Sequential and parallel letter processing in letter-by-letter dyslexia

Abstract

Four experiments are reported that focus on the issue of sequential vs. parallel letter processing in letter-by-letter (LBL) dyslexia; these were conducted on patient IH. Expt. 1 showed a large linear reduction of word naming times with an increase in the number of orthographic neighbours of the target (i.e., words of the same length differing by just one letter; N size). Given the large negative linear correlation existing between word length and N size, this result raises the possibility that the large word length effect diagnostic of LBL dyslexia may be, in fact, an artefact of uncontrolled N size. Expt. 2 falsified this possibility by showing that the word length effect is unaffected by whether N size is controlled for or not. This result also suggested that the facilitatory effect of increased N size in LBL dyslexia is based on the parallel processing of the constituent letters of the target. Further supporting a contribution of parallel letter processing to overt word recognition performance in the disorder, Expt. 3 showed significant but independent effects of word length and letter confusability (i.e., similarity of the constituent letters of the target word with other letters of the alphabet). The letter confusability effect therefore appears to rest on the parallel analysis of the letters in the target word. Finally, Expt. 4 showed that the facilitatory effect of N size is prevented with high letter-confusability targets. These observations suggest that LBL dyslexia rests on an impairment of letter encoding that results in an excessive level of background noise in the activation of lexical-orthographic representations when letters are processed in parallel. This prevents overt identification of the target and forces sequential letter processing in order to achieve this goal.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 45 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.