Predictive monitoring platform

Continuous mental health monitoring through conversational avatar, vocal biomarker analysis, semantic intelligence & wearable integration

PsyAvatar is a predictive monitoring platform that fuses AI-driven avatar interaction, clinically relevant acoustic biomarkers, real-time semantic analysis of free speech, and wearable physiology — delivering continuous, multimodal risk assessment for perinatal mental health.

⌚ Wearable Integration 📊 Continuous Monitoring 🩺 Clinical Decision Support
Live Patient Dashboard
EPDS Score 16/30
Voice Risk Index 0.78
HRV Coherence 42ms
Sleep Quality 2.1/5
Semantic Risk Moderate
⚠ Multimodal alert: Voice + HRV + Sleep convergence detected
Multimodal Fusion + Predictive Layer
No single biomarker is enough. PsyAvatar combines three independent signal layers into a unified predictive risk engine — moving from reactive screening to continuous, anticipatory care.
🧠

Psychometrics + Conversational AI

Edinburgh Postnatal Depression Scale delivered through conversational avatar, enriched with semantic analysis of free speech — 17 lexical categories, VADER sentiment, and clinical flags extracted in real time.

🎙

Clinically Relevant Acoustic Biomarkers

Objective signal layer capturing pitch variability, prosody, speech rate, pause duration, and jitter — acoustic features correlated with depression severity in peer-reviewed literature.

Wearable Physiological Data

Heart rate variability, sleep architecture, stress index, and activity patterns from consumer wearables — enabling continuous monitoring outside the clinic, between clinical visits.

📈

Predictive Risk Scoring

Multimodal fusion generates a composite risk score that evolves over time — detecting deterioration trajectories before clinical threshold is reached, enabling earlier intervention.

🔄

Continuous Monitoring

Voice diary + wearable data stream between visits. Clinicians see real-time trends, not isolated snapshots — transforming episodic screening into longitudinal surveillance.

🏥

Community-First Design

Built for GP surgeries, health visitor clinics, and home self-monitoring — reducing the burden on secondary care while improving detection rates in primary care and community settings.

Wearable Biomarkers and Continuous Monitoring
Consumer wearable data combined with conversational and acoustic signals improves early detection of perinatal depression — capturing physiology between clinical visits.
Garmin
Technology Partner
Currently testing with Garmin Index™ 2 smart scale for body composition and weight monitoring, and Garmin Vívosmart® 5 for continuous HRV, sleep, stress, and activity tracking. Data integration via Garmin Health API for real-time physiological data analysis and multimodal fusion.
Index™ 2 Vívosmart® 5 Health API
💓
Heart Rate Variability

Autonomic nervous system dysregulation — reduced HRV is a validated biomarker for depression and anxiety states.

🌙
Sleep Architecture

Total sleep time, deep sleep ratio, wake episodes — disrupted sleep predicts PPD onset with high sensitivity.

Stress Index

Sympathetic/parasympathetic balance derived from photoplethysmography — captures chronic stress accumulation.

🚶
Activity & Movement

Step count, sedentary time, movement variability — psychomotor retardation is an early indicator of depressive episodes.

⚖️
Body Weight & BMI

Continuous weight monitoring tracks postpartum recovery trajectory — abnormal weight patterns correlate with disordered eating and depressive states.

🔴
Pain Score

Composite pain index derived from physiological and behavioural signals — chronic pain is a comorbidity and predictor of perinatal depression onset.

Multimodal Fusion Layer

Wearable physiology + voice biomarkers + semantic analysis converge into a single predictive score — no single modality achieves comparable sensitivity alone.

Wearable Data Stream — 7 Day View
Avg HRV (RMSSD)28ms ↓
Deep Sleep0.8h ↓
Stress Index74/100 ↑
Daily Steps2,140 ↓
Trend alert: HRV declining 3 consecutive days + sleep disruption
Clinically Relevant Acoustic Biomarkers
Voice carries objective physiological signals. PsyAvatar adds an acoustic signal layer to complement psychometric and physiological assessment — not to replace clinical judgment, but to enrich it.
F₀
Pitch Variability
Fundamental frequency range narrows in depressive states — reduced melodic contour correlates with anhedonia.
λ
Prosody
Intonation patterns, stress patterns, and rhythm — monotone speech is associated with severity of depressive episodes.
σ
Speech Rate
Words per minute, articulation rate, and syllable duration — psychomotor retardation manifests as slowed speech.
Pause Duration
Unfilled pauses, response latency, inter-utterance gaps — increased pausing reflects cognitive load and rumination.
J
Jitter
Cycle-to-cycle frequency perturbation — vocal micro-tremor increases under autonomic stress.
S
Shimmer
Amplitude perturbation — voice instability correlates with emotional dysregulation and anxiety.
H
HNR
Harmonics-to-noise ratio — vocal quality deterioration tracks with depressive severity.
E
Energy
Voice volume, loudness variability — reduced vocal energy reflects psychomotor slowing.

PsyAvatar's acoustic biomarker engine adds an objective signal layer to clinical assessment. Voice features are not used in isolation for diagnosis — they are combined with psychometric, semantic, and physiological data to generate a multimodal risk profile that supports, rather than replaces, clinical decision-making.

Real-Time Clinician Dashboard
All data streams — conversational avatar responses, vocal biomarkers, semantic analysis, and wearable physiology — converge into a single real-time clinician dashboard. Every patient interaction, every voice diary entry, every wearable signal update is immediately visible, enabling clinical action in real time rather than at the next scheduled visit.
01

Patient Triage

Automatic risk stratification across your entire caseload. Patients are ranked by composite risk score — critical cases surface immediately, reducing time-to-intervention from weeks to hours.

02

Intervention Prioritisation

Deterioration trajectories are flagged before clinical thresholds are breached. Clinicians see which patients are worsening, not just who is currently symptomatic — enabling proactive, not reactive, care.

03

Workforce Burden Reduction

Continuous remote monitoring replaces repeated in-person screening. Clinicians focus clinical time on patients who need it most — reducing routine screening burden by up to 60%.

Evidence-Based, Clinically Grounded
PsyAvatar is built on validated clinical instruments and supported by ongoing pilot studies with partners in UK and Italy.
10
Pilot Study Participants
Ongoing pilot with perinatal women — measuring voice-EPDS correlation, usability, and acceptability of conversational AI screening in community settings.
17
Semantic Categories
Bilingual (EN/IT) lexical analysis based on VADER, NRC EmoLex, LIWC, Al-Mosaiwi absolutist dictionary, and De Choudhury PPD markers — all peer-reviewed sources.
EPDS
Gold Standard Psychometrics
Edinburgh Postnatal Depression Scale — sensitivity 86%, specificity 78% for major depression. WHO-recommended screening instrument used globally.
Preliminary Outcomes

Voice-EPDS score correlation under analysis. Initial observations suggest pitch variability and pause duration show meaningful differentiation between low-risk and moderate-risk profiles. Full validation study with SCID-5 comparison planned with NHS Trust partners in 2026–2027.

What Changes for Mothers and Clinicians
PsyAvatar doesn't just collect data — it drives measurable improvements in maternal mental health outcomes.

Earlier Diagnosis

Continuous monitoring detects deterioration weeks before standard 6-week postnatal check. Predictive risk scoring identifies at-risk mothers during the critical first 14 days — when intervention is most effective.

🛡

Reduced Complications

Earlier intervention reduces escalation to severe depression, psychosis, and hospitalisation. Lower rates of self-harm, bonding disruption, and infant developmental delay — measurable improvements in NICE Quality Standards.

👩‍👧

Better Mother-Baby Outcomes

Improved maternal mood correlates with stronger bonding, better breastfeeding rates, and healthier infant attachment. Intergenerational impact: untreated maternal depression affects child cognitive and emotional development for years.

Who Pays, Who Buys, What It Costs
Perinatal mental health is a clinical priority with clearly defined payers and procurement pathways.

The Problem in Numbers

$4.2B
Global digital mental health market (2025), growing 16.5% CAGR. Perinatal segment is the fastest-growing sub-vertical.
1 in 5
Mothers experience perinatal depression. 75% are undiagnosed. £8.1B annual cost to NHS and social care (PSSRU/LSE).
50%
Of cases are missed by current paper-based EPDS screening — single-point-in-time assessment is structurally inadequate.

Who Buys PsyAvatar

🏥
Hospitals & NHS Trusts

Perinatal mental health services, maternity units, community mental health teams. SaaS license per-trust.

🏢
Insurers & Health Plans

Private insurers (Bupa, AXA, Cigna) — continuous monitoring reduces claims from maternal mental health complications.

🌍
Welfare & Public Health

NHS England, DHSC, WHO maternal health programmes, LMIC governments — population-level screening at scale.

👶
Maternal Care Programs

HMOs, employer wellness programs, maternal digital health platforms — B2B2C integration via API.

👤
Patients (D2C)

Self-monitoring subscription for at-risk mothers — freemium model with premium wearable integration.

Multidisciplinary Team
Team multidisciplinare di clinici, psicologi e sviluppatori con competenze in psichiatria, psicologia clinica, medicina generale, digital health, AI e policy sanitaria europea.
🧠

Clinical Psychologists

Specialist expertise in perinatal mental health assessment and psychometric validation.

🩺

Psychiatrists & Physicians

Clinical oversight, study design, and medical advisory from university faculty and NHS-experienced clinicians.

💻

Engineers & Data Scientists

Full-stack development, AI/ML pipeline engineering, vocal biomarker extraction, and wearable data integration.

🏛

Health Policy & Regulatory

EU and UK policy, privacy, and AI Act experts.

From Reactive Screening
to Predictive Monitoring

PsyAvatar is redefining how perinatal mental health is detected, monitored, and managed.

Access Platform Contact Us