Mendi's neurofeedback technology is built on decades of peer-reviewed fNIRS research.
fNIRS stands for functional Near-Infrared Spectroscopy, a non-invasive brain imaging technology that measures brain activity by detecting changes in blood oxygenation. Mendi uses fNIRS to measure blood oxygenation in your prefrontal cortex: the part of your brain that controls focus.
Mendi emits safe near-infrared light at dual wavelengths: 660nm and 805nm that passes safely through the forehead, reaching prefrontal cortex tissue below the skull without radiation or electrodes.
As light passes through brain tissue, the amount absorbed differs between oxygenated and deoxygenated blood. Active brain regions draw more oxygenated blood, revealing changes in prefrontal cortex activation.
Mendi's processor converts raw light data into a live hemodynamic signal. That signal feeds directly into the training game, translating prefrontal hemodynamic changes into a live feedback signal.
Mendi emits safe near-infrared light at dual wavelengths: 660nm and 805nm that passes safely through the forehead, reaching prefrontal cortex tissue below the skull without radiation or electrodes.
As light passes through brain tissue, the amount absorbed differs between oxygenated and deoxygenated blood. Active brain regions draw more oxygenated blood, revealing changes in prefrontal cortex activation.
Mendi's processor converts raw light data into a live hemodynamic signal. That signal feeds directly into the training game, translating prefrontal hemodynamic changes into a live feedback signal.
Improvement in sustained focus
42%
Reduction in perceived stress
31%
Improved mood and emotional balance
16%
Better general wellbeing
51%
Results are based on aggregated in-app data and self-reported outcomes from thousands of users who completed eight weeks of Mendi neurofeedback training. Improvements were measured using validated scales (ACS, WHO-5, PSS-10, GAD-7, PHQ-9). Individual results may vary.
Mendi is built on fNIRS neurofeedback principles studied in independent research. Some of these studies used Mendi directly; the rest establish the science behind it.
Independent research confirms the Mendi sensor accurately measures real prefrontal-cortex activity. The signal you train with is validated.
Mendi's sensor validated for measuring working-memory load
Independent validation showing the mobile Mendi fNIRS device reliably detects differences in oxyhemoglobin response between low and high working-memory loads in the prefrontal cortex. It confirms the sensor measures what it claims to.
Mendi detected the cognitive cost of multitasking
31 undergraduates played Tetris under single-task versus multi-task conditions while wearing the Mendi headband. Subjective load rose and performance dropped under multitasking, the study shows Mendi tracks that load-versus-disengagement dynamic.
Mendi works on the same principle these studies test: fNIRS neurofeedback. Run on other equipment, they show the method itself works. It's the science Mendi is built on.
One note for all four: they used fNIRS devices other than Mendi, so they back the method rather than Mendi specifically. Designs and sample sizes vary, the specifics are on each card.
A single session of fNIRS-neurofeedback changed brain activity during a later working-memory task
A randomized sham-controlled study of the left dorsolateral prefrontal cortex. Versus sham, the training group showed enhanced DLPFC activity; behavioural working-memory performance did not significantly differ. The neural mechanism responded to a single session; behaviour did not transfer in this design.
Reduced impulsive errors after 8 sessions in highly impulsive adults
20 adults high in subclinical impulsivity were randomised to fNIRS-neurofeedback or an EMG-control, completing 8 sessions over 2 weeks in a VR classroom. The neurofeedback group showed significantly fewer commission errors and increased prefrontal oxygenation.
Practical and well-tolerated across a 12-session program in children
A 3-arm pilot (9 children per arm). After 12 NIRS-neurofeedback sessions, the group showed within-group attention improvements at 4 weeks and 6 months, with no significant between-group differences. Framed as feasibility, practical and well-tolerated, not efficacy.
A single training session produced working-memory gains that held at one week
A randomized sham-controlled study (60 adults) targeting the DLPFC. Versus sham, the training group showed increased DLPFC activation and improved spatial working-memory and attention, maintained at one week. A direct contrast to Yang: single-session results vary across studies.
The research here includes studies that used Mendi directly and studies on the same fNIRS method using other devices. Mendi is a wellness device, not a medical treatment. It isn't intended to diagnose, treat, cure, or prevent any condition.
Built for short and long sessions. Low-profile, ergonomic, and comfortable on the forehead.
Mendi is safe for regular use, and meets the electromagnetic interference standards.
Research-validated measurement, built for daily use.
Signal validated against research-grade fNIRS in peer-reviewed studies
Traditional fNIRS systems cost tens of thousands of dollars and require lab environments. Mendi brings the same measurement technology to a device you wear at home.
Lab only
$30,000–$100,000+ per system
Requires trained specialist to operate
Bulky cap, lab environment only
Daily use
Consumer-accessible price point, no subscription
Automatic setup: detects skin contact, starts recording
Lightweight, wearable for daily sessions
Field-ready
Mental workload assessment outside the lab
Longitudinal sustained attention studies
Prefrontal cortex cognitive performance tracking
Available to universities, clinical labs, and research institutions. We support co-research, data-sharing, and custom integrations.
Co-research agreements, joint publications, and data-sharing protocols for academic and clinical programs.
Structured field data collection, multi-participant management, and IRB-compatible governance.
API docs, SDK access, and direct engineering support for custom integrations and signal processing pipelines.
