Research | X-Sleep Lab

Research Direction 01

Memory Editing During Human Sleep

Aversive memories are not read-only files permanently etched in the brain. Instead, sleep provides a critical window of neuroplasticity where these experiences are actively updated and rewired. Leveraging this natural process, our lab uses targeted memory reactivation (TMR) and sleep learning to guide this nighttime editing. We deliver specific sensory cues to the sleeping brain to modulate memory reactivation and emotion reprocessing. This allows us to selectively decouple the negative affect from aversive memories while leaving the factual record intact. Ultimately, we aim to develop non-invasive, robust interventions to improve mental health.

Representative Publications:

Xia, T.*, Hu, X.* (2025). Memory editing during sleep: mechanisms, clinical applications, and technological innovations. Trends in Cognitive Sciences.
Xia, T., et al. (2024). Aversive memories can be weakened during human sleep via the reactivation of positive interfering memories. PNAS, 121(31), e2400678121.
Xia, T., et al. (2023). Updating memories of unwanted emotions during human sleep. Current Biology, 33(2), 309–320.

Research Direction 02

Non-invasive Closed-loop Sleep Brain-Computer Interface

Rather than a passive resting phase, sleep is a highly interactive cognitive state where crucial offline processing is orchestrated by signature rhythms like slow oscillations and sleep spindles. To harness this, our lab builds non-invasive, closed-loop sleep BCIs that move beyond conventional open-loop cueing. By tracking these brainwaves with millisecond precision, we replace random stimulation with phase-locked auditory interventions. This real-time control allows us to either present specific memory cues that shape targeted experiences, or play simple tones that directly manipulate sleep oscillations to impact broader cognitive functions, thereby actively guiding offline processing and effectively engineering sleep. Beyond basic research, this technology offers a powerful tool for targeted, therapeutic brain modulation.

Representative Publications:

Xia, T., Antony, J., Paller, K.A., & Hu, X.* (2023). Targeted memory reactivation during sleep influences social bias as a function of slow-oscillation phase and delta power. Psychophysiology.
Liu, J.*, Xia, T., et al. (2023). Item-specific neural representations during human sleep support long-term memory. PLoS Biology, 21(11): e3002399.

Research Direction 03

Lucid Dreaming & Consciousness During Sleep

Sleep is not a simple 'on/off' switch for consciousness, but rather a rich spectrum of cognitive abilities that peaks with lucid dreaming. In this paradoxical state, metacognitive awareness and conscious decision-making coexist with objectively verified sleep. We leverage this unique phenomenon as a rigorous experimental model to understand how consciousness works from within. Our goal is to map the exact neural mechanisms that allow subjective experience, reflective self-awareness, and volitional choices to emerge and fade across different sleep stages. Combined with our ongoing work on offline sensory processing, this research probes the very boundaries of the sleeping brain, revealing how self-awareness can dynamically surface even when the body remains fast asleep.

Representative Publications:

Chen, D., Xia, T., et al. (2024). Modulating social learning-induced evaluation updating during human sleep. npj Science of Learning, 9, 43.
Liu, J.*, Chen, D., Xia, T., et al. (2025). Slow-wave sleep and REM sleep differentially contribute to memory representational transformation. Communications Biology, 8, 1407.