Light Coils

MRI with fully optical data and power transmission

Light Coils are MRI receive coils that are powered, read out and detuned entirely over optical fibre, with no galvanic cables between the coil array and the scanner. We show in vivo human brain imaging at 3 T with SNR comparable to a conventional coaxial readout, and a 4-channel array multiplexed over a single fibre.

Zining Liu†, Morteza Teymoori†, Jakob Gerlach, Reza Aghabagheri, Henning Helmers, Michael Bock, Çağlar Ataman, Ali Caglar Özen
†Equal contribution. University of Freiburg (Radiology · IMTEK) and Fraunhofer ISE.

Why it matters

Faster MRI relies on dense receive arrays: more coil elements mean more SNR and higher parallel-imaging acceleration. But every element needs coaxial cables for signal, power and detuning, and in a dense bundle those cables couple to each other. That coupling raises noise correlation between channels, which costs SNR at high acceleration (a higher g-factor), and the cables also add bulk and RF-heating risk. Light Coils replace the cables with optical fibre, which removes cable cross-talk as a source of coupling.

Light Coils concept: optically connected receive-array modules, modular head arrays, and the optical power, data and detuning architecture
Fig. 1 The Light Coils concept. (a) Lightweight, optically connected receive-array modules close to the patient. (b) Modular "array-of-arrays" head configurations. (c) One laser powers the on-coil LNAs through a photonic power converter (PPC); MR signals are put onto light by electro-optic modulators and wavelength-multiplexed out of the MR room, while detuning is also controlled optically.

Key results

5–10 mW Optical power at the modulator input needed to match the SNR of a galvanic coax link (with 80–100 mW into the PPC)
in vivo 3 T Human brain imaging with a single-channel Light Coil: image quality and SNR comparable to the same coil read out over coax
>28 dB Inter-channel optical isolation for a 4-channel array, dense-WDM multiplexed over a single fibre
≤ −21 dB Noise correlation between channels, vs. down to −13.6 dB for a matched coax array; lower g-factors in parallel imaging
Image SNR versus PPC power, and in vivo brain images and SNR maps for the reference coil and the Light Coil
Fig. 4 Single-channel Light Coil. (a) Phantom image SNR vs. optical power into the PPC for four optical link powers; the dashed line is the galvanic reference. (b, c) In vivo brain images and SNR maps, reference coil (top) vs. Light Coil (bottom).
4-channel Light Coil array schematic, WDM isolation matrix, noise correlation matrices and g-factor maps
Fig. 5 4-channel Light Coil array. (a) Schematic. (b) WDM channel isolation/loss matrix. (c, d) Noise correlation for the Light Coil array vs. a matched coax reference array. (e) Phantom images and g-factor maps at R = 2 and R = 3 (g = 1.44 vs. 1.54 and 1.82 vs. 2.02).
In vivo axial brain images with the 4-channel Light Coil array: GRE at R = 1 to 4 and MPRAGE
Fig. 6 In vivo brain imaging with the 4-channel Light Coil array: 2D GRE without acceleration and with GRAPPA R = 2, 3, 4, and a T1-weighted 3D MPRAGE.

How it works

Optical power transmission setup, noise figure and gain versus optical power, sequence timing, and detuning comparison
Fig. 3 Optical power delivery and detuning. (a) Setup combining power-over-fibre, optical detuning and the analog optical data link. (b, c) Noise figure and gain of the PPC-powered LNA chain vs. optical power, with bench supply and battery references. (d) Sequence timing of the coil state. (e) Galvanic vs. 10 mW optical detuning vs. no detuning.

Read more

Figures reproduced from the arXiv preprint.

Contact

QR code linking to this page

Dr. Morteza Teymoori
Microsystems for Biomedical Imaging Laboratory
IMTEK, University of Freiburg, Germany

Interested in collaborating, or want to talk about optical coils? Email me at Morteza.Teymoori@imtek.uni-freiburg.de or connect on LinkedIn.