Brain–computer interfaces (BCIs) hold great promise for restoring communication and movement, but long-term stability remains a major obstacle. Most implantable BCIs are placed directly on or in brain tissue, where mechanical mismatch and immune responses can gradually degrade signal quality.
A research team from Tsinghua University and collaborating institutions has now demonstrated an alternative strategy: recording brain activity from the lateral ventricle, a cerebrospinal fluid–filled cavity deep within the brain. The study introduces a lateral ventricular brain–computer interface (LV-BCI) that combines a minimally invasive surgical route with an expandable, flexible electrode inspired by the structure of traditional Chinese lanterns.
Delivered through a pathway similar to routine external ventricular drainage, the folded electrode expands once inside the ventricle and gently conforms to the ventricular wall. This design allows the device to remain mechanically compliant while maintaining close contact for signal acquisition.
In rat experiments lasting up to six months, the ventricular interface showed signal bandwidth comparable to standard subdural electrocorticography (ECoG) electrodes, but with superior long-term stability. Visual and auditory evoked responses remained consistent over time, while signals from cortical implants gradually declined.
Immunohistological analysis revealed a key advantage: unlike cortical electrodes, which triggered persistent microglial activation, the ventricular implant induced only a transient immune response that returned to baseline levels within weeks. The cerebrospinal fluid environment and the flexible electrode architecture appear to reduce chronic tissue irritation.
The researchers further tested the system’s ability to decode cognition using a memory-guided T-maze task. By analyzing sequences of neural microstates recorded before movement, the LV-BCI predicted whether rats would turn left or right with accuracies as high as 98 percent—significantly outperforming cortical electrodes. The results suggest enhanced sensitivity to deep brain circuits involved in memory and decision-making, such as the hippocampus.
By demonstrating stable, high-performance neural recording from within the ventricular system, this work establishes the lateral ventricle as a viable new access route for brain–computer interfaces. The authors note that future efforts will focus on scaling the design for human anatomy, improving imaging compatibility, and carefully evaluating cerebrospinal fluid dynamics and long-term safety.
This ventricular approach could complement existing cortical interfaces and expand the design space of implantable BCIs for both clinical and neuroscience applications.
National Science Review
Experimental study