Aphasia: progress in the last quarter of a century

Hillis AE. Neurology 2007; 69: 200-213. (Views and reviews).
Summary: Dax, Broca and Wernicke made observations that Geschwind then summarized and reclassified in the 1960's. His BDAE held throught the CT/ MRI era.

In the 1980s, PET, fMRI and MEEG revolutionized our thinking. They showed bilateral involvement during language tasks, although the left hemisphere has more activation. These activated areas include remote locales such as inferior and anterior temporal cortex and basal ganglia and thalamus. The cognitive processes involved in language tasks in increasingly complex models were mapped out, involving PDP systems, linguistics, mathematics, and neuroscience. Third, analysis of dementia allowed analysis of semantic dementia after lobar degeneration, which does not occur after stroke.

Minutiae discovered include the constant finding of activation of the midfusiform area during reading, and the left midfusiform area during naming and other lexical tasks. Inhibition with cortical stimulation among presurgical epileptic patients of this area or the posterior basal temporal area disrupts kanji (represents meanings not sounds) and picture naming (semantic to phonologic conversions are necessary) but not comprehension of kana (sounds of words), copying or tool use (Usui, Brain 2003). Infarction of this area caused impaired oral reading and naming.

Classic aphasiology is reviewed. In Broca's aphasia, "nonfluency"can refer to decreased length of phrases, impaired melodies and articulatory agility, decreased words per minute, or agrammatism. Most patients have islands of preserved fluency especially of overlearned speech. This confounds classification. Comprehension of complex forms or noncanonical phrases (eg. passive voice) or comparisons of word order or meaning are damaged (Is a horse larger than a dog?) as is spelling of familiar words, and use of phonics to spell unfamiliar words. She criticizes the argument htat a single underlying impairment can cause Broca's syndrome, but allows that it is a vascular syndrome. She also discusses the conundrum of Broca's area lesion causing speech apraxia, and warns against excessive inflexibility in "localization," since individuals vary in their makeup. Hypoperfusion of surrounding tissue without infarction is another source of error.

Wernicke's aphasia consists of fluent jargons, neologisms, and lack of awareness of the deficit. Melody is preserved. Writing parallels speech (meaningless). It relates to posterior division of the MCA .

ACA lesions cause TCM aphasia (or watershed strokes between ACA and MCA).

Conduction aphasia has fluent accurate spontaneous speech with phonemic paraphasias, with conduit 'approche. Author rebuts Geschwind that the arcuate fasciculus is important, citing lesions in SMG or deep parietal white matter and on lesion studies showing lack of reliability of arcuate fasciculus lesions causing conduction aphasia.

Author discusses pure alexia and takes the classic disconnection explanation, emphasizing the co-occurrence with optic aphasia and the similarity of the neurologic lesions. She discusses semantic errors in such patients noting that they are partial.
Hillis discusses aphasia secondly, from the pov of the distributed semantic networks implied in addition to the previously described vascular syndromes. These account for category selectivity in naming, eg.

Treatment is discussed briefly, but author notes that intense treatment 4 days per week for a short time is more effective than spread out sessions over alonger period (Bhogal, Stroke, 2003).