Purchase this article with an account.
Nishmar Cestero, Abigail Noyce, Barbara Shinn-Cunningham, David Somers; MVPA reveals specialization and generality of sensory-biased regions of frontal cortex. Journal of Vision 2016;16(12):1072. doi: 10.1167/16.12.1072.
Download citation file:
© 2017 Association for Research in Vision and Ophthalmology.
Discrete visual- and auditory-biased attention structures have recently been identified in caudolateral frontal cortex (Michalka 2015). Two visual-biased regions: superior and inferior precentral sulcus (sPCS and iPCS) are interleaved with two auditory-biased regions: transverse gyrus intersecting precentral sulcus (tgPCS) and caudal inferior frontal sulcus (cIFS). Using multivariate pattern analysis (MVPA), we tested whether structures preferentially active during visual or auditory working memory (WM) contain information about memory content. We collected fMRI while subjects (n = 12) performed visual and auditory 2-back tasks. Visual stimuli were faces (male or female in separate blocks); auditory stimuli were animal vocalizations (cat or dog). Visual- and auditory-biased frontal regions were defined by contrasting visual with auditory WM activation. We performed two cortical surface-based MVPA analyses using the CosmoMVPA toolbox and a support vector machine to test within-modality decodability. Two separate classifiers attempted to categorize visual and auditory WM blocks into stimulus sets (visual: male or female faces; auditory: cat or dog vocalizations). In the first analysis, a whole-brain surface searchlight approach defined 100-vertex circular neighborhoods for each participant. Both visual and auditory classification was better than chance in caudolateral prefrontal cortex, pre-supplementary motor area, and anterior insula. Additionally, visual classification was decoded by activity in intraparietal sulcus, while auditory categories were predicted by superior temporal gyrus and angular gyrus activity. The strong classifier performance in caudolateral frontal cortex led us to further investigate our four frontal, sensory-biased regions, defining them as ROIs for MVPA classification. Classifiers for visual-biased sPCS and iPCS robustly decoded both visual and auditory categories, while auditory-biased tgPCS and cIFS ROIs robustly supported auditory categorization, but only weakly supported visual categorization. These results suggest that visual-biased caudolateral frontal structures may be part of a more domain-general cognitive circuit, while auditory-biased structures may be more domain specialized.
Meeting abstract presented at VSS 2016
This PDF is available to Subscribers Only