Biomedical modeling: choose a text and test a compartment balance
An equation can balance perfectly while representing the wrong system. Begin with what crosses a boundary, what is stored, which quantities are measured and what a prediction is meant to support. This route compares books by the work they help you do, then tests a deliberately simple model using invented data.
Course topic: Introduction to Biomedical Engineering
Study route and exercise: University Reading List editorial project. How we curate · Sources checked:
Scope and responsibility
An undergraduate study route in introductory biomedical engineering, not a device-building protocol or clinical recommendation. The example is a paper-only mixing reservoir, with no patient, medicine, dose or apparatus specification. Saterbak means the selected 2017 second-edition print product; Saltzman means second edition 2015. The free 2011 edited volume is a later research-reading companion, not a complete free beginner textbook or evidence of current practice. Activities and answers are original editorial work, not official assignments or a claim of full-book/classroom testing.
Choose your starting point
I can follow calculus but need engineering balances
Choose Saterbak as the main route. Start with units and a system boundary before the case-study chapter; use the worked model to test whether each term has the right physical meaning.
I need to connect engineering to biological context
Choose Saltzman when molecular/cellular foundations and physiological organization are your larger gap. Select its balances and removal chapters after the biological orientation rather than purchasing another survey for the same job.
I already have the foundations and need free research cases
Use the IntechOpen chapter as an evidence-audit exercise after the core mathematics. Separate what its authors fitted, what they tested and what would justify transfer. Arrange a legitimate library copy of a foundation text if balances are still unfamiliar.
Compare the three books
Bioengineering Fundamentals
Ann Saterbak; Ka-Yiu San; Larry V. McIntire · Edition: 2 · 2017
Quantitative first-course spine
- Choose it when
- You want a repeatable way to set up mass, energy, charge and momentum accounts rather than memorize separate subfield formulas.
- Use another route when
- You need an anatomy-led survey first or expect a short calculation course to establish clinical/device competence.
- Selected reading
- Second edition: chapters 1–2 for calculations and conservation framework,3 for mass, then selected 4–6; use 7 after these tools.
- Limits and trade-offs
- The selected hardcover is 2017, ISBN 9780134637433. Pearson’s newer digital-product dates and access identifiers do not silently turn it into a different print edition.
Try before choosing
From chapters 1–3, draw a well-mixed reservoir with an inlet and outlet. State whether concentration or mass is accumulated, then give units for every balance term.
Check your work: In the dM/dt amount balance, every term has units mass/time. After dividing by constant V, dC/dt has units concentration/time; concentration alone is not a rate. Equal volume flows preserve volume but need not preserve solute mass. State the mixing assumption.
Biomedical Engineering: Bridging Medicine and Technology
W. Mark Saltzman · Edition: 2 · 2015
Biological bridge with engineering applications
- Choose it when
- You want the biological hierarchy explained before interpreting a physiological engineering model.
- Use another route when
- You only need repeated balance practice or already have an assigned quantitative spine that meets the immediate task.
- Selected reading
- Second edition: chapter 1 for orientation;2–6 for biological context as needed;7–9 for balances, circulation and molecular removal.
- Limits and trade-offs
- Cambridge distinguishes the 2015 hardback ISBN 9781107037199 from the later digital product. Public contents and selected resources do not mean that the complete textbook is freely readable.
Try before choosing
After chapter 7 and the chapter 9 overview, make a two-column vocabulary map: the physical quantity being tracked and the biological interpretation that would require extra evidence.
Check your work: Distinguish a conserved amount, its concentration and a removal rate. A reservoir analogy alone does not establish that a living organ is well mixed or has a constant parameter.
Biomedical Engineering - From Theory to Applications
Reza Fazel-Rezai (editor) · 2011
Free historical research-case companion
- Choose it when
- You already understand balances and want to interrogate a contributed modeling study’s evidence.
- Use another route when
- You need a uniformly paced beginner course or current treatment/device guidance.
- Selected reading
- 2011 volume: chapter 3, “Physiological Cybernetics”, focusing on its separation of model construction and testing; other contributed chapters are optional, not a prescribed sequence.
- Limits and trade-offs
- The chapter’s age, advanced methods and clinical subject matter require careful scope control. Use it to inspect an argument; do not import its application claims as current guidance. Check individual chapter licences and credits.
Try before choosing
Locate the model-building/test-data discussion in chapter 3. Write an evidence ledger with the question, fitted quantities, withheld observations and stated transfer limits; no biological experiment is required.
Check your work: Name which observations were used in fitting and which were reserved for testing. If a claimed use goes beyond that evidence, label the gap instead of treating a plotted fit as validation.
Check your preparation
Can you distinguish amount from concentration?
If1 L contains 2 g of an arbitrary marker, concentration is 2 g/L and amount 2 g. A flow in L/min times concentration in g/L is a mass rate in g/min.
Can you read a first-order balance?
For constant V, V dC/dt=Q(Cin−C) describes this idealized mixed reservoir. You should be able to differentiate an exponential and check the initial value; a general physiology course is not implied.
Can you separate fitting from a new test?
Choosing parameters to reproduce two observations uses those observations. Testing a third, held-out prediction asks a different question and still needs an uncertainty/measurement model before a real-world claim.
Six steps with a concrete output
Outputs marked ready: 0 / 6
These are your own study marks, not an assessment. They are not sent to this site.
-
1. State the intended question
Selected reading: Saterbak 1–2; Saltzman 1.
Make this: Write a one-sentence prediction question and draw the boundary. List quantities inside and outside; state why a one-compartment approximation is being tried.
Check your work: The output is a concentration over time, not a judgment about a patient. Assumptions can be challenged separately from algebra.
-
2. Make the units ledger
Selected reading: Saterbak 1–3; Saltzman 7.
Make this: Define M,V,C,Q and Cin with units; write both the amount balance and C=M/V. Specify initial conditions and whether volume changes.
Check your work: All terms in dM/dt have units g/min. Only after stating constant V may you replace d(VC)/dt by V dC/dt.
-
3. Derive and check the prediction
Selected reading: Saterbak 3; Saltzman 7–9 selectively.
Make this: Solve the original reservoir problem below. Check the derivative, initial value, limiting value and integrated input/output amounts.
Check your work: The solution satisfies the equation and both amount accounting and units. A computational output without these checks is incomplete.
-
4. Identify what was fitted
Selected reading: IntechOpen 3 model-building/test discussion.
Make this: Complete the follow-up using the two supplied synthetic observations. Mark the observations used for parameter selection and keep the third value aside until the prediction is written.
Check your work: Perfect interpolation of two fitted points is not counted as an independent test. Every free parameter is named.
-
5. Challenge one assumption
Selected reading: Saltzman 9 overview; Saterbak 2–3; FDA guidance scope.
Make this: Write two competing reasons for a held-out discrepancy and one observation that could distinguish them. Keep model discrepancy, parameter uncertainty and measurement error separate.
Check your work: An unexplained residual is not proof of a specific biological mechanism. No certification or clinical validity is inferred from this classroom model.
-
6. Assemble a reproducible model memo
Selected reading: Return to the selected main text and source notes.
Make this: Produce a boundary diagram, units table, derivation, fitted-parameter record, held-out comparison and a one-paragraph limit on intended use. Include the exact editions consulted.
Check your work: Another reader can reproduce the numbers and identify which claim the evidence supports. Missing access to a full chapter is recorded, not disguised as a completed reading.
Worked exercise
Original model: marker accumulation in a virtual reservoir
An invented, well-mixed reservoir has constant volume 1.00 L. Equal inlet and outlet flows are 0.100 L/min. The inlet concentration of an arbitrary marker is 2.00 g/L and the initial reservoir concentration is zero. There is no generation, consumption or loss except the stated outlet. Find C(t), C(10 min), and the amounts entering, remaining and leaving in the first 10 min.
Assumptions
Treat the stated values as exact teaching inputs. Mixing is instantaneous, the outlet concentration equals the reservoir concentration, and volume, flow and inlet concentration are constant. This virtual reservoir represents no patient or actual device. Model time is in minutes; the parameter Q/V is in min⁻¹.
- Let M=VC. The amount balance is dM/dt=Q Cin−Q C. Since V is constant, dC/dt=(Q/V)(Cin−C)=0.100(2.00−C). Each amount-rate term is in g/min.
- With C(0)=0, C(t)=2.00(1−e^(−0.100t)) g/L. Substitution gives dC/dt=0.200e^(−0.100t), matching the balance.
- At10 min, C=2(1−e⁻¹)≈1.26424 g/L. The remaining amount is VC≈1.26424 g. Input amount is 0.100×2.00×10=2.00 g.
- Integrated outlet amount is 2.00−1.26424≈0.73576 g; equivalently integrate Q C(t). Initial amount 0 plus input 2.00 equals remaining 1.26424 plus outlet 0.73576.
Answer and interpretation
C(t)=2(1−e^(−0.1t)) g/L and C(10)≈1.264 g/L. By 10 min,2.000 g has entered,1.264 g remains and 0.736 g has left. Equal volumetric flows do not imply equal solute mass flows while concentration changes.
Test the result
C(0)=0, C increases toward 2 g/L, and the time constant V/Q is 10 min. As Q tends to zero, accumulation over a fixed finite time tends to zero. These are equation/accounting checks, not evidence that a real biological system meets the assumptions.
Change an assumption
A perfect fit is not the held-out prediction
Keep a constant-volume, initially marker-free, single-compartment exponential model C(t)=A(1−e^(−kt)). Now treat A and k as unknown positive parameters, replacing the original fixed inlet concentration and flow-to-volume ratio. Two invented exact observations are C(10)=1.20 and C(20)=1.80 g/L. Fit A and k, predict C(30), then compare a held-out invented exact observation 2.25 g/L.
- Put x=e^(−10k), with 0<x<1. Then C(20)/C(10)=(1−x²)/(1−x)=1+x. Thus 1.80/1.20=1.5 gives x=0.5.
- A=1.20/(1−0.5)=2.40 g/L and k=−ln(0.5)/10≈0.069315 min⁻¹. Both fitted observations are reproduced exactly.
- The held-out prediction is C(30)=2.40(1−0.5³)=2.10 g/L. The supplied held-out value 2.25 differs by 0.15 g/L. Under the stipulated exact-data model, that discrepancy cannot be ignored as rounding.
Answer and interpretation
The fitted plateau is 2.40 g/L, k≈0.069315 min⁻¹ and the predicted third value 2.10 g/L. The model matches its fitted points yet fails the held-out exact observation. This does not uniquely identify why the assumptions fail.
Test the result
Keep fitting and checking roles distinct. In actual measurements, an independently justified uncertainty model would be needed before judging a residual. Recalibrating on the third observation would make it training evidence, not preserve it as a new test.
Common mistakes
Calling concentration the conserved quantity
Mass is accounted for. Concentration can change even with constant volume, and changing volume requires differentiating VC rather than silently holding V fixed.
Treating a fit as proof of mechanism
Different assumptions or parameter choices may reproduce a short record. Separate algebraic consistency, parameter estimation and testing against observations not used for fitting.
Transferring historical cases without a use boundary
An older research chapter can teach evidence scrutiny. Its publication, a free download or a good graph does not establish current clinical practice, device safety or regulatory acceptability.
Before moving on
- I can derive the amount balance with a boundary, units and initial condition, and explain each simplifying assumption.
- I can reproduce both original exercises, name fitted versus held-out observations and state what a discrepancy does not prove.
- I can explain why I chose one main book, what the free case companion supplies, and which additional biological or mathematical preparation remains.
Sources and edition checks
Sources checked:
- Pearson: second-edition print identity, digital formats and contents
Pearson lists the selected hardcover as a 2017 update, with a separate digital identifier/date. The seven-chapter contents support the conservation-law route; print ownership is not inferred to include every digital product.
- Cambridge: second-edition contents and format-specific dates
Cambridge’s second-edition contents identify chapters 7–9 and distinguish 2015 hardback from later digital publication. Contents/previews support scope selection, not a claim of complete free access or full-text review.
- IntechOpen: 2011 edited volume, chapter access and publication metadata
The publisher identifies an edited 2011 volume with individually accessible contributions. This is an advanced research companion with uneven chapter demands, not a newly updated introductory course.
- IntechOpen: Physiological Cybernetics, model fitting and testing discussion
The chapter explicitly discusses model construction and separate testing. Its historical clinical examples are not copied into the exercise or used as current recommendations; the reservoir numbers and questions are invented.
- FDA: November 2023 guidance on credibility of physics-based computational models
The November 2023 US FDA guidance concerns credibility of physics-based models in device submissions. It provides context for asking about intended use and evidence; this study exercise establishes no submission compliance.
These are proposed study activities, not official course assignments or a claim that every book has been read cover to cover.