Slide Slam K4
Learning to write shapes literate speech perception
Alexis Hervais-Adelman1, Uttam Kumar2, Ramesh K. Mishra3, Viveka N. Tripathi4, Anupam Guleria2, Jay P. Singh4, Falk Huettig5; 1University of Zurich, 2Centre for Biomedical Research, Lucknow, India, 3University of Hyderabad, Hyderabad, India, 4University of Allahabad, Allahabad, India, 5Max Planck Institute for Psycholinguistics
Previous research suggest that literacy, specifically learning alphabetic letter-to-phoneme mappings, modifies online speech processing, and enhances brain responses to speech in auditory areas associated with phonological processing. However, alphabets are not the only orthographic systems in use in the world, and hundreds of millions of individuals speak languages that are not written using alphabets. In order to make claims that literacy per se has broad and general consequences for brain responses to speech, one must seek confirmatory evidence from non-alphabetic literacy. To this end, we conducted a longitudinal fMRI study in India probing the effect of literacy in Devanagari, an abugida, which encodes speech as consonant-vowel symbols, on functional connectivity and cerebral responses to speech in 91 variously literate Hindi-speaking individuals. Twenty-two completely illiterate participants underwent six months of reading and writing training, while two control groups, one literate (N=26) and one illiterate (N=12) underwent no training. All three groups returned for a follow-up scan after six months. We find that Devanagari literacy does not correlate with BOLD response to sentences at the whole brain level or in regions of interest selected a priori for their role in phonological processing, the planum temporale (PT) and posterior superior temporal gyrus (pSTG). At baseline there was evidence in favour of the null hypothesis of no relationship between literacy and auditory brain responses (N=91, correlation between word reading scores and BOLD: PT, Kendall's tau = .007, BF10 = 0.137; pSTG, Kendall's tau = -.030, BF10 = 0.149). After training, despite improvements in reading scores and letter recognition, there was evidence against increases in BOLD response to speech in either PT or pSTG (BF10s < 0.3). However, literacy is associated with increased functional connectivity between pSTG and dorsal sensorimotor cortex consistent with the Graphomotor Frontal Area (GMFA) during speech processing (Baseline, N=91, correlation between literacy and pSTG-GMFA connectivity significant a cluster-mass corrected p<.05, 604 voxels), which increased significantly after literacy training, in the trainee group only (Time*Group interaction: F(2,57) = 4.602, p = .015). Taken together, the results suggest that learning to read does not in and of itself alter brain response to speech, but that learning to map sounds onto written symbols potentially fosters phonological-graphomotor connections. We propose that neural orthography on speech processing effects must be considered with regard to the properties of the orthography at hand, in particular, the speech unit that is encoded in the characters - it is conceivable that learning to map subsyllabic segments to a visual code in alphabetic writing systems might require or induce modifications to auditory processing and representations of speech in order to support the phoneme-level manipulations and representations that are less relevant for a syllable based orthography. These findings show that a radical reconfiguration of the neurofunctional substrates of online speech processing is not a universal result of learning to read, and raise the possibility that writing, not only reading, may be instrumental in creating a functional scaffold that can modify literate speech perception.