Spectral clustering of low-frequency fluctuations in fMRI data reveal a distinct separation between the superior temporal sulcus and the superior temporal gyrus

Introduction Low Frequency Fluctuations (LFFs) are known to represent a large portion of the variance of the BOLD signal, especially in non-primary areas of cortex. Furthermore, such fluctuations generally have significant spatial coherence. Task-dependent condition-locked fMRI data has confirmed an important role of the superior temporal cortex in many language and hearing related processes. Within this area, many studies have claimed to identify activation distinct to superior temporal gyrus (STG) and superior temporal sulcus (STS), for different stimuli ([1],[2]). Using a data-driven clustering technique applied to LFFs, we investigated the spatial coherence structure of the upper temporal lobe and sought to substantiate the suggested functional distinction of gyrus and sulcus.