Modeling speech processing in case of neurogenic speech and language disorders: neural dysfunctions, brain lesions, and speech behavior

Computer-implemented neural speech processing models can simulate patients suffering from neurogenic speech and language disorders like aphasia, dysarthria, apraxia of speech, and neurogenic stuttering. Speech production and perception tasks simulated by using quantitative neural models uncover a variety of speech symptoms if neural dysfunctions are inserted into these models. Neural model dysfunctions can be differentiated with respect to type (dysfunction of neuron cells or of neural connections), location (dysfunction appearing in a specific buffer of submodule of the model), and severity (percentage of affected neurons or neural connections in that specific submodule of buffer). It can be shown that the consideration of quantitative computer-implemented neural models of speech processing allows to refine the definition of neurogenic speech disorders by unfolding the relation between inserted neural dysfunction and resulting simulated speech behavior while the analysis of neural deficits (e.g., brain lesions) uncovered from imaging experiments with real patients does not necessarily allow to precisely determine the neurofunctional deficit and thus does not necessarily allow to give a precise neurofunctional definition of a neurogenic speech and language disorder. Furthermore, it can be shown that quantitative computer-implemented neural speech processing models are able to simulate complex communication scenarios as they appear in medical screenings, e.g., in tasks like picture naming, word comprehension, or repetition of words or of non-words (syllable sequences) used for diagnostic purposes or used in speech tasks appearing in speech therapy scenarios (treatments). Moreover, neural speech processing models which can simulate neural learning are able to simulate progress in the overall speech processing skills of a model (patient) resulting from specific treatment scenarios if these scenarios can be simulated. Thus, quantitative neural models can be used to sharpen up screening and treatment scenarios and thus increase their effectiveness by varying certain parameters of screening as well as of treatment scenarios.