The Science Behind REWRITE

Vocal biomarker analysis is backed by decades of peer-reviewed research. Here's how it works, where the science comes from, and what we measure.

What Are Vocal Biomarkers?

Your voice carries hundreds of measurable signals beyond the words you say. Pitch, rhythm, micro-tremors, harmonic patterns, and spectral characteristics all shift with your emotional and cognitive state — often before you're consciously aware of the change.

These signals are called vocal biomarkers — measurable acoustic features that correlate with emotional states such as stress, confidence, and cognitive load. Unlike self-reported surveys, vocal biomarkers are embedded in natural speech and are difficult to consciously control, making them a valuable complement to traditional coaching tools.

REWRITE analyzes 151+ distinct vocal signals per session, drawing from the full acoustic spectrum to surface patterns worth reflecting on — then presents them as five intuitive dimensions in the Emotional Pentagon™.

Research Foundations

Vocal biomarker science has been studied across multiple disciplines — clinical psychology, neuroscience, computational linguistics, and affective computing. The following peer-reviewed work informs our approach:

Voice & Emotional State Detection

Schuller, B. et al. (2013). "Computational Paralinguistics: Emotion, Affect and Personality in Speech and Language Processing." Wiley. Foundational text on detecting emotional states from acoustic features in speech.
doi.org/10.1002/9781118706664

Stress Detection via Voice

Lu, H. et al. (2012). "StressSense: Detecting Stress in Unconstrained Acoustic Environments Using Smartphones." Dartmouth College / UbiComp 2012. Demonstrated that voice-based stress detection is feasible from everyday speech, even in unconstrained environments.
doi.org/10.1145/2370216.2370270

Vocal Markers of Depression & Mental Health

Cummins, N. et al. (2015). "A Review of Depression and Suicide Risk Assessment Using Speech Analysis." Speech Communication, Vol. 71. Meta-review of acoustic features (pitch, energy, formants, timing) as markers for emotional distress.
doi.org/10.1016/j.specom.2015.03.004

Confidence & Persuasion in Voice

Guyer, J. J. et al. (2021). "Speech Rate, Intonation, and Pitch: Investigating the Bias and Cue Effects of Vocal Confidence on Statement Credibility." Journal of Language and Social Psychology. Demonstrates acoustic correlates of perceived confidence.
doi.org/10.1177/0261927X18795924

Cognitive Load & Speech Patterns

Yin, B. & Chen, F. (2007). "Towards Automatic Cognitive Load Measurement from Speech Analysis." Human-Computer Interaction (HCI International). Early work linking speech patterns (pause duration, articulation rate) to cognitive load levels.
doi.org/10.1007/978-3-540-73216-7_111

Longitudinal Vocal Change

Cannizzaro, M. et al. (2004). "Voice Acoustical Measurement of the Severity of Major Depression." Brain and Cognition, Vol. 56(1). Demonstrated that vocal acoustic features change measurably over the course of treatment.
doi.org/10.1016/j.bandc.2004.05.003

Our Technology Partner: Emotion Logic

REWRITE's vocal analysis engine is built on technology from Emotion Logic, a voice analytics platform that has spent over a decade developing algorithms for detecting emotional states from speech signals.

Emotion Logic's engine processes voice input through proprietary spectral analysis, extracting biomarker patterns from features including:

  • Fundamental frequency (F0) variation — correlates with stress and emotional arousal
  • Jitter and shimmer — micro-perturbations linked to cognitive load and fatigue
  • Harmonic-to-noise ratio (HNR) — vocal clarity associated with confidence
  • Speech rate and pause patterns — rhythm changes indicating processing effort
  • Formant trajectories — articulatory shifts reflecting emotional engagement

These raw acoustic features are synthesized into the five dimensions of the Emotional Pentagon™ — making complex signal data immediately understandable.

The Emotional Pentagon™

REWRITE distills 151+ voice signals into five key dimensions, visualized as a pentagon shape that evolves over time:

⚡ Stress

Fight-or-flight activation detected through pitch variability and frequency perturbation. Lower scores indicate a calmer vocal state.

🛡️ Confidence

Vocal certainty measured via harmonic stability, pitch consistency, and speech rate control. Higher scores reflect more self-assured speech.

🔥 Engagement

Emotional presence and connection assessed through energy variation and prosodic dynamics. Higher engagement means you're present, not just going through the motions.

🧠 Cognitive Load

Mental processing strain detected via pause frequency, speech rate irregularity, and filled hesitations. Lower scores indicate clearer, less burdened thinking.

🎯 Authenticity

Alignment between what you say and how your voice behaves, assessed through consistency of vocal features across the session. Higher scores suggest greater emotional congruence.

Each session generates a new pentagon. Over time, you see your shape change — stress shrinking, confidence expanding, engagement deepening. That's transformation you can see, not just feel.

How We Process Your Voice

1. Capture

Your voice is recorded during coaching sessions through the REWRITE app. Audio is captured at high fidelity for acoustic analysis.

2. Analyze

Raw audio is processed through Emotion Logic's engine, extracting 151+ acoustic features in real time. Raw audio is discarded after analysis.

3. Synthesize

Extracted features are mapped to the five Emotional Pentagon™ dimensions using validated scoring algorithms.

4. Visualize

Results appear as your personal Emotional Pentagon™ — a shape you can read at a glance, compared against your own history.

Data handling: Raw voice audio is discarded immediately after analysis — only derived metrics (the pentagon scores and trend data) are retained to track your growth over time. You control your data, and you can request deletion at any time.

Data Security & Compliance

Your voice data is processed with the same rigor applied to the research behind it.

🔒 SOC 2 Type II Certified analysis platform
🇪🇺 GDPR Compliant EU data protection standards
🔐 AES-256 Encryption Data encrypted at rest & in transit

Important: REWRITE is a personal-growth and coaching-support tool. It is not therapy, diagnosis, or treatment. If you are in crisis, contact the 988 Suicide & Crisis Lifeline (call or text 988).

What We Don't Claim

Transparency matters. Here's what REWRITE does not do:

  • We don't detect lies. Voice analysis surfaces emotional patterns — it does not determine truth or deception.
  • We don't diagnose. REWRITE is not a medical or clinical tool. It does not diagnose depression, anxiety, or any condition.
  • We don't replace therapy. If you need professional mental health support, please work with a licensed therapist. REWRITE is a complement to coaching and self-reflection, not a substitute for care.
  • We don't guarantee outcomes. Vocal biomarkers provide data points for reflection. Personal growth is complex, and no single tool can promise transformation.

We believe honest tools build trust. That's why we show you exactly what we measure, how we measure it, and what it means.

Hear What Your Voice Has Been Saying

Start with three free sessions — no account, no credit card. Your voice tells a story. Let's listen.

Begin Your REWRITE →