However, suppose these classifiers were constructed from a small number of samples (e.g., tens of participants) from a single site. In such a case, it would not be possible to generalize to data obtained from other imaging sites. No one has ever developed a machine-learning-based brain network marker that is useful at any imaging site. One solution to this problem would be to collect vast quantities of data across many sites. However, significant site-related differences in fMRI data represent a formidable obstacle to such an undertaking.
The current study, published in the journal,
PLOS Biology
[1], addressed this difficulty using a novel harmonization method [2]. The research group developed a generalizable brain network marker for major depressive disorder (MDD), based on resting-state fMRI (rs-fMRI) data. This innovative study is expected to accelerate development of brain network markers for clinical applications based on fMRI. Ultimately, the research group expects that this new brain network marker could be adopted by public insurance and widely used in clinical practice in the near future.
The researchers specifically addressed the difficulty of developing a generalizable marker for MDD that would distinguish affected patients from healthy controls, based on resting-state functional connectivity [3] patterns. They used a discovery dataset with 713 participants from 4 imaging sites, removed site differences using their recently developed harmonization method,[2] and developed a brain network marker for MDD using a machine learning technique. The brain network marker achieved approximately 70% generalization accuracy for an independent validation dataset with 521 participants from 5 additional imaging sites. The capacity to generalize to a completely independent dataset acquired from multiple imaging sites is novel, and ensures scientific reproducibility and clinical applicability.
Toward practical clinical application of this brain network marker for MDD, in 2018, XNef company [4] commenced discussions with the Pharmaceuticals and Medical Devices Agency (PMDA) [5] in collaboration with Advanced Telecommunications Research Institute International (ATR), Hiroshima University, and the Japan Agency for Medical Research and Development (AMED), and reached a consensus on a development policy. Through continuing consultations, the group will prepare application documents for medical device development, apply for approval of the MDD marker in 2021, and obtain approval during 2022.
In the future, the research group will develop brain network markers for other psychiatric disorders, such as schizophrenia, anticipating that these brain network markers will contribute to accurate diagnoses in clinical situations.
###
Footnote
[1] Ayumu Yamashita, Yuki Sakai, Takashi Yamada, Noriaki Yahata, Akira Kunimatsu, Naohiro Okada, Takashi Itahashi, Ryuichiro Hashimoto, Hiroto Mizuta, Naho Ichikawa, Masahiro Takamura, Go Okada, Hirotaka Yamagata, Kenichiro Harada, Koji Matsuo, Saori C Tanaka, Mitsuo Kawato, Kiyoto Kasai, Nobumasa Kato, Hidehiko Takahashi, Yasumasa Okamoto, Okito Yamashita, and Hiroshi Imamizu. 2020. Generalizable brain network markers of major depressive disorder across multiple imaging sites.
PLoS Biology
. DOI: 10.1371/journal.pbio.3000966.
https:/
/
journals.
plos.
org/
plosbiology/
article/
related?id=
10.
1371/
journal.
pbio.
3000966
.
[2] Harmonization is a method that reduces intersite differences in fMRI data. The harmonization method is based on their 2019 publication, also in PLoS Biology: Ayumu Yamashita, Noriaki Yahata, Takashi Itahashi, Giuseppe Lisi, Takashi Yamada, Naho Ichikawa, Masahiro Takamura, Yujiro Yoshihara, Akira Kunimatsu, Naohiro Okada, Hirotaka Yamagata, Koji Matsuo, Ryuichiro Hashimoto, Go Okada, Yuki Sakai, Jun Morimoto, Jin Narumoto, Yasuhiro Shimada, Kiyoto Kasai, Nobumasa Kato, Hidehiko Takahashi, Yasumasa Okamoto, Saori C. Tanaka, Mitsuo Kawato, Okito Yamashita, Hiroshi Imamizu. 2019. Harmonization of resting-state functional MRI data across multiple imaging sites via the separation of site differences into sampling bias and measurement bias.
PLoS Biology
. DOI: 10.1371/journal.pbio.3000042.
https:/
/
journals.
plos.
org/
plosbiology/
article?id=
10.
1371/
journal.
pbio.
3000042
[3] Functional connectivity is quantified by a temporal correlation of resting-state fMRI blood-oxygen-level-dependent signals between pairs of brain regions.
[4] XNef company is a startup established in 2017 to develop clinical applications based upon research and development that addresses diagnosis and treatment of psychiatric disorders.
[5] PMDA (Pharmaceuticals and Medical Devices Agency) is a Japanese regulatory agency, working with the Ministry of Health, Labour and Welfare. Its mission is to protect public health by assuring the safety, efficacy, and quality of pharmaceuticals and medical devices. PMDA scientifically reviews marketing authorization applications for pharmaceuticals and medical devices, monitoring their post-marketing safety. PMDA is also responsible to provide relief compensation for sufferers from adverse drug reactions and infections caused by pharmaceuticals or biological products (
https:/
/
www.
pmda.
go.
jp/
english/
about-pmda/
outline/
0005.
html
).
This part of information is sourced from https://www.eurekalert.org/pub_releases/2020-12/abic-tca120120.php
