What we are building
Four apps, one shared methods stack.
Two apps are live: Grovalin for contextual learning, and Volitiq for clinic-grade motor phenotyping. Two more (WAVE and PULSE) are in active development under the HEARTH program, extending the lab's work from the clinic into the home. Each app draws from a shared methods stack (computer vision, deep learning, LLMs and NLP, and rigorous statistical modelling), chosen per problem rather than picked from a single playbook.
Grovalin
grovalin.com · contextual learning
The challenge. Children with autism often struggle to understand that the same word (e.g. "chair") applies to objects that look different. Traditional learning materials use generic stock images that don't match a child's actual home environment.
The approach. Grovalin turns familiar objects from a child's home into personalised learning material. Caregivers upload home videos or images; the platform extracts and segments objects, generates colour and structural variations, and surfaces them through child-facing modes (Explore, Game, Story, LifeSkills).
Grounded in VR cognitive-rehabilitation research showing significant improvements in contextual processing and cognitive flexibility in children with ASD (PMC3845243).
Volitiq
volitiq.com · motor phenotyping
The challenge. Motor differences are documented in 34–80% of autistic individuals, but measuring them objectively is difficult. Traditional motor assessment requires expensive equipment and trained specialists, gating research to a small set of clinics.
The approach. Volitiq turns any device with a camera into a research-grade motor assessment tool. Computer vision analyses motor movements from standard webcam video; researchers can capture quantitative data about movement patterns without specialised equipment, making motor research accessible.
Informed by digital phenotyping literature on postural sway and atypical head movement patterns in ASD toddlers (PMC6242931). Statistical layer feeds directly into the GAMM + GP-normative pipeline used across the lab.
The HEARTH program
Passive sensing for the home
HEARTH (Home Environment Ambient sensing for Real-Time Health) is a research program for passive, contact-free monitoring of older adults and people with disabilities. It rejects the trade-off that current home monitoring asks of vulnerable people: wear a device, accept a camera, step on a mat in order to be cared for.
HEARTH operates two complementary modalities. WAVE reads the body from radio. PULSE reads it from light. Their physical principles set their blind spots against each other in useful ways.
WAVE
WiFi Ambient Vitality and Event-detection
Every modern WiFi link carries a continuous estimate of the wireless channel, called Channel State Information (CSI), that is sensitive to every scattering object in the link's environment, the human body included. WAVE reads CSI to recover motion, posture, gait, falls, occupancy, and coarse respiratory rate, with no camera, no wearable, and no microphone.
Built on the cross-modal supervision paradigm behind Person-in-WiFi and DensePose-from-WiFi. Forward-compatible with IEEE 802.11bf (ratified October 2024), so the same algorithms run on commodity 802.11bf silicon as it ships.
PULSE
Passive Unobtrusive Light-based Signal Extraction
Each heartbeat sends a bolus of hemoglobin into cutaneous microcirculation, producing a small but predictable modulation of skin colour (0.5 to 2% of mean reflectance) that is below the threshold of human perception but recoverable from ordinary RGB video. PULSE extracts a continuous pulse waveform from face or hand and derives heart rate, heart rate variability, and respiratory rate.
Where WAVE penetrates walls but cannot resolve a heartbeat, PULSE resolves the heartbeat exquisitely but needs the person in the camera's field of view. Operated together, they cover each other's blind spots. Built on state-of-the-art rPPG methods (PhysFormer, CodePhys, Spiking-PhysFormer).
Roadmap
What's next
Beyond the four apps above, the lab is exploring further AI tools in adjacent areas.
Speech Pattern Analysis
AI-powered analysis of speech patterns, prosody, and communication style to support research and therapy.
Emotion Recognition Support
Tools to help children learn to recognise and understand emotions through interactive AI-guided practice.
Help us build the future
Whether you are a developer, researcher, clinician, or family member, your contribution matters.