Jargon in reading aloud sparing Arabic digits but not number words

• ¹ University of Padova, Neurosciences Department, Italy
• ² San Camillo Hospital Foundation IRCSS, Venezia Lido, Italy
• ³ Ospedali Riuniti, Trieste, Italy

A person with aphasia, DRG, showed a severe phonological/neologistic jargon when reading aloud but in a lesser degree in naming and spontaneous speech. Reading digits was free of phonological errors. Case report: DRG, 58 y.o.; education = 8 years. Vascular lesion: left occipito-temporal and right temporo-parietal (CT). No auditory cortical ERP were found. Neuropsychological examination: DRG showed the typical pattern of cortical deafness. He could not perceive any verbal and non-verbal sound. He did not blink to very loud sounds. His spontaneous speech was relatively understandable, fluent and loud, with phonemic paraphasias and a few neologisms. He had no problems in visual recognition and written calculation showed only occasional errors in complex operations. His reading comprehension was preserved while reading aloud was severely neologistic (Table 1). Experimental investigation: DRG's naming and reading aloud of words were compared at an interval of nine months. The amount of correct items and of non target-recognizable neologisms is reported in Table 1. A worse performance in reading aloud was consistent over time and tasks. Reading aloud of numbers, up to six digits long, presented in the Arabic code (e.g., "9", "267") and in the alphabetic code (e.g., "nine", "two hundred sixty seven"), resulted in a different performance. Reading aloud of alphabetically presented words was successful in only 28/110 items (25%), with neologisms being the most frequent error type. Reading Arabic numbers was successful in 25/55 items (45%); no phonological/neologistic error was produced, errors being of the "lexical" type, thus numbers were substituted by other phonologically correct numbers. Discussion: DRG showed a previously unreported dissociation in speech output: naming and connected spontaneous speech were less affected by phonological disturbances than reading aloud. The opposite dissociation, reading aloud superior to naming, was described by Semenza et al. (1992), also in a case of cortical deafness. These authors explained their case suggesting a difficulty in activating the speech output lexicon from the semantic system and by a disturbed print to sound conversion (sublexical route). Reading aloud could be performed either via the direct, lexical, non semantic, route or by summation of residual capacities of disturbed access to speech output lexicon from the semantic system and disturbed print to sound conversion. In contrast, DRG would be disturbed in the direct lexical route and in print to sound conversion: lexical activation of the speech output lexicon from the semantic system would not be sufficient to read correctly aloud without the help of the other two routes. Lack of phonological errors in reading Arabic numbers would be explained by the fact that Arabic numbers, consistently with recent literature (Bencini et al., 2011; Semenza et al., 2014; Dotan and Friedmann, 2015), would directly activate the whole phonological form rather than needing grapheme to phoneme conversion. This case provides insights about the processing dynamics among the three reading routes.