Multimodal Biosignal Preprocessing Pipeline

Reproducible data engineering across EEG, ECG, and wearable HAR datasets for downstream self-supervised learning.

Last updated on April 12, 2026

This project implements a fully reproducible pipeline for downloading, preprocessing, and validating five open-access multimodal time-series datasets (PAMAP2, WISDM, mHealth, EEGMMIDB, and PTB-XL), in preparation for downstream self-supervised learning workflows.

The pipeline scored 3rd out of 17 submissions in a competitive technical assessment for a Research Assistant post at Imperial College London, with the preprocessing plan rated the most thoroughly reasoned of all submissions.

  • EEG
  • ECG
  • Wearable HAR
  • MNE
  • NumPy
  • Self-supervised learning

Datasets and processing

Dataset Signal Processing
PAMAP2, WISDM, mHealth Wearable activity recognition (HAR) Harmonised to a common 20 Hz representation with a shared six-channel accelerometer/gyroscope schema and unified class label taxonomy, enabling a single model to train across datasets.
EEGMMIDB EEG (PhysioNet EEG Motor Movement/Imagery Database) Preprocessed using MNE, with event-aligned 4-second epochs extracted from motor imagery runs.
PTB-XL 12-lead ECG Ingested via the PhysioNet AWS S3 mirror, bandpass filtered, and split into patient-safe train, validation, and test folds.

Outputs

All outputs are stored as float32 NumPy arrays in a consistent [N, C, T] format alongside structured metadata CSVs covering subject provenance, label mappings, sampling rates, channel schemas, and QC flags.

Validation

A validation script checks:

  • Array integrity
  • Label distributions
  • Subject-level leakage controls
  • HAR harmonisation across datasets