The current understanding of mathematical and numerical processes indicates that mathematical abilities are not singular. Instead, they consist of multiple components that can be linked to various domain-general abilities. This understanding is derived from studies in typically developing children, adults, and children with learning disabilities [[1], [2], [3], [4], [5], [6]].

The main components of arithmetical cognition are commonly grouped into three categories: factual, procedural, and conceptual knowledge. Factual knowledge is the ability to retrieve fast and effectively from long-term memory answers to well-known simple operations. It has been tested mostly during the solution of single-digit addition and multiplication problems under limited time. Procedural knowledge refers to the ability to understand how to solve multistep operations, required in two-digit addition and subtraction, equations, and series. Conceptual knowledge includes mathematical rules such as the base 10 structure and parity [7].

Another categorization that differentiates between mathematical and numerical tasks is the triple code model [[8], [9], [10], [11], [12], [13]]. According to this model, two main codes of number representation can be activated during mathematical processing, contingent on the task performed and proficiency in that task. One of these is the semantic code, which is specific to mathematics and is activated to understand the magnitude of a given number or the relationship between numbers. The semantic code involves the analogic, non-symbolic magnitude representation of a number 3, i.e., the meaning of the number (e.g., •••). It is activated during comparison, mathematical and number line estimations and subtraction. The verbal code involves a format-specific representation of a number: the word frame (e.g., three). The verbal code is not number-specific and is shared with reading. It is activated during well-rehearsed mathematical tasks such as the retrieval of arithmetic facts or number reading [[8], [9], [10], [11], [12], [13]]. The third code is the visual Arabic code, which is specific to the format of representation.

The triple code model not only differentiates between mathematical tasks, but also differentiates between calculations. Single digits multiplication, and to some degree addition, are associated with the verbal code, and are based on a non-numerical associative network. By contrast, multi-digit subtraction is associated with the semantic code [[8], [9], [10], [11], [12], [13]].

Differentiated mathematical tasks are also contingent on different domain-general abilities [4,[14], [15], [16], [17], [18], [19]]. Domain-general ability reflects cognitive skills required for learning in general, but not specific to mathematics. Based on the triple code model, one of the main domain-general abilities is reading and language. In the triple code model, the verbal numerical representation is central to arithmetical thinking in school [[20], [21], [22]]. For example [22], tested the pathways to mathematics that were based on the triple code model, and discovered that the linguistic pathway is one of the strongest domain-general abilities needed in school mathematics. Specifically, linguistic abilities were the strongest and most significant predictor of calculation, geometry, and mathematical reasoning. Later, the same group [23] proposed that linguistic abilities are strongly associated with arithmetical fluency, mathematical rules (such as base 10) and complex calculation. However, this relation is modulated by the specific math task and the specific linguistic ability used. One of the linguistic abilities is RAN (Rapid Automatic Naming) that tests phonological abilities. In a meta-analysis, it was discovered that RAN produced stronger correlations with arithmetic calculation tasks than with general achievement tests; stronger correlations with single-digit calculation tasks than multidigit calculation tasks; and stronger correlations with math fluency tasks than math accuracy tasks [24].

The correlations between mathematical abilities and reading abilities are also very strong (r = 0.52) [25]. Word reading was more strongly associated with computation (r = 0.46) than with mathematics fluency (r = 0.36). However, in the same study it was discovered that this strong correlation between reading and math is based upon similar domain-general factors that influence mathematical processing and reading such as IQ, RAN, and working memory. The authors concluded that all the domain-general abilities were part of a general factor (G) that influences mathematical abilities (0.85) and reading abilities (0.76) [25]. Hence, it will be important to test, in the same study, many domain general factors such as RAN, intelligence and working memory in addition to reading.

The connection between mathematical processing and intelligence has been documented in several studies [[26], [27], [28], [29]]. Reasoning is essential for understanding numerical symbols, their relationships, and their applications in the number system, as well as the rules and principles for calculations.

Working memory plays a crucial role in mathematical processing. Many math tasks require individuals to actively maintain multiple concepts of mathematical expressions and intermediate numbers while solving problems [6,14,15,18,21,[30], [31], [32], [33], [34], [35], [36]].