CODO Monitor User Manual (English)

0. Purpose of this document

This manual explains how to operate CODO Monitor for real-time HRV monitoring from ECG, including adaptive (personalized) alerting, manual annotation/exclusion during care, and multi-scale visualization. CODO Monitor is described as a standalone executable compiled with MATLAB AppDesigner; the distributed package includes MATLAB Runtime, enabling cross-platform use (Windows/macOS) without a MATLAB license or separate Runtime installation.

Note: This manual is about software operation. It does not provide medical advice and must not be used as a substitute for clinical judgment.

1. Quick start

1.1 Before monitoring

1. Enter patient name or ID

2. Select the audio/ECG input source

3. (Optional) Select the interval (e.g., log save interval) and analysis window (1 min / 5 min)

4. Press Start

1.2 During monitoring

• Press Care/Data Exclusion when care/artifacts occur; press again when completed

• Add a Memo (event note)

• Choose indices to display; use log scale and frequency options as needed

• Press More Graphs to display additional figures

• Configure/enable Alert (and sensitivity)

• Use Save graphs / Save figure when you need snapshots

• Use Trend for long-term visualization

2. Main window

The main window includes:

• Upper panel: settings and control buttons

• Top-right: real-time ECG waveform with detected R-wave peaks (red dots)

• Six lower panels: selected HRV indices

  • Outliers exceeding personalized thresholds are highlighted in red with dashed red lines

  • Data recorded during “Care/Data Exclusion” periods are shown in blue 

3. Care/Data Exclusion and Memo (annotation)

3.1 Care/Data Exclusion (critical for robust alerts)

CODO Monitor provides a workflow-integrated manual annotation mechanism:

• During nursing care/movement/artifacts, you can flag periods via a UI button.

• These points remain visible (distinct color) but are excluded from the quartile calculations (Q1/Q3) used by the adaptive alert algorithm, preventing artifacts from corrupting personalized thresholds.

• Annotated intervals are still displayed without interpolation so temporal relationships can be inspected.

3.2 Memo

Enter event notes (e.g., medication, cry). Example memo sheet entries are shown in the supplemental output examples.

4. Selecting indices, log scale, and HF range

4.1 Supported indices (representative list)

CODO Monitor is described as supporting: pNN50, SDNN, RMSSD, SDNN/RMSSD; Poincaré SD1/SD2/SD1/SD2; and VLF/LF/HF and their ratios.

4.2 More Graphs window

The “More Graphs” window:

• is accessible via the More Graphs button,

• allows selecting a larger set of indices using checkboxes,

• provides options to adjust HF calculation range and switch to logarithmic scale for some frequency-domain indices.

4.3 Trend window

The “Trend” window displays long-term trends (e.g., multiple hours) with the personalized normal range indicated by red dashed lines.

5. Adaptive alert

5.1 Algorithm (IQR-based personalized thresholds)

Thresholds are defined using quartiles (Q1, Q3) and IQR:

• Low threshold = Q1 − α(Q3 − Q1)

• High threshold = Q3 + α(Q3 − Q1)
  CODO JMS_R3

α is user-selectable (paper implementation: 1.724 / 2.465 / 3.207), roughly corresponding to ~3σ / 4σ / 5σ assumptions; smaller α is more sensitive, larger α more conservative.

5.2 Visualization

Outliers are highlighted in red; care/exclusion points are in blue on the time-series plots.

6. Saving and export

6.1 Automatic export formats

All calculated indices, alert events, and manual annotations are automatically logged and exported in .xlsx and .mat formats.
The system “systematically logs all computed HRV indices, timestamps, alert occurrences, and manual annotations” in .xlsx.

6.2 Excel examples (supplemental)

• Exported Excel log example (2-min interval): includes data_time, care, memo, RRI, HR, …, Total Power, Total Power (HF<1Hz)

• Memo sheet example: Index / Time / Memo

• Alert sheet example: Time / per-index alert flags / per-index values

7. Troubleshooting (common)

• No ECG / unstable R peaks: verify correct input device/source, signal amplitude (avoid saturation), and noise/artifacts. Input ECG quality and reliable R-wave detection are prerequisites for correct analysis.

• Too many alerts: increase α (more conservative), ensure Care/Data Exclusion is used during artifacts.

• Export not created: check output folder permissions, disk space, and file naming constraints.