MCG Heart Scans

What the Signal Tells You

MCG heart scans — short for magnetocardiography — measure the tiny magnetic fields produced by the heart's electrical activity. The value is not the raw signal by itself. The value is what clinicians and cardiac teams can reliably infer after the signal is captured consistently and processed into interpretable outputs.

If you're exploring magnetocardiography as a practical modality, SandboxAQ's AQMed focuses on applying advanced sensing and computation to clinical workflows.

What an MCG heart scan measures

An MCG heart scan measures magnetic fields associated with cardiac electrical activity. Because those magnetic signals are extremely small, the system must capture the signal in a way that is stable and repeatable, then process it to reduce noise and artifacts.

A useful way to frame it:

ECG/EKG measures electrical potentials at the skin
MCG measures magnetic fields produced by that activity

Different signal type, different sensing approach, and different challenges in real-world capture.

What the signal can tell you

High-level interpretation typically comes from patterns and features extracted from the measured signal, not from a single "magic readout." In practice, teams care about whether the scan output is consistent enough to support decision-making and whether it adds value alongside existing cardiac tools.

The key question is not "can we capture the signal," but:

can we capture it consistently enough to compare across scans
can we process it in a way that preserves meaningful structure
can outputs be presented in a way clinicians can trust and understand

Why MCG is a signal-to-noise problem

Most of the difficulty in magnetocardiography comes down to separating a faint cardiac signal from the environment.

Real-world variables include:

electromagnetic interference
motion artifacts and positioning variability
differences in capture conditions between scans
workflow constraints that make "perfect capture" unrealistic

This is why end-to-end system design matters as much as the sensing method.

Where AI and modeling add real value

AI is most useful when it makes MCG more reliable and usable — for example:

denoising and artifact suppression
consistent feature extraction across scans
quality checks and confidence indicators so outputs are not treated as absolute

This is where model-driven inference earns its keep in physical sensing applications. SandboxAQ's work on Large Quantitative Models speaks to how quantitative modeling improves signal interpretation in exactly these kinds of complex, noisy environments.

Where the modality may add clinical value

The strongest framing is practical: MCG is evaluated where teams believe an additional signal modality could improve confidence, reduce ambiguity, or provide complementary information to existing approaches. Whether it adds value depends on how reliably it can be captured, interpreted, and integrated into real workflows.