Tricky Biomedical Engineering interview questions

75 questions with answers (BME-426 to BME-500). Try each one aloud before reading the answer.

Anatomy & Physiology for Engineers

BME-426 · Do nerves conduct electricity like copper wires?

Trap: The idea that nerves are passive cables. Axons are poor passive conductors. Action potentials are regenerated by ion channels along the membrane and travel at only about 1–120 m/s. In copper, signals propagate at a large fraction of the speed of light.

BME-427 · Is the heart's natural pacemaker in the ventricles?

Trap: Confusing the site of the main pumping chambers with the pacemaker site. The SA node, in the wall of the right atrium, sets normal rhythm at 60–100 bpm. Ventricular escape pacemakers fire at only about 20–40 bpm.

BME-428 · Is venous blood blue?

Trap: Veins look blue, so people assume the blood is blue. Deoxygenated blood is dark red. Veins appear blue because of how skin and tissue scatter and absorb different wavelengths of light.

BME-429 · Does a more negative resting potential make a neuron more excitable?

Trap: Assuming a larger magnitude means more "charged up". Hyperpolarization moves the membrane farther from threshold, so the cell becomes less excitable. Mild depolarization brings it closer to threshold.

BME-430 · Do we breathe mainly because oxygen levels fall?

Trap: The assumption that low O₂ is the main drive to breathe. The main drive is CO₂ and pH, sensed by central chemoreceptors. The hypoxic drive from peripheral chemoreceptors becomes significant only when PaO₂ falls below about 60 mmHg.

BME-431 · If SpO₂ is 98%, is oxygen delivery adequate?

Trap: Treating saturation as oxygen content. SpO₂ is the fraction of haemoglobin that is saturated. Severe anaemia or low cardiac output can leave oxygen delivery dangerously low despite normal saturation (DO₂ = CO × CaO₂).

Biomedical Instrumentation

BME-432 · Why is the ECG hard to measure if the heart is so strong?

Trap: Assuming a strong muscle produces a large surface signal. At the skin the ECG is only about 1 mV, and it is buried in larger interference: powerline common-mode voltage, EMG, baseline wander and electrode offsets of hundreds of mV. It needs high-CMRR, high-impedance amplifiers and careful electrodes.

BME-433 · Does more amplifier gain improve ECG quality?

Trap: Thinking gain improves SNR. Gain amplifies signal and in-band noise equally. Excessive early gain also saturates on electrode DC offset. Quality depends on low-noise design, CMRR, filtering and electrode contact.

BME-434 · A patient pulled from a house fire has SpO₂ of 99%. Are they well oxygenated?

Trap: Trusting a two-wavelength oximeter in CO poisoning. Carboxyhaemoglobin absorbs red light much like oxyhaemoglobin, so SpO₂ reads falsely high. Use multi-wavelength CO-oximetry on a blood sample.

BME-435 · Does an oscillometric cuff directly measure systolic and diastolic pressure?

Trap: Assuming it detects systolic and diastolic events like auscultation. It directly identifies MAP, at maximum oscillation amplitude. Systolic and diastolic values are estimated by proprietary algorithms, which can err in arrhythmia or stiff arteries.

BME-436 · Is using a large cuff on everyone safer?

Trap: The idea that "bigger is gentler and safer". A cuff that is too large underestimates BP and one that is too small overestimates it. The bladder width should be about 40% of arm circumference and its length about 80%.

BME-437 · Do Ag/AgCl electrodes eliminate motion artifact?

Trap: The belief that non-polarizable means artifact-free. They reduce electrode polarization artifacts, but skin stretch still changes skin potential and impedance. Skin preparation and mechanical stability are still required.

BME-438 · With a 120 dB CMRR amplifier, can you ignore electrode impedance imbalance?

Trap: Thinking CMRR alone governs interference rejection. Impedance imbalance against a finite common-mode input impedance converts common-mode voltage into a differential signal before the amplifier, and no CMRR can reject that. High input impedance and balanced contacts are needed.

Biosignal Processing (ECG, EEG, EMG)

BME-439 · Should a 50 Hz notch filter always be switched on?

Trap: Assuming the notch removes only noise. It also removes genuine signal content near 50 Hz and can ring after sharp QRS complexes and pacing spikes. Fix interference at its source first. Use adaptive or narrow notches only when needed.

BME-440 · Is sampling at exactly twice the highest frequency enough?

Trap: Applying Nyquist literally. Real anti-aliasing filters roll off gradually, and signals are not strictly band-limited, so practical rates are 2.5–10× the bandwidth or higher. Pacemaker spike detection needs much higher rates.

BME-441 · Can zero-phase filtering be used in a real-time bedside monitor?

Trap: Forgetting that forward–backward filtering needs future samples. It is non-causal and suits offline analysis only. Real-time systems use causal linear-phase FIR filters, which add a fixed group delay, or carefully designed IIR filters.

BME-442 · Does a longer FFT window always improve EEG spectral analysis?

Trap: Thinking frequency resolution is the only goal. Longer windows improve frequency resolution (Δf = 1/T) but blur changes over time in non-stationary signals. Choose window length to suit the stationarity of the phenomenon, or use time–frequency methods.

BME-443 · Is the EEG larger in amplitude than the ECG because the brain is so active?

Trap: Equating activity with signal size. Scalp EEG is about 10–100 µV, roughly 10–100× smaller than the ECG, because of skull attenuation and spatial averaging. ECG often appears as an artifact in EEG.

BME-444 · A monitor shows a flat line. Is it asystole?

Trap: Trusting one lead blindly. Check for lead or electrode disconnection, gain settings and a second lead first. Fine VF can look flat in a single lead. Confirm clinically before choosing treatment.

Medical Imaging

BME-445 · Why can't you take metal into an MRI room?

Trap: Assuming all metals are equally dangerous. Ferromagnetic objects become projectiles and implants can experience torque. Non-ferrous conductors are not attracted but can still heat from RF induction or cause artifacts. Items must be screened and labelled MR Safe, Conditional or Unsafe.

BME-446 · Is a higher-resolution imaging device always better?

Trap: Equating resolution with diagnostic value. Higher resolution can mean more dose (CT), longer scans and motion (MRI), more noise per voxel, higher cost and more data. The task and the contrast-to-noise ratio matter more.

BME-447 · Is MRI completely risk-free because it uses no ionizing radiation?

Trap: Treating non-ionizing as harmless. Risks include projectiles, implant interactions, RF burns, acoustic noise, peripheral nerve stimulation, cryogen quench and gadolinium contrast concerns, such as NSF in severe renal impairment.

BME-448 · Does sound travel faster in air than in soft tissue?

Trap: Intuition that "lighter means faster". Sound travels at about 343 m/s in air, about 1540 m/s in soft tissue and about 4000 m/s in bone. Speed depends on stiffness relative to density. Scanners assume 1540 m/s, which causes depth errors in other media.

BME-449 · Is the MRI magnet switched off when no scan is running?

Trap: Thinking of it like an X-ray tube. Superconducting magnets stay at field 24/7. Only the RF and gradients switch on during scans, so projectile risk is constant. Shutting the field down requires a quench.

BME-450 · Does a PET scan show detailed anatomy?

Trap: Assuming every scan shows anatomy. PET maps tracer distribution, meaning function and metabolism, with resolution of several mm and little anatomical detail. It is fused with CT or MRI for localization and attenuation correction.

BME-451 · Does doubling the mAs in CT halve the image noise?

Trap: Assuming noise scales linearly with dose. Noise ∝ 1/√(mAs), so doubling mAs reduces noise by only about 29% while doubling dose. Halving noise needs 4× the dose.

Biomaterials

BME-452 · Is titanium used for implants because it's the strongest metal?

Trap: Assuming strength is the reason. CoCr and some steels are stronger and stiffer. Titanium is chosen for its stable TiO₂ passive layer (biocompatibility and corrosion resistance), osseointegration, low density and a modulus closer to bone.

BME-453 · Does a bioinert material cause no reaction in the body?

Trap: Taking "inert" literally. Every implant adsorbs proteins within seconds and triggers some foreign body response. Bioinert materials typically end up encapsulated in a thin fibrous layer rather than bonded to tissue.

BME-454 · Is a stiffer bone plate always better for healing?

Trap: Thinking more rigidity always helps. Overly stiff fixation causes stress shielding and can suppress the micromotion that drives secondary healing. Fixation stiffness should match the healing strategy.

BME-455 · Do biodegradable implants simply vanish harmlessly?

Trap: Assuming degradation is benign and uniform. Polyesters can lose strength well before they lose mass. Bulk erosion can release a burst of acidic products, causing sterile inflammation, and the timing must match healing.

BME-456 · Is stainless steel immune to corrosion in the body?

Trap: Trusting the name. Chloride-rich body fluids can break down the passive film, causing pitting and crevice corrosion. That is why 316L is mainly used for temporary devices such as fracture plates and screws.

BME-457 · If a material passes cytotoxicity testing, is it biocompatible?

Trap: Treating one test as proof. Cytotoxicity is one endpoint. Biocompatibility requires a risk-based evaluation of all relevant endpoints for the specific contact type and duration, including manufacturing residues and sterilization effects.

BME-458 · Is a biocompatible material biocompatible in every application?

Trap: Treating biocompatibility as an intrinsic material property. It is application-specific. A material that integrates well in bone may be thrombogenic in blood or inflammatory in the eye.

Biomechanics

BME-459 · Is adult bone a static, inert structure?

Trap: Picturing a skeleton as dead scaffolding. Bone is living tissue that is constantly remodelled by osteoclasts and osteoblasts, renewing roughly 10% of the adult skeleton each year and adapting to load (Wolff's law).

BME-460 · Is the hip joint force during walking about equal to body weight?

Trap: Ignoring muscle forces. Muscles act through short moment arms, so their forces add to body weight. Hip contact forces during walking are typically around 2.5–4× body weight and higher on stairs.

BME-461 · Is blood viscosity constant, like water?

Trap: Assuming blood is Newtonian everywhere. Blood is shear-thinning. In vessels below about 0.3 mm, apparent viscosity falls further because a cell-free layer forms near the wall (Fåhræus–Lindqvist effect). In capillaries, single-file red cells behave differently again.

BME-462 · Is tendon a linear elastic spring?

Trap: Applying Hooke's law over the whole range. Tendon has a non-linear toe region, viscoelastic creep and relaxation, hysteresis and strain-rate–dependent stiffness and strength.

BME-463 · Can a muscle push a bone?

Trap: Thinking muscles work in both directions. Muscles can only actively contract, or pull. Movement in the opposite direction needs antagonist muscles, gravity or elastic recoil.

BME-464 · Does higher blood pressure in a vessel mean higher flow?

Trap: Confusing absolute pressure with pressure gradient. Flow depends on the pressure difference and the resistance (Q = ΔP/R). Vasoconstriction can raise upstream pressure while reducing flow.

Therapeutic & Life-Support Devices

BME-465 · Can a defibrillator restart a flatlined heart?

Trap: The movie version of defibrillation. Asystole is not shockable. Defibrillation stops chaotic electrical activity (VF and pulseless VT) so the natural pacemaker can resume. Asystole is treated with CPR and adrenaline, along with treating its cause.

BME-466 · Is a higher defibrillation energy always more effective?

Trap: Thinking "more joules is better". Excess energy increases myocardial injury and post-shock dysfunction. Waveform shape and current matter more than energy, so biphasic waveforms succeed at lower energies.

BME-467 · Does a pacemaker shock the heart like a defibrillator?

Trap: Confusing pacemakers with ICDs. Pacemakers deliver tiny pulses of a few volts and microjoules to trigger beats. Only ICDs deliver high-energy shocks, up to tens of joules.

BME-468 · Is it safest to ventilate everyone on 100% oxygen?

Trap: Thinking more oxygen is always better. Prolonged high FiO₂ causes oxygen toxicity, absorption atelectasis and oxidative injury. Titrate FiO₂ to target saturations, often about 92–96% for most patients.

BME-469 · Is a bigger tidal volume better ventilation?

Trap: Equating more volume with more support. Large tidal volumes overdistend alveoli and cause volutrauma. Lung-protective ventilation uses about 6 mL/kg predicted body weight and limits plateau pressure.

BME-470 · Can patients with pacemakers never have an MRI?

Trap: An outdated absolute rule. MR Conditional pacemaker and ICD systems can be scanned under specified conditions, with reprogramming and monitoring. Some legacy systems can also be scanned under strict protocols.

Tissue Engineering

BME-471 · Can you grow a whole organ just by seeding stem cells onto a scaffold?

Trap: Oversimplifying the triad. Organs need multiple cell types in a precise architecture, hierarchical vasculature, innervation, maturation and integration with the host. Simple constructs (skin, cartilage, bladder) have succeeded, but solid organs remain far harder.

BME-472 · Is higher scaffold porosity always better?

Trap: Maximizing porosity alone. Higher porosity improves transport and ingrowth but sharply reduces stiffness and strength. Load-bearing scaffolds must balance porosity against mechanical requirements.

BME-473 · Are autologous iPSC-derived cells risk-free?

Trap: Assuming "own cells" means no risk. Residual undifferentiated cells can form teratomas. Reprogramming and culture can introduce genetic or epigenetic abnormalities. Manufacturing is costly and slow, and immunogenicity is not completely excluded.

BME-474 · Does a decellularized scaffold trigger no immune response?

Trap: Assuming decellularization removes all antigens. Residual DNA, cell debris and xenogeneic epitopes (such as α-gal) can provoke inflammation. A commonly cited target is below 50 ng dsDNA per mg of dry ECM with fragments under 200 bp.

BME-475 · Does more growth factor give better regeneration?

Trap: Thinking dose and response are monotonic. Supraphysiological BMP-2 doses have caused ectopic bone, swelling and inflammation. Controlled, localized, sustained release at the lowest effective dose works better.

BME-476 · Do cells survive bioprinting unharmed?

Trap: Assuming printing is gentle. Shear stress in the nozzle, pressure, UV or photoinitiator exposure, and time outside culture reduce viability and function. Ink, nozzle and process parameters must be optimized and viability checked after printing.

Biosensors

BME-477 · Does a CGM measure blood glucose?

Trap: Assuming the number is blood glucose. CGMs measure interstitial fluid glucose, which lags blood by several minutes. The difference is largest during rapid changes, so confirm with fingerstick readings when symptoms don't match.

BME-478 · Is a more sensitive biosensor always better?

Trap: Confusing sensitivity (slope) with overall performance. A high slope may saturate early and reduce dynamic range. If noise rises proportionally it doesn't improve LOD, and selectivity, stability and clinical range matter equally.

BME-479 · Is the pH electrode slope always 59 mV/pH?

Trap: Forgetting the temperature term in the Nernst equation. The slope is 2.303RT/F: about 59.2 mV/pH at 25 °C and about 61.5 mV/pH at 37 °C. Real electrodes are also slightly sub-Nernstian and age, so they need temperature compensation and calibration.

BME-480 · In a sandwich immunoassay, does more analyte always give more signal?

Trap: Assuming a monotonic dose–response. At very high concentrations, excess antigen saturates both capture and detection antibodies separately (the high-dose hook effect), so the signal falls and gives falsely low results. Dilute the sample to check.

BME-481 · Is the LOD the lowest concentration you can report quantitatively?

Trap: Confusing detection with quantification. At the LOD (about 3σ) the analyte is only distinguishable from blank. Reliable quantitative reporting starts at the LOQ (about 10σ), or at the validated lower limit of the measuring range.

BME-482 · Can a graphene FET detect large proteins directly in undiluted blood?

Trap: Ignoring ionic screening. At physiological ionic strength the Debye length is below 1 nm, so the charges of large antibody–protein complexes are mostly screened. Dilution, small receptors or special interfacial polymers are needed.

Regulatory Standards & Quality

BME-483 · Is CE marking a quality certificate awarded by the EU?

Trap: Treating CE as an EU approval or quality seal. It is the manufacturer's declaration of conformity with the regulations, supported by Notified Body assessment for most device classes. The EU itself does not "approve" devices.

BME-484 · Does 510(k) clearance mean the device is FDA approved?

Trap: Using "cleared" and "approved" interchangeably. 510(k) gives clearance based on substantial equivalence. "Approval" refers to PMA, which requires independent evidence of safety and effectiveness.

BME-485 · Does ISO 13485 certification prove a product is safe?

Trap: Confusing system certification with product conformity. ISO 13485 certifies the quality management system. Product safety and performance are shown through technical documentation, standards testing, risk management and clinical evidence.

BME-486 · Is risk management finished at design freeze?

Trap: Treating the risk file as a one-time deliverable. ISO 14971 requires risk management across the whole lifecycle, including production and post-production information (complaints, vigilance, PMS), with updates to the risk file.

BME-487 · Can a warning in the IFU make an unacceptable risk acceptable?

Trap: Over-relying on labelling. Information for safety is the lowest-priority control and cannot replace feasible design changes or protective measures. Regulators expect inherent safety first, with residual risks disclosed.

BME-488 · Is software that only displays medical data never a medical device?

Trap: Judging by function rather than intended use. Pure storage, transfer or display can be excluded, depending on the jurisdiction. But software that analyses, interprets, alarms or drives clinical decisions is usually a medical device, for example under EU MDR Rule 11.

Clinical Engineering & Electrical Safety

BME-489 · Is low voltage always safe for patients?

Trap: Thinking voltage alone determines harm. Current through the heart matters. With a conductive catheter, millivolt-level potentials can drive microampere currents into the heart that cause VF.

BME-490 · Does a three-prong plug guarantee electrical safety?

Trap: Assuming the earth connection is always intact. A broken earth conductor, damaged cord or faulty outlet leaves Class I equipment unprotected without any visible sign. Periodic earth continuity and leakage testing are needed.

BME-491 · Should operating rooms always use GFCI protection to protect patients?

Trap: Assuming tripping is always the safest response. A GFCI cuts power on the first fault, which can stop life-support equipment mid-procedure. Isolated power systems with line isolation monitors keep power on and alarm instead, where interruption is unacceptable.

BME-492 · If a device passed electrical safety testing at acceptance, is it safe for life?

Trap: Ignoring degradation. Insulation ages, cords get damaged, fluids enter and repairs alter construction. Recurrent testing (for example to IEC 62353) and after-repair testing are required.

BME-493 · Is any leakage current a sign of a fault?

Trap: Thinking leakage is abnormal. Some leakage always flows through stray and filter capacitance and insulation. It is a hazard only when it exceeds IEC 60601-1 limits, which are much tighter for CF applied parts.

BME-494 · Is a given current equally dangerous at any frequency?

Trap: Ignoring frequency dependence. 50/60 Hz is near the most dangerous range for nerve and heart stimulation. Above about 100 kHz, as in electrosurgery, currents of amperes cause heating but no stimulation.

Healthcare Informatics & AI

BME-495 · Is a model with 99% accuracy excellent?

Trap: Accuracy under class imbalance. With 1% prevalence, predicting "negative" for everyone gives 99% accuracy and 0% sensitivity. Report sensitivity, specificity, PPV, NPV and PR-AUC.

BME-496 · Does a high AUC guarantee clinical usefulness?

Trap: Equating discrimination with utility. A model can rank well but be miscalibrated, perform poorly at the operating threshold, have low PPV at real prevalence or fail to change outcomes. Decision-curve analysis and prospective evaluation are needed.

BME-497 · Does removing names and record numbers make data anonymous?

Trap: Ignoring quasi-identifiers. Combinations of dates, ZIP codes, rare diagnoses, facial reconstruction from imaging or DICOM headers can re-identify patients. Use formal de-identification and risk assessment.

BME-498 · Once an AI device is cleared, will its performance stay constant?

Trap: Assuming validation lasts forever. Changes in populations, scanners, protocols and clinical practice cause data drift. Ongoing real-world performance monitoring and controlled updates are required.

BME-499 · Does more data always fix a biased model?

Trap: Thinking volume cures bias. More data with the same skew, label bias or proxy variables preserves the bias. Representative sampling, label auditing and subgroup evaluation are needed.

BME-500 · Is FHIR a database for storing health records?

Trap: Confusing an exchange standard with storage. FHIR defines resources and APIs for exchanging data. Servers may persist FHIR resources, but FHIR itself specifies interoperability, not a database design.

Other levels: Basic · Intermediate · Advanced · All Biomedical Engineering questions