ECG - Electrocardiography
Electrocardiography (ECG) records the electrical activity of the heart by measuring the potential difference across electrode pairs placed on the chest (or limbs). In cognitive monitoring, ECG is primarily used to derive heart rate (HR) and heart rate variability (HRV) - two well-established physiological correlates of stress and cognitive load.
Anatomy of the ECG Signal
A single cardiac cycle produces a characteristic waveform:
- P wave - atrial depolarisation; marks the start of the heartbeat.
- QRS complex - ventricular depolarisation; the large spike easily detected as the heartbeat event.
- T wave - ventricular repolarisation.
The R-peak (the tallest point in the QRS complex) is the primary fiducial point used to measure beat timing.
Derived Measures
Heart Rate (HR)
where \(\overline{RR}\) is the mean RR interval (time between successive R-peaks) in seconds.
Heart Rate Variability (HRV)
HRV quantifies the variation in beat-to-beat timing. A large HRV indicates flexible autonomic regulation (healthy, parasympathetically dominant); low HRV indicates rigid, sympathetically dominant autonomic state associated with stress, high workload, and fatigue.
Time-domain HRV features:
| Feature | Description |
|---|---|
| SDNN | Standard deviation of all NN (normal-to-normal) intervals |
| RMSSD | Root mean square of successive differences; primarily reflects parasympathetic activity |
| pNN50 | Percentage of successive NN intervals differing by more than 50 ms |
Frequency-domain HRV features:
| Band | Range | Interpretation |
|---|---|---|
| VLF | 0.003 – 0.04 Hz | Very low frequency; thermoregulation and other slow processes |
| LF | 0.04 – 0.15 Hz | Low frequency; mixed sympathetic and parasympathetic |
| HF | 0.15 – 0.4 Hz | High frequency; parasympathetic (vagal) activity; respiration-linked |
| LF/HF ratio | - | Sympatho-vagal balance; increases with sympathetic dominance (stress) |
HRV as a cognitive load marker
Reduced HF HRV is one of the most replicated psychophysiological markers of mental workload. During tasks with high NASA-TLX scores, HF power decreases and the LF/HF ratio increases, reflecting sympathetic activation.
Comparison with PPG
Both ECG and PPG capture cardiac information, but differ in measurement approach:
| Feature | ECG | PPG |
|---|---|---|
| Signal type | Electrical | Optical |
| Sensor placement | Chest electrodes | Wrist / finger optical sensor |
| Motion sensitivity | Moderate | High |
| HRV accuracy | High (gold standard) | Moderate; requires beat detection from peaks |
| Invasiveness | Electrode adhesion required | Wristband |
ECG is the gold standard for HRV analysis. PPG-derived HRV is acceptable for approximate monitoring but introduces additional error from beat detection variability.
Datasets Providing ECG
| Dataset | ECG Sensor | Sampling Rate | Physical Context |
|---|---|---|---|
| WAUC | Bioharness 3 | 250 Hz | Physical activity (bike, treadmill) |
Role in Brain Foundation Models
ECG is an underrepresented modality relative to EEG in large pre-training corpora. The Cross-Modal Alignment strategy transfers knowledge from the pre-trained EEG encoder to an ECG encoder, bootstrapping its representations for cognitive state prediction without requiring a separate large ECG pre-training corpus.
PhysioWave (2025) provides the first large-scale pre-trained model for ECG alongside EEG, using multi-scale wavelet-transformer pre-training on combined physiological signal corpora.