Skip to content

Rhythm Classification Task

Overview

The objective of rhythm classification is to detect and classify abnormal heart rhythms, also known as arrhythmias, directly from ECG signals.

Characteristics

There are a variety of heart rhythms that can be detected using ECG signals. In this task, we predominantly focus on detecting arrhythmias such as atrial fibrillation (AFIB) and atrial flutter (AFL). The following table summarizes characteristics of a few rhythms:

Sinus rhythm (SR) is a type of rhythm where the heart beats regularly and originates from the sinus node. SR is considered normal and is typically characterized by the following:

  • Regular rhythm
  • P waves present
  • QRS complexes usually < 120ms
  • Ventricular rate 60-100 bpm

Atrial fibrillation (AFIB) is a type of rhythm where the atria (upper chambers of the heart) beat irregularly and out of sync with the ventricles (lower chambers of the heart). AFIB is the most common type of rhythm and can lead to serious complications such as stroke and heart failure. AFIB is typically characterized by the following:

  • Irregularly irregular rhythm
  • No P waves
  • Variable ventricular rate
  • QRS complexes usually < 120ms
  • Fibrillatory waves may be present

Atrial flutter (AFL) is a type of rhythm where the atria (upper chambers of the heart) beat regularly but faster than normal. AFL is less common than AFIB and can lead to serious complications such as stroke and heart failure. AFL is typically characterized by the following:

  • Narrow complex tachycardia
  • Regular atrial activity at ~300 bpm
  • Loss of the isoelectric baseline
  • “Saw-tooth” pattern of inverted flutter waves in leads II, III, aVF
  • Upright flutter waves in V1 that may resemble P waves
  • Ventricular rate depends on AV conduction ratio

Supraventricular tachycardia (SVT) is a type of rhythm where the heart beats faster than normal due to abnormal electrical signals originating above the ventricles. SVT is typically characterized by the following:

  • Narrow complex tachycardia
  • Regular rhythm
  • P waves may be absent, hidden, or retrograde
  • QRS complexes usually < 120ms
  • Ventricular rate > 100 bpm

Sinus bradycardia (BRAD) is a type of rhythm where the heart beats slower than normal due to abnormal electrical signals originating from the sinus node. BRAD is typically characterized by the following:

  • Regular rhythm
  • P waves present
  • QRS complexes usually < 120ms
  • Ventricular rate < 60 bpm

Pre-trained Models

The following table provides the latest performance and accuracy results for rhythm models. Additional result details can be found in Model Zoo → Rhythm.

NAME DATASET FS DURATION # CLASSES MODEL PARAMS FLOPS METRIC
ARR-2-EFF-SM Icentia11K, PTB-XL, LSAD 100Hz 5s 2 EfficientNetV2 18K 1.2M 99.5% F1
ARR-4-EFF-SM LSAD 100Hz 5s 4 EfficientNetV2 39K 2.2M 95.3% F1

Target Classes

Below outlines the classes available for rhythm classification. When training a model, the number of classes, mapping, and names must be provided.

CLASS LABEL DESCRIPTION
NSR 0 Normal sinus rhythm
SBRAD 1 Sinus bradycardia
STACH 2 Sinus tachycardia
SARRH 3 Sinus arrhythmia
SVARR 4 Supraventricular arrhythmia
SVT 5 Supraventricular tachycardia
VTACH 6 Ventricular tachycardia
AFIB 7 Atrial fibrillation
AFLUT 8 Atrial flutter
VFIB 9 Ventricular fibrillation
VFLUT 10 Ventricular flutter
BIGU 11 Bigeminy (every other beat is PVC)
TRIGU 12 Trigeminy (every third beat is PVC)
PACE 13 Paced rhythm
NOISE 127 Noise

Class Mapping

Below is an example of a class mapping for a 2-class rhythm model. The class map keys are the original class labels and the values are the new class labels. Any class not included will be skipped.

{
    "num_classes": 2,
    "class_names": ["NSR", "AFIB"],
    "class_map": {
        "0": 0,  // Map None to None
        "7": 1,  // Map AFIB to AFIB
        // Skip remaining classes
    }
}

References