A CONNECTIONIST MODEL OF ALEXIA - COVERT RECOGNITION AND CASE MIXING EFFECTS

A connectionist model was developed to simulate the production of phon ological, semantic and orthographic lexical representations from an or thographic input code. When lesioned, the network displayed many of th e characteristics of the neurological disorder, purr alexia. These inc luded: greater accuracy in the recognition of letters compared to word s, and (in some cases) spared semantic categorization and lexical deci sion ability relative to naming. Performance was superior for high rel ative to low frequency words and (for some lesions) for 'high imageabi lity' words with more semantic referents. Similar to alexic patients, errors in the model tended to be visually rather than semantically rel ated to target words and disrupted input had a greater effect on namin g than on lexical decision performance. There were also tendencies app arent in the model which have not so far been reported for alexic pati ents but are found in other neurological patients, including category specificity effects and superior performance on superordinate relative to subordinate semantic categorization. Noise was added to the input units of the unlesioned and lesioned model in an attempt to simulate c ase mixing effects on both normal and alexic reading. The unlesioned m odel demonstrated similar effects to normal readers with mixed case st imuli (Besner & McCann, 1987; Mayall & Humphreys, 1996). Further, effe cts on the lesioned model were comparable to those found with alexic p atients (Bub & Arguin, 1995; Mayall & Humphreys, submitted). The simil arity of the performance characteristics of the model to those of pure alexic patients suggests that the architectures of the modelled and t he human reading system encompass common properties, and that covert r ecognition in pure alexia can be attributed to the architecture of the word processing system.