Reconstructing lost BOLD signal in individual participants using deep machine learning
Abstract The blood oxygen level-dependent (BOLD) signal in functional neuroimaging suffers from magnetic susceptibility artifacts and interference from metal implants. The resulting signal loss hampers functional neuroimaging studies and can lead to misinterpretation of findings. Here, we reconstructed compromised BOLD signal using deep machine learning. We trained a deep learning model to learn principles governing BOLD activity in one dataset and reconstructed artificially-compromised regions in another dataset, frame by frame. Strikingly, BOLD time series extracted from reconstructed frames were correlated with the original time series, even though the frames did not independently carry information about BOLD fluctuations through time. Moreover, reconstructed functional connectivity (FC) maps exhibited good correspondence with the original FC maps, indicating that the deep learning model recovered functional relationships among brain regions. We replicated this result in patients whose scans suffered signal loss due to intracortical electrodes. Critically, the reconstructions captured individual-specific information rather than group information learned during training. Deep machine learning thus presents a unique opportunity to reconstruct compromised BOLD signal while capturing features of an individual’s own functional brain organization..
| Medienart: |
|
|---|
Preprint| Erscheinungsjahr: |
2020 |
|---|---|
| Erschienen: |
2020 |
| Enthalten in: |
bioRxiv.org - (2020) vom: 08. Dez. Zur Gesamtaufnahme - year:2020 |
|---|
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Yan, Yuxiang [VerfasserIn] |
|---|
| Links: |
|---|
| doi: |
10.1101/808089 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
XBI000646717 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | XBI000646717 | ||
| 003 | DE-627 | ||
| 005 | 20230429082708.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 200312s2020 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1101/808089 |2 doi | |
| 035 | |a (DE-627)XBI000646717 | ||
| 035 | |a (DE-599)biorXiv10.1101/808089 | ||
| 035 | |a (biorXiv)10.1101/808089 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 082 | 0 | |a 570 |q DE-84 | |
| 100 | 1 | |a Yan, Yuxiang |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Reconstructing lost BOLD signal in individual participants using deep machine learning |
| 264 | 1 | |c 2020 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a Abstract The blood oxygen level-dependent (BOLD) signal in functional neuroimaging suffers from magnetic susceptibility artifacts and interference from metal implants. The resulting signal loss hampers functional neuroimaging studies and can lead to misinterpretation of findings. Here, we reconstructed compromised BOLD signal using deep machine learning. We trained a deep learning model to learn principles governing BOLD activity in one dataset and reconstructed artificially-compromised regions in another dataset, frame by frame. Strikingly, BOLD time series extracted from reconstructed frames were correlated with the original time series, even though the frames did not independently carry information about BOLD fluctuations through time. Moreover, reconstructed functional connectivity (FC) maps exhibited good correspondence with the original FC maps, indicating that the deep learning model recovered functional relationships among brain regions. We replicated this result in patients whose scans suffered signal loss due to intracortical electrodes. Critically, the reconstructions captured individual-specific information rather than group information learned during training. Deep machine learning thus presents a unique opportunity to reconstruct compromised BOLD signal while capturing features of an individual’s own functional brain organization. | ||
| 700 | 1 | |a Dahmani, Louisa |e verfasserin |4 aut | |
| 700 | 1 | |a Shen, Lunhao |e verfasserin |4 aut | |
| 700 | 1 | |a Peng, Xiaolong |e verfasserin |4 aut | |
| 700 | 1 | |a Wang, Danhong |e verfasserin |4 aut | |
| 700 | 1 | |a Ren, Jianxun |e verfasserin |4 aut | |
| 700 | 1 | |a He, Changgeng |e verfasserin |4 aut | |
| 700 | 1 | |a Jiang, Changqing |e verfasserin |4 aut | |
| 700 | 1 | |a Gong, Chen |e verfasserin |4 aut | |
| 700 | 1 | |a Tian, Ye |e verfasserin |4 aut | |
| 700 | 1 | |a Zhang, Jianguo |e verfasserin |4 aut | |
| 700 | 1 | |a Guo, Yi |e verfasserin |4 aut | |
| 700 | 1 | |a Lin, Yuanxiang |e verfasserin |4 aut | |
| 700 | 1 | |a Wang, Meiyun |e verfasserin |4 aut | |
| 700 | 1 | |a Li, Luming |e verfasserin |4 aut | |
| 700 | 1 | |a Hong, Bo |e verfasserin |4 aut | |
| 700 | 1 | |a Liu, Hesheng |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t bioRxiv.org |g (2020) vom: 08. Dez. |
| 773 | 1 | 8 | |g year:2020 |g day:08 |g month:12 |
| 856 | 4 | 0 | |u https://doi.org/10.1038/s41467-020-18823-9 |z lizenzpflichtig |3 Volltext |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1101/808089 |z kostenfrei |3 Volltext |
| 912 | |a GBV_XBI | ||
| 912 | |a SSG-OLC-PHA | ||
| 951 | |a AR | ||
| 952 | |j 2020 |b 08 |c 12 | ||
| 953 | |2 045F |a 570 | ||