Authors
Jiang Liu, Yuming He, Hua-Peng Liaw, Pietro Russo, Marios Gourdouparis, Xiongchuan Huang, Jie Chen, Yao-Hong Liu
Published in
IEEE transactions on microwave theory and techniques. Volume 74. Issue 8. Pages 7429-7440.
Abstract
This article presents a networked wireless telemetry for 384-sensor-pixel neural data acquisition, designed to support high throughput, low power consumption, and meter-scale communication distance. The proposed event-based Address-Event-μPacket (AEμP) encoder with differential binary phase shift keying (DBPSK) modulation addresses the challenging throughput-power trade-off for neural telemetry without compromising transmission range and neural signal fidelity. The encoder compresses data and minimizes protocol overhead as well as memory access. In addition, a duty-cycling IR-UWB transmitter is proposed as communication link for the event-based encoder. To support the multi-access operation, a duty-cycling frequency-locked loop (FLL) is adopted to provide fast-settling packet-level frequency control for frequency division multiplexing (FDM). The FLL achieves warm-start lock within 1 μs with 2.73-MHz RMS variation, meeting the frequency accuracy required for DBPSK at a 125-MHz symbol rate. The measured 3-access-node spectra comply with FCC/ETSI masks in 6-8.5 GHz. The neural signal integrity is validated by evaluating the RMS error between the reconstruction signals and the original in-vivo data captured by a high-pixel count Neuropixels probe. The presented AEμP achieves a 11.52× compression ratio. Fabricated in 65 nm CMOS with active area of 0.12 mm2, the presented telemetry is capable of transmitting 384 pixels of neural data with low compression error (3.6-μV RMS), at record-low power of 1 mW, and with a 6.6-meter air communication range.
PMID:
42893344
Bibliographic data and abstract were imported from PubMed on 18 Oct 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 18
- Comments 0