Research

We pursue one question — how the sleeping brain consolidates memory and recalibrates emotion — from converging angles. Our work is organized in two layers: a Foundation of basic circuit and computational studies, and a Translation arm pursuing two parallel axes toward clinical applications. Discoveries in basic circuits inform translational design; clinical observations send us back to the bench.

Foundation

Basic mechanisms of how sleep — particularly REM sleep — supports memory consolidation, emotional recalibration, and circuit-level integration of new neurons. This layer is the scientific core of the laboratory and the source from which all translational efforts flow.

Foundation

Adult-born neurons × REM consolidation

Question: How do new neurons born in the adult hippocampus integrate into existing memory circuits?

The dentate gyrus generates new neurons throughout life. Using cell-type-specific optogenetics during defined sleep stages, we showed that sparse activity of adult-born neurons during REM sleep is necessary for memory consolidation (Kumar et al., Neuron 2020). Building on this, chronic miniscope imaging revealed how small ensembles of adult-born neurons are transiently reactivated during REM sleep across days (Srinivasan et al., Nature Communications 2025).

Why it matters: Understanding integration of new neurons opens routes for regenerative therapy in neurodegenerative disease and the rehabilitation of memory after insult.

Key publications: see Publications →

Technology

Real-time AI sleep staging

Question: Can we detect sleep stages — and intervene — at second-scale resolution?

We developed an AI sleep-staging system that supports closed-loop, stage-specific intervention (Tezuka et al., Scientific Reports 2021; Koyanagi et al., Neuroscience Research 2022). Combined with custom miniaturized imaging and analysis tools (Vergara et al., Nature Communications 2025), this enables real-time, stage-specific imaging and intervention during sleep.

Why it matters: Stage-specific manipulation moves sleep research from description to causation, opening the way to apply these tools in broader research and clinical settings.

Key publications: see Publications →

Translation

Two parallel translational axes extend our basic findings toward applications for specific neurological and psychiatric conditions. This page provides scientific context only. It is not a channel for medical advice, and it is not a recruitment channel for any clinical study.

Translation · Axis 1

Circuit-level repair of the central nervous system

Question: Can newly born neurons be guided to integrate into existing circuits to replace function lost to injury or degeneration?

Building on our foundational work on adult-born neurons, we are developing approaches in which immature neurons are recruited into pre-existing neural circuits. The aim is to provide a biological substrate for repair following neurodegenerative disease and traumatic neural loss.

Status: Active research direction. Detailed mechanism, methods, and team information will be released as scientific publications and other formal disclosures become available.

Translation · Axis 2

Sound Exposure during Sleep (SES)

Question: Can targeted memory reactivation through sound during specific sleep stages support recovery from trauma-related memory states?

Building on basic findings that sound delivered during sleep can modulate fear memory (Purple et al., Scientific Reports 2017), we have studied Sound Exposure during Sleep (SES) — an approach that delivers individualized auditory cues during specific sleep stages.

We have successfully completed a medical-institution-led, First-in-Human feasibility study of SES (registered as jRCT1030230706) — an important milestone for the platform. As reported in a recent preprint, 13 patients provided informed consent and 6 completed overnight SES. The intervention proved well tolerated: slow-wave sleep was preserved and no adverse events were attributed to the auditory cues, supporting the feasibility and safety of delivering individualized sound during sleep in a clinical setting. Exploratory analyses are described in the preprint (Ino et al., medRxiv 2026).

With feasibility now established, we are building on these encouraging results to refine the SES protocol and to design the next phase of clinical research. SES remains an investigational platform; this site provides no medical advice. The registered study has concluded and is not currently recruiting participants.

Key publications: see Publications →

Methods we develop & deploy

In vivo

Miniaturized fluorescence microscope

Custom miniscope optimized for freely-moving recording in chronic sleep contexts.

Computation

Real-time sleep AI

Edge-deployable model for closed-loop, stage-specific intervention during sleep.

Genetic tools

Cell-type-specific opto / chemo

Transgenic mouse lines targeting adult-born neurons across the lifespan.