Fig. 1.

Given observed data, in the previous works, the improvement of the FBN estimation is mainly based on 1) high-quality data, 2) sophisticated preprocessing pipeline, and 3) reasonable priors. However, it is hard to obtain an “ideal” result, since the data acquisition is hard to control and the understanding of brain is limited. To alleviate this issue, in this paper, a basic idea is setting the FBN of the high-quality data as a “guider” to help the FBN estimation task, which can efficiently provide more useful information and thus can reduce the dependency for data. Specifically, in this paper, we employ the link-strength information of high-quality data for guiding the FBN estimation.