Graph-Based Decoding Model for Functional Alignment of Unaligned fMRI Data

Authors: Weida Li, Mingxia Liu, Fang Chen, Daoqiang Zhang2653-2660

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on five datasets suggest that the proposed method is not only superior to several state-of-the-art methods on temporally-aligned f MRI data, but also suitable for dealing with temporally-unaligned f MRI data.
Researcher Affiliation Academia 1College of Computer Science and Technology & MIIT Key Laboratory of Pattern Analysis and Machine Intelligence, Nanjing University of Aeronautics and Astronautics, Nanjing, China 2School of Information Science and Technology, Taishan University, Taian, China
Pseudocode Yes Algorithm 1 Graph-Based Decoding Model (GDM) Input: Aligning data {Xi RVi Ti}M i=1, the number of the shared features K, the energy {pi%}M i=1 to be kept, a specific Laplacian matrix L and kernel functions for each subject. 1: For each i, standardize Xi such that it has zero mean along the second dimension and the variance of each feature, i.e., voxel, is 1. 2: Generate {Ki}M i=1 via specified kernel functions. 3: Centralize Gram matrices: Ki Ki + T 2 i JTi Ti Ki JTi Ti T 1 i JTi Ti Ki T 1 i Ki JTi Ti 4: for i 1 to M do 5: Ki = Vi Di VT i by spectral decomposition. The eigenvalues in Di is in descending order. 6: Find Li such that the first Li diagonal elements of D 1 2 i contains approximately pi% energy. 7: Let ˆVi be the first Li columns of Vi . 8: Let ˆDi be the top left Li Li submatrix of Di . 9: end for 10: By spectral decomposition, ˆVT L ˆV = EΣET where the diagonal elements of Σ is ascending. 11: Let ˆE be the first K columns of E and then cut ˆE along the first dimension such that ˆEi RLi K. 12: For 1 i M, W i Φi ˆVi ˆD 1 i ˆEi .
Open Source Code No The paper does not provide any link or explicit statement about releasing source code.
Open Datasets Yes We utilize five datasets shared by openfmri.org and Chen et al. (Chen et al. 2015). The relevant information about each dataset is outlined in Table 1.
Dataset Splits Yes We follow the experiment setup with a cross-validation strategy in previous studies (Chen et al. 2014; Haxby et al. 2011), as illustrated by Fig. S1 and Fig. S2 in the Supplementary File. Specifically, except for the Raider, each subject s data is equally divided into two parts with each category being equally split, one is for alignment whereas the other is for training or testing a classifier. Switching the roles of the two parts and leave-k-subject-out strategy are adopted for cross-validation.
Hardware Specification No The paper does not provide specific hardware details used for running experiments.
Software Dependencies No Raw datasets are preprocessed by using FSL (fsl.fmrib.ox.ac.uk), following a standard process (i.e., slice timing, anatomical alignment, normalization, and smoothing). The default parameters in FSL were taken when the dataset does not provide. ... Like previous studies, ν-SVM is used for classification (Chang and Lin 2011). All methods are implemented by ourselves in Python.
Experiment Setup Yes The parameter settings for each dataset are briefly listed in Table 1. For a fair comparison, the parameter ν in ν-SVM (with a linear kernel) is fixed for all methods on each dataset. For six competing methods, we choose the optimal hyperparameters according to their original papers. For our GDM model, a linear kernel is fixed, while the influence of different kernels are shown in Figs. S3-S8 in the Supplementary File.