Degenerative diseases such as frontotemporal lobar degeneration and Alzheimer’s disease can cause selective language impairments that get worse over time, a condition known as Primary Progressive Aphasia (PPA). An inability to name objects, known as anomia, is a common symptom in PPA. Anomia could be caused by difficulty in (a) object recognition, (b) comprehension of word meaning, (c) access to word forms, or (d) phonology/articulation, all of which need to be intact in order to name objects aloud. Event-related potentials (ERPs) are uniquely suited for the study of object naming deficits: they provide excellent temporal resolution relevant for evaluating staged processes of naming, and the N400 ERP potential is thought to reveal the extent of predictive encoding of words.
Participants completed a task evoking key processes involved in object naming. Picture cues were followed by word targets; participants pressed one button when the word matched (i.e. named) the picture, and another button for mismatches. Mismatches were of two types: related mismatches from the same category as the picture, and unrelated mismatches from a different category. ERPs were examined in response to the target words. After ERP testing patients were shown the same set of pictures and were asked to name each one aloud, allowing for examination of effects on trials with named versus unnamed items.
Prior to our work, ERPs had not been examined in a PPA population. In the first study, we explored whether language dysfunction in PPA results in ERP abnormalities. In both controls and PPA patients the N400 was of larger amplitude to mismatched compared to matching targets (N400mismatch), reflecting predictive encoding of the object’s name. Although present in controls, PPA patients did not show a differential increase in amplitude to unrelated compared to related targets (N400differential), indicating decreased resolution in object-word association. The N400mismatch was found to relate to the phenomena of anomia as well. A subset of patients with severe anomia showed an N400mismatch to items they could name, but no N400 effects to items they could not name, revealing a lack of predictive encoding for those items. The N400mismatch was further shown to index semantic rather than post-semantic stages of naming, as shown by correlations with neuropsychological tests. This was further demonstrated in a patient with anomia due to post-semantic factors, who showed impaired oral naming but intact written naming. This patient showed an intact N400mismatch to unnamed items.
It was unclear to what extent single-word comprehension deficits contributed to the results from the first study. Results were reanalyzed with an expanded sample in order to address this question. Patients of the semantic subtype (PPA-S), who show profound word comprehension deficits, were shown to have difficulty in differentiating between members of the same category, as shown by abnormal N400 potentials, accuracy, and reaction times on trials with related targets. This difficulty was found to relate to anomia in PPA-S; accuracy was further reduced on matching and related trials with unnamed targets. Non-semantic subtypes of patients also showed no N400 effects to unnamed items, but no corresponding changes in accuracy, suggesting isolated failures in predictive encoding not due to comprehension failures.
The goal of the second study was to determine whether atrophy in the cortical epicenters of language impacted results from the picture-word task. Cortical thickness measurements were derived from magnetic resonance images. Atrophy in posterior temporal, but not anterior temporal or inferior frontal areas, was associated with diminished N400mismatch potentials. Atrophy in anterior temporal areas was instead associated with a reduction in accuracy on related trials.
Our results validate the use of N400 potentials for the study of anomia. These potentials reveal the extent of lexical activation in response to pictures, a critical piece of information in understanding anomia which cannot be examined based on behavior alone.