Dissociations between word and picture naming in Persian speakers with aphasia

• ¹ The university of Hong Kong, Division of Speech and Hearing Sciences, Hong Kong, SAR China
• ² Tehran University of Medical Sciences, Speech Therapy Department, Iran

Studies of patients with aphasia have found dissociations in their ability to read words and name pictures (Hillis & Caramazza, 1995; Hillis & Caramazza, 1991). Persian orthography is characterised by nearly regular orthography-phonology (OP) mappings however, the omission of some vowels in the script makes the OP mapping of many words less predictable. The aim of this study was to compare the predictive lexico-semantic variables across reading and picture naming tasks in Persian aphasia while considering the variability across participants and items using mixed modeling. Methods and Results A total of 21 brain-injured Persian-speaking patients suffering from aphasia were asked to name 200 normalized Snodgrass object pictures and words taken from Bakhtiar, Nilipour and Weekes (2013) in different sessions. The results showed that word naming performance was significantly better than object naming in Persian speakers with aphasia (p<0.0001). Applying McNemar’s test to examine individual differences found that 18 patients showed significantly better performance in word reading compared to picture naming, 2 patients showed no difference between naming and reading (i.e. case 1 and 10), and one patient (i.e. case 5) showed significantly better naming compared to reading χ (1)=10.23, p< 0.01 (see also Figure 1). A mixed-effect logistic regression analysis revealed that the degree of spelling transparency (i.e. the number of letters in a word divided by the number of its phonemes) had an effect on word naming (along with frequency, age of acquisition (AoA), and imageability) and picture naming (along with image agreement, AoA, word length, frequency and name agreement) with a much stronger effect on the word naming task (b= 1.67, SE= 0.41, z= 4.05, p< 0.0001) compared to the picture naming task (b= -0.64, SE= 0.32, z= 2, p< 0.05). Conclusion The dissociation between word naming and picture naming shown by many patients suggests at least two routes are available for lexical access for phonological output: a semantic route for picture naming and a route for mapping orthography to phonology directly. The main effect of spelling transparency along with other lexico-semantic variables (such as frequency, imageability and AoA) on word naming suggests an interaction between these routes (Bi, Han, Weekes, & Shu, 2007; Hillis & Caramazza, 1995), which makes the word naming less challenging than the picture naming task for most speakers with aphasia. However, the result from case 5 could suggest that the direct route for reading can be selectively impaired. Moreover, the mild effect of spelling transparency on picture naming even after other lexico-semantic variables (e.g. AoA, frequency, imageability, name agreement, imageability, and word length) have been explained, indicates that objects with more transparent spellings have a greater probability of being named accurately in aphasia. This finding raises the possibility that orthographic representations are activated to some degree during picture naming to help access the phonology of the word as long as the spelling of the word is transparent. This explanation is compatible with the summation account of lexical access in aphasia (Hillis & Caramazza, 1995).