers.[99] In particular, the non-conservative hydrodynamic forces which microswimmers experience call for a distinct navigation strategy than the conservative gravitational forces acting, e.g. on space vehicles. Recent work has explored optimal navigation problems of dry active particles (and particles in external flow fields) accounting for (i) and partly also for (ii). Specifically recent research has pioneered the use of reinforcement learning [173][174][175] such as determining optimal steering strategies of active particles to optimally navigate toward a target position [162][163][166][167] or to exploit external flow fields to avoid getting trapped in certain flow structures by learning smart gravitaxis.[176] Deep reinforcement learning has been used to explore microswimmer navigation problems in mazes and obstacle arrays [177] assuming global [163] or only local [164] knowledge of the environment. Analytical approaches to optimal active particle navigation [165][166] complement these works and allow testing machine-learned results.[166][167][156]
Applications
As is the case for microtechnology and nanotechnology in general, the history of microswimmer applications arguably starts with Richard Feynman’s famous lecture There's Plenty of Room at the Bottom.[178] In the visionary speech, among other topics, Feynman addressed the idea of microscopic surgeons, saying: "...it would be interesting in surgery if you could swallow the surgeon. You put the mechanical surgeon inside the blood vessel and it goes into the heart and <<looks>> around (of course the information has to be fed out). It finds out which valve is the faulty one and takes a little knife and slices it out. Other small machines might be permanently incorporated in the body to assist some inadequately-functioning organ." The concept of the surgeon one could swallow was soon after presented in the science-fiction movie Fantastic Voyage and in Isaac Asimov’s writings.[1]
Only a few decades later, microswimmers aiming to become true microscale surgeons evolved from an intriguing science-fiction concept to a reality explored in many research laboratories around the world, as already highlighted by Metin Sitti in 2009.[180][1] These active agents that can self-propel in a low Reynolds number environment might play a key role in the future of nanomedicine, as popularised in 2016 by Yuval Noah Harari in Homo Deus: A Brief History of Tomorrow.[181] In particular, they might become useful for the targeted delivery of genes [182] or drugs [183][184] and other cargo [185][186] to a certain target (e.g. a cancer cell) through our blood vessels, requiring them to find a good, or ideally optimal, path toward the target avoiding, e.g., obstacles and unfortunate flow field regions.[156]
Already in 2010, Nelson et al. reviewed the existing and envisioned applications of microrobots in minimally invasive medicine.[187] Since then, the field has grown, and it has become clear that microswimmers have much potential for biomedical applications.[1] Already, many interesting tasks can be performed in vitro using tailored microswimmers. Still, as of 2020, a number of challenges regarding in vivo control, biocompatibility and long-term biosafety need to be overcome before microswimmers can become a viable option for many clinical applications.[188][1]
A schematic representation of the classification of biomedical applications is shown in the diagram on the left below. This includes the use of microswimmers for cargo transport in drug delivery and other biomedical applications, as well as assisted fertilisation, sensing, micromanipulation and imaging. Some of the more complex microswimmers fit into multiple categories, as they are applied simultaneously for e.g., sensing and drug delivery.[1]
The design of an untethered microscopic mobile machine or microrobot to function in vivo with medical interventional capabilities should assume an integrated approach where design 3D body shape, material composition, manufacturing technique, deployment strategy, actuation and control methods, imaging modality, permeation of biological barriers, and the execution of the prescribed medical tasks need to be considered altogether, as illustrated in the diagram on the right above. Each of these essential aspects contains a special design consideration, which must be reflected at the physical design of the microrobot.[189]
See also
- Bioinspiration
- Bio-inspired robotics
- Bio-inspired engineering
- Gray goo
- Robotic sperm
- Soft robotics
- Squirmer
References
- ^ PMID 33261101. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- PMID 26699202.
- ^ PMID 29955099.
- S2CID 5125033.
- ^ S2CID 40955542.
- S2CID 199451792.
- PMID 33429648.
- PMID 32923590.
- S2CID 203945671.
- ^ S2CID 128326068.
- S2CID 221383266.
- S2CID 198797318.
- .
- .
- ^ PMID 33500976. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- ^ doi:10.1119/1.10903.
- PMID 25369018.
- .
- S2CID 62330062.
- PMID 24911046.
- PMID 32341723.
- PMID 30742410.
- PMID 23165991.
- S2CID 53644820.
- S2CID 22780031.
- S2CID 213942288.
- S2CID 214408287.
- PMID 29127414.
- ^ PMID 33299981.
- PMID 25851515.
- ^ S2CID 3025635.
- PMID 26878315.
- S2CID 210801365.
- PMID 32864116.
- PMID 33500976.
- S2CID 46918320.
- PMID 27991641.
- S2CID 221747106.
- PMID 30781488.
- S2CID 7924312.
- S2CID 247697846.
- PMID 27749832.
- PMID 32832367.
- S2CID 205280841.
- S2CID 104145610.
- S2CID 52310451.
- S2CID 52278805.
- S2CID 221418150.
- S2CID 214386444.
- S2CID 149445838.
- S2CID 210959593.
- S2CID 201837838.
- S2CID 139612870.
- S2CID 215751974.
- S2CID 139612870.
- S2CID 215751974.
- PMID 32343587. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- ^ S2CID 3932471.
- PMID 15939237.
- PMID 23166050.
- PMID 11412087.
- PMID 20080560.
- S2CID 52845451.
- PMID 20108965.
- PMID 31429577.
- ISBN 9781292235103.
- PMID 25684261.
- ^ doi:10.1063/1.4993441. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- PMID 17189361.
- S2CID 84413776.
- PMID 11356579.
- PMID 16095621.
- .
- PMID 7665498.
- PMID 21205908.
- PMID 15004231.
- S2CID 6459972.
- .
- PMID 23184985.
- .
- PMID 11889087.
- PMID 16103369.
- .
- PMID 12684369.
- S2CID 125623927.
- ^ S2CID 19603301.
- ^ PMID 26339314.
- ^ PMID 18649256.
- PMID 24535032.
- ISSN 2072-666X.
- ^ Dumé, Isabelle (2020) Microswimmers benefit from thermoelectric guidance Physics World.
- ^ PMID 15479099.
- ^ PMID 17105298.
- ^ PMID 17995374.
- ^ PMID 18416540.
- PMID 28983536.
- PMID 22631222.
- PMID 24827167.
- ^ S2CID 14940249.
- S2CID 55870000.
- PMID 38144701.
- .
- ^ S2CID 3909877.
- ^ PMID 9326605.
- S2CID 15834620.
- ^ S2CID 3932471.
- ^ S2CID 4173914.
- OCLC 56124142.
- S2CID 221921243.
- PMID 20126443.
- .
- ^ S2CID 119229534. Material was copied from this source, which is available under a Creative Commons Attribution 3.0 International License.
- S2CID 120382159.
- S2CID 14646885.
- S2CID 27500334.
- S2CID 229547570.
- S2CID 4718416.
- S2CID 119474335.
- S2CID 45431975.
- S2CID 10335181.
- S2CID 3200289.
- S2CID 13651250.
- ISSN 1433-7851.
- ^ PMID 30242284. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- ^ PMID 34362934. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- PMID 21573199.
- S2CID 189870.
- .
- PMID 27687580.
- S2CID 208201932.
- PMID 32839447.
- PMID 28798960.
- .
- PMID 32680965.
- S2CID 37057964.
- S2CID 206722021.
- PMID 25941383.
- PMID 32471993.
- PMID 20866681.
- PMID 28573270.
- S2CID 201748635.
- S2CID 139844553.
- S2CID 189929035.
- S2CID 103866599.
- S2CID 4461200.
- PMID 30746446.
- S2CID 201058417.
- PMID 28932674. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- PMID 31905650. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License
- PMID 24895215.
- PMID 27641944.
- S2CID 26622101.
- S2CID 4388403.
- S2CID 26622101.
- S2CID 4388403.
- ^ S2CID 234012727. Material was copied from this source, which is available under a Creative Commons Attribution 4.0 International License.
- PMID 29717736.
- PMID 30339172.
- ISBN 978-0-486-13507-6.)
{{cite book}}
: CS1 maint: location missing publisher (link - ISBN 9780511902680.
- PMID 26990103.
- ^ S2CID 4938282.
- ^ S2CID 52824752.
- ^ S2CID 199000857.
- ^ S2CID 203038971.
- ^ S2CID 202540395.
- ^ S2CID 197935446.
- S2CID 13849152.
- ISBN 978-1-107-17465-8.)
{{cite book}}
: CS1 maint: location missing publisher (link - S2CID 119100040.
- S2CID 15672131.
- S2CID 3948148.
- ISBN 978-0-262-35270-3.
- S2CID 214355969.
- S2CID 199543817.
- S2CID 13695532.
- S2CID 211133324.
- ISBN 978-0-429-50045-9.
- PMID 32438567.
- S2CID 205044764.
- ISBN 978-1-4735-4537-3.
- S2CID 95812709.
- PMID 28873304.
- PMID 22770233.
- PMID 25844893.
- PMID 29799053.
- PMID 20415589.
- PMID 33173743.
- ^ S2CID 199341199. Material was copied from this source, which is available under a Creative Commons Attribution 3.0 International License.