Sleep compression technology
A wearable system that detects slow-wave sleep via dry EEG and delivers low-intensity focused ultrasound to enhance glymphatic clearance—potentially achieving full restorative sleep in less time.
Slow-wave sleep is not uniformly efficient. If we can mechanically amplify the natural CSF oscillation window, we can compress the sleep requirement.
During deep sleep, the brain's glymphatic system opens, allowing cerebrospinal fluid to wash through perivascular spaces and clear metabolic waste. This process is gated by slow oscillations—waves that occur approximately every 15–25 seconds, varying by individual.
Natural short-sleepers with DEC2 and NPSR1 mutations achieve full cognitive restoration in 4–6 hours. Their sleep is not just shorter—it is qualitatively more efficient at waste clearance per hour. Somnison attempts to mechanically induce this phenotype by enhancing glymphatic influx during the precise window when it naturally occurs.
Closed-loop enhancement
The device does not blindly emit ultrasound. It listens for the precise neurophysiological signature of slow-wave sleep, then delivers calibrated mechanical stimulation timed to the user's natural CSF oscillation rhythm.
Detection
Four dry spring-pin EEG electrodes (Fp1, Fp2, F3, F4) continuously monitor delta-band power. A power-ratio classifier distinguishes SWS from lighter sleep stages. Impedance is checked in real time; if frontal electrodes fail, the system degrades gracefully to temporal channels.
Latency: < 30s classification window
Reference: Fz or linked mastoids (A1/A2)
Gating
Upon SWS onset, the system extracts the user's natural CSF oscillation period from the EEG envelope—typically 15–25 seconds. It predicts the next inflow window and fires a brief ultrasound burst during the low-resistance phase, when glymphatic channels are most open.
Envelope: adaptive period (user-specific)
Burst: ~1.5s ON, remainder OFF (~7.7% duty)
Target: MCA M1 segment, transtemporal window
Enhancement
Low-intensity focused ultrasound interacts with the natural arterial pulsation that drives glymphatic influx. The mechanism may involve acoustic radiation force adding directional momentum to perivascular fluid, or cellular mechanotransduction (TRPV4–AQP4) reducing hydraulic resistance to flow. Both effects amplify the existing clearance process without microbubbles or blood–brain barrier disruption.
MI: ~0.9 (FDA non-significant risk)
Delivery: bilateral temporal, 15° anterior, 10° superior
Coupling: integrated hydrogel pad (no user-applied gel)
Peer-reviewed foundation
Every claim is supported by published research. No animal studies using microbubbles are cited as primary evidence for the wearable protocol.
Demonstrated that natural sleep expands the brain's interstitial space by 60%, enabling convective exchange between CSF and interstitial fluid, and doubling β-amyloid clearance relative to wakefulness.
PubMed →Identified large CSF oscillations at 0.05 Hz during NREM sleep in humans, coupled to delta slow-wave EEG activity—establishing the precise mechanical window for intervention.
PMC Full Text →Non-microbubble FUS (650 kHz, 0.2 MPa, 7.7% duty cycle) significantly enhanced glymphatic transport across all anesthesia levels. Histologically safe at tenfold below FDA diagnostic limits.
PubMed →Low-intensity transcranial FUS protocol drives CSF clearance via meningeal lymphatics without exogenous agents (no microbubbles, nanoparticles, or drugs). Effective awake, asleep, or semi-conscious. MI = 0.9.
PubMed →Unilateral LIFU restored glymphatic CSF influx bilaterally within 10–20 minutes, peaking at 30–40 minutes post-stimulation—demonstrating rapid onset and non-localized effects.
PubMed →Planar ultrasound at 3.68 mW/cm² enhanced glymphatic influx via TRPV4 mechanotransduction and AQP4 water channel opening—without microbubbles. Authors explicitly proposed wearable integration.
PMC Full Text →DEC2-P384R and NPSR1-Y206H mutations accelerated glymphatic clearance, significantly reducing amyloid and tau pathology despite 25–40% shorter sleep duration—proving efficient sleep is biologically possible.
PubMed →Systematic review of 11 human RCTs and 44 animal studies confirmed therapeutic efficacy across multiple domains with no major adverse effects. Reversible minor symptoms resolved spontaneously.
PubMed →Meta-analysis of 34 studies found zero major adverse events across 690 participants. Most common effects: scalp sensations (9.01%), sleepiness (7.17%), neck pain (5.93%). All resolved without intervention.
Springer →Technical parameters
Prototype specifications for the Phase 0 proof-of-concept device. Subject to change during preclinical validation.
Research updates
We are building the first closed-loop sleep compression device. If you are a sleep researcher, neurologist, or technical contributor, we would like to hear from you.
MRG. Not a medical device.
research@somnison.com
Location
Lab address TBD
IRB Status
Not yet submitted
Webhook
Connected