Waveform Asymmetry as a Biomarker of Neural Aging

Code and analysis pipeline for:

Waveform Asymmetry as a Biomarker of Neural Aging: Spatial Degradation of Oscillatory Cycle Shape Across Two Independent Cohorts Theodor Spiro (Vaika Inc., East Aurora, NY, USA) Under review at Frontiers in Aging Neuroscience Preprint: paper/preprint.pdf · DOI: 10.5281/zenodo.19912202 Contact: tspiro@vaika.org

---

Paper summary

The shape of each neural oscillatory cycle - quantified by peak-trough asymmetry (PTA), the ratio of rise to fall time - may reflect excitatory-inhibitory balance in the generating circuit. We measured broadband PTA across five frequency bands in resting-state EEG from 215 adults (LEMON, age 20-77) and replicated all findings in 608 adults (Dortmund Vital Study, age 20-70), including 208 with 5-year longitudinal follow-up.

Three findings:

1. Asymmetry is band-specific and spatially structured - slow rhythms show excitatory-type asymmetry, fast rhythms show inhibitory-type asymmetry, with a spatial double dissociation between posterior alpha and central beta.
2. Beta-band asymmetry decreases with age (LEMON: r = -0.326, d = 0.69; Dortmund: r = -0.314, d = 0.75), surviving control for aperiodic slope. Effect size exceeds classical alpha slowing. Longitudinal data confirmed the predicted direction (mean Δ = -0.017, r = 0.45 test-retest).
3. Spatial organization degrades with age, and theta spatial entropy predicts memory performance independent of age.

Bottom line: Waveform shape provides a power-independent measure of neural aging that complements existing spectral biomarkers, accessible from standard 3-minute EEG recordings.

---

Quick start

git clone https://github.com/mool32/waveform-asymmetry-aging.git
cd waveform-asymmetry-aging
pip install -r requirements.txt

# Main analysis script - PTA computation across LEMON dataset
python src/phase1/run_pta.py

# Verify every statistic cited in the paper
python scripts/verify_numbers.py

Main analysis entry points

Script	Purpose
src/phase1/run_pta.py	Main PTA pipeline (LEMON, broadband across 5 bands)
scripts/analyze_dortmund_correlations.py	Dortmund replication (N=608)
scripts/analyze_longitudinal.py	5-year longitudinal analysis (N=208)
scripts/run_age_prediction.py	Age-prediction Ridge models (PTA vs spectral)
scripts/generate_paper_figures.py	Regenerate all 6 main + 2 supplementary figures
scripts/verify_numbers.py	Cross-check every number in the manuscript

---

Repository structure

.
├── README.md                         # This file
├── paper/
│   ├── preprint.pdf                  # Full preprint with figures (24 pp)
│   ├── manuscript_submission.md      # Source markdown
│   └── frontiers/                    # Frontiers submission package
├── src/
│   ├── phase1/
│   │   ├── run_pta.py                # ★ Main PTA pipeline
│   │   ├── run_eceo_entropy.py       # Eyes-closed vs open
│   │   ├── statistics.py             # Statistical tests
│   │   └── visualize.py              # Plotting
│   ├── utils/
│   │   └── asymmetry.py              # Core PTA computation
│   └── phase4/
│       └── meta_analysis.py          # Cross-scale meta-analysis (33 systems)
├── scripts/
│   ├── analyze_longitudinal.py
│   ├── analyze_dortmund_correlations.py
│   ├── run_age_prediction.py
│   ├── generate_paper_figures.py
│   └── verify_numbers.py
└── results/
    ├── phase1_full/                  # LEMON outputs (summary.json)
    ├── dortmund_replication/         # Dortmund cross-sectional
    ├── dortmund_longitudinal/        # 5-year follow-up
    ├── eceo_entropy/                 # Eyes-closed/open
    └── figures/                      # Generated figures (PDF/PNG)

---

Datasets

• LEMON (MPI Leipzig Mind-Brain-Body) - N=215, age 20-77
• Dortmund Vital Study (OpenNeuro ds005385) - N=608, age 20-70, including N=208 with 5-year longitudinal follow-up

Both datasets are openly licensed; see source links for terms.

---

Citation

If you use this code or build on these findings, please cite:

@article{Spiro2026waveform,
  author  = {Spiro, Theodor},
  title   = {Waveform Asymmetry as a Biomarker of Neural Aging:
             Spatial Degradation of Oscillatory Cycle Shape Across
             Two Independent Cohorts},
  year    = {2026},
  doi     = {10.5281/zenodo.19912202},
  url     = {https://github.com/mool32/waveform-asymmetry-aging}
}

---

Requirements

Python 3.10+. Dependencies in requirements.txt: numpy, scipy, matplotlib, mne, pandas, statsmodels, scikit-learn.

---

Contact

Theodor Spiro - Vaika Inc., East Aurora, NY, USA Email: tspiro@vaika.org

---

License

MIT License - free to reuse with attribution.