PT - JOURNAL ARTICLE AU - Işıl Uluç AU - Lisa Alexandria Velenosi AU - Timo Torsten Schmidt AU - Felix Blankenburg TI - Parametric representation of tactile numerosity in working memory AID - 10.1523/ENEURO.0090-19.2019 DP - 2020 Jan 09 TA - eneuro PG - ENEURO.0090-19.2019 4099 - http://www.eneuro.org/content/early/2020/01/09/ENEURO.0090-19.2019.short 4100 - http://www.eneuro.org/content/early/2020/01/09/ENEURO.0090-19.2019.full AB - Estimated numerosity perception is processed in an approximate number system (ANS) that resembles the perception of a continuous magnitude. The ANS consists of a right lateralized frontoparietal network comprising the lateral prefrontal cortex (LPFC) and the intraparietal sulcus. Although the ANS has been extensively investigated, only few studies focus on the mental representation of retained numerosity estimates. Specifically, the underlying mechanisms of estimated numerosity working memory (WM) is unclear. Besides numerosities, as another form of abstract quantity, vibrotactile WM studies provide initial evidence that the right LPFC takes a central role in maintaining magnitudes. In the present fMRI MVPA study in numerosity WM, we designed a delayed-match-to-numerosity paradigm to test what brain regions retain approximate numerosity memoranda. In line with parametric WM results, our study found numerosity-specific WM representations in the right LPFC as well as in the supplemental motor area and the left premotor cortex extending into the superior frontal gyrus, thus bridging the gap in abstract quantity WM literature.Significance Statement While the perception of approximate numerosities has been extensively investigated, research into the mnemonic representation during working memory (WM) are relatively rare. Here, we present the first study to localize WM information for approximate numerosities using functional magnetic resonance imaging (fMRI) in combination with multivariate pattern analysis (MVPA). Extending beyond previous accounts that used either a priori brain regions or electrocorticography (EEG) with poor spatial resolution and univariate analysis methods, we employed an assumption-free, time-resolved, whole-brain searchlight MVPA approach to identify brain regions which code approximate numerosity WM content. Our findings, in line with previous work, provide preliminary evidence for a higher level, modality- and format-independent abstract quantitative WM system which resides within the right lateral PFC.