Neural dynamics of speech perception and production: From speaker normalization to apraxia of speech

This dissertation seeks to enhance understanding of speech mechanisms by employing computational modeling in two key areas: understanding how the brain builds speaker-independent representations of heard speech sounds and why apraxic speakers are unable to effectively generate speech motor programs.;The first portion of the dissertation introduces the Neural Normalization Network model (NormNet) that has been developed to explain how the human brain is able to convert speaker-dependent acoustic information into speaker-independent language representations. NormNet is part of an emerging model of auditory streaming and speech categorization. Multiple strip representations and asymmetric competitive circuits are both used in the auditory streaming and speaker normalization parts of the model, thereby suggesting that these two circuits arose from similar neural designs.;NormNet is able to explain and accomplish speaker normalization by generating pitch-independent representations of speech sounds while preserving information about speaker identity. The speaker-independent representations are categorized into unitized speech items, which input to sequential working memories whose distributed patterns can be rapidly categorized into syllable and word representations and stably remembered by Adaptive Resonance Theory circuits. Model simulations use synthesized steady-state vowels from the Peterson and Barney (1952) database. The model achieves accuracy rates similar to those achieved in human listeners.;The second portion of the dissertation investigates how brain lesions in patients with apraxia of speech (AOS) give rise to different behavioral characteristics. AOS is a disorder of the planning and/or programming of speech production without comprehension impairment and without weakness in the speech musculature. The DIVA model (Directions into Velocities of Articulators) and the GODIVA model (Gradient Order DIVA) provide a framework for theorizing about two possible subtypes of AOS. The first subtype is hypothesized to arise from damage to the inferior frontal sulcus region (IFS). This damage would result in fluent productions of erroneous or misplaced speech sounds. The second subtype is hypothesized to arise from damage to the frontal operculum region (FO). This damage would result in poorly articulated approximations of the desired syllables. These hypotheses are tested by investigating damage scenarios in DIVA and GODIVA. The results are compared to an apraxic patient case study.