Computational protocol: Classification of Autism Spectrum Disorder Using Random Support Vector Machine Cluster

Similar protocols

Protocol publication

[…] Because of the lower signal-to-noise ratio, all the fMRI images are preprocessed by the Data Processing Assistant for Resting-State fMRI (DPARSF) software (Cui et al., ) (http://d.rnet.co/DPABI/DPABI_V2.3_170105.zip). The whole preprocessing procedures include:(1) Converting DICOM to NIFTI, (2) Removing the first 10 time points, (3) Slicing timing, (4) Realigning, (5) Normalizing with the echo-planar imaging (EPI) template, (6) Smoothing, (7) Removing the linear trend, (8) Temporal filtering, (9) Removing covariates. [...] Our brain is consisted of different regions. Although each region carries out its own tasks, they often interconnect with each other, and the connections form the brain network. Graph could be employed to represent networks. A graph has two major components, nodes and links. fMRI image could be divided into 90 regions based on the Automatic Anatomical Labeling (AAL) (Tzourio-Mazoyer et al., ) atlas. We make these regions as the nodes of brain network, so there are 90 nodes. An average of the time series of a region's all voxels can be used to capture the mean time series of this region (Khazaee et al., ). The average time series of every two brain regions can be employed to calculate the Pearson correlation coefficient which is made as the link of the brain network (Khazaee et al., ), so there are 4,005 (90*89/2) weighted edges. Then we take the absolute value of the correlation coefficient and set a suitable threshold for the connectivity matrix to get an adjacency matrix. In this paper, the threshold is 0.25. Graph metrics calculated for this paper included: degree, shortest path, local efficiency and clustering coefficient. Thus, results have 90°, 4,005 shortest paths, 90 local efficiency, and 90 clustering coefficients. These graph metrics were the subsequent experimental features. […]

Pipeline specifications

Software tools DPABI, AAL
Applications Magnetic resonance imaging, Functional magnetic resonance imaging
Organisms Homo sapiens
Diseases Nervous System Diseases, Q Fever