Functional Brain Areas Mapping in Patients with Glioma Based on Resting-State fMRI Data Decomposition

What Can Resting-State fMRI Data Analysis Explain about the Functional Brain Connectivity in Glioma Patients?

“…Supporting this statement, they also found that the same sequence achieved better concordance both with the tb-fMRI maps and the motor cortex template. This is particularly significant considering that other studies using longer TR [23] only report discrete overlap between the two fMRI approaches (Dice coefficient ~0.3). It must be noted that constrained versions of ICA, where a priori spatial information concerning the regions known to be functionally involved in the network are fed to the algorithm, show improved concordance with tb-fMRI [23].…”

“…This is particularly significant considering that other studies using longer TR [23] only report discrete overlap between the two fMRI approaches (Dice coefficient ~0.3). It must be noted that constrained versions of ICA, where a priori spatial information concerning the regions known to be functionally involved in the network are fed to the algorithm, show improved concordance with tb-fMRI [23]. Diagram representing the number of papers reviewed for the present work, subdivided on the basis of their goal, functional network studied, and analytical method used, reaching a total of 1452 patients examined .…”

“…Using a discriminability-index-based component identification algorithm for the automatic recognition of the LN among the ICA components, Lu and colleagues [27] were able to achieve significant results in terms of automatic identification of the network in healthy controls and glioma patients and high sensitivity in the overlap of the reconstructed maps with the DES points: 87% when considering a 1 cm radius from the stimulated region against 47.8% obtained through SBA. Another study using both ICA and semi-blind ICA [23] to identify the LN, despite exhibiting more variable results depending on the infiltration of the tumor in the language areas, showed even higher concordance than the motor one with the corresponding tb-fMRI activation outcomes (Dice coefficient ~0.5 when language areas were not directly involved by the lesion and ~0.2 otherwise).…”

“…Using ICA, the MN has been successfully identified in patients affected by supratentorial glioma in different recent studies [23,24]. However, Voets and colleagues [24] reported that this approach was not able to isolate the MN in every patient, failing in ~14% of the cases.…”

“…Being able to assess FC properties in more homogeneous groups of patients will improve the power of the rs-fMRI analysis to determine specific alterations associated with different characteristics of the tumors. [22][23][24][25][26][27][28][29][30]32,33,35,[38][39][40][43][44][45][46][47][49][50][51][52][53][54][55][56], grouped by field of application and reported in order of citation. This information was not available for the other 7 studies [31,34,36,37,41,42,48].…”