Cortical indices of picture name encoding times in fluent and stuttering adults
Maxfield, Nathan D
MetadataShow full item record
Speakers generate utterances through a system of psycholinguistic processes, including lexical selection, grammatical encoding, and phonological encoding. Efficiency in these processes facilitates verbal fluency. Since adults who stutter are chronically dysfluent, the purpose of this dissertation was to assess whether utterance planning operates inefficiently for them. Based on converging evidence, three utterance planning processes were targeted. First, we assessed the rate at which adults who stutter access constituent sounds of words during utterance planning. Second, we examined the rate at which they access linguistic knowledge involved in generating grammatical structures. Third, we measured the rate at which adults who stutter access linguistic knowledge involved in generating phonological plans for connected speech. To make these assessments, we tested adults who stutter and fluent control subjects in three lexical decision tasks. For each task, they tacitly labeled pictures and made binary decisions about target properties of the labels. For task one, participants judged whether labels ended with phoneme /n/ or /d/. For task two, participants judged which of two grammatical determiners the labels elicited. For task three, participants judged whether resyllabified labels had one or two syllables. The binary decisions for each task were mapped onto a "go-nogo" response paradigm, for which participants pressed a button in response to one decision but not the other. During testing, event-related potentials were recorded at the scalp. Exploratory analysis revealed both expected and unexpected ERP components. The time-course of an expected component, "nogo N200", was measured as an index of when, following picture onset, participants retrieved response-inhibiting targets in each task. The amplitude of an unexpected ERP, "left temporal negativity", was measured as an index of heightened processing activity. "Nogo N200" latency measurements revealed that a large subgroup of dysfluent participants retrieved word-final sounds (task 1), on the average, 95 milliseconds earlier than fluent participants. Furthermore, all three tasks elicited left temporal scalp activations that were larger for dysfluent than for fluent participants. Results are discussed in terms of the amount of control adults who stutter have over the timed retrieval of phonological segments, and the amount of cognitive activity they generate when processing language.