Whole‐brain functional MRI registration based on a semi‐supervised deep learning model

Abstract

Purpose Traditional registration of functional magnetic resonance images (fMRI) is typically achieved through registering their coregistered structural MRI. However, it cannot achieve accurate performance in that functional units which are not necessarily located relative to anatomical structures. In addition, registration methods based on functional information focus on gray matter (GM) information but ignore the importance of white matter (WM). To overcome the limitations of exiting techniques, in this paper, we aim to register resting‐state fMRI (rs‐fMRI) based directly on rs‐fMRI data and make full use of GM and WM information to improve the registration performance. Methods We provide a robust representation of WM functional connectivity features using tissue‐specific patch‐based functional correlation tensors (ts‐PFCTs) as auxiliary information to assist registration. Furthermore, we propose a semi‐supervised deep learning model that uses GM and WM information (GM ts‐PFCTs and WM ts‐PFCTs) during training as a fine tweak to improve registration accuracy when such information is not provided in new test image pairs. We implement our method on the 1000 Functional Connectomes Project dataset. To evaluate our method, a group‐level analysis was implemented in resting‐state brain functional networks after registration, resulting in t maps. Results Our method increases the peak t values of the t maps of default mode network, visual network, central executive network, and sensorimotor network to 21.4, 20.0, 18.4, and 19.0, respectively. Through comparison with traditional methods (FMRIB Software Library(FSL), Statistical Parametric Mapping _ Echo Planar Image(SPM_EPI), and SPM_T1), our method achieves an average improvement of 67.39%, 12.96%, and 25.14%. Conclusion We propose a semi‐supervised deep learning network by adding GM and WM information as auxiliary information for resting‐state fMRI registration. GM and WM information is extracted and described as GM ts‐PFCTs and WM ts‐PFCTs. Experimental results show that our method achieves superior registration performance.

Visualizing Cerebral Small Vessel Degeneration During Aging and Diseases Using Magnetic Resonance Imaging

“…Compared to 7 T, this system has significantly less weight and occupies smaller spaces and thus is much easier to be installed in hospitals. The advantages of less peripheral nerve stimulation and lower SAR may also increase patient compliance 98 . The disadvantages of lower SNR may be compensated by advanced sequence design, image reconstruction, and analytical methods.…”

Independent component analysis: a reliable alternative to general linear model for task-based fMRI

“…As far as pre-processing of fMRI data is concerned, many improvements have been made in recent decades. Moreover, motion artifact correction ( 1 – 3 ), slice time correction ( 4 , 5 ), spatial smoothing ( 6 – 9 ), and registration ( 10 – 12 ) have been extensively studied, all of which demonstrate a high signal-to-noise ratio of pre-processed fMRI datasets.…”

Custom 3D fMRI Registration Template Construction Method Based on Time-Series Fusion

“…Registration based on individual space effectively avoids this problem and is more suitable for fMRI data that usually help determine active brain areas by comparing the differences between the sequence data of the same subject. However, in the current registration methods based on individual space, most of the reference templates are randomly selected from data sequences, selected by researchers according to experience, or directly selected from the first effective time-point data, which leads to such methods being too dependent on researchers and having strong randomness [ 15 ]. Therefore, the registration of fMRI data needs a reasonable fMRI template based on individual space as a reference image for intra-individual registration.…”