Neurotassel elastocapillary filament arrays
Applications
Neurotassel self-assembled filament arrays
Guan and colleagues’ March 2019 Science Advances paper reports flexible filaments that gather into a stiff, implantable bundle when withdrawn from molten polyethylene glycol (PEG). The primary affiliations include China’s National Center for Nanoscience and Technology and the Chinese Academy of Sciences’ Institute of Neuroscience. This is China-secondary coverage in a US-first catalog, not Stanford NeuroRoots or a renamed NET probe.
Variant-specific structure
| Configuration | Published structure |
|---|---|
| 16-channel example | 12 µm-wide, 3 µm-high filaments; 10 µm-diameter recording sites; assembled PEG fiber about 55 µm diameter |
| 128/256 channels | 10 × 1.5 µm filament cross-section; 10 µm-diameter sites arranged in a V before assembly |
| 512/1,024 channels | 3 × 1.5 µm filament cross-section; semicircular-arch sites before assembly |
| 128 / 1,024 assembled bundles | About 80 / 100 µm diameter, respectively |
| Optical combination | Separate 61-channel Neurotassel assembled around a sharpened optical fiber |
The freestanding section transitions from plane to mesh to filaments. A chromium/gold conductor is insulated between polyimide layers, with sites exposed at the front. Platinum electrodeposition reduces impedance from greater than 1 MΩ to approximately 50 kΩ at 1 kHz for the reported recording preparation. Bare gold, platinum-treated contacts and different filament variants are not interchangeable specifications.
Delivery and connection
The paper describes drawing the device from molten PEG4000 at 120°C into ambient air. PEG solidifies to stiffen the bundle before insertion; body fluids dissolve it afterward. The bath temperature is not an instruction to insert hot polymer into tissue.
16-channel devices use flip-chip-bonded flexible circuits and an Omnetics connector. A distinct 61-channel flexible circuit/PCB supports the optical combination. Methods describe probe-station electrical measurements for the 128-1,024-channel variants. They do not establish a completed chronic, freely behaving 1,024-channel acquisition package.
Yield and evidence boundary
The paper reports greater-than-95% electrical integrity after implantation for the tested preparation. Separately, fabrication/electrical characterization across channel counts reported greater-than-80% yield and 54 ± 15 kΩ average impedance after platinum deposition. These are different denominators and stages.
The mouse recording application uses 16-channel arrays for learning and weeks-long recording. The main text calls the 1,024-channel recordings preliminary and refers to Supplementary Figure S16. That supplementary file was not available as a readable PDF during this catalog pass, so its cohort, connected-channel count, duration and unit yield are not asserted here.
The authors identify on-chip amplification/multiplexing to reduce external leads and biomimetic interfaces as future work. This paper is not evidence of human use or proof that later development stopped.
Model boundary
No full model is supplied. The micrographs show the plane-mesh-filament transition and assembly, but the published summary dimensions do not define the complete mask, mesh junctions, individual lengths, assembled contact positions or connector outline. A neat cylinder of evenly spaced contacts would invent the post-assembly geometry.