Cognitive loci of impairments in picture naming by aphasic subjects

In order to identify the process or processes responsible for impaired naming by aphasic patients, ten aphasic adults and ten normal adults performed three independent tasks--picture naming, modified Sternberg picture recognition, and modified Sternberg random shape recognition (Sternberg, 1966). Response times and error percentages were the dependent variables. Independent variables in naming were stimulus codability measured in bits of uncertainty (two levels) and number of naming trials (three trials). Independent variables in the recognition tasks were uncertainty (two levels), number of stimuli to be remembered (two or four stimuli) and response type ("yes" or "no"). The results showed that uncertainty had significant effects on naming but not on recognition performance. The aphasic group produced significantly longer naming response times regardless of uncertainty level. The differences between groups were much greater for high-uncertainty pictures (1100 msec) than for low-uncertainty pictures (270 msec). A comparison of estimates of word retrieval times showed that the two subject groups differed significantly for high-uncertainty but not for low-uncertainty items. It was concluded that (1) use of the term "word retrieval problem" rather than "loss of memory problem" was justified to describe the major component in the aphasic naming impairment for high-uncertainty items, and (2) when naming low-uncertainty items these aphasic subjects did not demonstrate a word retrieval problem. The results indicate that treatment procedures designed to improve naming should be process rather than content oriented.