Novel matched stimuli for assessment of lexical semantics

Background: Diagnosis of semantic impairment in stroke and progressive neuro-cognitive conditions is typically facilitated using tests of word comprehension, such as word-picture matching. Many of these tests are not controlled for psycholinguistic variables or the semantic relationships between competitor stimuli and involve pictures which are not controlled for ease of access. Semantic assessment also demands additional cognitive resources, such as explicit decision-making and suppression of semantic competitors. These factors may all confound test performance and subsequent diagnosis.

Aims: To develop novel semantic test stimuli for three new semantic processing assessments, which are controlled for psycholinguistic variables, semantic relationship between stimuli, and visual similarity between images presented simultaneously. An additional aim included matching stimuli for these variables across three tests: semantic priming, word-picture verification, and word-picture matching, to allow direct comparison of performance on tests that differ in terms of the additional cognitive demands involved, with priming entailing implicit semantic processing.

Methods & Procedures: In phase one, novel stimuli were developed for the three semantic processing tests. Existing databases were searched for values to match stimuli psycholinguistic variables. In phase two, new data were collected from control participants regarding the semantic and visual similarity of stimuli presented simultaneously.

Outcomes & Results: Data for three sets of target and distractor stimuli are presented, which are psycholinguistically matched within and between the three semantic tests for concreteness, imageability, age of acquisition, frequency, word length, and emotional valence. “Semantic” relationships between pairs of stimuli are differentiated by semantic similarity (dog-cat) or association (dog-lead). Visual similarity is controlled between images presented in an array.

Conclusions: The data provided ensure that test performance across three semantic tasks, differing in additional cognitive demands, can be directly compared in people with potential semantic deficits. This is the first such study to provide control of stimuli within and across a range of semantic tests. Patterns of performance via test reaction time and accuracy data may reveal semantic deficit or competence, contributing to more accurate diagnosis and appropriate therapy choice.