Response selection is critical for successful functioning in the real world. When this ability is impaired (through brain damage or normal cognitive aging), numerous difficulties occur, such as perseverations and other action slips. The present proposal attempts to develop a general model of response selection to account for how stimulus inputs are transformed to in goal-based actions, and how properties of the stimuli and responses interact to affect behavior. The model we implement is based on dynamic neural field theory (DNFT) and operates in real time on subsymbolic representations. Moreover, the model readily simulates neural activation and thus makes predictions for neuroimaging experiments. We focus on two factors that have received scant attention in the response selection literature but are critical for understanding central operations and applying theory to real-world situations. We choose these factors because they are the points at which the proposed model makes clear predictions that are not made by traditional accounts of response selection. Moreover, they present the opportunity to establish new, theory- constraining lines of behavioral research. We examine these factors throughout practice, because practice shapes response selection and changes the underlying operations. Our DNFT model captures these changes and makes specific predictions regarding how practice should alter performance and neural activation. The first factor involves the pairings of input and output modalities. Previous work indicates that response selection processes are profoundly affected by how the modalities are paired. We explore how modalities pairings affect learning with a series of transfer experiments that identify the locus of the changes in the task representation associated with practice. These experiments will test the DNFT account and establish a theoretical framework of central cognitive operations, which are critical for intelligent, goal-oriented behavior. The second critical factor is metrics. Our model predicts that stimulus and response metrics should produce interactions in the behavioral data, and this prediction is borne out in our initial pilot data. While it is well-known that metrics can affect stimulus or response processing separately, interactions between the two factors are difficult for traditional symbolic accounts of response selection to accommodate. These findings can provide direct evidence for metrically coded stimulus-response translation processes. The proposed research will produce new approaches for evaluating and understanding central operations. We focus on the performance of normal, healthy adults in experimental settings, but our findings will have broad implications for both applied and basic research. The development of a theory of central operations will motivate translational research relating to executive dysfunction, a hallmark of many psychiatric and neurological disorders, including Alzheimer's disease, schizophrenia, and frontal lobe damage.
The present proposal aims to develop a theory of response selection, with particular emphasis on explaining the important effects of practice and the pairings of input modality (e.g., eyes versus ears) and output modality (e.g., voice vs. hands). Response selection processes are fundamental to goal-directed human behaviors and are frequently the locus of mental illness. The new knowledge and methodological innovations emerging from this project will facilitate translational research relating to aging and clinical syndromes such as executive dysfunction.
|Freedberg, Michael; Glass, Brian; Filoteo, J Vincent et al. (2016) Comparing the effects of positive and negative feedback in information-integration category learning. Mem Cognit :|
|Freedberg, Michael; Wagschal, Tana T; Hazeltine, Eliot (2014) Incidental learning and task boundaries. J Exp Psychol Learn Mem Cogn 40:1680-700|
|Buss, Aaron T; Wifall, Tim; Hazeltine, Eliot et al. (2014) Integrating the behavioral and neural dynamics of response selection in a dual-task paradigm: a dynamic neural field model of Dux et al. (2009). J Cogn Neurosci 26:334-51|