# Semantic Processing of Arabic Numbers Across Tasks

## TL;DR

The study shows that numbers are processed automatically like words and pictures, but use different brain networks for quantity understanding.

## Contribution

It reveals distinct neural mechanisms for numerosity versus general semantics using ERP repetition effects across tasks.

## Key findings

- N400 repetition effects for numbers were consistent across form matching, arithmetic, and quantity tasks.
- Scalp distribution of N400 for numbers differed from that for words, indicating distinct brain networks.
- Numerosity processing appears automatic but involves different neural pathways than general semantics.

## Abstract

Numbers are used in a variety of ways and in different contexts—as labels, markers of ordinality, or indicators of quantity, in addresses and phone numbers and in mathematical equations. This raises the question of whether knowledge access from numbers involves similar or distinct mechanisms across these uses and how it compares to accessing knowledge from words and pictures. To investigate this, we presented double‐digit numbers in three tasks designed to target different types of information: a matching task requiring access only to number form, a divisor task situating numbers in the context of basic arithmetic, and a quantifier task using numbers to represent everyday quantities. We measured event‐related potentials (ERPs), focusing on the N400 component, which has been linked to access from long‐term semantic memory, and looking at the impact of repetition as an implicit probe of facilitated knowledge retrieval. Our results revealed reliable N400 repetition effects for numbers across all tasks, suggesting that numbers are linked to associated representations of numerosity in a relatively automatic manner, using similar mechanisms as the access of semantics from words and pictures. However, consistent with claims that representations of numerosity involve different brain networks compared to general semantics, the scalp distribution of the N400 repetition effect for numbers, which was consistent across our three tasks, differed from that to words in the present experiment and from that observed in prior work using numbers to access general semantics.

Using event‐related potentials (ERPs) in a repetition paradigm, we found that accessing numerosity information from Arabic numbers produces consistent N400 effects across tasks involving simple form matching, arithmetic, and quantity judgments, indicating relatively automatic processing. However, differences in N400 scalp distributions between numbers and words suggest that distinct brain networks are involved in representing numerosity versus general semantics.