Author:

Abstract:

We continuously shift our attention between items in the visual environment, which can be based on either task relevance (top-down) or sudden saliency of an unexpected stimulus (bottom-up). As this is a fundamental process for guiding goal-directed behavior, it is to be expected that this property is evolutionary conserved in primates. In humans top-down directed spatial attention shifts most strongly activate the medial superior parietal lobe (mSPL) (Vandenberghe et al. 2001, Yantis et al. 2002). In order to identify a potential monkey homologue of this mSPL region, we scanned (at 3 T) monkeys (N=3) and humans (N=31) performing the same covert selective spatial attention task. We used an eventrelated fMRI design whereby periods of shifts and sustained attention were interleaved (Caspari et al., 2015; Molenberghs et al., 2007). Two pairs of shapes were presented on the horizontal meridian (9.25 deg), each containing a relevant and irrelevant shape. Subjects fixated centrally and responded manually when the relevant stimulus dimmed. An event consisted of the replacement of the current stimulus pair by the next. In 1/3 of the trials, this change elicited a covert spatial shift in attention as the relevant stimulus was replaced by an irrelevant one. Monkeys were scanned (1.25 mm isotropic) using implanted phased-array Rx coils and MION, and humans (2.75 x 2.75 x 3.5 mm) using BOLD. First, we topologically compared shift-selectivity across species using standard GLM analyses. Next, we performed a novel data-driven Inter Species Beta Correlation (ISBC) analysis using 10 dissociable task conditions modeled by the GLM. These included shift (L/R), sustained attention (L/R), dimming (L/R and relevant/irrelevant) and null events (L/R). The GLM-contrast showed highly similar shift responses across species in the medial superior and inferior parietal lobe, with mSPL as strongest shift-selective region in humans. Monkeys recruited more profoundly frontal regions during shifting compared to humans. The ISBC method revealed a subset of parietal and frontal voxels in monkeys correlating with activity averaged over the shift-selective human mSPL-cluster. Medial parietal cortical areas V6/V6A in monkey correlated best with the shift-selective local maxima within human mSPL (based on the GLM contrast), compared to the cytoarchitectonic SPL subdivisions SPL5 and SPL7 from Caspers et al. 2007. Finally, functional connectivity of human shift-selective mSPL at rest resembled known anatomical connections of V6/V6A in monkey, indicating that both species recruit evolutionary largely conserved anatomical regions for spatial attention shifting within medial SPL.