Abstract
We recently described neural populations tuned to object number (numerosity) forming a topographic map in the human right parietal lobe (Harvey et al, Science 2013; Harvey et al, PNAS 2015). Previous studies describe activations in many parietal and frontal areas during numerosity tasks. We hypothesize that there is an extended numerosity-processing network comprising multiple numerosity maps, analogous to the multiple visual field maps that process visual space. Methods: We acquired ultra high-field fMRI (7T) data while showing visual stimuli of changing numerosity. We used several stimulus conditions that distinguish numerosity selectivity from selectivity to co-varying low-level stimulus features. We analyzed responses using population receptive field modeling (Dumoulin and Wandell, 2008, Neuroimage) to summarize each recording site's response as a tuning function with a particular preferred numerosity and tuning width. Results: We find several maps of gradually changing numerosity. Our previously described map lies medial to the intraparietal sulcus and posterior to the postcentral sulcus. We find two further parietal numerosity maps in the postcentral sulcus, one superior and one inferior to the junction with the intraparietal sulcus. In the occipital lobe, we find one map at the superior end of the parieto-occipital sulcus, with another anterior/inferior to the lateral occipital sulcus and posterior/inferior to the inferior temporal sulcus. Finally, we find a frontal numerosity map at the junction of the pre-central and superior frontal sulci. Progressions of preferred numerosity and tuning width are remarkably similar between maps, although we find some differences in lateralization. Conclusion: Numerosity maps group similar numerosity preferences together, and gradually change numerosity preferences across the cortical surface. However, several broadly distributed areas process numerosity, each with similar response properties. We speculate these areas may support distinct cognitive roles of numerosity processing, such as multiple object tracking, distribution of attention across objects, and decision-making.
Meeting abstract presented at VSS 2016