A Korean research team's study that drew international attention after claiming brain activity could be directly observed with magnetic resonance imaging (MRI) has ultimately been retracted. Published in the international journal Science in 2022 and hailed as an innovative breakthrough, it was withdrawn after reproducibility disputes and additional verification revealed unexpected signal characteristics, leading to the judgment that it was difficult to maintain the initial conclusions.
Science said on the 26th that a joint research team including Professor Park Jang-yeon of Sungkyunkwan University, Professor Gwak Ji-hyun of Seoul National University, and Professor Lee Jong-ho of Seoul National University voluntarily retracted their paper published in Aug. 2022.
In the paper, the team presented a method, DIANA (Direct Imaging of Neuronal Activity), that could capture signals on the order of ms (milliseconds, one-thousandth of a second) in time and at the level of cellular layers that make up the brain in space. They said the process by which neural signals are transmitted could be tracked without inserting electrodes into the brain, using only conventional functional magnetic resonance imaging (fMRI) equipment. In experiments on mice, they observed the process in which signals move from the thalamus, the sensory hub, to the somatosensory cortex in units of one-fifth of a millisecond. It was eight times faster than existing techniques.
The technology raised expectations that it would bring major advances in understanding brain diseases and cognitive function. According to Google Scholar, it has been cited 108 times so far, and in 2023 it was included in the "Top 100 Outstanding National R&D Achievements" selected by the Ministry of Science and ICT.
However, in May 2023, Sungkyunkwan University Distinguished Professor Kim Seong-gi published a study on the preprint site bioRxiv reporting that the DIANA results could not be reproduced, igniting a full-fledged controversy. In Aug. of the same year, Science publicly shared readers' concerns that "the results are difficult to reproduce using the methods presented in the paper" and expressed concern. In particular, questions were raised about how MRI data were selected, and multiple research teams conducted experiments on animals and humans but failed to obtain similar results.
The team then conducted additional verification at Science's request. But in the process, they found a new signal pattern that did not match their previous claims. It was confirmed that signals that looked like neural activity in the brain might in fact be "noise" generated by MRI equipment. In the end, the team acknowledged flaws in the core conclusions and chose to retract the paper.
In a written statement, Professor Park Jang-yeon said, "We posted to bioRxiv on the 4th a summary of the newly discovered signals, the replication experiment results, and our current academic assessment," adding, "We are supplementing the content for submission to a journal."
According to the paper posted on bioRxiv, the team repeated experiments that delivered electrical stimulation to a mouse's forepaw under various magnetic field strengths. As a result, it emerged that what had been considered a DIANA signal could in fact be a special MRI signal pattern called "pseudo-steady state (PSS)." PSS signals arise from specific frequency offsets, and as multiple signals overlap, they can create patterns that look like brain activity at certain times and locations. Put simply, it may be an illusionary signal generated by the MRI equipment, not an actual brain response.
The team said, "If PSS oscillations are a major cause of the DIANA signal, it may be premature to interpret it as a neural response to sensory stimulation," while adding, "However, the consistency and response timing of the DIANA signal are hard to explain as mere coincidence. Further research is needed to clarify the relationship between the signal and actual brain activation."
Meanwhile, attention is turning to whether the retraction will bring disadvantages. A science community official said, "Follow-up actions vary depending on the reason for and the circumstances of the retraction. A voluntary retraction does not necessarily exempt one from discipline, nor does a journal-initiated retraction automatically entail penalties," adding, "It is hard to generalize because judgments differ by case."
References
Science (2025), DOI: https://doi.org/10.1126/science.aec1773
bioRxiv (2025), DOI: https://doi.org/10.1101/2025.08.31.673005