A cognitive model for multidigit number reading: Inferences from individuals with selective impairments

Publication date: Available online 14 November 2017
Source:Cortex
Author(s): Dror Dotan, Naama Friedmann
Reading multi-digit numbers aloud involves visual analysis of the digit string and oral production of the verbal number. To examine these processes in detail, we investigated the number processing abilities of seven individuals with different selective deficits in number reading. Some participants were impaired in visual analysis of digit strings – in encoding the digit order, in encoding the number length, or in parsing the digit string to triplets (e.g., 314987 → 314 and 987). Other participants were impaired in verbal production, making errors in the number structure (shifts of digits to another decimal position, e.g., 3,040 → 30,004). Their selective deficits yielded several dissociations: first, a double dissociation between visual analysis deficits and verbal production deficits. Second, several dissociations within visual analysis: a double dissociation between errors in digit order and errors in the number length; a dissociation between order/length errors and errors in parsing the digit string into triplets; and a dissociation between the processing of different digits – impaired order encoding of the digits 2-9, without errors in the 0 position. Third, within verbal production, a dissociation between digit shifts and substitutions of number words. On the basis of these selective impairments and previous findings, we propose a detailed cognitive model of number reading. The model postulates that within visual analysis, separate sub-processes encode the digit identities and the digit order, and additional sub-processes encode the number's decimal structure: its length, its triplet structure, and the positions of 0. Verbal production consists of one process that generates the verbal structure of the number, and another process that retrieves the phonological forms of each number word. We propose that the verbal number structure is first encoded in a tree-like structure and then linearized to a sequence of number-word specifiers, similarly to syntactic trees of sentences. A selective deficit in any of these processes, which we propose to term "dysnumeria", would cause difficulties in number reading.



http://ift.tt/2jug9RH