Processing of dialect-specific phonetic differences
Behavioral studies show that humans are quite sensitive to phonetic differences that are associated with gender, ethno-racial, and geographic differences. In a project led by William Idsardi and Thomas Purnell, we are investigating the general hypothesis that these phonetic differences are registered in early differential neural activation. This is assessed by subtracting the MEG waveform associated with one value (e.g. gender = male) from the waveform associated with the contrasting value (gender = female) in an oddball paradigm. We expect to find a mismatch negativity when a deviant token differs in a biosocially relevant manner from a string of preceding standard tokens (i.e. gender is changed).
Neural representation of syntactic vs. phonological information
In a study led by Dr. Lisa Sanders, we used behavioral and ERP measures to determine whether phonological violations (i.e. presenting listeners with the wrong allomorph of the plural suffix) and syntactic violations (deletions that result in number disagreement) are processed using the same neural system. In the behavioral task, listeners heard passages containing a small proportion of phonological and syntactic violations, and they were asked to press a button in response to hearing anything abnormal. Behaviorally, they were able to detect the syntactic violations, but not the phonological violations. In the ERP measurements, we found differences in the polarity, latency, and distribution of the earliest ERPs associated with phonological and syntactic violations, which suggests that a single morpheme may be processed using distinct neural populations.