An evidence accumulation model driven exploration of endogenously and exogenously triggered errors and their consequences

Evidence accumulation models (EAMs) are based upon cognitive-theories of speeded decision-making. These models propose that evidence accumulates for decision alternatives until the evidence for one alternative reaches some threshold, and a decision is triggered. EAMs have provided an accurate account of entire response time distributions and the differential probabilities of correct and error responses, across a range of speeded decision- making tasks. Consequently, errors and error-related phenomena have been central benchmarks for EAMs. In what has been a somewhat symbiotic process, errors have influenced EAM development. This thesis uses EAMs to investigate what we refer to as exogenous and endogenous errors and their post-error consequences. We use these terms in a novel way to make reference to the events that caused the errors; exogenous errors refer to errors that have been caused by factors external to the organism (e.g., an unexpected change in an environmental regularity), while endogenous errors refer to those errors that have been caused by the organism (e.g., an error of choice on the previous trial). Section One of this thesis focuses on exogenous errors caused by task irrelevant oddballs as they occur in the distraction paradigm (Schröger & Wolff, 1998). Our use of EAMs in this paradigm is novel. We are successful in our attempts to use EAMs with distraction paradigm data. We also extend EAM techniques to include another source of data, omissions, in an effort to gain further insight in to exogenous errors and the impact they have on behaviour. Section Two of this thesis focuses on endogenous errors and their post-error consequences as they occur in recognition memory paradigm experiments. While EAMs have acknowledged differences in the cause and relative speed of errors, post-error research has largely neglected these factors. We conduct an EAM inspired exploration of post-error consequences. We find evidence of different types of post-error consequences and conduct EAM analysis of this data. Our efforts serve to elucidate theoretical understandings and reconcile debate within the literature.