AI-Powered Mental Health Innovation

MATHI

Reducing Auditory Hallucinations Through Psychoacoustic Masking

A $14 device using Edge AI to help millions—no phone, no internet, complete privacy

60-80%
Prevalence
~$14
Device Cost
1-4 kHz
Target Band
100%
Private
Understanding the Condition

What are Auditory Hallucinations?

Auditory hallucinations are experiences of hearing speech when no external sound is present—distinct from illusions or simple tones

The Experience

People hear voice-like content—words, phrases, commentary, or commands—with speech-like timbre and temporal structure including syllabic rhythm.

May feel internal or external in location
Carries speech-like qualities and rhythm
Distinct from misinterpreting real sounds
The Scale

Lifetime occurrence is 60-80% in schizophrenia, with onset often beginning in late adolescence and frequent persistence or recurrence.

Affects majority of individuals with schizophrenia
Often begins in late adolescence
Persistence and recurrence are common
Three Dimensions of Burden
Impact determined by distress, uncontrollability, and interference—not just frequency

Distress

Emotional suffering and aversiveness that changes rapidly with stress, arousal, and attentional capture

Uncontrollability

Difficulty redirecting attention or dampening the experience, leading to feelings of helplessness

Interference

Impact on sleep, conversations, work, and daily functioning—disrupting quality of life

Comorbid conditions are common: Sleep disturbance, anxiety, depression, social withdrawal, and elevated self-harm risk frequently accompany auditory hallucinations. Any practical aid must respond to rapid within-person variability, not just average severity.

The Challenge

A Global Mental Health Crisis

The public health need is large and unequal—access to skilled care is critically limited

60-80% Prevalence

Auditory hallucinations affect the majority of individuals with schizophrenia, yet treatment remains out of reach for millions worldwide.

<1 per 100K

Severe shortage of psychiatrists in low-resource regions creates massive treatment gaps. Follow-up care is expensive and inaccessible.

High Costs

Direct costs (fees, travel), opportunity costs (missed work), and hidden expenses (data plans, maintenance) make care unsustainable.

Why Current Solutions Fail
Existing tools don't meet the needs of those who need them most

Expensive & Proprietary

Imported devices are priced beyond local means and tied to proprietary ecosystems unavailable in under-resourced settings

Connectivity Dependent

Phone-centric tools require data plans, maintenance, and battery management—adding hidden costs and complexity

Not Adaptive

Fixed presets with limited control don't adapt to moment-to-moment distress, context, or preference

Safety Concerns

Headphones reduce environmental awareness and can be uncomfortable over long sessions

Our Solution

MATHI: Psychoacoustic Masking with Edge AI

A clinic-independent, privacy-preserving device that uses targeted sound and adaptive AI to reduce the salience of auditory hallucinations

What is MATHI?

MATHI is a compact, standalone device that performs sensing, on-device personalization, and audio rendering locally—no phone or network required. It uses psychoacoustic masking principles to reduce the salience of voice-like hallucinations while preserving comfort, safety, and complete privacy.

~$14 Cost

Ultra-affordable for global deployment

100% Offline

Works anywhere, no connectivity needed

Complete Privacy

All processing on-device, no data shared

How It Works

MATHI combines established psychoacoustic principles with real-time Edge AI adaptation

Speech-Band Targeting
1-4 kHz Frequency Emphasis

Concentrates energy in critical frequency bands where speech information resides (≈1-4 kHz within 0.25-6 kHz passband), using ERB-scale filters to improve signal-to-masker ratio in the auditory system.

Temporal Modulation
4-16 Hz Amplitude Modulation

Shallow low-rate modulation (≈4-16 Hz with slight jitter) targets syllabic-rate speech cues while remaining unobtrusive through smooth transitions, competing with voice-like content without distraction.

Edge AI Personalization
Real-Time Adaptive Control

On-device Edge AI adjusts parameters based on distress ratings (0-10 scale) and acoustic context with <20ms latency. Closed-loop controller personalizes masking from brief user feedback while preserving safety constraints—all processing happens locally on the embedded processor.

Privacy-Preserving Design
On-Device Processing Only

All processing happens locally on the device. No raw audio is stored or transmitted. Only coded scalar states (parameters, ratings, exposure counters) are retained. No internet required—complete privacy guaranteed.

6-Stage Signal Processing Pipeline
From carrier generation to audio output
1
Carrier Generation
Pink/brown noise baseline
2
Spectral Shaping
ERB filter 1-4 kHz emphasis
3
Temporal Modulation
4-16 Hz AM with jitter
4
Level Control
Constant LAeq tracking
5
Safety Limiting
Exposure budgeting & limiting
6
Audio Output
Near-ear open-ear speaker
Key Innovation

Three Pillars of Innovation

What makes MATHI uniquely positioned to bridge the treatment gap

1. Technology
Evidence-Based Science
Grounded in psychoacoustic masking principles
Edge AI adapts to individual needs in real-time
Validated through rigorous AB/BA crossover study
2. Accessibility
Designed for Equity
~$14 cost makes it globally accessible
Works offline—no phone, network, or clinic needed
Low maintenance, usable in crowded homes
3. Privacy
Complete Data Protection
All processing happens on-device locally
No raw audio stored or transmitted ever
Open-ear speaker preserves situational awareness
Scientific Foundation

Evidence-Based Design

Grounded in established psychoacoustic principles and validated through rigorous research

Research Validation
Rigorous methodology with level-matched controls

Methodology

  • Randomized crossover trial with counterbalanced order
  • Level-matched adaptive vs. static baseline comparison
  • Controlled washout period between conditions
  • Prespecified distress reduction as primary outcome

Outcomes

  • Primary: Mean within-block distress change (0-10 NRS)
  • Secondary: Controllability, comfort, distraction
  • Device logs: Exposure and controller stability
Psychoacoustic Foundations
Theory-led rationale for sound-based intervention
Energetic Masking

Overlapping frequency content within critical bands reduces signal-to-masker ratio, making voice-like content less salient

Modulation Masking

Shallow low-rate AM interacts with slow envelope fluctuations carrying syllabic-rate cues in perception

Attention Anchoring

Stable external auditory anchor may reduce attentional capture by internal percepts (plausible interpretation)

UN Sustainable Development Goals

Aligned with Global Goals

MATHI directly contributes to achieving the United Nations Sustainable Development Goals

SDG 3
Good Health and Well-Being

MATHI promotes mental health and well-being by providing accessible, evidence-based support for individuals experiencing auditory hallucinations in schizophrenia.

Reduces distress and improves daily functioning for 60-80% of affected individuals
Provides continuous, on-demand mental health support without clinic visits
Evidence-based psychoacoustic approach validated through rigorous research
SDG 10
Reduced Inequalities

By prioritizing affordability, offline functionality, and privacy, MATHI bridges the mental health treatment gap in underserved and low-resource communities worldwide.

Ultra-affordable ~$14 cost makes technology accessible to all income levels
Works offline without phone or internet—ideal for regions with limited infrastructure
Addresses healthcare disparities where psychiatrist access is <1 per 100,000

By aligning with SDG 3 (Good Health and Well-Being) and SDG 10 (Reduced Inequalities), MATHI demonstrates how innovative technology can create meaningful social impact and contribute to a more equitable, healthier world for all.

Meet the Innovator

Suhan D

SD

Suhan D

Student Researcher & Innovator

CHIREC International School

Hyderabad, Telangana, India

suhandhandapani@gmail.com

Driven by a passion for accessible mental health technology and equitable healthcare solutions for underserved communities.

MATHI demonstrates how rigorous psychoacoustic theory combined with Edge AI can provide meaningful support for individuals living with schizophrenia—especially in low-resource settings where the need is greatest.

"Technology should empower those who need it most"

The Vision

Bringing accessible mental health technology to millions worldwide

Together, we can bridge the treatment gap and transform lives through innovation, accessibility, and compassion