Endovascular BCI (Stentrode)

Electrophysiology · reads

An endovascular brain-computer interface is an electrode array mounted on a stent and threaded through the jugular vein into the superior sagittal sinus, a large vein next to the motor cortex. From inside the vein it records electrocorticography, the electrical field produced by sensorimotor cortex activity, sensed through the vessel wall; this is a direct electrical measure of brain activity. The approach avoids open-brain surgery, but the implant is still an invasive vascular procedure.

Spatial resolution: 2.4 mm (sampling)
Temporal resolution: 4.08–4.85 ms (recoverable timing)
Field of view: 422 mm³ (sensing envelope)
Reads: electrocorticographic (ECoG) signals
Destructive?: minimally destructive
Prerequisites: other: Preimplant MRI assessment of venous anatomy, attempted-movement fMRI to confirm preserved motor activation, CT venography, and preoperative structural MRI target marking.; other: Preimplant dual antiplatelet therapy.; anesthesia: General anesthesia in an operating room.; surgery: Direct percutaneous internal-jugular-vein access; catheter delivery under 3D digital subtraction angiography; deployment in the superior sagittal sinus; lead tunneling to an IRTU placed in a subcutaneous pocket.

Description

An endovascular brain-computer interface is an electrode array mounted on a stent and placed inside a cerebral vein, where it records the electrical activity of the cortex from within the vessel wall. The array is delivered through the internal jugular vein into the superior sagittal sinus, a large vein adjacent to the precentral gyrus, and a lead is tunnelled under the skin to an electronics pocket. The approach is an alternative to placing electrodes in or on the dura by open-brain surgery. The implant is still an invasive vascular procedure. The signal is electrocorticography (ECoG): field potentials produced by sensorimotor cortex activity, sensed through the sinus wall, which in sheep is 200 to 600 µm thick. In sheep, electrodes 2.4 mm apart on the stent recorded distinguishable waveforms, with bandwidth and signal-to-noise ratio comparable to epidural and subdural arrays; spatial resolution there was limited by electrode spacing and size. In a ten-participant human analysis, the electrodes sat at different points around the stent's circumference, some closer to the cortex and others within the opposite wall of the sinus; a simplified model that placed the array along the middle of the sinus put it on average 6.1 ± 1.5 mm from the cortical surface.

Download the dataset (JSON, described by its JSON Schema).