Duke university visit

Early this month, an informal research meeting was held at Prof. Allen Song's laboratory at Duke University in North Carolina. Three people from GE Healthcare met with Allen, Prof. Dean Darnell, and a graduate student Devon Overson, to talk about MRI systems engineering to realize wireless receiver coils. This was a second visit in about two years. In 2022, when we were still wearing masks (picture on the left), we had a meeting on the "iPRES" coil, an RF receiver coil that doubles as a multi-channel shim array. The meeting this month, as before, was very productive. The Duke team has worked on wireless signal transmission for peripheral devices control for a number of years. It remains to be seen whether wire-free, battery-powered MRI RF coil can be realized without sacrificing the sensitivity to replace the traditional signal reception architecture.

ISMRM Abstracts, 2025

The meeting next year will be in Hawaii, a popular place, and the ISMRM organizers are sifting through more than 8000 abstracts submitted for presentation there. I have been critical to the abstract's HTML format, and a recently introduced sectioned Synopsis (in fact the presence of Synopsis at all). However, from an abstract reviewer's point of view, I must say these changes are helpful, to quickly get to know the work, and inspect details of selected figures. The organizers also did a good job in summarizing the review progress graphically in real time. This allows a reviewer to stay "normalized", by observing how the score distribution changes from start to finish as more than 50 assigned abstracts are scored. Certainly a lot of change has happened in ISMRM abstract processing since the early 2010s. I still hold up two things as suggestions to the Society: (1) The scores should be made available to the authors, and (2) All rejected abstracts should be given a chance to be archived, if not presented, on-line. That is democratization, not giving out low-performance MRI to those who deserve better.

MRI as a patient

I guess a web log dedicated to Magnetic Resonance Life is not complete without MR images of the author as a patient. For that purpose, I had an opportunity, albeit painful, to collect exactly such data in early November. The 1.5 T GE Signa Artist images shown above display "unremarkable" brain anatomical images with multiple contrasts following a standard protocol in Albany Medical Center. The total exam lasted for 21 minutes, and I must say the image qualities are pretty good for 1.5 T. The Center used apparent diffusion coefficient (ADC) to look for brain vessel blockage (infarct), and susceptibility-weighted imaging (SWI) for brain blood leakage (hemorrhage).

Three patent applications of summer of 2024

Three patent applications were published in July, that are all related to what is called concomitant gradient correction. This refers to magnetic field vector components that are unwanted but occurring as a byproduct of the desired magnetic fields in MRI. The three patents each address measurement, hardware-based correction, and software correction. The problem of concomitant field first came to me in early 2000s when I was a graduate student working on microTesla MRI with superconducting sensors. That work was reported here in a JMR article: https://pubmed.ncbi.nlm.nih.gov/19664947/. At that time the issue was too weak a main magnetic field; now the problem is due to too strong a gradient field. Different regime, but the same physics.

Three papers of summer of 2024

I was silent for 3 months! In those summer months, 3 papers of mine came out, in Journal of Applied Physics, IEEE Transactions on Magnetics, and Magnetic Resonance Imaging. They report studies related to, and motivated by, MRI gradient coils. The three papers, respectively, provide a first systematic mathematical formulation of magneto-mechanical coupling, derive a versatile, surface eddy current equation, and describe a practical method to reduce acoustic noise in high-performance MRI. They all center around solving critical engineering problems in high-field, high-gradient MRI, in order to make strong magnetic fields more accessible and exploitable for advanced brain imaging. Apart from the subject matter, a not-so-coincidental feature that is common to all three papers is that they have a non-sectioned, single-paragraph abstract! They are also in journals where authors don't have to pay to publish.

Head-only then and now

In this year's ISMRM meeting, accessibility, Helium-free magnet, and high-performance gradients were at the top of many people's mind. In 2016, GE's 3T head-only scanner already combined all three. The figure above compares the 2019 ISMRM floor (right) showcasing GE's then-latest ESP head-only scanner, with this year's floor in Singapore (left) where MAGNUS 3T was unveiled with much acclaim. ESP didn't pan out; time will tell if MAGNUS will. 

About Low-field: This year's ISMRM was flooded with low-field MRI presentations in all formats. Anyone who was in the MRI field long enough will know that performance is the key. Did the bar go down in the name of democratization? Will AI magically lift the SNR? Or is MRI undergoing a transition from a physics and engineering wonder to commodity electronics? Maybe all of the above. Something good may well come out of this, but I sincerely hope that the trend is guided by unselfish, conflict-free, balanced and broad view of MRI's place in medical imaging. 

Digital posters for ISMRM 2024

 

ISMRM announced proceedings of this year's scientific meeting on its website. The digital posters are now presented as an author-recorded video (in most cases, a power point with narration) in a common wrapper which contains links to section headings. This reduces the video screen size but helps audience keep track of the contents in context. Two such posters of mine are shown in screen captures above. Availability of remote participation was certainly accelerated by the pandemic, and may contribute to lower in-person attendance from regions far away. Good luck with all those who travel to Singapore in May!

DIANA questions, raised scientifically

 A first scientific counterargument against DIANA was published recently in an open-access journal: A different interpretation of the DIANA fMRI signal | Science Advances. The three main points are the following: (1) The authors did reproduce the original MRI data and therefore the original report by Toi et al. in Science is likely true on observed facts. (2) On the other hand, the authors could explain their own observations better without invoking neuronal response, but rather based on instrumental artifacts and coincidental circumstances. (3) Regarding membrane potential-induced T2 change, the authors believe that the originally reported in-vitro T2 change is rather caused by cell swelling than membrane potential. The authors did leave several questions unanswered that need to be answered in order to reconcile their assertion with the large amount of experimental data reported in Toi et al.

It is notable that the authors humbly titled their paper as "A different interpretation of DIANA fMRI" and showed reservation of judgment and desire to be constructive. The paper ends with a respectable remark, "Although the goal of probing neural activity with millisecond temporal resolution using fMRI remains unmet, we hope that the contribution of Toi et al. will inspire further efforts to achieve this ambitious and worthy objective."

The questions raised by this paper are certainly serious and substantial. But they were brought up in a way to allow professional debate and objective mutual verification. I hope this development, along with the growing body of positive data from Prof. Park's team, will soon bring clarity to this tantalizing DIANA puzzle.

Installation finished

Brigham and Women's hospital in Boston now has the latest Magnus gradient system which started to generate beautiful brain images!

[Update] An opening ceremony for the Magnus imaging facility was held on March 28th in a nearby Brigham and Women's hospital building with the senior leadership of the hospital. It was a nice showcase of an exemplary co-work among the government, industry, and academia. It is only the beginning!

[Link] A website (https://brighamandwomens.theopenscholar.com/magnus/) introduces the State government-funded Magnus program.

Installation under way

 

Installation of a brand-new Magnus gradient system is under way in Brigham and Women's hospital in Boston. This was funded by the Massachusetts Life Sciences Center (link to grant news). Prof. Westin, the PI on the right in the picture, indicated that the users will be split half and half between clinical and physics/engineering research. Brigham and Women's hospital has many campuses across the area. The Fenwood location, where the 2023 neuroimaging conference was held, had received the country's second clinical 7T MRI machine from Siemens in 2017. (That's why the name may sound familiar to those involved in the 7T commotion in Korea.)

ISMRM abstract announcement

This year's ISMRM abstract announcement came a week earlier than usual. In my count, about 143 x 36 = 5148 abstracts were accepted, as compared to 141 x 38= 5358 last year. I see submission numbers going beyond 8600 so the acceptance rate is perhaps about 60%, quite low compared to the past. The question of fairness and transparency naturally arises as more people get rejection -- prominently missing any explanation. When Dr. Derek Jones, now president of ISMRM, visited us last year I did ask about providing some kind of scores or justification for abstract decisions. He was under impression that rejected ones are of obvious low quality, and the question may be more on oral vs posters. I think this position needs re-visiting, as (i) the field is getting ever more diversified, making judgement of others' work difficult, and (ii) unlike APS, for example, ISMRM abstract submission requires considerable manual work (on formatting) so the rejected authors deserve some sort of feedback at the minimum.

Personally, two first-authored abstracts were accepted for digital posters as requested. They both involve eddy current. Video recording should happen in the next several weeks. More on that part later!